Green Polymer Chemistry New Products, Processes, and Application 9780841233881, 0841233888, 9780841233898

Citation preview

Green Polymer Chemistry: New Products, Processes, and Applications

ACS SYMPOSIUM SERIES 1310

Green Polymer Chemistry: New Products, Processes, and Applications H. N. Cheng, Editor Southern Regional Research Center U.S. Department of Agriculture, Agricultural Research Service New Orleans, Louisiana

Richard A. Gross, Editor Rensselaer Polytechnic Institute Troy, New York

Patrick B. Smith, Editor Michigan State University Midland, Michigan

Sponsored by the ACS Division of Polymer Chemistry, Inc.

American Chemical Society, Washington, DC Distributed in print by Oxford University Press

Library of Congress Cataloging-in-Publication Data Names: Cheng, H. N., editor. | Gross, Richard A., 1957- editor. | Smith, Patrick B. (Materials scientist), editor. | American Chemical Society. Division of Polymer Chemistry. Title: Green polymer chemistry : new products, processes, and applications / H.N. Cheng, editor (Southern Regional Research Center, U.S. Department of Agriculture, Agricultural Research Service, New Orleans, Louisiana), Richard A. Gross, editor (Rensselaer Polytechnic Institute, Troy, New York), Patrick B. Smith, editor (Michigan State University, Midland, Michigan) ; sponsored by the ACS Division of Polymer Chemistry, Inc. Description: Washington, DC : American Chemical Society, [2018] | Series: ACS symposium series ; 1310 | Includes bibliographical references and index. Identifiers: LCCN 2018049618 (print) | LCCN 2018050363 (ebook) | ISBN 9780841233881 (ebook) | ISBN 9780841233898 Subjects: LCSH: Polymers. | Polymerization. | Green chemistry. Classification: LCC QD381 (ebook) | LCC QD381 .G726 2018 (print) | DDC 660.028/6--dc23 LC record available at https://lccn.loc.gov/2018049618

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences—Permanence of Paper for Printed Library Materials, ANSI Z39.48n1984. Copyright © 2018 American Chemical Society Distributed in print by Oxford University Press All Rights Reserved. Reprographic copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Act is allowed for internal use only, provided that a per-chapter fee of $40.25 plus $0.75 per page is paid to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. Republication or reproduction for sale of pages in this book is permitted only under license from ACS. Direct these and other permission requests to ACS Copyright Office, Publications Division, 1155 16th Street, N.W., Washington, DC 20036. The citation of trade names and/or names of manufacturers in this publication is not to be construed as an endorsement or as approval by ACS of the commercial products or services referenced herein; nor should the mere reference herein to any drawing, specification, chemical process, or other data be regarded as a license or as a conveyance of any right or permission to the holder, reader, or any other person or corporation, to manufacture, reproduce, use, or sell any patented invention or copyrighted work that may in any way be related thereto. Registered names, trademarks, etc., used in this publication, even without specific indication thereof, are not to be considered unprotected by law. PRINTED IN THE UNITED STATES OF AMERICA

Foreword The ACS Symposium Series was first published in 1974 to provide a mechanism for publishing symposia quickly in book form. The purpose of the series is to publish timely, comprehensive books developed from the ACS sponsored symposia based on current scientific research. Occasionally, books are developed from symposia sponsored by other organizations when the topic is of keen interest to the chemistry audience. Before agreeing to publish a book, the proposed table of contents is reviewed for appropriate and comprehensive coverage and for interest to the audience. Some papers may be excluded to better focus the book; others may be added to provide comprehensiveness. When appropriate, overview or introductory chapters are added. Drafts of chapters are peer-reviewed prior to final acceptance or rejection, and manuscripts are prepared in camera-ready format. As a rule, only original research papers and original review papers are included in the volumes. Verbatim reproductions of previous published papers are not accepted.

ACS Books Department

Contents Preface .............................................................................................................................. xi 1.

Green Polymer Chemistry: Pipelines Toward New Products and Processes ..... 1 H. N. Cheng, Richard A. Gross, and Patrick B. Smith

Novel Bioengineering Approaches 2.

Bioorthogonal Approaches To Prepare Specifically Modified Functional Proteins ................................................................................................................... 15 Seiichi Tada, Hideyuki Miyatake, Takanori Uzawa, and Yoshihiro Ito

3.

Engineering the Microbial Cell Membrane To Improve Bioproduction .......... 25 Laura R. Jarboe, Jeffery B. Klauda, Yingxi Chen, Kirsten M. Davis, and Miguel C. Santoscoy

4.

Microbial Secretion System of Lactate-Based Oligomers and Its Application .............................................................................................................. 41 Camila Utsunomia and Seiichi Taguchi

New Enzymatic Methodologies 5.

Structural and Mutational Analysis of Polyethylene Terephthalate–Hydrolyzing Enzyme, Cut190, Based on Three-Dimensional Docking Structure with Model Compounds of Polyethylene Terephthalate .... 63 Takeshi Kawabata, Masayuki Oda, Nobutaka Numoto, and Fusako Kawai

6.

Conjugates Based on Enzyme-Metal-Organic Frameworks for Advanced Enzymatic Applications ......................................................................................... 77 Qian Liu and Cerasela Zoica Dinu

7.

Protease-Catalyzed Polymerization of Tripeptide Esters Containing Unnatural Amino Acids: α,α-Disubstituted and N-Alkylated Amino Acids ...................................................................................... 95 Kousuke Tsuchiya and Keiji Numata

vii

New Materials Based on Polysaccharides 8.

Sustainable Development of Polysaccharide Polyelectrolyte Complexes as Eco-Friendly Barrier Materials for Packaging Applications .......................... 109 Kai Chi and Jeffrey M. Catchmark

9.

Effects of Monomer Compositions and Molecular Weight on Physical Properties of Alginic Acid Esters ........................................................................ 125 Yusuke Matsumoto, Daisuke Ishii, and Tadahisa Iwata

10. Preparation of Hydrophobically Modified Cashew Gum Through Reaction with Alkyl Ketene Dimer ..................................................................................... 137 Atanu Biswas, Sanghoon Kim, Megan Buttrum, Roselayne F. Furtado, Carlucio R. Alves, and H. N. Cheng

Bio-Related Polyesters, Polyamides, and Polyurethanes 11. Salicylic Acid-Based Poly(anhydride-esters): Synthesis, Properties, and Applications .......................................................................................................... 149 Yue Cao and Kathryn E. Uhrich 12.

L-Lysine-Modified

Poly(ester-urethane) Based on Polycaprolactone for Controlled Release of Hydrocortisone ............................................................... 163 Karla A. Barrera-Rivera and Antonio Martínez-Richa

13. Rational Synthesis of Biobased Hyperbranched Poly(ester)s for Sustained Delivery ................................................................................................................. 177 Tracy Zhang, Bob A. Howell, Steven J. Martin, Brandon Zhu, Daniel Zhang, and Patrick B. Smith 14. Aromatic Bioplastics with Heterocycles ............................................................. 201 Sumant Dwivedi and Tatsuo Kaneko

Bio-Based Phenolics and Composites 15. Ferulic Acid- and Sinapic Acid-Based Bisphenols: Promising Renewable and Safer Alternatives to Bisphenol A for the Production of Bio-Based Polymers and Resins ............................................................................................ 221 Louis Hollande and Florent Allais 16. Application of Bio-Based Epoxy Resin as the Matrix for Composites ............ 253 Liang Yue 17. Strategic Assemblies of Modified Xylochemicals for New Bio-Based Polymers and Composites ................................................................................... 265 Joseph F. Stanzione III, Elyse A. Baroncini, Alexander W. Bassett, and Silvio Curia

viii

18. Enhancing the Sustainability of High-Performance Fiber Composites .......... 281 Christopher N. Kuncho, Wenhao Liu, Johannes Möller, Julia Kammleiter, Julia Stehle, Akshay Kokil, Emmanuelle Reynaud, and Daniel F. Schmidt

Bio-Based Monomers and Resulting Products 19. Divinylglycol, a Glycerol-Based Monomer: Valorization, Properties, and Applications .......................................................................................................... 299 Léa Bonnot, Christophe Len, Etienne Grau, and Henri Cramail 20. Levulinic Acid as Sustainable Feedstock in Polymer Chemistry .................... 331 Manuel Hartweg and C. Remzi Becer 21. Levulinic Acid: A Valuable Platform Chemical for the Fermentative Synthesis of Poly(hydroxyalkanoate) Biopolymers ........................................... 339 Richard D. Ashby and Daniel K. Y. Solaiman 22. Bioadvantaged Nylon from Renewable Muconic Acid: Synthesis, Characterization, and Properties ....................................................................... 355 Sanaz Abdolmohammadi, Nacú Hernández, Jean-Philippe Tessonnier, and Eric W. Cochran 23. Bio-Based Monomers as a New Route to Improved Performance in Thermosetting Resins: Examples from Cyanate Ester Studies ....................... 369 Andrew J. Guenthner, Benjamin G. Harvey, and Matthew C. Davis

Bio-Based Solvents and Additives 24. Investigation of PolarClean and Gamma-Valerolactone as Solvents for Polysulfone Membrane Fabrication ................................................................... 385 Xiaobo Dong, Halle D. Shannon, and Isabel C. Escobar 25. Flame Retardants from Renewable Sources: Food Waste, Plant Oils, and Starch .................................................................................................................... 405 Bob A. Howell, Yoseph G. Daniel, and Eric A. Ostrander Editors’ Biographies .................................................................................................... 423

Indexes Author Index ................................................................................................................ 427 Subject Index ................................................................................................................ 429

ix

Preface Green polymer chemistry is now a global pursuit and comprises diverse disciplines, such as organic synthesis, polymer chemistry, material science, microbiology, molecular biology, catalysis, enzymology, environmental science, analytical chemistry, and chemical engineering. This field is equally active in the United States as well as Europe and Asia. Researchers, students, and people new to this field value a forum to meet and share ideas; this can take the form of a symposium dedicated to this field, or a special book that features the latest work done by leading practitioners. For many years, we have been organizing an international symposium on “Green Polymer Chemistry: Biobased Materials and Biocatalysis” (or variations of this theme) at the American Chemical Society (ACS) national meetings at 3-year intervals. The symposium series has been very successful and serves to bring together a community of scientists with different backgrounds but with common research interests. In the latest symposium, which we organized in August 2017 in Washington, D.C., we had a total of 84 presentations and 16 posters (one of the largest symposia in the meeting). The symposium was structured into 10 sessions: • • • • • • • • • •

Bio-Based Materials: Industrial Perspectives Developments in Biocatalysts Green Biocatalytic Transformations Chemical Catalytic Routes to Bio-Based Materials New Reaction Strategies and Materials Polysaccharide-Based Materials Plant Oils and Ferulate-Based Materials Bio-Based Thermosetting Resins Therapeutics and Opto-Electronics Further Applications of Bio-Based Materials

Many of the leading researchers in this field accepted the invitation to speak, and they reported exciting findings in various areas, including new bio-based source materials, green conversion methods, new or improved processing methodologies, and green polymer-related products. Because of the enthusiastic reception at Washington, D.C., we decided to edit a symposium book that provides a written forum for our colleagues to share their latest research work. We are very pleased to have 25 chapters in this book, which that cover a representative range of topics in green polymer chemistry. For convenience, this book is organized into seven sections: novel bioengineered approaches; new enzymatic methodologies; new materials based xi

on polysaccharides; bio-related polyesters, polyamides, and polyurethanes; bio-based phenolics and composites; bio-based monomers and resulting products; and bio-based solvents and additives. We thank the authors for their contributions and for their patience during the peer review and copyediting processes. Thanks are also due to the many reviewers, who volunteered their time and talent to read the manuscript and make valuable suggestions. This book is targeted for scientists, engineers, and students, who are involved or interested in green polymer chemistry. Hopefully, they are intrigued by the range and the depth of the topics covered. This can also be a textbook for a course on green polymer chemistry and a reference book for people who need information on specific topics involving green chemistry and sustainability. We gratefully acknowledge the assistance of the ACS Books staff (particularly Arlene Furman, Amanda Koenig, and Sara Tenney). The 2017 Green Polymer Chemistry symposium was jointly sponsored by ACS Divisions of Polymer Chemistry (POLY) and Polymeric Materials: Science & Engineering (PMSE), with the nominal co-sponsorship of the ACS Divisions of Agricultural and Food Chemistry (AGFD) and Cellulose and Renewable Materials (CELL). We also thank BASF for additional financial support.

H. N. Cheng Southern Regional Research Center U.S. Department of Agriculture, Agricultural Research Service 1100 Robert E. Lee Boulevard New Orleans, Louisiana 70124, United States

Richard A. Gross Department of Chemistry and Chemical Biology Rensselaer Polytechnic Institute 110 8th Street Troy, New York 12180, United States

Patrick B. Smith Michigan State University 1910 West St. Andrews Road Midland, Michigan 48640, United States

xii

Chapter 1

Green Polymer Chemistry: Pipelines Toward New Products and Processes H. N. Cheng,*,1 Richard A. Gross,2 and Patrick B. Smith3 1Southern

Regional Research Center, USDA Agriculture Research Service, 1100 Robert E. Lee Boulevard, New Orleans, Louisiana 70124, United States 2Department of Chemistry and Chemical Biology, Rensselaer Polytechnic Institute, 110 8th Street, Troy, New York 12180-3590, United States 3Michigan State University, 1910 West St. Andrews Road, Midland, Michigan 48640, United States *E-mail: [email protected].

The quest for green chemistry that reduces or eliminates the use or generation of hazardous substances in polymer products and processes represents an opportunity for polymer scientists and engineers. Indeed, the need to promote safety, sustainability, and environmental stewardship, and to reduce plastic waste and energy usage is widely recognized. As a result, intense research activities are ongoing worldwide to develop green polymer products and processes. Advances have been made in many areas, including new source materials, green conversion methods, new or improved processing, and green polymer-related products. In this article, an overview is provided of green polymer chemistry, particularly emphasizing the different ways whereby new products and processes have been made. Selected examples are provided, particularly from the chapters included in this symposium volume and from the presentations given during a Green Polymer Chemistry symposium at the American Chemical Society (ACS) national meeting in Washington, D.C. (August 2017).

© 2018 American Chemical Society

Introduction Green polymer chemistry entails the development of polymer-related products and processes that reduce or eliminate the use or generation of hazardous substances (1–9). Over the years, different methods have been adopted to produce green polymers (3–15). These include the use of agro-based starting materials, biocatalysts (such as enzymes and microbial cells), bio-based monomers or building blocks, and biodegradable and/or recyclable polymers. Improvements in processes have also been made in order to minimize energy usage, byproduct yield, and hazardous solvents. For convenience, we have previously (11, 14, 15) grouped the major modes of green polymer chemistry into eight categories (Table 1). These categories are compatible with the principles of green chemistry as proposed initially by Anastas and Warner (1, 2).

Table 1. Major Applications of Green Polymer Chemistry. Adapted with permission from ref. (15). (Copyright 2015 American Chemical Society). Major Pathways

Examples

Green catalysts

Biocatalysts, such as enzymes and whole cells

Diverse feedstocks

Bio-based, renewable feedstock and building blocks (e.g., carbohydrates, peptides/proteins, triglycerides, lignin, terpenes). Bio-fillers in composites. Bio-based small molecules (including CO2) as monomers

Degradable polymers

Agro-based polymer backbones. Biodegradable polymers (e.g., some polyesters and polyamides)

Recycling of polymer products and catalysts

Many degradable polymers can potentially be recycled. Immobilized enzymes may be reused

Energy generation or minimization of use

Biofuels. Energy-saving processes (e.g., reactive extrusion, jet cooking, microwave)

Optimal molecular design and activity

Atom economy. Polymers with designed structures or functions. Improved catalytic activity. Improved biocatalysts via genetic and metabolic engineering.

Avoidance of hazardous chemicals

Use of safer chemicals and solvents. Use of water, ionic liquids, or reactions without solvents

Improved syntheses and processes

Reaction efficiency. Reduction of hazards and toxicity. Waste minimization. Byproduct reduction

Although the concept of green polymer chemistry emerged relatively recently, it has quickly become an active field of research and development (R&D) due to several factors, such as the desire to decrease the dependence on petroleum as raw material for polymers, the relatively stable prices for agro-based products, the awareness of limited natural resources and the need for sustainability, the danger of excessive plastic waste, the biodegradable and environmental benefits of biobased materials, and the increased economic opportunities for farmers and rural 2

communities through increased usage of agricultural products. A large number of books (3–9) and reviews (10–15) have been written on this topic. In this article, the authors have endeavored to give an overview of green polymer chemistry, with a particular emphasis on the product development aspects of this topic. Specific references have been made to the chapters included in this symposium volume (16–39), which were presented at the ACS National Meeting in Washington, D.C. in August 2017.

Pipelines to Green Products As noted earlier (15), the pipelines toward the generation of green polymeric products can be envisioned as in Figure 1.

Figure 1. A simplified scheme that illustrates the different aspects of the development of green polymeric products. Adapted with permission from ref. (15) (Copyright 2015 American Chemical Society).

In Figure 1 the source can be diverse feedstocks (e.g., bio-based monomers or building blocks, and agro-based raw materials, such as carbohydrates, proteins, triglycerides, lignin, and terpenes) or recycled polymers. The conversion may involve polymerization of monomers or building blocks, polymer modification or derivatization, reaction with a catalyst (e.g., enzymatic, microbial, inorganic, or organic), or transformation through molecular biology (with new or improved enzymes, or metabolic engineering). In the processing area, reaction efficiency, byproduct and waste reduction, energy savings, hazard reduction, and green solvents are some of the opportunities. At the end of R&D, of course, it is desirable to come up with products; these can be polymers or their derivatives, biopolymers (including bioplastics), composites, polymer additives, and other polymer-related materials. Product development may also involve regulatory compliance, intellectual property protection, application development (e.g., formulation, optimization, stabilization, evaluation), and a range of scale-up and engineering issues, where green chemistry can also play a major role. In the following write-up, examples are given of each of the above four aspects of development of the green polymer-related products. 3

Source Materials for Green Polymer Chemistry A major source of materials for green polymer chemistry consists of the agro-based raw materials, such as carbohydrates, peptides/proteins, triglycerides, lignin, terpenes, and others. All these materials have been widely used for green product development. Numerous books (40–49) and review articles (50–56) have appeared on this subject. In this book, several polysaccharides were utilized as starting materials by Chi and Catchmark (22), Matsumoto et al. (23), and Biswas et al. (24). Chitosan was converted into a flame retardant by Howell et al. (39). Protein modification was reported by Tada et al. (16), and polymerization of peptides by Tsuchiya and Numata (21). Polymerization of a large number of lignin-based phenolic materials was reviewed by Stanzione et al. (31). Other source materials include bio-based monomers and building blocks. Phenolics seem to be among the more popular bio-based raw materials. In this book, four chapters used bio-based phenolics as the starting materials for polymerization. For example, Hollande and Allais (29) employed ferulic acid and sinapic acid as alternatives to bisphenol A for the production of bio-based epoxies, epoxy amine resins, cyclocarbonates, and non-isocyanate polyurethane. Yue (30) made epoxy resins with alkyl diphenolate esters and glycidyl ether of eugenol. Guenthner et al. (37) took resveratrol and carvacrol and produced thermosetting resins from them. Stanzione et al. (31) synthesized monomers, oligomers, and resins from guaiacyls, coniferyls, coumaryls, and other bio-based materials, such as cardanol and furandicarboxylic acid. Two chapters in this book are concerned with triglyceride oils as starting materials. In the first, Kuncho et al. (32) used anhydride-cured epoxidized linseed oil in the preparation of glass fiber composites. In the second, Howell et al. (39) converted castor oil into a bio-based flame retardant. Four chapters report on the use of non-aromatic bio-based acids. Thus, levulinic acid was converted by Hartweg and Becer (34) into polyamidelactams and by Ashby and Solaiman (35) into poly(hydroxyalkanoates). Abdolmohammadi et al. (36) utilized muconic acid and its derivatives to produce unsaturated polyamides. Howell et al. made flame retardants from tartaric acid (39). Other examples of bio-based source materials include: Utsunomia and Taguchi’s chapter on lactic acid oligomers as a building block for poly(ester-urethanes) (18); the chapter by Bonnot et al. (33) on divinyl glycol (a glycerol-based monomer) for polycondensation and polyaddition reactions; and the conversion of isosorbide to bio-based flame retardants (39). Green Conversion Methods The polymerization of bio-based monomers or building blocks constitutes the largest number of green conversion methods. Among the chapters in this book, polycondensation reactions are the most common. Thus, several chapters deal with polyesters (27, 31, 33, 35, 36), poly(anhydride esters) (25), poly(ester-urethanes) (26), polyureas and polyimides (28, 34), polyamides (28), polycyanurates (37), 4

epoxy resins (29–33), and cyclocarbonates and non-isocyanate urethane oligomers (29). Polyaddition reactions are represented in two chapters (30, 31). In the first, lignin-based methacrylates were polymerized via controlled reversible addition-fragmentation chain transfer (RAFT) (31). In the second, divinyl glycol was polymerized through its vinyl functionality by acyclic diene metathesis (ADMET) and thiol-ene addition (33). Polymer modification reactions are represented by two chapters, both pertaining to polysaccharides (23, 24). In one chapter, alginic acid was esterified with alkyl chains with different chain lengths (23). In the second chapter, cashew gum was derivatized with alkyl ketene dimer to impart a hydrophobic character to the polysaccharide (24). Enzyme-catalyzed reactions are reported in three chapters. Kawabata et al. (19) modified the enzyme cutinase genetically to improve its degradative activity toward poly(butylene succinate-co-adipate). Tsuchiya and Numata (21) used the protease enzyme to polymerize tripeptide esters containing unnatural amino acids. Liu and Dinu (20) reported a new method to immobilize enzymes by using metalorganic frameworks. The synthesis of proteins through biorthogonal techniques can also be considered a green reaction. Tada et al. (16) used this technique to produce two types of proteins: adhesive growth factors and PEGylated proteins. Green Processing Processing and scale-up are important steps in the development of a product. With green processing, further considerations are needed to design and carry out a process that reduces pollution and chemical hazards and minimizes risk to human health and the environment without sacrificing economic viability and efficiency. In many of the chapters in this book, some efforts seemed to have been made to optimize the reaction processes. Four chapters, in particular, place special emphasis on processing. For processes involving microbial biocatalysis, one of the factors that inhibits the optimization of the process is the damage done on the lipid-rich microbial membrane. In their chapter, Jarboe et al. (17) reviewed their successful studies to engineer the microbial cell membrane in order to improve the bioproduction of fuels and chemicals. In the microbial production of lactate-based oligomers, the amount of the product secreted by a micro-organism is an important factor in product yield. In their chapter, Utsunomia and Taguchi (18) described their efforts to enhance the secretory production of lactate oligomers. In the conversion of glucose to muconic acid, Abdolmohammadi et al. (36) reported a unique approach of utilizing an engineered strain of Saccharomyces cerevisiae yeast for the conversion of sugar into muconic acid. The muconic acid was then converted into trans-3-hexenedioic acid through electrochemical hydrogenation. Both acids are useful monomers for the synthesis of polymers. For the fabrication of polysulfone membrane, a suitable solvent is needed. Petroleum-derived solvents tend to be toxic; thus, a bio-derived solvent with 5

low toxicity would be an ideal substitute. Dong et al. (38) discovered that a mixture of methyl-5-(dimethylamino)-2-methyl-5-oxopentanoate (PolarClean) and γ-valerolactone provided the best results. Both solvents were reported to be water-soluble, eco-friendly, and biodegradable.

Green Products In R&D planning, research teams often have specific products in mind. Therefore, it is not surprising that new (or potential) products are mentioned in many chapters of this book. For illustration, these included new proteins for medical and other uses (16), improved microbial systems for enhanced product yields (17, 18), cutinase with improved hydrolytic activity toward polyesters (19), a new type of enzymes immobilized on metal-organic frameworks (20), and polypeptides containing unnatural amino acids (21). In the polysaccharide area, polyelectrolyte complexes were reported as eco-friendly barrier materials for packaging applications (22), alginic esters as a potential bioplastic with good thermal and mechanical properties (23), and hydrophobically modified cashew gum, possibly as polymeric surfactant, encapsulant, and thickening agent (24). Among the polyesters, the salicylic acid-containing materials reported by Cao and Uhrich (25) seemed to be very promising in biomedical applications. The lysine-modified, polycaprolactone-based poly(ester-urethane), developed by Barrera-Rivera and Martinez-Richa (26), was targeted for controlled release of hydrocortisone. Zhang et al. (27) prepared hyperbranched polyesters for sustained delivery of active ingredients. Dwivedi and Kaneko (28) developed a large number of heterocyclic and/or aromatic polyamides, polyureas, and polyimides as bioplastics. In the phenolics area, a number of bio-based materials were exploited as building blocks and replacements for bisphenol A, such as ferulic and sinapic acid-based bisphenols (29), eugenol (30), guaiacyls, coniferyls, coumaryls, cardanol, and furandicarboxylic acid (31). The resulting materials can be potentially used as final products or as components in composites. The anhydride-cured epoxidized linseed oil was also reported to be a suitable material for the preparation of high-performance fiber composites (32). Quite a few polymeric products were produced from the bio-based monomers reported in this book. Thus, divinyl glycol was converted via polycondensation and polyaddition reactions to various products (33). Levulinic acid was polymerized to polyamide-lactams (34) and poly(hydroxyalkanoates) (35), muconic acid and its derivatives into polyamides (36), resveratrol and carvacrol into polycyanurates and thermosetting polymer networks (37). Finally, two bio-derived solvents (PolarClean and γ-valerolactone) were found to be suitable for polysulfone membrane fabrication (38), and a number of flame retardants were synthesized from bio-based and nontoxic building blocks, such as tartaric acid, chitosan, castor oil, and isosorbide (39).

6

Conclusions Green polymer chemistry appears to have gained momentum in the past 10 years, as shown by the increasing R&D activities in academia, industry, and government laboratories. In the final analysis, these developments are driven by economics and sustainability. In broad terms, sustainability covers the management of natural resources, environmental protection, reduction in plastic waste, and avoidance of hazardous chemicals, in order to ensure that future generations of mankind can continue to enjoy an acceptable quality of life on earth. In polymer product development, it is useful, therefore, to design the entire life cycle of a product—from starting materials, conversion, process, use, and disposal, such as to maximize the “green” aspects (as shown in Table 1) and to minimize waste and hazards. Thus, the entire pipeline for product/process development (as exemplified by Figure 1) needs to be managed in order to obtain optimal results. However, the economic considerations cannot be ignored, and new products and processes must be competitive in terms of product cost and market demand. Fortunately, the polymer chemistry community is fully aware of the challenges and opportunities posed by green chemistry and sustainability. Great progress has already been made across the product/process pipelines, as shown by the R&D efforts illustrated in this article (and described in more detail in individual chapters of this book). With the continued interest and engagement on the part of industry and entrepreneurial businesses, further advancements can be made in product and process development in the future. Hopefully, eco-friendly “green” polymeric products will gain increasing popularity in the marketplace of tomorrow.

Acknowledgments Thanks are due to the chapter authors of this symposium volume for their contributions and for their cooperation during the peer review process. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture (USDA). USDA is an equal-opportunity provider and employer.

References 1. 2. 3. 4.

Anastas, P. T.; Warner, J. C. Green Chemistry: Theory and Practice; Oxford Univ. Press: Oxford, U.K., 2000. Horváth, I. T.; Anastas, P. T. Chem. Rev. 2007, 107 (6), 2169–2173. Stevens, F. S. Green Plastics: An Introduction to the New Science of Biodegradable Plastics; Princeton Univ. Press: Princeton, NJ, 2002. Green Polymer Chemistry: Biocatalysis and Biomaterials; Cheng, H. N., Gross, R. A., Eds.; ACS Symposium Series 1043; American Chemical Society: Washington, DC, 2010.

7

5.

6.

7.

8. 9. 10. 11.

12. 13. 14.

15.

16.

17.

18.

19.

Green Polymerization Methods: Renewable Starting Materials, Catalysis and Waste Reduction; Mathers, R. T., Meier, M. A. R., Eds.; Wiley-VCH: Weinheim, Germany, 2011. Green Polymer Chemistry: Biocatalysis and Materials II; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1144; American Chemical Society: Washington, DC, 2013. Green Polymer Chemistry: Biobased Materials and Biocatalysis; Cheng, H. N., Gross, R. A.;, Smith, P. B., Eds.; ACS Symposium Series 1192; American Chemical Society: Washington, DC, 2015. Green Polymers and Environmental Pollution Control; Khalaf, M. N.; CRC Press: Boca Raton, FL, 2016. Enzymatic Polymerization towards Green Polymer Chemistry; Kobayashi, S., Uyama, H., Kadokawa, J., Eds.; Springer: Heidelberg, Germany, 2018. Williams, C. K.; Hillmyer, M. A. Polymer Rev. 2008, 48, 1–10. Green Polymer Chemistry: Biocatalysis and Biomaterials; Cheng, H. N., Gross, R. A., Eds.; ACS Symposium Series 1043; American Chemical Society: Washington, DC, 2010; pp 1–14. Gandini, A. Green Chem. 2011, 13, 1061–1083. Mulhaupt, R. Macromol. Chem. Phys. 2013, 214, 159–174. Green Polymer Chemistry: Biocatalysis and Materials II; Cheng, H. N.; Gross, R. A.; Smith, P. B.; Eds.; ACS Symposium Series 1144; American Chemical Society: Washington, DC, 2013; pp 1–12. Green Polymer Chemistry: Biobased Materials and Biocatalysis; Cheng, H. N.; Gross, R. A.; Smith, P. B.; Eds.; ACS Symposium Series 1192; American Chemical Society: Washington, DC, 2015; pp 1–13. Tada, S.; Miyatake, H.; Uzawa, T.; Ito, Y. Bioorthogonal approaches to prepare specifically modified functional proteins. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 2, pp 15−24. Jarboe, L. R.; Klauda, J. B.; Chen, Y.; Davis, K. M.; Santoscoy, M. C. Engineering the microbial cell membrane to improve bio-production. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 3, pp 25−39. Utsunomia, C.; Taguchi, S. Microbial secretion system of lactate-based oligomers and its applications. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 4, pp 41−60. Kawabata, T.; Oda, M.; Numoto, N.; Kawai, F. Structural and mutational analysis of PET-hydrolyzing enzyme, Cut190, based on 3D docking structure with model compounds of PET. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 5, pp 63−75.

8

20. Liu, Q.; Dinu, C. Z. Conjugates based on enzyme-metal-organic frameworks for advanced enzymatic applications. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 6, pp 77−93. 21. Tsuchiya, K.; Numata, K. Protease-catalyzed polymerization of tripeptide esters containing unnatural amino acids: α,α-disubstituted and N-alkylated amino acids. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 7, 95−105. 22. Chi, K.; Catchmark, J. M. Sustainable development of polysaccharide polyelectrolyte complexes as eco-friendly barrier materials for packaging applications. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 8, pp 109−123. 23. Matsumoto, Y.; Ishii, D.; Iwata, T. Effects of monomer compositions and molecular weight on physical properties of alginic acid esters. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 9, pp 125−135. 24. Biswas, A.; Kim, S.; Buttrum, M.; Furtado, R. F.; Alves, C. R.; Cheng, H. N. Preparation of hydrophobically modified cashew gum through reaction with alkyl ketene dimer. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 10, pp 137−146. 25. Cao, Y.; Uhrich, K. E. Salicylic acid-based poly(anhydride-esters): synthesis, properties and applications. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 11, pp 149−162. 26. Barrera-Rivera, K. A.; Martinez-Richa A. L-lysine-modified poly (ester-urethane) based on PCL, for controlled release of hydrocortisone. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018, Chapter 12, pp 163−175. 27. Zhang, T.; Howell, B. A.; Martin, S. J.; Zhu, B.; Zhang, D.; Smith, P. B. Rational synthesis of biobased hyperbranched poly(esters) for sustained delivery. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 13, pp 177−199. 28. Dwivedi, S.; Kaneko, T. Aromatic bioplastics with heterocycles. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H.

9

29.

30.

31.

32.

33.

34.

35.

36.

N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 14, pp 201−218. Hollande, L.; Allais, F. Ferulic acid- and sinapic acid-based bisphenols: promising renewable and safer alternatives to bisphenol A for the production of nio-based epoxies, epoxy-amines resins, cyclocarbonates and NIPU oligomers. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 15, pp 221−251. Yue, L. Application of bio-based epoxy resin as the matrix for composites. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 16, pp 253−263. Stanzione, J. F.; Baroncini, E. A.; Bassett, A. W.; Curia, S. Strategic assemblies of modified xylochemicals for new bio-based polymers and composites. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 17, pp 265−279. Kuncho, C. N.; Liu, W.; Möller, J.; Kammleiter, J.; Stehle, J.; Kokil, A.; Reynaud, E.; Schmidt, D. F. Enhancing the sustainability of high performance fiber composites. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 18, 281−295. Bonnot, L.; Len, C.; Grau, E.; Cramail, H. Divinylglycol, a glycerol-based monomer: valorization, properties and applications. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 19, pp 299−330. Hartweg, M.; Becer, C. R. Levulinic acid as sustainable feedstock in polymer chemistry. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 20, pp 331−338. Ashby, R. D.; Solaiman, D. K. Y. Levulinic acid: A valuable platform chemical for the fermentative synthesis of poly(hydroxyalkanoate) biopolymers. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 21, pp 339−354. Abdolmohammadi, S.; Hernández, N.; Tessonnier, J.-P.; Cochran, E. W. Bioadvantaged Nylon from renewable muconic acid: synthesis, characterization and properties. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 22, pp 355−367.

10

37. Guenthner, A. J.; Harvey, B. G.; Davis, M. C.. Bio-based monomers as a new route to improved performance in thermosetting resins: examples from cyanate ester studies. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 23, pp 369−381. 38. Dong, X.; Shannon, H. D.; Escobar, I. C. Investigation of PolarClean and gamma-valerolactone as solvents for polysulfone membrane fabrication. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC, 2018; Chapter 24, pp 385−403. 39. Howell, B. A.; Daniel, Y. G.; Ostrander, E. A. Flame retardants from renewable sources: food waste, plant oils and starch. In Green Polymer Chemistry: New Products, Processes, and Applications; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; ACS Symposium Series 1310; American Chemical Society: Washington, DC; Chapter 25, pp 405−421. 40. Natural Fibers, Biopolymers, and Biocomposites; Mohanty, A. K.; Misra, M.; Drzal, L. T.; CRC Press, Boca Raton, FL, 2005. 41. Polymer Biocatalysis and Biomaterials; Cheng, H. N., Gross, R. A., Eds.; ACS Symposium Series 900; American Chemical Society: Washington, DC, 2005. 42. Wool, R.; Sun, X. S. Bio-Based Polymers and Composites; Elsevier, Burlington, MA, 2005. 43. Polymer Biocatalysis and Biomaterials II; Cheng, H. N., Gross, R. A., Eds; ACS Symposium Series 999; American Chemical Society: Washington, DC, 2008. 44. Biobased Monomers, Polymers, and Materials; Smith, P. B.; Gross, R. A., Eds., ACS Symposium Series 1105; American Chemical Society: Washington, DC, 2012. 45. Degradable Polymers and Materials: Principles and Practice (2nd Edition); Khemani, K.; Scholz, C., Eds.; ACS Symposium Series 1114; American Chemical Society: Washington, DC, 2012. 46. Bio-Based Plastics: Materials and Applications; Kabasci, S., Ed.; Wiley: Chichester, U.K., 2014. 47. Greene, J. P. Sustainable Plastics; Wiley: Hoboken, NJ, 2014. 48. Biodegradable and Biobased Polymers for Environmental and Biomedical Applications; Kalia, S., Avérous, L., Eds.; Scrivener, Salem, MA, 2016. 49. Ashter, S. A.; Technology and Applications of Polymers Derived from Biomass; William Andrew/Elsevier: Cambridge, MA, 2017. 50. Meier, M. A. R.; Metzgerb, J. O.; Schubert, U. S. Chem. Soc. Rev. 2007, 36, 1788–1802. 51. Bhardwaj, R.; Mohanty, A. K. J. Biobased Mater. Bioenergy 2007, 1, 191–209. 52. Raquez, J.-M.; Deléglise, M.; Lacrampe, M.-F.; Krawczak, P. Progr. Polym. Sci. 2010, 35, 487–509. 53. Chen, G-Q.; Patel, M. K. Chem. Rev. 2012, 112, 2082–2099.

11

54. Babu, R. P.; O’Connor, K.; Seeram, R. Prog. Biomater. 2013, 2, 8. 55. Reddy, M. M.; Vivekanandhan, S.; Misra, M.; Bhatia, S. K.; Mohanty, A. K. Prog. Polym. Sci. 2013, 38, 1653–1689. 56. Zhu, Y.; Romain, C.; Williams, C. K. Nature 2016, 540, 354–362.

12

Novel Bioengineering Approaches

Chapter 2

Bioorthogonal Approaches To Prepare Specifically Modified Functional Proteins Seiichi Tada,1 Hideyuki Miyatake,2 Takanori Uzawa,1,2 and Yoshihiro Ito*,1,2 1Emergent

Bioengineering Materials Research Team, RIKEN Center for Emergent Matter Science, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan 2Nano Medical Engineering Laboratory, RIKEN Cluster for Pioneering Research, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan *E-mail: [email protected],

Techniques to incorporate noncanonical amino acids at a specific position of a protein extend conventional protein engineering into bioorthogonal protein engineering. By employing such bioorthogonal protein techniques, we have engineered two types of bioorthogonal proteins. The first type consists of adhesive growth factors, which impart biological activity on the surface of medically oriented materials, including metals. To this end, noncanonical amino acids found in underwater adhesive proteins were incorporated into growth factors to enable surface coatings. The other type is genetically PEGylated proteins to potentially realize protein drugs. Two different lengths of poly(ethylene glycol) (PEG) were incorporated into specific sites using PEG-conjugated transfer RNA through cell-free translation.

Introduction Chemists can precisely (even enantio-selectively) synthesize small molecules; however, it remains difficult to synthesize large molecules, including functional biological macromolecules with a high level of complexity. For example, solid-phase synthesis methods enable synthesis of oligonucleotides, © 2018 American Chemical Society

oligopeptides, and oligosaccharides, although there is a limitation on their chain lengths because of the exponential accumulation of reaction losses. This limitation is partly overcome by extension of conventional protein engineering to bioorthogonal protein engineering. Currently, preparation of many types of complex macromolecules as synthetic polymer–biomacromolecular conjugates and biologically processed biomacromolecules is possible (1). Here we describe three methods to engineer bioorthogonal proteins, as shown in Figure 1. The first method involves enzymatic modification of a genetically engineered protein with a noncanonical component (2, 3). The second method entails enzymatic ligation of a noncanonical peptide that was separately prepared by solid-phase chemistry (4). The third method involves ribosomal synthesis using misacylated transfer RNA (tRNA) (5–8).

Figure 1. Synthetic strategies of bioorthogonal protein engineering.

We used tyrosinase (3) and sortase (4) as examples of enzymatic modification and ligation, respectively. Using these methods, we incorporated 3,4-dihydroxyphenylalanine and phosphorylated serine into growth factor proteins. Both noncanonical amino acids play important roles in adhesion of the mussel foot protein (9) and salivary statherin (10), respectively. The synthesized proteins served as adhesives on material surfaces and induced cell differentiation or enhanced cell growth significantly on the modified surface. Furthermore, the long-lasting activity and high local concentration of the growth factors on the surface induced significant biological effects on cells (11). As an example of a bioorthogonal protein using the misacylated tRNA technique, we employed a ribosome translation system together with a tRNA that was acylated with poly(ethylene glycol) (PEG)–containing amino acids (12). We genetically incorporated different lengths of PEG at a position of stop codon or four-base codon via cognate anticodon tRNAs, which were acylated with PEG-containing amino acids. 16

Adhesive Growth Factors A wide variety of different biomaterials are currently employed for medical applications. Research towards the development of new biomaterials traditionally focused on the preparation of functional materials capable of simple adhesion of cells or the connection of tissues to metals and ceramics. However, there is growing interest in the development of biomaterials that involve the immobilization of growth factors, which would allow these artificial materials to regulate specific cellular functions, including gene expression processes associated with cell growth and differentiation (13). Binding growth factors have been designed and applied for two purposes. One is the delivery of growth factors to specific sites or enhancement of local concentrations at specific sites. For this purpose, binding growth factors are injected directly into animals or humans (14). The second purpose involves surface modification of scaffolds for implantation of regenerated tissues based on tissue engineering (2). For the latter application, various types of materials have been modified. However, although numerous studies have reported the use of metallic materials in medical devices such as artificial joints, dental implants, and stents, there are very few reports describing surface modification of metal or ceramic materials with biological agents (15). In recent years, an increasing number of reports have been published on natural adhesives (16–18). To design such proteins as a strategy for surface modification of metals and ceramics, biomimetic approaches inspired by underwater adhesive proteins have been used with biological materials, as shown in Figure 2. One of the underwater adhesive proteins is the mussel foot protein (Mfp), which is involved in a sticky pad at the end of threads that adhere firmly to rocks (or any other hard surface). The active sequence has 3,4-dihyroxyphenylalanine (DOPA). The salivary statherin protein is another underwater adhesive protein, which is a multifunctional molecule that possesses high affinity for calcium phosphate minerals such as hydroxyapatite (HA) to maintain the appropriate mineral solution dynamics of enamel. The active site of the statherin contains phosphorylated serines. The active sites for the adhesion of both proteins consist of posttranslationally modified amino acids. Therefore, both modifications cannot be incorporated directly into a protein using conventional protein engineering techniques. Bioorthogonal approaches were required to prepare proteins with site-specific incorporation of these noncoding amino acids. Thus motivated, we prepared pS-incorporated bone morphogenetic protein-4 (BMP-4) and DOPA-incorporated insulin-like growth factor-1 (IGF-1) using enzymatic approaches as shown in Figure 3.

17

Figure 2. Combination of growth factors with the active sites of underwater adhesion proteins, such as the Mfp and salivary statherin (2–4). (A) Structural models of Mfp-3-1b and Mfp-5. The DOPA (posttranslationally modified tyrosine [Y]) and lysine (K) residue content in these proteins is 20–30% and 10–20%, respectively. (B) A structural model of IGF-1 conjugated with the adhesive sequence of XKXKX (where X represents DOPA) by a bioorthogonal technique (2). (C) A structural model of human salivary statherin (UniProtKB ID: P02808, 62 amino acids). Three phosphoserine (pS) residues are depicted. (D) Structural model of human BMP-4 conjugated with pSs. Two pS residues were conjugated to the C-terminus of hBMP-4 by the sortase reaction (4). All models were generated by using the I-TASSER server (19).

Enzymatic Modification: Adhesive IGF-1 Incorporation of DOPA into IGF was achieved by tyrosinase, an enzyme that converts tyrosine to DOPA (Figure 3) (3). In the procedure, a tyrosine-lysine-tyrosine-lysine-tyrosine sequence was genetically appended to the C terminus of IGF-1. The resultant protein is referred to as IGF-Y. Subsequently, IGF-Y was treated with tyrosinase, and the tyrosine residues were converted to DOPA residues. The final product is referred to as IGF-X. The binding affinities of the different IGF-1 derivatives towards titanium (Ti) were investigated using a quartz crystal microbalance (QCM) with dissipation monitoring, as shown in Figure 4. The binding affinity of IGF-X was significantly higher than that of IGF-Y at pH 8.5. Furthermore, the bound IGF-X did not dissociate even after it had been washed with phosphate buffered saline. 18

A cell growth assay using NIH 3T3 cells in the presence of soluble IGF-Y and bound IGF-X were performed on Ti. The bound IGF-X produced a significant enhancement of cell growth when compared with soluble IGF.

Figure 3. Design of adhesive BMP-4 and IGF-1 synthesized using a combination of recombinant DNA technology and enzyme treatments.

Figure 4. Bound amounts of IGF-Y (non–tyrosinase-treated IGF-1 derivative) and IGF-X (tyrosinase-treated IGF-1 derivative) on titanium at pH 8.5 as assessed using a QCM. These data represent the mean values of three independent experiments ± standard deviation (SD). ** p < 0.01. Reproduced with permission from ref. (3). Copyright 2016. John Wiley & Sons. 19

Enzymatic Ligation: Adhesive BMP-4 Tyrosinase can effectively oxidize tyrosine to DOPA; however, if a tyrosine residue is located in the active site of a protein, then tyrosinase converts this tyrosine to DOPA, which may lead to loss of biological activity. In such a case, the tyrosinase method is not suitable for modification of particular tyrosine residues in an adhesive domain. Therefore, the ligation method is desired in some cases to specifically incorporate a noncoding amino acid. We used sortase A to ligate human BMP (hBMP) with a pS-containing peptide (4). Figure 3 shows the preparation of the protein derivative. The growth factor portion and the phosphorylated peptide portion were separately prepared using gene engineering and a solid-phase method, respectively, and then the products were ligated using sortase A. Figure 5 shows that the prepared phosphorylated peptide-carrying human BMP-4 (hBMP4-pSpS) had higher binding affinity to HA than the nonphosphorylated construct (hBMP4-SS). In addition, phosphorylated BMP-4 significantly induced bone formation activity on HA. Taking these results into consideration, the technique to engineer bioorthogonal growth factors with pS is expected to contribute to the preparation of bioactive materials for bone regeneration.

Figure 5. Evaluation of phosphorylated hBMP4. (A) Binding of engineered hBMP4 derivatives to HA (hydroxyapatite). hBMP4-SS (nonphosphorylated serine-carrying hBMP4 derivative) or hBMP4-pSpS (phosphorylated serine-carrying BMP4 derivative) was incubated with HA beads. The amount of bound proteins was measured using an anti-BMP4 antibody. Mean (n = 3) ± SD is plotted. * p < 0.01 versus hBMP4-SS. (B) Biological activity was detected using reverse transcription polymerase chain reaction analysis of osteogenic marker expression. C3H10T1/2 cells were cultured on HA beads treated with or without recombinant hBMP4 derivatives. Reproduced with permission from ref. (4). Copyright 2011. Springer Nature. 20

Genetic Incorporation: Genetically PEGylated Protein Synthetic polymer–protein hybrids were developed as therapeutic proteins or bioreactor enzymes. Because PEG is nontoxic, nonimmunogenic, and highly soluble in water and organic solvents, many PEG-conjugated (PEGylated) proteins have been developed and shown to be very beneficial (7). For specific chemical modification, a genetic encoding approach was reported in 1989 as a method to site-specifically incorporate noncanonical amino acids into peptides or proteins (20, 21). The method utilizes the UAG codon (i.e., the amber nonsense codon), which normally directs the termination of protein synthesis, to instead encode a noncanonical amino acid that is loaded onto the cognate tRNA. Some researchers insert a noncanonical and reactive amino acid at a site of a protein using the misacylated tRNA method and posttranslationally link a PEG molecule to reactive amino acids using click chemistry (22). We reported direct incorporation of a longer PEG using PEG-acylated tRNA and named this method genetic PEGylation (12). Shozen et al. also incorporated a PEG chain into a polypeptide using a misacylated tRNA that recognizes a four-base anticodon (23). The advantage of using genetic PEGylation is that the addition of PEG occurs at predetermined, specific locations in the peptide. The placement of PEG might provide the best improvement in proteolytic stability and biological activity (24). Use of genetic PEGylation theoretically allows the precise insertion of even two or more PEG chains, of differing lengths, into specific positions in individual proteins. Thus motivated, we attempted to site-specially incorporate one and two PEGs by adding tRNAs carrying PEGs of various lengths. By changing the anticodon on the tRNAs, we incorporated PEGs at the amber codon or a four-base codon, as shown in Figure 6.

Figure 6. Incorporation of different-sized PEGs into a polypeptide by in vitro translation. Reproduced from ref. (7). Copyright 2015. Royal Society of Chemistry. 21

Figure 7. Yield efficiency (logarithms of relative amount of PEGylation product to an internal standard peptide, 3×FLAG) of varying lengths of PEGylation (PEG4, PEG8, PEG12, and PEG24) using the amber (UAG) or frameshift (CGGG) codon. fM in the peptide sequence indicates a formylmethionine. The two magenta bars represent that three- and twofold concentrations of tRNA were used in the translation system for PEG12-amber and PEG8-CGGG, respectively. Reproduced from ref. (7). Copyright 2015. Royal Society of Chemistry.

Figure 8. Incorporation of different-sized PEGs into a peptide using two different codons. PEG4 and PEG8 were incorporated using amber (UAG) or frameshift (CGGG) codons. (N) and (C) indicate the N and C termini of the peptide, respectively. Reproduced from ref. (7). Copyright 2015. Royal Society of Chemistry. 22

We incorporated a series of PEGs (PEG4, PEG8, PEG12, and PEG24) into a peptide using genetic PEGylation and found that the incorporation of the longer PEG chains reduced the yield of the translation product (Figure 7). The PEGlength dependency of the translation products of PEG8 to PEG24 may be explained by the previously reported steric hindrance between PEG and the ribosome and between PEG and EF-Tu (25). We then site-specifically incorporated two PEGs into a peptide via the two codons (amber and four-base codons) and observed a position-dependent size preference of PEGs used in the peptide synthesis by the ribosome (Figure 8). The site-specific incorporation of one or two different-sized PEGs into one peptide was demonstrated, although the PEGylation efficiency depended on the length of the PEG, the codon, and the incorporation sites. This method may therefore have applications in the precise synthesis of bioconjugate drugs.

Conclusion We have provided a summary of new bioorthogonal approaches for site-specific incorporation of noncanonical amino acids, which play a key role in imparting nonnatural function to proteins. In the future, a combination of these approaches will provide new design tools for protein engineering and biomaterials.

Acknowledgments This work was partially supported by the Japan Society for the Promotion of Science KAKENHI (grant 15H01810 and 22220009). We thank the Edanz Group (www.edanzediting.com/ac) for editing a draft of this manuscript.

References 1. 2.

3. 4. 5.

6.

7.

Ito, Y. ChemBioChem 2012, 13, 1100–1102. Zhang, C.; Miyatake, H.; Ito, Y. In Advances in Bioinspired and Biomedical Materials; Ito, Y., Chen, X., Kang, I.-K., Eds.; American Chemical Society: Washington, DC, 2017; Vol. 1 pp 83–91. Zhang, C.; Miyatake, H.; Wang, Y.; Inaba, T.; Wang, Y.; Zhang, P.; Ito, Y. Angew. Chem., Int. Ed. 2016, 55, 11447–11451. Sakuragi, M.; Kitajima, T.; Nagamune, T.; Ito, Y. Biotechnol. Lett. 2011, 33, 1885–1890. Tada, S.; Abe, H.; Ito, Y. In Green Polymer Chemistry: Biocatalysis and Materials II; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; American Chemical Society: Washington, DC, 2013; pp 223–233. Tada, S.; Uzawa, T.; Ito, Y. In Green Polymer Chemistry III: Biobased Materials and Biocatalysis; Cheng, H. N., Gross, R. A., Smith, P. B., Eds.; American Chemical Society: Washington, DC, 2015; pp 169–180. Zang, Q.; Tada, S.; Uzawa, T.; Kiga, D.; Yamamura, M.; Ito, Y. Chem. Commun. 2015, 51, 14385–14388. 23

8.

9. 10. 11.

12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

25.

K. C., T. B.; Tada, S.; Zhu, L.; Uzawa, T.; Minagawa, N.; Luo, S.-C.; Zhao, H.; Yu, H.-h.; Aigaki, T.; Ito, Y. Chem. Commun. 2018, 54, 5201–5204. Lee, H.; Dellatore, S. M.; Miller, W. M.; Messersmith, P. B. Science 2007, 318, 426–430. Raj, P. A.; Johnsson, M.; Levine, M. J.; Nancollas, G. H. J. Biol. Chem. 1992, 267, 5968–5976. Ito, Y. In Biological Interactions on Materials Surfaces: Understanding and Controlling Protein, Cell and Tissue Responses; Puleo, D. A., Bizios, R., Eds.; Springer: New York, 2009; pp 173–197. Tada, S.; Andou, T.; Suzuki, T.; Dohmae, N.; Kobatake, E.; Ito, Y. PLoS One 2012, 8, e49235. Ito, Y. Soft Matter 2008, 4, 46–56. Park, S.-H.; Uzawa, T.; Hattori, F.; Ogino, S.; Morimoto, N.; Tsuneda, S.; Ito, Y. Biomaterials 2018, 161, 270–278. Zhou, D.; Ito, Y. RSC Adv. 2013, 3, 11095–11106. Hennebert, E.; Maldonado, B.; Ladurner, P.; Flammang, P.; Santos, R. Interface Focus 2015, 5, 20140064. Wilker, J. J. Science 2015, 349, 582–583. Messersmith, P. B. Science 2008, 319, 1767–1768. Yang, J.; Zhang, Y. Nucleic Acids Res. 2015, 43, W174–W181. Noren, C. J.; AnthonyCahill, S. J.; Griffith, M. C.; Schultz, P. G. Science 1989, 244, 182–188. Bain, J. D.; Glabe, C. G.; Dix, T. A.; Chamberlin, A. R.; Diala, E. S. J. Am. Chem. Soc. 1989, 111, 8013–8014. Deiters, A.; Cropp, T. A.; Summerer, D.; Mukherji, M.; Schultz, P. G. Bioorg. Med. Chem. Lett. 2004, 14, 5743–5745. Shozen, N.; Iijima, I.; Hohsaka, T. Bioorg. Med. Chem. Lett. 2009, 19, 4909–4911. Lawrence, P. B.; Gavrilov, Y.; Matthews, S. S.; Langlois, M. I.; ShentalBechor, D.; Greenblatt, H. M.; Pandey, B. K.; Smith, M. S.; Paxman, R.; Torgerson, C. D.; Merrell, J. P.; Ritz, C. C.; Prigozhin, M. B.; Levy, Y.; Price, J. L. J. Am. Chem. Soc. 2014, 136, 17547–17560. Mittelstaet, J.; Konevega, A. L.; Rodnina, M. V. J. Am. Chem. Soc. 2013, 135, 17031–17038.

24

Chapter 3

Engineering the Microbial Cell Membrane To Improve Bioproduction Laura R. Jarboe,*,1 Jeffery B. Klauda,2 Yingxi Chen,1 Kirsten M. Davis,1 and Miguel C. Santoscoy1 1Department

of Chemical and Biological Engineering, Iowa State University, 2114 Sweeney Hall, 618 Bissell Road, Ames, Iowa 50011, United States 2Department of Chemical and Biomolecular Engineering, University of Maryland, 4418 Stadium Drive, College Park, Maryland 20742, United States *E-mail: [email protected].

Inhibition of the microbial biocatalyst often limits the attainment of the yields, titers, and rates needed for economic viability production of biorenewable fuels and chemicals. In many cases, this toxicity can be attributed to damage of the lipid-rich microbial membrane. Just as the composition of a reaction vessel can be altered to improve its resistance to corrosion, the composition of the microbial cell membrane can be altered in order to decrease its vulnerability to this damage. Contrastingly, in some cases the membrane can be weakened in order to increase the space available for intracellular accumulation of a product, or the overall abundance of the membrane can be increased in order to serve as a sink for a membrane-associated product molecule. This chapter reviews efforts to engineer the microbial cell membrane, with a focus on engineering strategies that improve bioproduction of fuels and chemicals.

Introduction Microbial Inhibition Often Limits Bioproduction Inhibition of the microbial biocatalyst, either by the product or by compounds in the biomass-derived sugar stream, often limits the attainment of economic viability for bioproduction processes (1–6). This toxicity is often attributed © 2018 American Chemical Society

to damage of the cell membrane (7–18). For the macro scale, if a reaction vessel is corroded by its contents, a typical strategy would be to change the vessel composition. The feasibility of membrane engineering for improving fermentative production of biochemicals has been established and is reviewed here. This membrane engineering is broken into broad membrane components of the lipid tails, the lipid heads, and the membrane-associated proteins and sugars. A variety of attractive biobased products, such as butanol, fatty acids, and styrene, are toxic to the microbial biocatalyst (19–21). Hydrolysis and pyrolysis are attractive means of depolymerizing lignocellulosic biomass into fermentable monomers, but the utilization of these sugars is hindered by the presence of compounds that are inhibitory to the biocatalyst (3, 22–24). This feedstock-associated inhibition means that the growth and metabolic activity of the biocatalyst are also limited. Thus, both ends of biocatalyst metabolism are impacted by inhibition: the utilization of cheap, biomass-derived monomers is limited by inhibitory compounds in the feedstock, while commercially viable production of biorenewable fuels and chemicals is limited by product toxicity. Multiple studies have demonstrated that improving microbial tolerance of such inhibitors can improve bioproduction (4–6). However, the development of resistant organisms often relies on evolutionary or library-based methods (10, 25–31). The use of evolutionary or library-based methods for increasing tolerance is most effective when production of an inhibitory compound is the only means for the microbial biocatalyst to maintain redox balance and/or adenosine triphosphate (ATP) production, and thus growth serves as a selection marker for production and tolerance (25, 32–34). However, evolving a strain for increased tolerance of an exogenously-provided inhibitor with the goal of improving production of this compound has been met with mixed results. Thus, as we strive to improve production of inhibitory compounds that are not coupled with microbial growth, methods of strain improvement that do not rely on evolution or selection-based methods are needed. Inhibition Is Often Due To Membrane Damage The microbial cell membrane plays a vital role in many cellular processes. Membrane damage is often evident as decreased integrity and/or perturbed fluidity. Decreased integrity of the membrane can be measured via the release of metabolites, such as Mg2+ (7, 8) or lipoproteins (35) from the cell, or as the entry of membrane-impermeable molecules, such as the SYTOX nucleic acid dye, into the cell (9, 36). Integrity of the inner and outer membranes can also be quantified individually, as well as visualization of membrane deformation in the presence of membrane-damaging molecules, such as butanol (35). Fluidity is measured via the membrane polarization, which can be quantified with an amphiphilic fluorescent probe such as 1,6-diphenyl-1,3,5-hexatriene (DPH) (8, 37, 38). An increase in membrane polarization indicates increased membrane rigidity and decreased polarization indicates increased membrane fluidity (39). Too large of a change in either direction can be problematic. These types of membrane damage have been noted during characterization of bioproduction (8, 36, 40, 41). For example, membrane fluidity was measured 26

during fatty acid challenge and production (8, 37) and membrane integrity was measured during the challenge with a variety of organic acids (16) and alcohols (17). Finally, a titer-dependent decrease in membrane integrity was observed during styrene production (9). Membrane Engineering Can Combat Membrane Damage If the microbial biocatalyst is envisioned as a miniature reaction vessel, then the cell membrane corresponds to the walls of this reaction vessel. Corrosion of a reaction vessel by its contents is problematic and typically, an alternative vessel whose walls are resistant to this corrosion would be sought. The conclusion of this analogy is that the composition of the microbial cell membrane should be altered in order to increase resistance to membrane damage. The alterations required for increasing resistance are expected to vary according to the chemical properties of the inhibitor; they are not an expectation of a single strategy for addressing all membrane-damaging inhibitors. An extreme version of this approach is to provide cells with an exogenous membrane via encapsulation in a polymeric shell (42). This chapter focuses on changes to the cell membrane in order to improve the structural properties related to bioproduction. We consider three distinct aspects of the membrane structure (Figure 1): the membrane-associated lipids, including but not limited to the lipid tails; the head of the lipid molecule; and the proteins and sugars within and associated with the membrane. The discussion focuses on engineering strategies that have been demonstrated to improve bioproduction, with key examples summarized in Table 1.

Figure 1. Membrane engineering strategies can target the lipid tails, the lipid head groups, and the membrane-associated proteins and sugars. 27

Table 1. Examples of genetically implemented membrane modification strategies that improved bioproduction Organism

Engineering Strategy

Effect

E. coli

Decreased incorporation of medium chain fatty acids into the membrane via deletion of aas (40)

Increased titer of total free fatty acids in rich media

E. coli

Enabled production of novel lipid tails via tuned recombinant expression of cis/trans isomerase (Cti) (37)

Increased titer of octanoic acid in minimal media, increased titer of styrene

E. coli

Increased relative abundance of PE phospholipid head, decreased relative abundance of PG, via increased expression of pssA (43)

Increased titer of total free fatty acids in minimal media

E. coli

Increased expression of long-chain fatty acid importer FadL to enable recovery of fatty acids for membrane assembly (44)

Increased titer of C14 fatty acids in minimal media

E. coli (45), H. campaniensis (46)

Tuned expression of actin-like cytoskeleton protein MreB to weaken the cytoskeleton and enable enlarged bacterial size

Increased PHB titer and yield

E. coli

Decreased expression of cell wall synthesis genes, such as murE and murD, to weaken the cell wall (47)

Increased conversion efficiency of glucose to PHB

E. coli

Increased expression of membrane-bending proteins, such as MtlA and ALmgs, and membrane synthesis proteins, such as PlsB and PlsC (48)

Increased titer and specific production (mg/g) of beta-carotene in rich media

B. subtilis

Increased abundance of CL via expression tuning of pgsA and clsA and altered distribution of CL by expression tuning of ftsZ (49)

Increased titer of hyaluronic acid in rich media

Lipid Engineering The lipid tails in E. coli consist of straight-chain cis-mono-unsaturated fatty acids (C16:1 and C18:1) (Figure 2), straight-chain saturated fatty acids (C12:0, C14:0, C16:0, and C18:0), and the cyclopropane fatty acids C17cyc and C19cyc. Two commonly used metrics for summarizing the lipid tail distribution are the relative abundance of unsaturated fatty acids and the average lipid length. The relationship between membrane lipid composition and tolerance of membranedamaging compounds, such as ethanol, has been known for more than 40 years (50). However, efforts to engineer the lipid distribution are more recent. 28

Figure 2. Structures of select molecules related to membrane engineering. (1) octanoic acid (C8), (2) styrene, (3) n-butanol, (4) beta-carotene, (5) ergosterol, (6) ubiquinone, (7) oleic acid, C18:1, cis, (8) elaidic acid, C18:1, trans, (9) linoleic acid, (10) phosphatidylethanoloamine, (11) diacylglycerol, (12) ceramide, (13) DGGGP, (14) COE1-5C.

Altering the relative distribution of saturated and unsaturated lipid tails in the E. coli membrane has been shown to increase tolerance toward ethanol (51), fatty acids (36), and hexane (52). This strategy was effective in decreasing the membrane leakage caused by fatty acids, but fatty acid production was not increased (36). Tuning of the relative abundance of saturated and unsaturated lipid tails has also been demonstrated in S. cerevisiae (38), ranging from 20% of acyl chains being unsaturated to 80% being unsaturated. The increase in the relative abundance of unsaturated lipids was demonstrated to increase the membrane fluidity (38), similar to the familiar tendency of cooking oils to trend to higher fluidity as the degree of unsaturation increases. 29

Studies of fatty acid production have concluded that the incorporation of shorter fatty acids as phospholipid tails is detrimental to membrane function. Elimination of the pathway responsible for this incorporation improved fatty acid tolerance and production in rich media (40). S. cerevisiae cells challenged with exogenous octanoic acid (C8) (Figure 2) were observed to show a dose-dependent increase in the average length of membrane lipids (7). Media supplementation with oleic acid (C18:1) (Figure 2) was shown to increase growth in the presence of C8 and to decrease the associated loss of membrane integrity, though attempts to engineer S. cerevisiae for sufficiently increased membrane oleic acid content without this exogenous supplement were not successful (7). A strain of S. cerevisiae expressing a mutant form of the Acc1 acetyl-CoA carboxylase was found to have altered membrane lipid composition, both in the form of increased average length and an increase in saturated lipid tails (53). This strain was found to have increased tolerance of exogenous hexanoic acid, C8, 2-propanol, and n-butanol, as well as increased membrane integrity during C8 challenge (53). Other genetic manipulation targets for altering the degree of lipid unsaturation and average lipid length have been demonstrated in S. cerevisiae (54). Cyclopropane fatty acids are often implicated in characterization of inhibited or evolved E. coli strains (55–58). A variety of studies have described engineering of the cyclopropane fatty acid content, including fatty acid tolerance in E. coli (59) and enabling the production of cyclopropane fatty acids in yeast (7). Strains deficient in cyclopropane fatty acid production have increased sensitivity to stressors such as heat, pressure, and inorganic acid stress (60, 61), but increasing the production of cyclopropane fatty acids usually does not improve resistance to these stressors (7, 8). It has been demonstrated that enabling E. coli to isomerize some of the native cis unsaturated fatty acid (CUFA) lipid tails to trans unsaturated fatty acids (TUFA) (Figure 2) increased tolerance to several membrane-damaging compounds and conditions (62). TUFA production was enabled by expression of the Pseudomonas putida cis/trans isomerase enzyme (Cti). Strains with a TUFA/CUFA ratio of approximately 0.085 had the largest increase in specific growth rate in the presence of octanoic acid and also showed an approximate 40% increase in fatty acid titers in minimal media. This membrane engineering strategy was found to primarily impact membrane fluidity. Specifically, the membrane polarization increased, indicating an increase in membrane rigidity, which presumably mitigates the fluidizing effect of the fatty acids. Further characterization showed that this strain also had increased specific growth rate relative to the non-TUFA producing control when challenged with butanol, styrene, and 42 °C, as well as a significant increase in styrene production titers. Some organisms also produce sphingolipids, which contain an amine group and an alcohol group (63). Ceramides are a subset of sphingolipids (Figure 2). Strains of Zygosaccharomyces bailii that were treated with myriocin, known to decrease the production of sphingolipids, showed increased sensitivity to acetic acid, formic acid, and lactic acid (64), and ceramides were found to have a dramatic effect on the properties of in silico membranes (65). Genetic engineering strategies for reducing the relative abundance of sphingolipids in S. cereivisae have been demonstrated, with a negative impact on cell viability (54). 30

In addition to the fatty acid tails of the phospholipid molecules, microbial membranes may also contain other types of lipid molecules. E. coli was observed to have a drastic increase in abundance of ubiquinone (Figure 2) in the cell membrane during osmotic challenge, and in vitro characterization demonstrated that increased ubiquinone content was associated with increased liposome stability (66). Ubiquionone is a 1,4-benzoquinone with a side chain of isoprenoid subunits and thus does not fit the standard conception of a lipid, but it is membrane-associated hydrocarbon. Sterols are steroids with a hydroxyl group at position three of the A ring, one of the four rings that define the steroid group. Strains of S. cerevisiae evolved for thermotolerance or butanol tolerance (67) had altered distribution and/or abundance of certain sterols, such as ergosterol (Figure 2). In addition to the genetic modification strategies described above, supplementation with exogenous lipid-type molecules have provided further proof of concept for engineering of the membrane at the lipid composition level. Media supplementation with the polyunsaturated omega-6 18-carbon fatty acid linoleic acid increased the fluidity of S. cerevisiae during beta-carotene production, resulting in an increase in beta-carotene titers (68). Provision of exogenous COE1-5C (Figure 2), an oligo-polyphenylene-vinylene conjugated oligoelectrolyte, increased the specific growth rate of E. coli during challenge with relatively high (3.5 vol%) concentrations of butanol (69). Supplementation of E. coli with cedar wood oil significantly increased the membrane fluidity and increased production of menaquinone (18). Lipid Head Group Engineering Lipid tails make up the bulk of the membrane interior and are attached to head groups. E. coli natively produces three different types of phospholipid head molecules: phosphatidylethanoloamine (PE) (Figure 2), phosphatidylglycerol (PG), and cardiolipin (CL). Each of these three molecules has different chemical properties. Perturbing the relative abundance of these head groups in E. coli by altering the expression of the genes encoding their biosynthesis pathways can substantially change the membrane composition and properties. For example, a strain engineered for increased PE content showed a substantial increase in octanoic acid tolerance and fatty acid production (43). Further characterization of this strain showed increased resistance to the membrane permeabilization and intracellular acidification caused by exogenously provided octanoic acid and a greater than 20% change in membrane surface potential (43). Finally, this engineered strain showed increased tolerance of other inhibitors relevant to cost-effective bioproduction, such as the biomass-derived inhibitors furfural and acetic acid. Studies performed in B. subtilis have shown that altering the relative abundance of the native phospholipid head groups increased secretion of alpha-amylase (70), increased osmotic tolerance (71), and increased sensitivity to certain antimicrobial compounds (72). Alteration of the native phospholipid head distribution in S. cerevisiae increased growth and decreased the relative abundance of peroxidized lipids during growth in the presence of lactic acid at 31

low pH (73). E. coli was engineered for production of nonnative glycolipid head group diacylglycerol (DAG) (Figure 2) in the absence of production of the PE phospholipid head (74). The engineered strains were shown to have altered cell length and altered resistance to some membrane-damaging antibiotics (74). The membranes of Archaea differ from Bacteria and Eukarya in a variety of ways. One of these differences is the connection of the lipid tail to the glycerol-phosphate backbone with an ether linkage, as opposed to the ester linkage used by bacteria and eukaryotes. Additionally, the lipid tails of the archaeal membrane consist of isoprenoids instead of fatty acids, where the isoprenoids contain multiple unsaturated bonds and methyl groups. E. coli was engineered to express an isoprenoid ether lipid biosynthesis pathway, resulting in production of digeranylgeranylglyceryl phosphate (DGGGP) (Figure 2) (75). More recently, E. coli was engineered to produce the ether phospholipids archaetidylethanolamine (AE) and archaetidylglycerol (AG) (76, 77). E. coli strains producing the AG-containing membrane were demonstrated to have increased cell length, heat tolerance, and tolerance to freeze/thaw (76). Membrane Proteins and Lipopolysaccharide Engineering In addition to lipid molecules, the microbial cell membrane is rich in proteins and a variety of sugar-type molecules. These are also important aspects of the membrane properties and therefore, the membrane function. Proteins involved in membrane assembly, repair, and structural properties have also proven to be a very promising method for altering membrane properties and improving bioproduction. It has been demonstrated multiple times that enabling export of inhibitory products can improve production of those compounds (78, 79). Tolerance of inhibitory compounds can also be increased by preventing their entry into the cell. For example, deletion of the OmpF porin increased octanoic acid tolerance (80). This strategy represents an increase in tolerance by decreasing the intracellular concentration of the inhibitor and would not fall under what we classify here as “membrane engineering.” However, identification and engineering of product transporters becomes increasingly important as membrane robustness is increased. Products that previously exited the cell by passing through the membrane can possibly accumulate to toxic levels in strains subjected to improvement by membrane engineering. More simply, the desire to produce “stronger walls” leads to a need for “bigger doors.” In contrast to the standard example of increasing exporter abundance in order to increase production, increasing the abundance of the long-chain fatty acid importer FadL was shown to increase fatty acid production (44). This increased expression of FadL not only increased fatty acid titers, but also increased the membrane lipid mass per dry cell weight by more than 10% (44). Presumably, this increased expression of FadL allowed cells to recoup some of the long-chain fatty acids needed for membrane assembly that were being lost via the fatty acid efflux system. Unlike many of the bioproduced bulk fuels and chemicals, which are produced internally and then excreted from the microbial cell factory, some bioproducts accumulate inside the cell as inclusion bodies or accumulate in the membrane. 32

In these cases, membrane engineering can be used to increase the storage capacity for the product and thereby improve bioproduction. Using polyhydroxybutyrate (PHB) as a model inclusion body, it was demonstrated that tuning of the expression of the actin-like protein MreB was able to increase PHB production in both E. coli (45) (Figure 3) and Halomonas campaniensis (47). In E. coli, mreB expression was placed under control of SulA (45), where SulA inhibits cell division. The altered expression of MreB weakened the cystoskeleton, resulting in larger cells with increased sphericity (45). In H. campaniensis, mreB expression occurred from a temperature-sensitive plasmid (47). MreB has also been described as a frequent location of point mutations when E. coli was evolved for osmotolerance (81).

Figure 3. TEM morphology of PHB-producing E. coli with expression tuning of mreB to increase the available cell volume for inclusion bodies accumulation. Left, control strain. Right, engineered strain. Scale bar, 5 μm. Reproduced with permission from (45). Copyright 2015 Elsevier.

The idea of weakening the cell wall to increase the available space for inclusion bodies was expanded with additional genetic targets in E. coli (46). Increased expression of certain cell wall synthesis genes increased the Young’s Modulus and thickness of the cell membrane, resulting in a decrease in PHB content per dry mass. Decreased expression of these genes increased PHB abundance. Finally, work in H. campaniesis (47) showed that tuned expression of cell division protein FtsZ increased cell volume by increasing cell length, increasing PHB production. The work described above aimed to increase the volume available for inclusion bodies. Some metabolic products partition into the cell membrane itself, and increasing the abundance of the membrane could increase the production of these compounds. This approach was used to increase production of beta-carotene by E. coli (48). Specifically, increased expression of membrane-bending proteins and the membrane synthesis pathway resulted in an increase in both the bulk beta-carotene titer and the amount of beta-carotene produced per biomass (48). The sugars and proteins produced on or near the cell surface also contribute to the physical properties and performance of the organism (82). Strains of E. 33

coli containing a single amino acid mutation in the carbon storage regulator CsrA showed increased membrane integrity in the presence of biomass-derived pyrolytic sugars (83). This change in membrane integrity was associated with a twofold decrease in the abundance of extracellular proteins, but no change in the abundance of extracellular polysaccharides (83). The surface display of proteins with various physical properties, such as hydrophobicity, altered the tolerance of S. cerevisiae to various inhibitors, such as nonane (84). Perturbation of the core oligosaccharide lipopolysaccharide biosynthesis pathway in E. coli was demonstrated to alter the cell surface hydrophobicity, outer membrane permeability, and sensitivity to various antibiotics (85).

In Vitro and in Silico Characterization The work described thus far demonstrates the potential for membrane engineering to improve bioproduction and the range of membrane composition and properties that can currently be achieved. However, characterization of membranes in vitro and in silico can provide further insight for engineering strategies, particularly in terms of membrane compositions that are difficult to attain in vivo. As described above, exogenous supplementation with the oligopolyphenylene-vinylene conjugated oligoelectrolyte COE1-5C increased the growth of E. coli during challenge with butanol (69). COE1-5C was selected as an exogenous membrane insertion molecule based on molecular dynamic simulations of lipid bilayers in the presence of butanol (69). As also described above, altering the abundance of the native phospholipid head groups in E. coli increased tolerance and production of fatty acids (43). However, interpretation of the changes in the membrane composition was complicated by the fact that in addition to changes in the head group distribution, the lipid tails also had altered distribution. Molecular dynamic simulations allowed separation of these changes, leading to the conclusion that the increased PE content was responsible for increased membrane thickness, but that the changes in headgroup and lipid distribution both contributed to the increased hydrophobic core thickness (43). Molecular dynamics simulations have also led to increased insight in terms of the effect of membrane composition on interactions with various membrane-damaging compounds (64, 86–89). Experimental characterization of in vitro systems also provides insight into membrane engineering strategies. Disruption of supported bilayers of varying composition can be observed over time (90). This type of finely detailed experimental characterization of supported bilayers can lead to increased discrimination of the mode of action of different membrane-damaging compounds (91). Advances are also being made in the assembly of lipid vesicles for experimental characterization (92, 93).

34

Concluding Remarks The work reviewed here demonstrates the potential of membrane engineering to improve bioproduction. This engineering can have a variety of intentions, such as to combat the damage imposed by a specific molecule, to weaken the membrane to increase the available space for intracellular species, or to increase the amount of available membrane for product accumulation. The design, build, test, learn cycle (26) for membrane engineering can be further strengthened by characterization of evolved strains and proteins that currently have unknown function (94). Continued efforts to combine experimental in vitro systems, whole-cell systems, and in silico modeling are expected to lead to further gains in this area.

Acknowledgments This work was supported in part by the NSF Center for Biorenewable Chemicals Engineering Research Center (EEC-0831570), Energy for Sustainability program (CBET-1604576), and the USDA National Institute of Food and Agriculture program (2017-67021-26137). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. We thank the editors, H.N. Cheng, Richard A. Gross, and Patrick B. Smith, for the invitation to contribute to this work.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

Jarboe, L. R.; Liu, P.; Royce, L. A. Curr. Opin. Chem. Eng. 2011, 1, 38–42. Julsing, M. K.; Kuhn, D.; Schmid, A.; Buehler, B. Biotechnol. Bioeng. 2012, 109, 1109–1119. Mills, T. Y.; Sandoval, N. R.; Gill, R. T. Biotechnol. Biofuels 2009, 2article 26. Jia, H.; Yanshuang, F.; Feng, X.; Li, C. Front. Bioengin. Biotechnol. 2014, 2 article 44. Mukhopadhyay, A. Trends Microbiol. 2015, 23, 498–508. Wang, S. Z.; Sun, X. X.; Yuan, Q. P. Bioresour. Technol. 2018, 258, 302–309. Liu, P.; Chernyshov, A.; Najdi, T.; Fu, Y.; Dickerson, J.; Sandmeyer, S.; Jarboe, L. Appl. Microbiol. Biotechnol. 2013, 97, 3239–3251. Royce, L. A.; Liu, P.; Stebbins, M. J.; Hanson, B. C.; Jarboe, L. R. Appl. Microbiol. Biotechnol. 2013, 97, 8317–8327. Lian, J.; McKenna, R.; Rover, M. R.; Nielsen, D. R.; Wen, Z.; Jarboe, L. R. J. Ind. Microbiol. Biotechnol. 2016, 43, 595–604. Atsumi, S.; Wu, T.-Y.; Machado, I. M. P.; Huang, W. C.; Chen, P.-Y.; Pellegrini, M.; Liao, J. C. Mol. Syst. Biol. 2010, 6article 449. Bui, L. M.; Lee, J. Y.; Geraldi, A.; Rahman, Z.; Lee, J. H.; Kim, S. C. J. Biotechnol. 2015, 204, 33–44. Reyes, L. H.; Abdelaal, A. S.; Kao, K. C. Appl. Environ. Microbiol. 2013, 79, 5313–5320. Reyes, L. H.; Almario, M. P.; Winkler, J.; Orozco, M. M.; Kao, K. C. Metab. Eng. 2012, 14, 579–590. 35

14. Romano, K. A.; Frasciello, M. P.; Sudano, A. M.; Robinson, P.; Sueck, S. E.; Wiltsey, C. T.; Creamer, S. E.; Tawes, B. R.; Caputo, G. A.; Hecht, G. B. Abstr. Gen. Meet. Am. Soc. Microbiol. 2011, 111, 2765–2765. 15. Si, H.-M.; Zhang, F.; Wu, A.-N.; Han, R.-Z.; Xu, G.-C.; Ni, Y. Biotechnol. Biofuels 2016, 9 article 114. 16. Zaldivar, J.; Ingram, L. O. Biotechnol. Bioeng. 1999, 66, 203–210. 17. Zaldivar, J.; Martinez, A.; Ingram, L. O. Biotechnol. Bioeng. 2000, 68, 524–530. 18. Liu, Y.; Ding, X. M.; Xue, Z. L.; Hu, L. X.; Zhang, N. J.; Wang, Z.; Yang, J. W.; Cheng, Q.; Chen, M. H.; Zhang, Z. Z.; Zheng, Z. M. World J. Microbiol. Biotechnol. 2017, 33article 52. 19. McKenna, R.; Nielsen, D. Metab. Eng. 2011, 13, 544–554. 20. Ezeji, T.; Milne, C.; Price, N. D.; Blaschek, H. P. Appl. Microbiol. Biotechnol. 2010, 85, 1697–1712. 21. Brynildsen, M. P.; Liao, J. C. Mol. Syst. Biol. 2009, 5article 277. 22. Liu, Z. L. Appl. Microbiol. Biotechnol. 2011, 90, 809–825. 23. Lian, J. N.; Chen, S. L.; Zhou, S. A.; Wang, Z. H.; O’Fallon, J.; Li, C. Z.; Garcia-Perez, M. Bioresour. Technol. 2010, 101, 9688–9699. 24. Jarboe, L. R.; Wen, Z.; Choi, D. W.; Brown, R. C. Appl. Microbiol. Biotechnol. 2011, 91, 1519–1523. 25. Jin, T.; Chen, Y.; Jarboe, L. R. Evolutionary Methods for Improving the Production of Biorenewable Fuels and Chemicals. In Biotechnology for Biofuels Production and Optimization; Eckert, C., Trinh, C., Eds.; Elsevier, 2016. 26. Jin, T.; Lian, J.; Jarboe, L. R., Ethanol: A Model Biorenewable Fuel. In Industrial Biotechnology; Lee, S. Y., Stephanopoulos, G., Nielsen, J. B., Wittmann, C., Eds.; Wiley: 2017. 27. Alper, H.; Moxley, J.; Nevoigt, E.; Fink, G. R.; Stephanopoulos, G. Science 2006, 314, 1565–1568. 28. Reyes, L. H.; Almario, M. P.; Kao, K. C. PloS One 2011, 6, e17678. 29. Royce, L. A.; Yoon, J. M.; Chen, Y.; Rickenbach, E.; Shanks, J. V.; Jarboe, L. R. Metab. Eng. 2015, 29, 180–188. 30. Tomko, T. A.; Dunlop, M. J. Biotechnol. Biofuels 2015, 8 article 165. 31. Woodruff, L. B. A.; Boyle, N. R.; Gill, R. T. Metab. Eng. 2013, 17, 1–11. 32. Jarboe, L. R.; Grabar, T. B.; Yomano, L. P.; Shanmugan, K. T.; Ingram, L. O. Development of ethanologenic bacteria. In Biofuels. Advances in Biochemical Engineering/Biotechnology; Olsson L., Ed.; Springer: Berlin, Heidelberg, 2007; Vol. 108. 33. Jantama, K.; Zhang, X.; Moore, J. C.; Shanmugam, K. T.; Svoronos, S. A.; Ingram, L. O. Biotechnol. Bioeng. 2008, 101, 881–893. 34. Zhang, X.; Jantama, K.; Moore, J. C.; Shanmugam, K. T.; Ingram, L. O. Appl. Microbiol. Biotechnol. 2007, 77, 355–366. 35. Fletcher, E.; Pilizota, T.; Davies, P. R.; McVey, A.; French, C. E. Appl. Microbiol. Biotechnol. 2016, 100, 9653–9659. 36. Lennen, R. M.; Pfleger, B. F. PloS One 2013, 8. 37. Tan, Z. G.; Yoon, J. M.; Nielsen, D. R.; Shanks, J. V.; Jarboe, L. R. Metab. Eng. 2016, 35, 105–113. 36

38. Degreif, D.; de Rond, T.; Bertl, A.; Keasling, J. D.; Budin, I. Metab. Eng. 2017, 41, 46–56. 39. Mykytczuk, N. C. S.; Trevors, J. T.; Leduc, L. G.; Ferroni, G. D. Prog. Biophys. Mol. Biol. 2007, 95, 60–82. 40. Sherkhanov, S.; Korman, T. P.; Bowie, J. U. Metab. Eng. 2014, 25, 1–7. 41. Lennen, R. M.; Kruziki, M. A.; Kumar, K.; Zinkel, R. A.; Burnum, K. E.; Lipton, M. S.; Hoover, S. W.; Ranatunga, D. R.; Wittkopp, T. M.; Marner, W. D., II; Pfleger, B. F. Appl. Environ. Microbiol. 2011, 77, 8114–8128. 42. Sakkos, J. K.; Kieffer, D. P.; Mutlu, B. R.; Wackett, L. P.; Aksan, A. Biotechnol. Bioeng. 2016, 113, 513–521. 43. Tan, Z. G.; Khakbaz, P.; Chen, Y. X.; Lombardo, J.; Yoon, J. M.; Shanks, J. V.; Klauda, J. B.; Jarboe, L. R. Metab. Eng. 2017, 44, 1–12. 44. Tan, Z. G.; Black, W.; Yoon, J. M.; Shanks, J. V.; Jarboe, L. R. Microb. Cell Fact. 2017, 16 article 38. 45. Jiang, X. R.; Wang, H.; Shen, R.; Chen, G. Q. Metab. Eng. 2015, 29, 227–237. 46. Zhang, X. C.; Guo, Y.; Liu, X.; Chen, X. G.; Wu, Q.; Chen, G. Q. Metab. Eng. 2018, 45, 32–42. 47. Jiang, X. R.; Yao, Z. H.; Chen, G. Q. Metab. Eng. 2017, 44, 30–37. 48. Wu, T.; Ye, L. J.; Zhao, D. D.; Li, S. W.; Li, Q. Y.; Zhang, B. L.; Bi, C. H.; Zhang, X. L. Metab. Eng. 2017, 43, 85–91. 49. Westbrook, A. W.; Ren, X.; Moo-Young, M.; Chou, C. P. Biotechnol. Bioeng. 2018, 115, 216–231. 50. Ingram, L. O. J. Bacteriol. 1976, 125, 670–678. 51. Luo, L. H.; Seo, P.-S.; Seo, J.-W.; Heo, S.-Y.; Kim, D.-H.; Kim, C. H. Biotechnol. Lett 2009, 31, 1867–1871. 52. Oh, H. Y.; Lee, J. O.; Kim, O. B. Appl. Microbiol. Biotechnol. 2012, 96, 1619–1627. 53. Besada-Lombana, P. B.; Fernandez-Moya, R.; Fenster, J.; Da Silva, N. A. Biotechnol. Bioeng. 2017, 114, 1531–1538. 54. Lindahl, L.; Santos, A. X. S.; Olsson, H.; Olsson, L.; Bettiga, M. Sci. Rep. 2017, 7 article 41868. 55. Zu, T. N. K.; Athamneh, A. I. M.; Wallace, R. S.; Collakova, E.; Senger, R. S. J. Bacteriol. 2014, 196, 3983–3991. 56. Kanno, M.; Katayama, T.; Tamaki, H.; Mitani, Y.; Meng, X.-Y.; Hori, T.; Narihiro, T.; Morita, N.; Hoshino, T.; Yumoto, I.; Kimura, N.; Hanada, S.; Kamagata, Y. Appl. Environ. Microbiol. 2013, 79, 6998–7005. 57. Chang, Y. Y.; Cronan, J. E. Mol. Microbiol. 1999, 33, 249–259. 58. Brown, J. L.; Ross, T.; McMeekin, T. A.; Nichols, P. D. Int. J. Food Microbiol. 1997, 37, 163–173. 59. Royce, L. A.; Boggess, E.; Fu, Y.; Liu, P.; Shanks, J. V.; Dickerson, J.; Jarboe, L. R. PloS One 2014, 9, e89580. 60. Chen, Y. Y.; Gaenzle, M. G. Int. J. Food Microbiol. 2016, 222, 16–22. 61. Shabala, L.; Ross, T. Res. Microbiol. 2008, 159, 458–461. 62. Tan, Z.; Yoon, J. M.; Nielsen, D. R.; Shanks, J. V.; Jarboe, L. R. Metab. Eng. 2016, 35, 105–113. 37

63. Megyeri, M.; Riezman, H.; Schuldiner, M.; Futerman, A. H. J. Mol. Biol. 2016, 428, 4765–4775. 64. Lindahl, L.; Genheden, S.; Eriksson, L. A.; Olsson, L.; Bettiga, M. Biotechnol. Bioeng. 2016, 113, 744–753. 65. Wang, E.; Klauda, J. B. J. Phys. Chem. B 2017, 121, 10091–10104. 66. Sevin, D. C.; Sauer, U. Nat. Chem. Biol. 2014, 10, 266–272. 67. Ghiaci, P.; Norbeck, J.; Larsson, C. Biotechnol. Biofuels 2013, 6 article 101. 68. Liu, P. T.; Sun, L.; Sun, Y. X.; Shang, F.; Yan, G. L. J. Ind. Microbiol. Biotechnol. 2016, 43, 525–535. 69. Hinks, J.; Wang, Y. F.; Matysik, A.; Kraut, R.; Kjelleberg, S.; Mu, Y. G.; Bazan, G. C.; Wuertz, S.; Seviour, T. ChemSusChem 2015, 8, 3718–3726. 70. Cao, H. J.; van Heel, A. J.; Ahmed, H.; Mols, M.; Kuipers, O. P. Microb. Cell Fact. 2017, 16 article 56. 71. Lopez, C. S.; Alice, A. F.; Heras, H.; Rivas, E. A.; Sanchez-Rivas, C. Microbiol. SGM 2006, 152, 605–616. 72. Salzberg, L. I.; Helmann, J. D. J. Bacteriol. 2008, 190, 7797–7807. 73. Berterame, N. M.; Porro, D.; Ami, D.; Branduardi, P. Microb. Cell Fact. 2016, 15article 39. 74. Wikstrom, M.; Kelly, A. A.; Georgiev, A.; Eriksson, H. M.; Klement, M. R.; Bogdanov, M.; Dowhan, W.; Wieslander, A. J. Biol. Chem. 2009, 284, 954–965. 75. Lai, D.; Lluncor, B.; Schroder, I.; Gunsalus, R. P.; Liao, J. C.; Monbouquette, H. G. Metab. Eng. 2009, 11, 184–191. 76. Caforio, A.; Siliakus, M. F.; Exterkate, M.; Jain, S.; Jumde, V. R.; Andringa, R. L. H.; Kengen, S. W. M.; Minnaard, A. J.; Driessen, A. J. M.; van der Oost, J. Proc. Natl. Acad. Sci. U.S.A. 2018, 115, 3704–3709. 77. Caforio, A.; Jain, S.; Fodran, P.; Siliakus, M.; Minnaard, A. J.; van der Oost, J.; Driessen, A. J. M. Biochem. J 2015, 470, 343–355. 78. Dunlop, M. J.; Dossani, Z. Y.; Szmidt, H. L.; Chu, H. C.; Lee, T. S.; Keasling, J. D.; Hadi, M. Z.; Mukhopadhyay, A. Mol. Syst. Biol. 2011, 7 article 487. 79. Kell, D. B.; Swainston, N.; Pir, P.; Oliver, S. G. Trends Biotechnol. 2015, 33, 237–246. 80. Rodriguez-Moya, M.; Gonzalez, R. J. Proteomics 2015, 122, 86–99. 81. Winkler, J. D.; Garcia, C.; Olson, M.; Callaway, E.; Kao, K. C. Appl. Environ. Microbiol. 2014, 80, 3729–3740. 82. Zhao, W. Q.; Yang, S. S.; Huang, Q. Y.; Cai, P. Colloids Surf., B 2015, 128, 600–607. 83. Jin, T.; Rover, M. R.; Petersen, E. M.; Chi, Z.; Smith, R. G.; Brown, R. C.; Wen, Z.; Jarboe, L. R. J. Ind. Microbiol. Biotechnol. 2017, 44, 1279–1292. 84. Perpina, C.; Vinaixa, J.; Andreu, C.; del Olmo, M. Appl. Microbiol. Biotechnol. 2015, 99, 775–789. 85. Wang, Z.; Wang, J. L.; Ren, G.; Li, Y.; Wang, X. Y. Mar. Drugs 2015, 13, 3325–3339. 86. Konas, R. M.; Daristotle, J. L.; Harbor, N. B.; Klauda, J. B. J. Phys. Chem. B 2015, 119, 13134–13141. 38

87. Konas, R. M.; Daristotle, J. L.; Harbor, N. B.; Klauda, J. B. Biophys. J. 2015, 108, 244A–244A. 88. Lindahl, L.; Genheden, S.; Faria-Oliveira, F.; Allard, S.; Eriksson, L. A.; Olsson, L.; Bettiga, M. Microb. Cell 2018, 5, 42–55. 89. Terama, E.; Ollila, O. H. S.; Salonen, E.; Rowat, A. C.; Trandum, C.; Westh, P.; Patra, M.; Karttunen, M.; Vattulainen, I. J. Phys. Chem. B 2008, 112, 4131–4139. 90. Setiawan, I.; Blanchard, G. J. J. Phys. Chem. B 2014, 118, 3085–3093. 91. Yoon, B. K.; Jackman, J. A.; Kim, M. C.; Sut, T. N.; Cho, N. J. Langmuir 2017, 33, 2750–2759. 92. Exterkate, M.; Caforio, A.; Stuart, M. C. A.; Driessen, A. J. M. ACS Synth. Biol. 2018, 7, 153–165. 93. Markones, M.; Drechsler, C.; Kaiser, M.; Kalie, L.; Heerklotz, H.; Fiedler, S. Langmuir 2018, 34, 1999–2005. 94. Jarboe, L. R. Curr. Opin. Biotechnol. 2018, 53, 93–98.

39

Chapter 4

Microbial Secretion System of Lactate-Based Oligomers and Its Application Camila Utsunomia*,1 and Seiichi Taguchi2,3 1Institute

of Life Technologies, HES-SO Valais Wallis, Route du Rawyl 64, P.O.B 2134, CH-1950 Sion, Switzerland 2Department of Chemistry for Life Sciences and Agriculture, Faculty of Life Sciences, Tokyo University of Agriculture, 1-1-1 Sakuragaoka, Setagaya-ku, Tokyo 156-8502, Japan 3CREST-JST, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan *E-mail: [email protected].

Early microbial secretion system studies of D-lactate (LA)-based oligomers (D-LAOs), co-oligomers of D-LA and D-3-hydroxybutyrate (3HB), are reviewed. D-LAOs were found to be produced and spontaneously secreted by an engineered Escherichia coli using sugar as a carbon source. Aiming for practical applications, the initial low secretory production of D-LAOs could be remarkably increased by the addition of diethylene glycol into the culture medium which functioned as a chain transfer agent. The passage of D-LAOs across the cell membrane was reported to occur via multiple routes including membrane protein transporters. Furthermore, with the successful conversion of D-LAOs to lactide, a green shortcut route in the process of poly(D-lactide) (PDLA) production was created. Also, D-LAOs could be used as a building block for the synthesis of an LA-based poly(ester-urethane). Thus, D-LAOs are promising green compounds with diverse applications, and with production systems expected to be expanded to various organic acid oligomers.

© 2018 American Chemical Society

Introduction At present, value-added products such as bio-based materials have been extensively produced through biorefinery technologies from various renewable feedstocks. Polyhydroxyalkanoates (PHAs) are bacterial storage polyesters that are used in the production of bio-based and biodegradable/biocompatible plastics and have attracted considerable attention as an alternative to petroleum-derived plastics. A typical PHA, poly(3-hydroxybutyrate) [P(3HB)], was discovered by Dr. Maurice Lemoigne from the Institute of Pasteur, France, in the 1920s (1). Since then, more than 160 different monomeric constituents have been identified in the PHA family, thus giving rise to polymers with diverse properties (2, 3). In the green polymerization system, much attention has been focused on the creation of designed PHA synthases to produce new polymers incorporating unnatural monomers, as well as on the tailor-made biosynthesis of PHAs with designable properties. To this end, for the past few decades beneficial mutations closely related to the activity and substrate specificity of several kinds of PHA synthases have been addressed via an evolutionary engineering approach (4, 5). This evolutionary lineage of enzymes can be considered the most important key point for accomplishing the synthetic biology of polymeric material manufacturing. Thus, the incorporation of new and unnatural monomers into the polymeric backbone is of great interest. A typical example is lactic acid. Given that a D-specific lactate (LA)-polymerizing enzyme (LPE) has been artificially evolved from one of the PHA synthases, LA-based polymers could be synthesized in vivo by using LPE (6–8). This major breakthrough opened the door to expanding the diversity of monomers that can be incorporated as unnatural new building blocks into polymeric backbones (9). As for optimization of PHA production, cell volume is a critical limitation factor for the microbial factory (10). As an attempt to mitigate this limitation, many efforts have been made to increase the bacterial cell sizes, including the overexpression of a gene involved in cell division (11) and the deletion of others responsible for maintaining the cell shape (12). Although PHA accumulation could be significantly increased, the cell membrane is still a barrier for large PHA production. As such, secretion of polymerized products should be very attractive for an unlimited and continuous synthesis of PHA. In this chapter, the secretion system for low molecular weight polymer or oligomer of LA-based polyesters (successfully developed by our group) will be described. The chapter will cover topics from the discovery of the oligomer secretion phenomenon to the demonstration of its various applications, including the establishment of a shortcut process for polylactide (PLA) production and the synthesis of a poly(ester-urethane).

Discovery of LA-Based Oligomers Secretion by Escherichia coli As mentioned previously, in 2008 the first microbial platform for the production of D-LA-based polyesters was established, thanks to the development of an engineered PHA synthase (PhaC), LPE (7). Enantiomerically pure polyesters (13), which are copolymers of D-LA and D-3-hydroxybutyrate (3HB) 42

known as P(LA-co-3HB), were successfully synthesized in Escherichia coli (7) and Corynebacterium glutamicum (14). Furthermore, it was found that when poly(D-lactide) (PDLA) polymers were produced, they tended to have relatively low molecular weights (< 104) (14, 15). Recently, the low mobility of PDLA chains polymerized by LPE (PhaC1PsSTQK) was reported as the critical factor that led to the biosynthesis of low molecular weight PDLA (16). The inverse relationship between LA fraction and polymer molecular weight triggered the interest in developing a bacterium-based system for PDLA oligomer production. In addition, based on the knowledge that compounds such as lipids (17), polysaccharides (e.g., cellulose, xanthan, and gellan) (18), amino acids (19), and proteins (20) are bacterially secreted, the initial highlight of this research was the possibility of D-LA oligomers secretion by E. coli. Typically, PDLA and PLLA are synthesized via multiple steps, including the generation of D-LA oligomers in a ring-opening polymerization process. Therefore, the direct secretion of D-LA oligomers would work as a biotechnological shortcut route in the production process by eliminating the necessity of LA purification and oligomerization; thus, reducing the energy, time, and costs of PLA production. Therefore, the establishment of a secretion system of D-LA oligomers would have the potential to greatly contribute to the industrial production of PDLA and its copolymers. The ability of the P(LA-co-3HB)-producing E. coli to secrete oligomers into the medium was initially investigated by Utsunomia et al. (21) (Figure 1). To this end, the recombinant E. coli BW25113 expressing the plasmid pTV118NpctphaC1Ps(ST/FS/QK)AB harboring pct to supply LA-CoA, phaA and phaB to supply 3HB-CoA, and the ST/FS/QK mutant (evolved LPE) of PHA synthase gene (phaC1Ps) from Pseudomonas sp. 61−3 (22), were grown in complex medium containing glucose. After 48 h cultivation, as an attempt to extract the hypothetical extracellular oligomers, the cell-free culture supernatant was subjected to a two-phase extraction, whereby longer-size oligomers, which are more hydrophobic, would migrate to the organic phase (Figure 1). Proton Nuclear Magnetic Resonance (1H NMR) analysis of the extracted fraction revealed nearly identical resonances assigned to the protons of LA and 3HB to that of the polymer P(LA-co-3HB) (Figure 2). The signals were slightly shifted downfield possibly due to the difference in molecular weight between polymers and oligomers. In addition,1H−1H Correlation Spectroscopy (COSY)-NMR showed the existence of eight cross signals which were ascribed to the hydroxyl terminal structures of oligomers with a random distribution of LA and 3HB (23). Thus, 0.4 g/L of LA-based oligomers containing 63 mol% LA, designated as D-LAOs, were surprisingly secreted and successfully extracted from the culture supernatant. Moreover, Electrospray Ionization-Time-of-Flight-Mass Spectrometry (ESI-TOF-MS) analysis of the oligomers detected a bimodal distribution with periodic m/z values in the range of approximately 400−1400 (23). This finding indicated that: (i) D-LAOs can be synthesized during the production of LA-based polymer; and (ii) the synthesized D-LAOs can be spontaneously secreted from the cells. To achieve large-scale applications of D-LAOs, improving the secretory production by E. coli was mandatory. It was thought that by finding a way to efficiently reduce the polymer molecular weight, the synthetic system would 43

be shifted into a microbial factory of oligomers. In the termination step of the polymer synthesis, a chain transfer (CT) reaction, in which the growing polymer chain is transferred to a compound naturally acting as a CT agent (such as water), is suggested to take place (24). Compounds bearing hydroxyl groups, such as polyethylene glycol (PEG) of various molecular weights (25) and short chain alcohols (26), are known to perform as exogenous CT agents by increasing the frequency of the CT reaction and reducing the molecular weight of microbial polyesters (Figure 3). Therefore, it was hypothesized that the synthetic capacity of D-LAOs could be enhanced by increasing the frequency of the CT reaction during the polymerization.

Figure 1. Method used for investigating the extracellular production of LA oligomers by engineered E. coli.

Four compounds known to act as CT agents, PEG200 (average molecular weight = 200), diethylene glycol (DEG), ethylene glycol (EG), and ethanol, were added at the beginning of the bacterial cultivation together with glucose as a carbon source (21) (Figure 3). After 48 h cultivation, the cell-free culture supernatant was analyzed by High performance liquid chromatography (HPLC)to measure the concentration of extracellular oligomers. The developed method was based on the analysis of the supernatant either before and after a treatment with HCl (23). As confirmation of the hypothesis, indeed the addition of CT agents in the bacterial cultivation improved the secretory production of oligomers (Figure 4). The highest production (8.3 ± 1.5 g/L−1) of D-LAOs containing 86.0 ± 4.5 mol% LA was achieved with 5% DEG supplementation, corresponding to 57% of the theoretical carbon yield (21). The role of DEG as a CT agent was also supported by the significant reduction in the molecular weight of the intracellularly accumulated polymers. The maximum Mn reduction of 79% was achieved with the addition of 5% DEG (vol %), which correlates with the highest production and secretion of oligomers. 44

Figure 2. 1H NMR spectrum of LA-based oligomers (D-LAOs) extracted from the culture supernatant of engineered E. coli. The introduction of DEG into the microbial system successfully enlarged the secretory production of D-LAOs rising up the question: How does DEG cause such an effect? One possibility would be the occurrence of cell lysis provoked by DEG, thus increasing extracellular D-LAOs. Microscopic observations revealed that E. coli cells remained intact upon DEG addition at various concentrations. In addition, the cells were also subjected to 4′,6-diamidino-2-phenylindole (DAPI)/ propidium iodide (PI) dual staining, an assay which can distinguish between dead cells and living cells. As a result, no significant difference between the population of cells from cultivations without and with 5% DEG was observed (21). Thus, it was concluded that the increased amount of D-LAOs in the medium was due to the spontaneous secretion performed by living E. coli and not caused by cell lysis. On the other hand, if DEG operated as a CT agent in the synthesis of D-LAOs, extracellular oligomers were likely to be capped with DEG at their carboxyl terminal. To evaluate the mode of DEG conjugation, extracted D-LAOs obtained with DEG supplementation were subjected to various NMR analyses. 1H−1H COSY-NMR and 1H−13C HMQC-NMR demonstrated that DEG has a covalent bond with the carboxyl terminal of D-LAOs (21). Furthermore, 1H−1H Diffusion-Ordered Spectroscopy (DOSY-NMR) revealed that the diffusion coefficient of DEG was similar to that assigned to the oligomer but lower than that of free DEG (Figure 5). Therefore, the observed DEG was suggested to be entirely bound to D-LAOs. The oligomers with DEG at the carboxyl terminal were defined as D-LAOs-DEG, while the oligomers biosynthesized without DEG (i.e., free-form D-LAOs) were termed as D-LAOs*. ESI-TOF-MS analysis 45

of the extracted D-LAOs-DEG detected periodic m/z values in the range of approximately 400−800 (corresponding to ~4- to 10-mers), indicating that shorter oligomers were synthesized relative to those obtained without DEG. Furthermore, a series of peaks corresponding to D-LAOs* were not detected, evidencing that D-LAOs produced with the addition of DEG are nearly fully modified with it at the carboxyl terminal. Thus, DEG actually increased the frequency of CT reaction and, consequently, the production of oligomers (21).

Figure 3. Increasing the frequency of chain transfer (CT) reaction via the addition of exogenous CT agents as a strategy to enhance the extracellular production of D-LAOs.

DEG supplementation remarkably increased the ratio of the extracellular fraction oligomers over the intracellular fraction, indicating that D-LAOs-DEG were more efficiently secreted than D-LAOs*. Thus, not only the production but also the secretion efficiency was found to be improved, presumably due to effects of DEG addition on the concentration, length, and hydrophobicity of oligomers. The higher production of oligomers by the action of DEG (8.3 ± 1.5 g/L−1) may increase the mass transfer rate across the cell membrane driven by the concentration gradient of the oligomers. Also, the molecular weight distributions of D-LAOs* and D-LAOs-DEG showed that shorter oligomers were secreted with DEG supplementation. Moreover, D-LAOs-DEG were thought to be less hydrophilic than D-LAOs* because of the absence of an ionic carboxyl terminus. These factors are likely to positively influence the transport of D-LAOs-DEG to the extracellular environment (21). To date, the effect of the terminal structure on cell membrane-penetration efficiency remains unknown. 46

Figure 4. Effect of the independent addition of four CT agents (PEG200, DEG, ethylene glycol, and ethanol) on the concentration of oligomeric LA in the culture supernatant. Values are presented as the mean of biological triplicates. Error bars indicate standard deviation.

Figure 5. 1H−1H DOSY-NMR spectrum of extracellular oligomers produced with the addition of DEG in comparison with free DEG. 47

Exploration of Membrane Transporters Related to the Secretion of LA-Based Oligomers The identification of D-LAOs secretion route was an important research target to elucidate the mechanism behind this newly found phenomenon and to further enhance the secretory production of D-LAOs. As D-LAOs are unnatural products to E. coli, it was not straightforward to predict by which means these oligomers are secreted. In related studies on the transport of organic acids in E. coli, the uptake of D, L-lactate and glycolate was reported to be mediated by the membrane carriers L-lactate-permease (LldP) and glycolate permease (GlcA) (27). The secretion of free fatty acids (FFA), on the other hand, has been proposed to occur via multiple pathways. Meng et al. reported the secretion of FFA either by active transport assisted by the ATP-binding cassette (ABC) transporter MsbA or by simple diffusion in E. coli (28). In addition, after a screening of membrane-associated proteins potentially involved in the export of FFA, Lennen and co-authors suggested various multi-protein machineries mediating the export of FFA in E. coli (29). On the basis of this background, D-LAOs were likely to be secreted via membrane proteins and/or simple diffusion. The starting point of the study of D-LAOs secretion mechanism was to address the existence of membrane transporters of D-LAOs* (oligomers biosynthesized without DEG addition) in E. coli. There was no rational prediction on D-LAOs* transporters based on biological machinery, such as the operon structure and the gene regulation network, as D-LAOs* are artificial metabolites. Therefore, a cause–consequence model was created to conduct this investigation. If an E. coli knockout mutant of a gene encoding a membrane protein is used for the secretory production of D-LAOs* and this protein is related to the transport of oligomers, decreased secretion and increased intracellular oligomer accumulation were expected. Thus, a loss-of-function phenotype of D-LAOs* secretion would be observed (Figure 6).

Figure 6. Strategy for identifying membrane transporters related to D-LAOs* secretion based on a loss-of-function screening. 48

A total of 209 proteins involved in the transport of organic compounds that are classified into the bacterial secretion system (30), ABC transporters (31), two-component system (32), phosphotransferase system (PTS) (33), and other transporters, were selected as potential candidates of membrane proteins involved in D-LAOs* secretion. Many transportation machineries are composed by multiple proteins. To disrupt such multi-component transporters, one gene encoding a putative essential component was deleted. Thus, E. coli single-gene deletants selected from the Keio collection (34), which are deficient mutants of a membrane protein transporter, and the wild-type BW25113 (parent strain), were transformed with pTV118NpctphaC1Ps(ST/FS/QK)AB and subjected to the loss-of-function screening of D-LAOs* secretion ability—that is, decreased secretion and increased D-LAOs* intracellular accumulation (35). A total of 55 strains were found to secrete a significantly decreased amount of D-LAOs* (< 0.3 g/L-1) compared to the parent strain (Figure 7). Furthermore, 7 out of 55 deletants also presented increased level of D-LAOs* intracellular accumulation. As a result of the initial screening, seven membrane proteins were selected for further investigation. The function and localization of the candidates of D-LAOs* transporters in the membrane are presented in Table 1. Among them two are porins, OmpF and OmpG. These proteins are associated with the outer membrane, and they contain large, open, water-filled channels that nonspecifically control and enable the spontaneous diffusion of ions, and small hydrophilic nutrient molecules through bacterial outer membranes (36, 37). The genes encoding OmpF and OmpG were then overexpressed to verify whether the opposite effect, an increased secretion of D-LAOs, would be observed. The cultivation of OmpG-overexpressing strain (ompGox) yielded a larger amount of extracellular D-LAOs compared to the parental strain. As expected, Reverse Transcription Polymerase Chain Reaction (RT-PCR) analysis revealed that the expression level of ompG was significantly higher in the strain ompGox (Figure 8A). It was also shown that the increased secretion of D-LAOs by this strain was not due to reinforcement in glucose consumption. On the other hand, in the cultivation of the strain overexpressing ompF, the extracellular production of D-LAOs* was decreased as well as ompF expression level compared to the parental strain (Figure 8A). OmpF is well-known to be one of the most abundant proteins and one of the major porins in E. coli (38, 39). Thus, the unsuccessful overexpression of ompF is likely to be due to a repression in the high basal expression level of ompF in the presence of abundant OmpF. These results suggested that porins facilitate the passage of D-LAOs* across the outer membrane. In particular, OmpF might be one of the major porins responsible for mediating the spontaneous secretion of D-LAOs* by E. coli, due to its high basal expression level. Aditionaly, OmpF was previously demonstrated to be involved in the transport of fatty acids, compounds which are structurally related to D-LAOs*. After a comparative proteomic analysis of E. coli under octanoic acid stress, OmpF was suggested to facilitate the transport of short-chain fatty acids into the cells (40). Also, the deletion of ompF was shown to increase membrane integrity and fatty acid tolerace/production in E. coli, possibly because the path for fatty acids’ re-entry was blocked (41). OmpG, on the other hand, is likely to play a secondary role in the secretion of D-LAOs*. 49

Figure 7. Extracellular production of D-LAOs* by the single-gene deletion mutants used in the loss-of-function screening of D-LAOs secretion.

With respect to the other five selected candidates, they were all inner membrane-associated proteins with no function in common (Table 1). This could be evidence that multiple secretion routes across the inner membrane may exist. The gene mngA encodes a protein that works as a single-component transporter, and the effect of its overexpression was also verified. The strain overexpressing mngA, the mngAox, produced greater amounts of extracellular D-LAOs* compared to the parent, and the successful gene overexpression was also confirmed by RT-PCR (Figure 8B). Aside from its role in the uptake of 2-O-α-mannosyl-D-glycerate (42, 43), it was reported that MngA (formerly HrsA) can be a positive porin regulator under certain conditions (44). However, no considerable upregulation of ompF and ompG was observed in the strain mngAox. Therefore, the positive effect of mngAox on D-LAOs production was unlikely to be due to an elevation of porins expression levels, but rather due to direct enhancement of the D-LAOs export. Additionally, it was observed that the increased secretion of D-LAOs by ompGox and mngAox did not result in oligomers with molecular weights significantly different from those secreted by the parental strain (35). This indicated that the overexpression of ompG and mngA should reinforce the intrinsic secretion route, rather than opening a new pathway, which would allow the secretion of longer-chain D-LAOs. The other selected candidates, ArgT, MacA, CitA, and CpxA, function together with other membrane components forming complex transportation machineries; this will be evaluated in future (35). 50

Table 1. Function and Membrane Localization of the Seven Membrane Proteins Candidates Selected from the Loss-of-Function Screening Orf

Protein function

Localization

ompF

Outer membrane porin 1a (Ia;b;F).

Outer membrane.

ompG

Outer membrane porin G.

Outer membrane.

citA

Sensory histidine kinase in two-component regulatory system with CitB. Citrate uptake.

Inner membrane (periplasmic and cytoplasmic domain).

argT

Member of a multi-protein transporter. Binds lysine/arginine/ornithine for uptake.

Periplasm. The multi-protein transporter is located in the periplasm and inner membrane.

mngA

2-O-α-mannosyl-D-glycerate (MG) specific enzyme consisting of three domains IIA/IIB/IIC. MG uptake.

Two cytoplasmic domains and one inner membrane domain.

macA

Membrane fusion protein (MFP) component. Macrolide transporter with MacB.

Periplasm. The multi-protein transporter is located in the periplasm and inner membrane.

cpxA

Sensory histidine kinase in two-component regulatory system with CpxR. Regulates a vast number of genetic loci in response to periplasmic stress.

Inner membrane.

Based on the aforementioned results, a model of D-LAOs secretion was proposed (35) (Figure 9). The passage of D-LAOs across the outer membrane could occur via passive diffusion mediated by porins, such as OmpF and OmpG. In sequence, diverse inner membrane-associated proteins, including MngA, CitA, ArgT, MacA, and CpxA, are suggested to transport D-LAOs* through the inner-membrane. In addition, the possibility of D-LAOs* secretion via simple diffusion was not excluded from this model. Thus, the secretion of D-LAOs* by E. coli was identified to be unspecific and to occur via multiple routes. However, this is a freshly initiated investigation, and further studies still have to be conducted to clarify the secretion mechanism either of D-LAOs* or the conjugated form D-LAOs-DEG. By these means, the higher secretion efficiency of D-LAOs-DEG compared to D-LAOs* could be elucidated.

51

Figure 8. RT-PCR analysis of overexpression of the genes encoding the selected membrane candidates and D-LAOs* extracellular production by the overexpressors. (A) Overexpression of ompF and ompG. (B) Overexpression of mngA.

52

Figure 9. Proposed model for the mechanism of D-LAOs secretion across the membrane of E. coli.

Practical Applications of LA-Based Oligomers As stated early on, the main motivation behind the creation of a microbial secretion system of LA-based oligomers was to improve the price competitiveness of PDLA and its copolymers over conventional polyesters. Therefore, to validate the potential of this novel compound, the feasible application of D-LAOs in the synthesis of PDLA and of an LA-based poly(ester-urethane) was evaluated.

D-LAOs as a Substrate for the Production of High-Molecular Weight PDLA PLA is one of the major commercially available thermoplastics derived from renewable carbohydrate-rich products (45). The production capacity of PLAs reached 195,000 tons in 2013, representing about 3.8% of the total production of biobased plastics (partially or 100% derived from renewable resources) (46), and by 2020 the production is expected to reach about 450,000 tons (47). PLAs are compostable and biocompatible materials that have been extensively applied in the biomedical field (48) and are being increasingly used in commodity applications, such as packaging (49) and textiles (50). In the industry, high-molecular weight PLA is mainly produced via ring-opening polymerization (ROP) through the formation of the cyclic dimer (lactide) of lactic acid (Figure 10). This method is based on the original Cargill-Dow patented process (51), which combines a solvent-free and a distillation process for producing PLA with controlled molecular weights. In this procedure, lactic acid is mainly produced from carbohydrate fermentation, and the purification of lactic acid from culture medium is typically performed by calcium salt precipitation and subsequent acidification, which generates salt as a byproduct (52). For the production of LA oligomer, the purified lactic acid is conventionally condensed at an increased temperature, and water is removed 53

under vacuum for several hours (53). The synthesis of lactide occurs via thermal depolymerization of the LA oligomers via a metal-catalyzed backbiting reaction of the –OH end groups. Finally, lactides are polymerized into high-molecular weight PLA via metal-catalyzed ROP. Although ROP is the main route, it is a multi-step chemo-bio process which requires additional purification steps, so it is considered relatively complex and expensive. Among the steps, the production of lactide from lactic acid, which comprises the oligomerization of lactic acid and the synthesis of lactide from LA oligomers, contributes to 30% of the PLA cost (53).

Figure 10. The conventional and the new shortcut route (represented by the double-line arrow), which eliminates the steps of lactic acid purification and polycondensation, for PDLA production.

Therefore, the utilization of microbially secreted D-LAOs as a substrate for the synthesis of lactide, which can be used for subsequent PDLA production, was attempted. It was thought that the secretion system of D-LAOs would greatly contribute to the industrial production of PLA as a “biosynthetic shortcut” to provide LA oligomers in the chemo-bio process (Figure 10). For evaluating the possibility of lactide synthesis, D-LAOs-DEG (68 mol% LA, degree of polymerization (DP) of approximately trimer to 7-mer) and D-LAOs* (61 mol% LA, DP ~ trimer to 16-mer), extracted from E. coli culture supernatant, was subjected to a thermal depolymerization with zinc oxide as a catalyst, and the vaporized fraction was recovered by condensation (54). The 1H NMR spectrum of the sample generated from D-LAOs-DEG exhibited signals, identical to those of standard D-lactide (Figure 11). D-LAOs* and chemically synthesized L-LA oligomers (DP ~ 6- to 16-mers), used as an experimental control, were also converted into lactides using the same procedure. These results proved that D-LAOs were a substrate for lactide conversion. 54

Figure 11. 1H NMR spectrum of the sample from the depolymerization reaction of D-LAOs-DEG using zinc oxide as catalyst. The signals corresponding to lactide are assigned. Nevertheless, the lactide formation over the substrate consumption of D-LAOs-DEG (13%) and D-LAOs* (25%) were considerably lower than that of synthetic L-LAOs (77%). The presence of 3HB in D-LAOs backbone was assigned as the main reason for such a difference, due to the following reasons: 3HB decreases the frequency of LA-LA dyad and 3HB may act as a stopper of the backbiting reaction. Accordingly, the production of LA-enriched oligomers was pursued to further increase the efficiency of the lactide synthesis (54). Previously, the use of an E. coli dual-gene knockout mutant (ΔpflA and Δdld), known as JWMB1, and xylose as a carbon source, successfully increased the LA fraction in P(LA-co-3HB) (55). The deletion of pflA is known to eliminate formate generation from acetyl-CoA, channeling the flux toward lactic acid (56), and the dld mutation prevents lactic acid oxidation into pyruvate, improving the intracellular availability of this organic acid (57). The LA enrichment caused by the consumption of xylose has been assigned to the different capacities of regenerating NADH and NADPH in the metabolism routes of xylose and glucose (58, 59). Indeed, by using ΔpflA/Δdld E. coli mutant grown on xylose, 8.1 ± 2.9 g/L-1 of D-LAOs-DEG with up to 97 ± 1 mol% LA was successfully produced. After extraction, 4.5 g/L-1 of D-LAOs-DEG containing 83 mol% LA (recovery of 56%) were obtained. When E. coli was further cultured under microaerobic conditions, although the total oligomer production and LA fraction were decreased, the recovery (65%) and LA fraction (89 mol%) of the extracted 55

D-LAOs-DEG were rather improved. Microaerobic cultivations are well-known promoters of lactic acid production (60, 61) and LA fraction enrichment in LA-based polyester (13). In the production of LA-based oligomers, on the other hand, the microaerobic condition was suggested to increase the hydrophobicity of the oligomers, which is determined by their molecular weight and monomer composition, thus improving the recovery of oligomers with higher LA fraction (54). The LA fraction enrichment in D-LAOs-DEG increased the lactide yield from 13% up to 18%; however, it was still lower than that conventionally obtained from synthetic L-LAOs. During the lactide synthesis, it was observed that short oligomers were lost by vaporization during heating, which prevented their conversion into lactide (54). In the industrial process, LA oligomers with molecular weights around 400–2500 g/mol, which correspond to a DP of approximately 5- to 34-mer, are used in the synthesis of lactide (53). Therefore, further increase in LA fraction and molecular weight of the secreted oligomers are suggested as reasonable approaches to improve the conversion efficiency of D-LAOs-DEG into lactide. The proof-of-concept that D-LAOs* and D-LAOs-DEG can be converted into lactide demonstrated that the microbial secretion system of D-LAOs can be used as a greener shortcut route for PDLA production (Figure 11).

D-LAOs-DEG as a Building Block for LA-Based Poly(ester-urethane) Polyurethanes are some of the versatile plastics used for a wide range of applications (62). These plastics are the sixth most produced in the world, which represents 14 million tons of polyurethanes per year (63). Generally, the polymer synthesis is carried out via a polyaddition reaction of an alcohol to an isocyanate compound (64). D-LAOs-DEG bearing hydroxyl group at both terminals would be a good target in such a reaction. In fact, hydroxyl-terminated LA oligomers have been found as suitable compounds to yield LA-based poly(ester-urethane)s (64, 65). Therefore, for expanding the range of applications, the use of D-LAOsDEG as a building block for the production of LA-based poly(ester-urethane) was evaluated (Figure 12) (66). D-LAOs-DEG containing 45 mol% LA and an estimated average molecular weight of 500 g/mol were polymerized by polyaddition reaction with 4,4′-diphenylmethane diisocyanate (MDI) as a chain extender. The reaction generated products with an average molecular weight (Mw, SEC) of 3300, ranging from approximately 1000 to 8000 g/mol (Figure 13A). Furthermore, the sample was subjected to Fourier transform infrared chromatography (FT-IR) analysis, and the spectrum showed the characteristic absorptions at 1532 cm-1 and 3334 cm-1 due to NH deformation and stretching vibrations, respectively, evidencing the urethane bond formation (Figure 13B) (67). The results clearly demonstrated that D-LAOs-DEG could be assembled into an LA-based poly(ester-urethane) (66). Nevertheless, the polymer molecular weight could not be further increased. It was presumed that such a limitation was mainly due to two factors: oligomer 56

molecular weight and purity of the biosynthesized D-LAOs-DEG. Therefore, increasing the molecular weight of the secreted D-LAOs-DEG will be an important research target, along with the optimization of the D-LAOs-DEG purification method.

Figure 12. Synthesis scheme of an LA-based poly(ester-urethane) from D-LAOs-DEG and MDI via polyaddition reaction.

Figure 13. Analyses of the resulting product from the polyaddition reaction of D-LAOs-DEG with MDI. (A) Molecular weight distribution by size exclusion chromatography (SEC). (B) FT-IR spectrum.

Future Prospects This chapter described the first finding of the secretion of oligomers including LA in microorganisms. Such an interesting phenomenon appeared in Escherichia coli, which could be a good model for better understanding the mechanisms for the generation and secretion of organic acid oligomers from the viewpoints of both academic research and industrial applications. This finding will hopefully open the door to the optimization of such a process to become cost-competitive with other 57

commodity polymers, enabling these oligomers to become promising value-added products. To further this research, the following three directions should be explored: (1) the exploration of the bioactivity of these compounds in living organism communities involved in various environments; (2) their utilization as building blocks for engineering functionalized materials; and (3) the optimization of the secretory production system of oligomers. For this purpose, research and engineering of membrane proteins related to oligomers’ secretion should be essential. In these directions, our research has already laid the groundwork for further progress. This is a typical synthetic biology study in order to create new oligomers as well as a tailor-made secretory production of the target oligomers from renewable feedstock based on a microbial platform. Hopefully the findings shown in this chapter may suggest other research areas where the same approach might be used.

References 1. 2. 3. 4. 5. 6. 7.

8. 9. 10. 11. 12. 13. 14.

15. 16. 17. 18.

Lemoigne, M. Bull. Soc. Chem. Biol. 1926, 8, 770–782. Doi, Y.; Steinbüchel, A. In Biopolymers, vol. 3a (Polyesters I) and 3b (Polyesters II); Wiley-VCH, Weinheim, Germany: 2001. Anderson, A. J.; Dawes, E. A. Microbiol. Rev. 1990, 54, 450–472. Stubbe, J.; Tian, J.; He, A.; Sinskey, A. J.; Lawrence, A. G.; Liu, P. Annu. Rev. Biochem. 2005, 74, 433–480. Taguchi, S.; Doi, Y. Macromol. Biosci. 2004, 4, 145–156. Nomura, C. T.; Taguchi, S. Appl. Microbiol. Biotechnol. 2007, 73, 969–979. Taguchi, S.; Yamada, M.; Matsumoto, K.; Tajima, K.; Satoh, Y.; Munekata, M.; Ohno, K.; Kohda, K.; Shimamura, T.; Kambe, H. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 17323–17327. Taguchi, S. Polym. Degrad. Stab. 2010, 95, 1421–1428. Matsumoto, K.; Taguchi, S. Curr. Opin. Biotechnol. 2013, 24, 1054–1060. Jiang, X.-R.; Chen, G.-Q. Biotechnol. Adv. 2016, 34, 435–440. Wang, Y.; Wu, H.; Jiang, X.; Chen, G.-Q. Metab. Eng. 2014, 25, 183–193. Jiang, X.-R.; Wang, H.; Shen, R.; Chen, G.-Q. Metab. Eng. 2015, 29, 227–237. Yamada, M.; Matsumoto, K.; Nakai, T.; Taguchi, S. Biomacromolecules 2009, 10, 677–681. Song, Y.; Matsumoto, K.; Yamada, M.; Gohda, A.; Brigham, C. J.; Sinskey, A. J.; Taguchi, S. Appl. Microbiol. Biotechnol. 2012, 93, 1917–1925. Shozui, F.; Matsumoto, K.; Motohashi, R.; Sun, J.; Satoh, T.; Kakuchi, T.; Taguchi, S. Polym. Degrad. Stab. 2011, 96, 499–504. Matsumoto, K.; Iijima, M.; Hori, C.; Utsunomia, C.; Ooi, T.; Taguchi, S. Biomacromolecules 2018 DOI:10.1021/acs.biomac.8b00454. Doshi, R.; Nguyen, T.; Chang, G. Proc. Natl. Acad. Sci. U.S.A. 2013, 110, 7642–7647. Sutherland, I. W. Int. Dairy J. 2001, 11, 663–674. 58

19. Krämer, R. FEMS Microbiol. Rev. 1994, 13, 75–93. 20. Mergulhao, F. J. M.; Summers, D. K.; Monteiro, G. A. Biotechnol. Adv. 2005, 23, 177–202. 21. Utsunomia, C.; Matsumoto, K.; Taguchi, S. ACS Sustainable Chem. Eng. 2017, 5, 2360–2367. 22. Yamada, M.; Matsumoto, K.; Shimizu, K.; Uramoto, S.; Nakai, T.; Shozui, F.; Taguchi, S. Biomacromolecules 2010, 11, 815–819. 23. Sun, J.; Matsumoto, K. I.; Tabata, Y.; Kadoya, R.; Ooi, T.; Abe, H.; Taguchi, S. Appl. Microbiol. Biotechnol. 2015, 99, 9555–9563. 24. Madden, L. A.; Anderson, A. J.; Shah, D. T.; Asrar, J. Int. J. Biol. Macromol. 1999, 25, 43–53. 25. Shi, F.; Gross, R. A.; Rutherford, D. R. Macromolecules 1996, 29, 10–17. 26. Hiroe, A.; Hyakutake, M.; Thomson, N. M.; Sivaniah, E.; Tsuge, T. ACS Chem. Biol. 2013, 8, 2568–2576. 27. Núñez, M. F.; Kwon, O.; Wilson, T. H.; Aguilar, J.; Baldoma, L.; Lin, E. C. C. Biochem. Biophys. Res. Commun. 2002, 290, 824–829. 28. Meng, X.; Shang, H.; Zheng, Y.; Zhang, Z. Biotechnol. Lett. 2013, 35, 2099–2103. 29. Lennen, R. M.; Politz, M. G.; Kruziki, M. A.; Pfleger, B. F. J. Bacteriol. 2012, JB-01477. 30. Green, E. R.; Mecsas, J. Microbiol. Spectrum 2016, 4 DOI:10.1128/ microbiolspec.VMBF-0012-2015. 31. Moussatova, A.; Kandt, C.; O’Mara, M. L.; Tieleman, D. P. Biochim. Biophys. Acta, Biomembr. 2008, 1778, 1757–1771. 32. Capra, E. J.; Laub, M. T. Annu. Rev. Microbiol. 2012, 66, 325–347. 33. Escalante, A.; Cervantes, A. S.; Gosset, G.; Bolívar, F. Appl. Microbiol. Biotechnol. 2012, 94, 1483–1494. 34. Baba, T.; Ara, T.; Hasegawa, M.; Takai, Y.; Okumura, Y.; Baba, M.; Datsenko, K. A.; Tomita, M.; Wanner, B. L.; Mori, H. Mol. Syst. Biol. 2006, 2, 2006.0008. 35. Utsunomia, C.; Chiaki, H.; Matsumoto, K.; Taguchi, S. J. Biosci. Bioeng. 2017, 124, 635–640. 36. Nikaido, H. J. Bioenerg. Biomembr. 1993, 25, 581–589. 37. Köster, S.; van Pee, K.; Yildiz, Ö. Methods Enzymol. 2015, 557, 149–166. 38. Delihas, N.; Forst, S. J. Mol. Biol. 2001, 313, 1–12. 39. Pratt, L. A.; Hsing, W.; Gibson, K. E.; Silhavy, T. J. Mol. Microbiol. 1996, 20, 911–917. 40. Rodríguez-Moyá, M.; Gonzalez, R. J. Proteomics 2015, 122, 86–99. 41. Tan, Z.; Black, W.; Yoon, J. M.; Shanks, J. V.; Jarboe, L. R. Microb. Cell Fact. 2017, 16, 38. 42. Sampaio, M.-M.; Chevance, F.; Dippel, R.; Eppler, T.; Schlegel, A.; Boos, W.; Lu, Y.-J.; Rock, C. O. J. Biol. Chem. 2004, 279, 5537–5548. 43. Jacobson, G. R.; Saraceni-Richards, C. J. Bioenerg. Biomembr. 1993, 25, 621–626. 44. Utsumi, R.; Horie, T.; Katoh, A.; Kaino, Y.; Tanabe, H.; Noda, M. Biosci., Biotechnol., Biochem. 1996, 60, 309–315. 45. Vink, E. T.; Davies, S. Ind. Biotechnol. 2015, 11, 167–180. 59

46. Aeschelmann, F.; Carus, M. Ind. Biotechnol. 2015, 11, 154–159. 47. Prieto, A. Microb. Biotechnol. 2016, 9, 652–657. 48. Lasprilla, A. J.; Martinez, G. A.; Lunelli, B. H.; Jardini, A. L.; Filho, R. M. Biotechnol. Adv. 2012, 30, 321–328. 49. Auras, R.; Harte, B.; Selke, S. Macromol. Biosci. 2004, 4, 835–864. 50. Gupta, B.; Revagade, N.; Hilborn, J. Prog. Polym. Sci. 2007, 32, 455–482. 51. Gruber, P. R.; Hall, E. S.; Kolstad, J. J.; Iwen, M. L.; Benson, R. D.; Borchardt, R. L. Continuous process for manufacture of lactide polymers with controlled optical purity. U.S. Patent 5,142,023, 1992. 52. Li, Q.-Z.; Jiang, X.-L.; Feng, X.-J.; Wang, J.-M.; Sun, C.; Zhang, H.-B.; Xian, M.; Liu, H.-Z. J. Microbiol. Biotechnol. 2016, 26, 1–8. 53. Van Wouwe, P.; Dusselier, M.; Vanleeuw, E.; Sels, B. ChemSusChem 2016, 9, 907–921. 54. Utsunomia, C.; Matsumoto, K.; Date, S.; Hori, C.; Taguchi, S. J. Biosci. Bioeng. 2017, 124, 204–208. 55. Nduko, J. M.; Matsumoto, K.; Ooi, T.; Taguchi, S. Appl. Microbiol. Biotechnol. 2014, 98, 2453–2460. 56. Zhu, J.; Shimizu, K. Appl. Microbiol. Biotechnol. 2004, 64, 367–375. 57. Zhou, L.; Zuo, Z.-R.; Chen, X.-Z.; Niu, D.-D.; Tian, K.-M.; Prior, B. A.; Shen, W.; Shi, G.-Y.; Singh, S.; Wang, Z.-X. Curr. Microbiol. 2011, 62, 981–989. 58. Chin, J. W.; Khankal, R.; Monroe, C. A.; Maranas, C. D.; Cirino, P. C. Biotechnol. Bioeng. 2009, 102, 209–220. 59. Lim, S.-J.; Jung, Y.-M.; Shin, H.-D.; Lee, Y.-H. J. Biosci. Bioeng. 2002, 93, 543–549. 60. Tarmy, E. M.; Kaplan, N. O. J. Biol. Chem. 1968, 243, 2579–2586. 61. Mat-Jan, F.; Alam, K. Y.; Clark, D. P. J. Bacteriol. 1989, 171, 342–348. 62. Zia, K. M.; Bhatti, H. N.; Bhatti, I. A. React. Funct. Polym. 2007, 67, 675–692. 63. Kopczyńska, P.; Datta, J. Polym. Int. 2016, 65, 946–954. 64. Hiltunen, K.; Seppälä, J. V.; Härkönen, M. J. Appl. Polym. Sci. 1997, 63, 1091–1100. 65. Ren, J.; Wang, Q. F.; Gu, S. Y.; Zhang, N. W.; Ren, T. B. J. Appl. Polym. Sci. 2006, 99, 1045–1049. 66. Utsunomia, C.; Saito, T.; Matsumoto, K.; Hori, C.; Isono, T.; Satoh, T.; Taguchi, S. J. Polym. Res. 2017, 24, 167. 67. Saito, T.; Aizawa, Y.; Tajima, K.; Isono, T.; Satoh, T. Polym. Chem. 2015, 6, 4374–4384.

60

New Enzymatic Methodologies

Chapter 5

Structural and Mutational Analysis of Polyethylene Terephthalate–Hydrolyzing Enzyme, Cut190, Based on Three-Dimensional Docking Structure with Model Compounds of Polyethylene Terephthalate Takeshi Kawabata,1 Masayuki Oda,2 Nobutaka Numoto,3 and Fusako Kawai*,4 1Institute

of Protein Research, Osaka University, Suita, Osaka 565-0871, Janan 2Graduate School of Life and Environmental Sciences, Kyoto Prefectural University, Kyoto, Kyoto 606-8522, Japan 3Medical Research Institute, Tokyo Medical and Dental University, Tokyo 113-8510, Japan 4Center for Fiber and Textile Science, Kyoto Insitute of Technology, Kyoto, Kyoto 606-8585, Japan *E-mail: [email protected].

The cutinase-like enzyme, Cut190, from Saccharomonospora viridis AHK190 can degrade the inner block of polyethylene terephthalate (PET) in the presence of Ca2+, and its mutant (Cut190*), S226P/R228S, exhibited increased activity and higher thermostability. The crystal structures of the Cut190 S226P mutant in the absence and presence of Ca2+ were determined and revealed a large conformational change induced upon Ca2+-binding. However, the substrate-bound three-dimensional (3D) structures of Cut190 remained unknown. In this study, to determine the substrate-binding site and improve the enzyme activity, we first built 3D structures of a PET model compound bound to the crystal structures, using the distance restraints between the scissile carbonyl group of the compound and the catalytic site of the enzyme. We then mutated the putative substrate-binding sites predicted

© 2018 American Chemical Society

from the models and experimentally determined the enzymatic activities of the mutants for the model substrate poly(butylene succinate-co-adipate). The mutated sites with decreased activity were consistent with the putative-binding sites predicted by the 3D model from the Ca2+-bound crystal structure, suggesting that the structure of the Ca2+-bound state represents the active state. We also generated the inactive mutant of Cut190*, Cut190*S176A, in which the active serine, Ser176, was mutated to Ala, and determined its crystal structure in complex with Ca2+. Three Ca2+ molecules were bound to Cut190*S176A at sites 1-3. The roles of Ca2+-binding at sites 1-3 were suggested.

Introduction Cutinases from a variety of microbial sources (fungal and bacterial) have been identified and characterized, and several crystal structures of homologues have been deposited in the Protein Data Bank (PDB), such as the fungal cutinases from Fusarium solani pici (PDB ID: 1CUS), Aspergillus oryzae (PDB ID: 3GBS), and Cryptococcus sp. (PDB ID: 2CZQ) as well as the bacterial cutinases from Thermobifida alba (PDB ID: 3VIS and 3WYN), a leaf compost (PDB ID: 4EB0), T. fusca (PDB ID: 4CG1, 4CG2, and 4CG3), and T. cellulosilytica (PDB ID: 5LUI, 5LUJ, 5LUK, and 5LUL). Cutinases naturally function in the invasion of phytopathogenic microorganisms into plants by attacking the plant’s layer of cutin (1). Cutinases have been utilized for various industrial purposes as biocatalysts together with lipases, and both are members of a lipase superfamily (2–5). Cutinases have an open active site, in contrast to lipases with a lid-covered active site. The open active site enables easy access of rigid polymer substrates or complex substrates, such as cutin, to the active serine in the catalytic triad. This ease of access is considered to be the reason why plastic polyesters, such as polyethylene terephthalate (PET), can be hydrolyzed by cutinases. All of the PET hydrolases reported so far are cutinases, regardless of whether they can degrade cutin. Cutinases are attracting increasing attention with regard to enzymatic recycling of waste plastics in the future. We have cloned a few cutinases from thermophilic actinomycetes, Thermobifida alba AHK119 and Saccharomonospora viridis AHK190 (6–8). X-ray crystallographic analyses of Est119, one of the tandem cutinases in T. alba AHK119, and Cut190 from S. viridis AHK190 have been reported, respectively (9, 10). The crystal structures of cutinases such as Est119 (9) and Cut190 (10), TfCut2 from T. fusca KW3 (11), LC-cutinase (metagenome from leaf compost) (12), and Thc_Cut2 from T. cellulosilytica (13) suggested that they are quite similar to each other. However, the Cut190 structure is unique, as its three-dimensional (3D) structures differ between the inactive form (Ca2+-free) and the active form (Ca2+-bound), whereas the other cutinases retain the same structures with and without Ca2+. The reported Ca2+-binding sites of Cut190 are also different from the others, suggesting that the role of Ca2+-binding in 64

Cut190 could be unique. Ca2+ is not a prosthetic group in the active site and not directly involved in the catalytic reaction but requisite for activation and thermostabilization of the enzyme. Poly(butylene succinate-co-adipate) (PBSA) is conveniently used as a polymer substrate for assaying Cut190 activity, and Cut190 significantly hydrolyzes amorphous PET film (8). Here we built 3D structural models of partial structures of PET and the model substrate PBSA bound to the crystal structures under the distance restraints between the scissile carbonyl group of the substrate and the catalytic site of the enzyme. As the S176A mutant lost the activity completely, as described previously (8), Ser176 would be indispensable for the enzyme catalytic function, and surrounding amino acids in the active site may be relevant to the substrate docking. We built models using a combination of our programs because docking the large chemical compound with the distance restraint is difficult for most of the academic-free molecular docking programs. Based on the docking models, we mutated the residues that were predicted to participate significantly in the substrate binding and catalysis and confirmed their expression levels and activities for the model substrate PBSA. Because the Cut190 mutant, S226P/R228S, displayed the highest thermostability (8), we used the S226P/R228S mutant (designated as Cut190* hereafter) as the template for further mutations in this study. We successfully obtained a more robust Cut190 mutant and identified the amino acids relevant to the substrate binding and required for the catalysis. Mutational results endorsed with the prediction by the simulation for docking of the big and long substrate. We also determined the crystal structure of Cut190*S176A in complex with Ca2+ and identified that three Ca2+ molecules were bound to Cut190*S176A at sites 1-3. The possible roles of Ca2+-binding at sites 1-3 will be discussed.

Docking Calculations of the Model Substrate with Cut190* The open (Ca2+-bound) and closed (Ca2+-free) forms of Cut190* were constructed by introducing the R228S mutation into crystal structures of Cut190 S226P (PDB ID: 4WFK and 4WFI) (14). As the substrate, “BABSBA” was first used for the docking calculations because it is considered to be the core structure of PBSA, which is a polymerized compound (Figure 1), and it was also used as the substrate for the activity assays described later. The docking calculations were performed for both Ca2+-free and Ca2+-bound Cut190* (S226P/R228S mutant), built from the crystal structures of the Cut190 S226P mutant (10). Since the residue Arg228 was exposed to the solvent, the side chain of this arginine could be replaced with that of serine without affecting the conformations of the other amino acids. The binding pockets in both the Ca2+-free and Ca2+-bound Cut190* were detected using the pocket-finding program Ghecom (15). As shown in Figure 2, a pocket region was found around the active site Ser176 in both structures. However, their shapes considerably differed. The pocket in the Ca2+-bound structure (Figure 2b) has a long and narrow shape, whereas that in the Ca2+-free structure (Figure 2a) has a quite round shape. The difference could be due to the 65

large conformational difference of Phe106. The substrate BABSBA was docked into the detected pocket under the distance restraints between the scissile carbonyl group of the substrate and the catalytic site of the enzyme, using the programs Fkcombu (16) and MyPresto (17). Figure 3 shows the models with the lowest and top 10 lowest binding energies, obtained by the optimization of the Generalized Born/Surface Area (GB/SA) energy (18). The 10 lowest energy models share several common interacting sites around the restrained catalytic site (Ser176). The residues that interacted with the substrate are summarized in Table 1. The numbers (from 0 to 10) shown in the table are the numbers of conformations that interacted with the residues among the 10. The residues with large numbers for both structures (Gly105, Phe106, Gln138, Ser176, Met177, Trp201, Ile224) are probably interacting residues. Table 1 also shows that some residues only interacted with BABSBA in one state. For example, Ser112, His175, His254, Phe255, and Asn258 only interacted in the Ca2+-bound state. This is due to the large conformational change induced by Ca2+-binding, as revealed by the crystal structure analysis (10). Since the enzymatic activity of Cut190 is activated by Ca2+-binding, the 3D model using the Ca2+-bound form would be more helpful to design engineered Cut190 proteins. An enlarged schematic view of the active site of the Ca2+-bound model is shown in Figure 4. The model substrate of PET, “TETETET”, was also used for docking calculations. Figure 5 shows the binding site of TETETET in the lowest binding energy model of its complex with Ca2+-bound Cut190*, in comparison with that of BABSBA. The binding conformation of TETETET resembles that of BABSBA, and its predicted interactive residues are also similar to those with BABSBA. This is in accordance with Cut190* hydrolyzing both PET and PBSA (8). The docking calculations revealed that approximately three to five monomer units, such as TET(ET) and BAB(SB), fill the active site and no more units are accommodated, suggesting that the model compounds are sufficient for an explanation of the enzyme–substrate interactions. In addition, the extended polymer chain from the active site would not bind to the enzyme surface, as also found with the polyhydroxybutyrate depolymerase from Penicillium funiculosum (19). It should be noted that the docking structures of the two polyesters were similar, which is understandable since Cut190* hydrolyzes both PBSA and PET.

66

Figure 1. Chemical structures of (a) the PBSA and PET molecules, (b) a partial model of PBSA, BABSBA, and (c) a partial model of PET, TETETET. The carbon atoms enclosed within red dotted circles are the scissile carbonyl groups of the model compounds. Reproduced with permission from ref (14). Copyright 2017 Elsevier. 67

Figure 2. Pocket regions of Cut190*. Green spheres are pocket probes. (a) Ca2:-free Cut190* (based on 4WFI); (b) Ca2:-bound Cut190* (based on 4WFK). Reproduced with permission from ref (14). Copyright 2017 Elsevier. (see color insert)

Figure 3. Structure models of Ca2+-free Cut190* (a, b) and Ca2+-bound Cut190* (c, d) in complex with BASBA. Models with the lowest GB/SA binding energy (a, c) with the top 10 lowest GB/SA binding energies (b, d). Reproduced with permission from ref (14). Copyright 2017 Elsevier. (see color insert) 68

Figure 4. Views around the active sites of the 3D model of Ca2+-bound Cut190* and BABSBA (the front (a) and the side (b)). The mutated residues are shown by sticks (groups 1, 2, and 3 colored blue, pink, and green, respectively). The dotted circle shows the oxyanion hole between Phe106 and Met177. Reproduced with permission from ref (14). Copyright 2017 Elsevier. (see color insert)

Figure 5. Comparison of the structure models of Ca2+-bound Cut190* in complex with TETETET (a) and BABSBA (b). Reproduced with permission from ref (14). Copyright 2017 Elsevier. (see color insert) 69

Table 1. Putative Residues Interacting with BABSBA in the Model Structure of Cut190*a. Reproduced with permission from ref (14). Copyright 2017 Elsevier. CaCa2+-free Cut190*

Ca2+-bound Cut190*

Ca2+-free Cut190*

Ca2+-bound Cut190*

Gly105

9

9

His175

0

7

Phe106

10

10

Ser176

10

10

Thr107

9

6

Met177

10

9

Ala108

5

10

Glu184

6

1

Ser109

0

1

Trp201

10

10

Gly111

0

1

Asn202

8

1

Ser112

0

7

Leu203

8

9

Arg135

10

2

Asp204

9

1

Leu136

3

0

Ile224

10

10

Asp137

10

0

His254

0

10

Gln138

10

10

Phe255

0

10

Pro139

9

3

Asn258

0

8

Gly140

2

1

Ile259

0

3

Arg142

1

0

a Numbers of putative BABSBA conformations bound to the residues among the 10 candidates are shown. When the distance between the heavy atom of BABSBA and that of a residue is within 4 Å, the residue is regarded as interacting with BABSBA.

Construction of Mutant Derivatives of Cut190* and Their Activities Based on the model structure of Cut190* in complex with the substrate, we selected certain residues to analyze their effects on the catalytic activity and divided them into three groups: (1) vicinity of the oxyanion hole: Trp201, Phe106, Thr107, Met177, and Gln138; (2) vicinity of the catalytic triad: Ser176-Asp222-His254: Ile224 and Thr223; and (3) amino acids presumably interacted with the Ca2+-bound form but not with the Ca2+-free form opposite the position of the oxyanion hole: Ser176: His175, Phe255, Met258, and Ser112 (14). The 3D structures of these sites with the model of the substrate are summarized in Figure 4. We replaced these residues mainly with Ala, or with other amino acids if necessary, as explained later. In addition to the single residue mutants, we also prepared mutants with simultaneous mutations with I224A. Some mutants were expressed at low levels, possibly because the mutation affected the protein folding. The enzyme kinetic parameters toward PBSA are summarized in Table 2. 70

Table 2. Enzyme Kinetic Parameters of Cut190* and Its Mutant Proteins toward PBSA. Reproduced with permission from ref (14). Copyright 2017 Elsevier. Enzyme

Km (mM)

Vmax (nkat mg–1)

kcat (s–1)

kcat / Km (mM–1 s–1)

Cut190*

0.089 ± 0.001

909 ± 1.0

27 ± 0.2

308 ± 1.0

W201A

0.65 ± 0.01

96.2 ± 2.2

2.9 ± 0.07

4.44 ± 0.17

F106A

0.016 ± 0.001

21.6 ± 2.2

0.65 ± 0.03

40.5 ± 0.50

F106Ya

0.080 ± 0.005

666 ± 1.5

20 ± 0.05

263 ± 11

M177A

no activity

T107A

0.036 ± 0.001

285 ± 1.0

8.7 ± 0.05

239 ± 5.5

Q138A

0.048 ± 0.001

2140 ± 11

65 ± 0.4

1360 ± 2.0

Q138L

no activity

Q138D

0.21 ± 0.001

2000 ± 11

61 ± 0.3

292 ± 0.80

I224A

0.15 ± 0.003

5060 ± 7.0

150 ± 0.2

1000 ± 3.0

I224A/T223V

not determined due to low expression of mutant

I224A/Q138A

0.24 ± 0.002

Group 1

Group 2

5000 ± 7.0

150 ± 0.2

626 ± 1.8

Group 3

a

F255A

not determined due to low expression of mutant

H175A

no activity

N258A

0.065 ± 0.001

853 ± 0.5

26 ± 0.01

395 ± 5.5

N258A/I224A

0.15 ± 0.001

1000 ± 0.5

30 ± 0.05

198 ± 0.50

S112A

0.071 ± 0.001

1000 ± 1.2

30 ± 0.1

426 ± 6.6

low expression level.

Among the group 1 mutations, the W201A mutation remarkably decreased the Vmax value and increased the Km value. The F106A mutation also decreased the activity, while the F106Y mutation retained most of the activity, indicating that the aromatic ring of residue 106 is indispensable. The substitution of Met177 with Ala abolished the catalytic activity, indicating that the residue plays a significant role in supporting a polymer chain. The T107A mutation slightly decreased the activity. The enzymatic activity of the enzyme with the Q138A mutation was significantly increased, while those of the enzymes with the Q138L and Q138D mutations were lost and almost unchanged, respectively. The I224A mutant in group 2 showed significantly increased Km, Vmax, kcat, and kcat/Km values. With the expectation of improved substrate binding by increased hydrophobicity, the I224A/T223V mutant was constructed, but its 71

expression was quite low. In addition, the I224A/Q138A mutant was constructed to determine whether the double mutation could increase the activity. As the result, the Vmax was similar to that of I224A and the Km was larger than those of Q138A and I224A. We performed the same modeling as described previously (Ca2+-bound structure and PBSA), using two mutants. Q138A shows lower GB/SA binding energies (18) and larger surface-binding areas than Cut190*, while those of I224A were approximately similar to those of Cut190*. In group 3, the F255A mutation decreased the expression level. The H175A mutation completely lacked activity, although its expression level was similar to that of Cut190*. The N258A and S112A mutations slightly increased the activities. Taken together, group 1 and Ile224 are requisite to the enzyme structure and activity and the models of the Ca2+-bound structure are more plausible than those of the Ca2+-free structure. The results of the group 3 mutations also supported that the Ca2+-bound structure represents the active state, as proposed previously (8). The indole ring of Tryp201 and the phenyl groups of Phe106 in group 1 are supposed to play an important role for substrate binding, especially for the terephthalate binding. In addition, His175 and Phe255 in group 3 are important for the substrate binding and activity. These four aromatic residues may prefer to bind an aromatic polyester substrate in the active site (14). Subsequently, a robust mutant, Q138A, was obtained.

Novel Ca2+-Binding Sites Are Identified in the Crystal Structure The inactive mutant of Cut190*, Cut190*S176A, was overexpressed and purified well, which was successfully crystalized. The crystal structure, in which two molecules are contained in the asymmetric unit, was determined at 1.6 Å resolution (PDB ID: 5ZNO). The structure clearly demonstrated that each protomer binds three Ca2+ ions (sites 1, 2, and 3), which are confirmed by the strong electron density irrelevant for water molecule, and/or coordinated geometries and distances between amino acids or water molecules (20) (Figure 6). The overall structure is almost the same as that of Ca2+-bound Cut190 S226P previously reported (10) (root-mean-square deviation of 0.27 and 0.25 Å for each protomer) (21). Site 1 is the same site as for the structure of Cut190 S226P. The Ca2+ ion is coordinated by the main-chain carbonyls of Ser76, Ala78, and Phe81 and three waters forming an octahedral geometry. This coordination geometry is almost identical to the case of Cut190 S226P with the exception of one more well-coordinated water molecule. Site 2 is contributed by the residues of the edge of β7, β8, and β9 of Glu220, Asp250, and Glu296, respectively, which is the same as the Ca2+-binding site indicated in Est119 (PDB ID: 3WYN). An additional three coordinated waters form heptacoordinated geometry. The coordinated water molecule forms a hydrogen-bond network with the adjacent protein molecule in the crystal. Site 3 is located at the loop between β6 and 7, which has not been documented to date. The side-chains of Asp204 and Thr206 and the main-chain carbonyl of Thr206 contribute to the Ca2+-binding, and an additional four waters create the octacoordinated geometry. 72

Figure 6. The crystal structure of Cut190*S176A in complex with Ca2+(5ZNO). The orange spheres indicate Ca2+molecules. (see color insert)

In addition, the structure reveals one more bound Ca2+ ion at the interface region of both protomers, but the Ca2+ ion interacts with the protein residues through coordinated water. This Ca2+ ion would be bound by the crystal packing and contribute to stabilizing the crystalline state of the protein. Although a potential ligand molecule of mono(ethylene terephthalate) was added in crystallization, no electron densities for the ligand were observed. To determine the binding stoichiometry of Ca2+ to Cut190*S176A, we measured a mass of Cut190*S176A-Ca2+ complex by native mass spectrometry (21). In the mass spectrum of Cut190*S176A, the ion series showed the mass of 29,163 Da, corresponding to the theoretical mass for Cut190*S176A of 29,165 Da. When Cut190*S176A reacted with CaCl2, the additional peaks were observed with a mass increase of 38 Da, 76 Da, and 115 Da, corresponding in mass to noncovalent binding of one, two, and three Ca2+, which supported the result of X-ray crystallography. We also observed the same results in His-tagged Cut190*S176A. These data endorse the result of X-ray crystallography described previously. On the other hand, the folding thermodynamics of Cut190*S176A showed the dependence of the denaturation temperature on a Ca2+ concentration, due to the enthalpy and entropy change (22). Molecular dynamics simulations indicated that the Ca2+-bound structure fluctuated less than the Ca2+-free structure, supporting that the Ca2+-bound structure is more stable than the Ca2+-free structure, and the active state of Cut190. The roles of sites 1-3 for activation and thermostabilization were determined by mutational analyses of amino acids involved in sites 1-3 (23). 73

Conclusion The present chapter describes the 3D modeling of the protein structures for Cut190* with two kinds of model compounds: BABSBA for PBSA and TETETET for PET. Based on the 3D modeling, we predicted the amino acids relevant to the binding of the model compounds and mutated the predicted amino acids to alanine. When the expression and/or the activity of a mutant was decreased or lost, an amino acid mutated to alanine is considered to be related to the substrate binding. Expression and kinetic values of the constructed mutants were in well accordance with the prediction, based on the 3D modeling. Two mutants, Q138A and I224A, showed the improved kinetic values. On the other hand, the crystal structure of Cut190*S176A, an inactive mutant of Cut190*, in complex with Ca2+, was newly solved, indicating that three Ca2+ ions are bound to sites 1-3 (Protein ID: 5ZNO). The mass spectrometry of Cut190*S176A endorsed the X-ray crystallography of Cut190*S176A in complex with Ca2+. The roles of sites 1-3 were suggested for activation and thermostabilization.

Acknowledgments This work was partially supported by the Platform Project for Supporting Drug Discovery and Life Science Research (Platform for Drug Discovery, Informatics, and Structural Life Science) from Japan Agency for Medical Research and Development (AMED), and by Platform Project for Supporting Drug Discovery and Life Science Research (Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS)) from AMED under Grant Number JP17am0101001. The authors thank the Institute for Fermentation, Osaka, Japan for financial support to Masayuki Oda.

References 1. 2. 3. 4. 5. 6. 7.

8.

Purdy, R. E.; Kolattukudy, P. E. Biochemistry 1975, 14, 2832–2840. Carvalho, C. M.; Aires-Barros, M. R.; Cabral, J. M. Biotechnol. Bioeng. 1999, 66, 17–34. Jaeger, K.-E.; Dijkstar, B. W.; Reetz, M. T. Annu. Rev. Microbiol. 1999, 53, 315–351. Pio, T. F.; Macedo, G. A. Adv. Appl. Microbiol. 2009, 66, 77–95. Chin, I. S.; Abdul Murad, A. M.; Mahadi, N. M.; Nathan, S.; Abu Bakar, F. D. Protein Eng. Des. Sel. 2013, 26, 369–375. Hu, X.; Thumarat, U.; Zhang, X.; Tang, M.; Kawai, F. Appl. Microbiol. Biotechnol. 2010, 87, 771–779. Thumarat, U.; Kawabata, T.; Nakajima, M.; Nakajima, H.; Sugiyama, A.; Yazaki, K.; Tada, T.; Waku, T.; Tanaka, N.; Kawai, F. J. Biosci. Bioeng. 2015, 120, 491–497. Kawai, F.; Oda, M.; Tamashiro, T.; Waku, T.; Tanaka, N.; Yamamoto, M.; Mizushima, H.; Miyakawa, T.; Tanokura, M. Appl. Microbiol. Biotechnol. 2014, 98, 10053–10064. 74

9. 10. 11. 12. 13.

14. 15. 16. 17. 18. 19.

20. 21.

22. 23.

Kitadokoro, K.; Thumarat, U.; Nakamura, R.; Nishimura, K.; Karatani, H.; Suzuki, H.; Kawai, F. Polym. Degrad. Stabil. 2012, 97, 771–775. Miyakawa, T.; Mizushima, H.; Ohtsuka, J.; Oda, M.; Kawai, F.; Tanokura, M. Appl. Microbiol. Biotechnol. 2015, 99, 4297–4307. Roth, C.; Wei, R.; Oeser, T.; Then, J.; Föllner, C.; Zimmermann, W.; Sträter, N. Appl. Microbiol. Biotechnol. 2014, 98, 7815–7823. Sulaiman, S.; You, D. J.; Kanaya, E.; Koga, Y.; Kanaya, S. Biochemistry 2014, 53, 1858–1869. Ribitsch, D.; Hromic, A.; Zitzenbacher, S.; Zartl, B.; Gamerith, C.; Pellis, A.; Jungbauer, A.; Lyskowski, A.; Steinkellner, G.; Gruber, K.; Tscheliessnig, R.; Acero, E. H.; Guebitz, G. M. Biotechnol. Bioeng. 2017, 114, 2481–2488. Kawabata, T.; Oda, M.; Kawai, F. J. Biosci. Bioeng. 2017, 124, 28–35. Kawabata, T. Proteins 2010, 78, 1195–1211. Kawabata, T.; Nakamura, H. J. Chem. Inf. Model. 2014, 54, 1850–1863. Fukunishi, Y.; Mikami, Y.; Nakamura, H. J. Phys. Chem. B 2003, 107, 13201–13210. Shimba, N.; Kamiya, N.; Nakamuira, H. J. Chem. Inf. Model. 2016, 56, 2005–2012. Hisano, T.; Kasuya, K.; Tezuka, Y.; Ishii, N.; Kobayashi, T.; Shiraki, M.; Oroudjev, E.; Hansma, H.; Iwata, T.; Doi, Y.; Saito, T.; Miki, K. J. Mol. Biol. 2006, 356, 993–1004. Harding, M. M. Crystallography 2001, 57, 401–411. Numoto, N.; Kamiya, N.; Bekker, G.-J.; Yamagami, Y.; Inaba, S.; Ishii, K.; Uchiyama, S.; Kawai, F.; Ito, N.; Oda, M. Biochemistry In press. DOI:10.1021/acs.biochem.8b00624. Inaba, S.; Kamiya, N.; Bekker, G.-J.; Kawai, F.; Oda, M. J. Therm. Anal. Cal. In press. DOI:10.1007/s10973-018-7447-9. Oda, M.; Yamagami, Y.; Inaba, S.; Oida, T.; Yamamoto, M.; Kitajima, S.; Kawai, F. Appl. Microbiol. Biotechnol. Under review.

75

Chapter 6

Conjugates Based on Enzyme-Metal-Organic Frameworks for Advanced Enzymatic Applications Qian Liu1 and Cerasela Zoica Dinu*,1 1Department of Chemical and Biomedical Engineering, Benjamin M. Statler College of Engineering and Mineral Resources, West Virginia University, Morgantown, West Virginia 26506, United States *E-mail: [email protected].

Enzymes are biocatalysts that are widely studied for various applications—from catalysis to industrial manufacturing to health care. However, the implementation of enzymes for these applications is currently restricted due to their instability in synthetic operational conditions, short shelf life, and limited applicability for multiproduct catalysis as a result of their unique substrate specificity. Enzyme immobilization has offered an alternative and effective strategy to overcome such restrictions. In this review, metal-organic frameworks (MOFs) are introduced as a novel platform for enzyme immobilization. Aspects related to MOFs’ structure, synthesis, and interaction with enzymes are discussed. Implementation strategies for enzyme-MOF conjugates and specific examples of applications are also highlighted.

Introduction Enzymes are promising candidates for food production (1–4), the industrial manufacturing of textiles (5–7), paper and pulp (8–10), pharmaceuticals (11–14), and chemicals (15–17), as well as for mitigating contamination (18–20) due to their high efficiency, selectivity, and specificity, as well as the mild operational and reaction conditions that enzymes require and their low environmental © 2018 American Chemical Society

and physiological toxicity, all relative to artificial catalysts. Nonetheless, the large-scale implementation of enzymes is currently hindered by their instability in in vitro operational conditions, especially considering that their lifetime is mostly preserved in water-based environments and in controlled temperature and pH conditions (21, 22). Immobilizing enzymes on solid supports has offered an alternative for user-controlled strategies capable of allowing enzyme integration in synthetic applications (23, 24). Polymers (25–28), hydrogels (29, 30), and various inorganic materials (31–34) were considered as supports for the formation of highly functional enzyme-based conjugates. Compared with their free forms, the immobilized enzymes in such conjugates were shown to possess improved storage capabilities and increased thermal and solvent stability (35–37). Performance enhancement was reported when enzymes were immobilized onto supports with controlled physicochemical properties (38, 39). Specifically, the physical (e.g., pore structure, curvature, aspect ratio, water affinity) and chemical (e.g., organic groups present on the surface, chemical stability, oxidation states) properties of the solid supports were shown to significantly affect the enzyme-interface reactions and to influence individual biocatalyst performance as measured by the yield of product generation. However, drawbacks (e.g., enzyme leaching, low loading at the support interface, low activity upon an enzyme’s active-site deformation, or enzyme denaturation) restrict mass transfer. Difficulties for large-scale applications continue to restrict the implementation of such conjugates (40–43). New supports are needed for enzyme immobilization. Such supports should not only ensure robust biocatalysis capable of operating in a variety of pHs and temperature conditions, but should also be designed to support integration and implementation in a variety of synthetic environments.

Metal-Organic Frameworks: An Adaptable Platform for User-Controlled Applications Metal-organic frameworks (MOFs) are a novel class of crystalline materials formed by linking metals containing units with organic linkers through strong bonds (reticular synthesis) (44). The combination of inorganic and organic components endows MOFs with diversity in geometry, topology, and functionality (45–47). MOF-synthesis strategies are generally based on user-controlled hydrothermal methods (48) and room temperature mixing (49), and are shown to determine a wealth of combinations of organic and inorganic units as well as their integration through self-assembly, all while taking advantage of the physical and chemical properties of the individual components. Microwave-assisted synthesis was also used and was shown to produce MOFs of reduced sizes in less time (50) and with increased catalytic (51) and adsorption capabilities (52, 53). The extra energy was the result of increasing the temperature. Synthesis time was also shown to significantly affect the morphology (54) and crystal 78

structure (55) of these MOFs. For example, the research showed that UiO-66, an MOF made up of [Zr6O4(OH)4] clusters with 1,4-benzodicarboxylic acid struts and several of its derivatives, could be synthesized at room temperature, albeit with a high occurrence rate of defects. The amount of defects could, however, be reduced by increasing the synthesis temperature (55). Additionally, when common benzenedicarboxylate was replaced with the 2-aminoterephthalic acid ligand, the resulting UiO-66-NH2 derivative was shown to possess extra functionalities, such as enhanced photoactivity (56) and catalytic activity in an aldol condensation reaction (57), as well as an increased capability for removing NO2 (58)—all relative to its UiO-66 counterpart. It has also been reported that the composition and porosity of UiO-66 can be tuned by manipulating the individual modulators used during its synthesis (59, 60) to achieve enhanced gas adsorption (61) as well as increased separation and catalytic activity (62). The specific uniformity of MOFs as a result of their tight synthesis conditions as well as their controllable pore size and structure were shown to allow for encapsulation of different analytes, thus resulting in the formation of MOF-based composites. Such composites subsequently benefited from increased stability while allowing application-specific versatility in catalysis or biomedicine. Zhang et al. (63), for example, encapsulated uniform NiPt-alloy nanoclusters within the opening porous channels of MIL-101(Cr) formed through the catalytic dehydrogenation of hydrazine borane (N2H4BH3) and hydrous hydrazine (N2H4·H2O). The authors showed superior H2 synthesis and complete H2 evolution (100% H2 selectivity). Morris et al. (64) created a new and modular class of nanostructures for nucleic acid-based assembly strategies. The authors showed that by using a controlled click reaction between the surface of the lab-synthesized UiO-66-N3 and specifically designed oligonucleotides, enhanced cellular uptake of the DNA-UiO-66-N3 could be achieved. Complementarily, Zhang et al. (65) reported on an aptamer-embedded Zr-based MOF composite for the detection of several analytes, including thrombin, kanamycin, and carcinoembryonic antigen. Lastly, Chen et al. (66) successfully immobilized insulin in an acid-resistant crystalline mesoporous MOF (i.e., NU-1000). The authors showed loadings of up to 40% in only 30 min of MOF exposure; it was also revealed that the synthesized MOFs were effective at preventing the degradation of insulin in the presence of pepsin, a digestive enzyme.

MOFs as Supports for the Formation of Enzyme-Conjugates Recently, MOFs have garnered attention as supports for enzyme immobilization due to research speculating that the variable framework structure, which could theoretically assume an infinite number of geometries as a result of the organic/inorganic materials and linker combinations, will allow for novel, enzymatic-based system design while ensuring their improved stability and catalytic efficiency. The main strategies used for the formation of such enzyme-MOF conjugates generally rely on surface adsorption, covalent binding, and pore and in situ encapsulation (Figure 1). 79

Figure 1. Schematic illustration of the strategies used for the formation of enzyme-MOF conjugates. A model enzyme and a model MOF support are shown; for scale, the pore of the model MOF is of similar size to that of the model enzyme. Surface adsorption is the most direct method for enzyme immobilization and the formation of enzyme-MOF conjugates. The process relies on physical interactions based on the van der Waals forces, as hydrophobic and/or electrostatic forces with such interactions control the overall binding efficiency of the two systems. The method was shown to be feasible for all kinds of enzyme-MOF combinations and does not account for enzyme specificity or the synthetic conditions encountered at MOF preparation (67–69). For example, specific studies have shown that the catalytic yield, reusability, and storage stability could be significantly enhanced for conjugates prepared by physical interactions; such conjugates also displayed minimal enzyme leaching. Liu et al. (70) showed that a trypsin–NBD (4-chloro-7-nitrobenzofurazan)-UiO-66 bioreactor formed by physical adsorption of the enzyme exhibits high proteolytic performance while also displaying high reusability (up to five consecutive uses), as well as consistency and stability for a long period (up to 30 days). Other study (71) has also shown that porcine pancreatic lipase (PLL) that is physically adsorbed onto UiO-66(Zr), UiO-66-NH2(Zr), MIL-53(Al), and carbonized MIL-53(Al) retained high catalytic performance and in such conjugates, minimum leaching was observed. MIL-53 is made up of Sc- and O-nodes (ScO6) with 1,4-benzodicarboxylic acid struts between them. MIL-53 is made up of Sc- and O-nodes (ScO6) with 1,4-benzodicarboxylic acid struts between them. Covalent binding relies on a chemical bond between the enzyme and the given MOF support. For this, functional groups (such as amino, carboxylate, or hydroxyl groups) derived intrinsically or through user-specific modifications on 80

the MOF’s surface could be coupled with the reactive groups on the surface of the enzyme (68, 72–74). Enzyme-MOF conjugates obtained through covalent immobilization strategies were shown to have good reusability for protein digestion and retained almost all of their activity even after four individual digestion cycles (75). Further, covalent binding of Lipase B Candida Antarctica (CAL-B) onto 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide- (EDC) or N,N′-Dicyclohexylcarbodiimide-(DCC) activated three-dimensional IRMOF-3 composed of Zn4O nodes and dicarboxylate organic linkers was shown to lead to increased enzymatic activity. Specifically, the immobilized enzyme achieved 103-fold higher activity than the free enzyme in the solution. The high activity was attributed to the zero-bound interaction formed between the enzyme and the MOF support, which limited enzyme denaturation and active site deformation at the support interface (76). Enzyme encapsulation into user-designed MOF structures was also considered due to (1) the high pore-volume aspect ratio and overall void space of the MOFs that could potentially enable high loadings while conferring increased stability and reusability; (2) uniform and controllable pore-size synthesis that could provide size selectivity for enzyme-driven specific reactions; and (3) controlled encapsulation that could provide a protective screen of sorts from physical or chemical denaturation. Lykourinou et al. (77) demonstrated the feasibility and efficiency of encapsulating microperoxidase-11 (MP-11) in mesoporous MOFs (Tb-TATB). The authors showed that the conjugates exhibited high catalytic activity, recyclability, and solvent adaptability. Further, the results also demonstrated that such a strategy could be extended to enzymes with larger molecular dimensions than the pore size of the MOFs used with subsequent characterization showing conformational changes occurring as the enzyme translocated the crystalline structure of the MOF (78). Complementarily, Feng et al. (79) used H3BTTC or H3TATB as linking agents with MCl3 (M = Al, Fe, V, Sc, and In) as base atoms to develop a series of ultra-large mesoporous cages with a 5.5-nm diameter to lead to metal ions that afforded octahedral crystals of PCN-332(M) (M = Al, Fe, Sc, V, and In) or PCN-333(M) (M = Al, Fe, and Sc). Such MOFs were then used as single-molecular traps and were shown to display high loading capabilities and recyclability for the encapsulation of enzymes for horseradish peroxidase (HRP), Cyt c, and MP-11. The displayed catalytic activity of the encapsulated enzymes was comparable to the activity of their free counterparts; further, an increase in stability under harsh conditions (e.g., imposed by exposure to tetrahydrofuran solutions) was noted. Moreover, in the same group, a hierarchical, porous MOF structure was used for a tandem nanoreactor (80). In the reactor formed from PCN-888, the relatively large, 6.2-nm cavity accommodated glucose oxidase (GOx) entrapment, while an intermediate cavity of 5.0 nm was used for HRP encapsulation. Complementarily, a relatively small cavity of 2.0 nm was integrated to remain open during the catalytic reactions of GOx and HRP to serve for substrate diffusion. A PCN-888 structure was formed from linking Al3+ and a heptazine-based tritopic ligand (HTB) through a coordination reaction. The tandem nanoreactor was not only shown to display excellent catalytic performance for the enzyme-initiated chain reactions, but was 81

also shown to considerably limit enzyme leaching for any of the two entrapped biocatalysts. Lastly, in situ processing allowed for enzyme immobilization during MOF synthesis itself; however, immobilization was only shown to be feasible under mild MOF preparation conditions (i.e., MOF synthesis at room temperature and in water-based solvents). During such processes, the enzyme and MOF precursor were generally mixed together to allow for the enzyme to be directly embedded within the MOF crystals (68, 72). ZIF-8 is considered the most common MOF to be used for such a strategy as its zeolitic imidazolate framework (with sodalite topology formed by linking 2-methylimidazole and Zn ions through coordination bonds) allows for synthesis in mild conditions. Cyt c is a highly conserved monohemic protein that serves as an electron shuttle between the respiratory complexes III and IV in the inner mitochondrial membrane (IMM) while also being involved in the generation, oxidation, and trapping of reactive oxygen species (ROS, (81)); Lyu et al. (82) first proposed the in situ encapsulation of Cyt c with ZIF-8 with the assistance of surfactants. Highly enhanced activity relative to the free enzyme counterparts was achieved. The produced Cyt c-ZIF-8 conjugates also exhibited convenient, fast, and highly sensitive detection of trace amounts of explosive organic peroxides in the solution. Wu et al. (83) showed that when ZIF-8 was used for in situ encapsulation of GOx-HRP conjugates, a cascade of enzymatic chain reactions was initiated with the yield of reaction being similar to that generated by the two free enzymes in the solution. Further, the authors showed that the MOF structure provided a three-dimensional microenvironment that ensured both enzymes’ protection as well as preservation of their activity while minimizing individual leaching. Lastly, Liang et al. (84) showed that by encapsulating bovine serum albumin (BSA) in different MOFs including ZIF-8, Cu3(BTC)2 (HKUST-1), Eu2(1,4-BDC)3(H2O)4, Tb2(1,4-BDC)3(H2O)4, and Fe(III) dicarboxylate MOF (MIL-88A)), conjugates with high efficiency and improved catalytic yield could be achieved.

Improved Performance of the Enzyme-MOF Conjugates The synthesis of enzyme-MOF conjugates was shown to not only result in enhanced enzymatic activity and environmental resistance, but also to allow for increased recyclability of the immobilized enzymes. For example, Lyu et al. (82) embedded Cyt c into ZIF-8. The in situ encapsulation showed that at the same enzyme concentration the embedded Cyt c displayed a 10-fold increase in its activity relative to the free counterpart exposed to the same experimental conditions (Figure 2a). The authors also demonstrated that the conjugates’ activity was enhanced, with specific results showing that the mixture of Cyt c and ZIF-8 had almost the same activity as the free Cyt c. On the contrary, the presence of polyvinylpyrrolidone (PVP) and Zn2+ in the solution only increased the activity of Cyt c by 79% and 57%, respectively. In control, the presence of 2-methylimidazole or ZIF-8 in the solution had no effect on the activity of Cyt c. Also, in control ZIF-8 had no catalytic activity toward the enzymatic substrates. 82

Figure 2. (a) The relative peroxidase activity of Cyt c, Cyt c-ZIF-8 conjugates, polyvinylpyrrolidone-Cyt c mixture, Cyt c-Zn ion mixture, Cyt c-2-methylimidazole mixture, and Cyt c-ZIF-8 mixtures. Reprinted with permission from ref. (82). Copyright 2014 American Chemical Society. (b) Thermal stability and (c) long-term storage stability in dry powder form (black bars represent the OPAA-PCN-128y conjugates while red bars represent the free OPAA respectively) as measured by the conversion of the diisopropyl fluorophosphate (DFP) substrate. Reprinted with permission from ref. (85). Copyright 2016 American Chemical Society. (d) Reaction rates of MP-11-Tb-mesoMOF and MP-11-MCM-41 at different cycles. Reprinted with permission from ref. (77). Copyright 2011 American Chemical Society.

The enhanced stability of the immobilized enzyme was demonstrated when a conjugate of organophosphorus acid anhydrolase (OPAA) and porous nanocage (PNC)-128y was used (85). OPAA is a prolidase enzyme that catalyzes the hydrolysis of P−F, P−O, P−CN, and P−S bonds commonly found in toxic organophosphorus compounds and G-type chemical agents. The PCN-128 MOF structure is constructed from H4ETTC11 (4′,4¢¢¢,4¢¢¢′′,4¢¢¢′¢¢¢-(ethene-1,1,2,2tetrayl)tetrakis-(([1,1′-biphenyl]-4-carboxylic acid))) and eight connected Zr6 clusters and has a pore size of 4.4 nm (86). The resulting OPAA-PNC-128y conjugates displayed higher thermal and storage stability relative to free OPAA in the solution when the hydrolytic degradation of diisopropyl fluorophosphate (DFP) was investigated as well as over a wide range of incubation and reaction temperatures (Figure 2b). Specifically, free OPAA incubated at 55 °C showed a significant loss of activity, while OPAA-PCN-128y conjugates yielded around 90% activity. When the temperature was further increased to 70 °C, 83

the OPAA-PCN-128 conjugates retained a conversion capability of almost 75%, which was in contrast to the conversion capability retained by the free enzyme that was shown to be completely denatured in the given experimental conditions. Additionally, the storage stability of the OPAA-PCN-128 conjugates was evaluated by determining the activity of the conjugates upon storing them in air-dried conditions. Analysis showed that OPAA-PCN-128y remained capable of catalyzing the hydrolysis of 90% of the present DFP even after 3 days of dry storage. On the contrary, the lyophilized, free OPAA exhibited only 30% catalysis capability (Figure 2c). It was concluded that the enhanced stability of the OPAA-PCN-128 conjugates resulted from PCN-128 protecting the enzyme from exposure to extreme environments. Increased resistance to organic solvents, metal ions, and/or digestion was also shown for enzyme-MOF based conjugates. Specifically, studies showed that the HRP-ZIF-8 conjugate exhibited high activity, and that exposure to trypsin, boiling water, and N,N-dimethylformamide (DMF) did not significantly affect the enzyme’s performance. This was attributed to the protective effect of the ZIF-8 matrix 1. He et al. (87) also demonstrated such protective effects by encapsulating lipase in ZIF-8 structures; the resulting lipase-ZIF-8 conjugates showed significantly enhanced tolerance to various metal ions (e.g., Zn2+, Mn2+, Ni2+, and Cu2+) as well as exposure to high temperatures. The relative activities of lipase-ZIF-8 conjugates decreased by only 9.6%, 14.8%, 18.4%, and 21.5% relative to activity losses of 46.2%, 48.9 %, 40.1%, and 58.6%, respectively, which were recorded for the corresponding free lipase. Also, lipase-ZIF-8 conjugates retained 50% of original activity at 50 °C after 50 min of incubation and 63% of original activity at 70 °C after 5 min of incubation, compared to the complete loss of activity for free lipase QLM. Lastly, the recyclability/reusability of enzyme-MOF conjugates was also demonstrated. The pioneering work of Lykourinou et al. (77) showed that by encapsulating MP-11 into Tb-TATB and using mesoporous silica material MCM-41 as a counterpart (Figure 2d), the reaction rate of the conjugates fluctuated from 5.40×10-5 to 8.34×10-5 mM/s in the first six cycles. Also, the conjugates only lost about 53% of their activity, even after the seventh cycle. In comparison, the activity of MP-11-MCM-41 conjugates decreased abruptly, with results showing more than a 60% activity loss after only the first cycle and only 28% of the original activity remaining by the third cycle. The retained activity was attributed to the strong hydrophobic interactions between the Tb-meso MOF framework and the MP-11 molecules, which trap the enzyme into the hydrophobic MOF cages thus preventing its escape while ensuring a protective shield against denaturation. Further, Liu et al. (71) established a novel enzyme-MOF system for the clinical synthesis of warfarin. Their strategy was based on immobilizing lipase onto UiO-66 (Zr). The resulting enzyme-MOF conjugate showed high activity and reusability, with about 76% warfarin formation for the first cycle which reduced to 58% by the fifth cycle, all relative to the conversion ability of the free lipase. The enhanced recyclability was presumably due to the well-known stabilization effect induced by the physical interaction between the enzyme and MOF supports (88). 84

Selected Applications of Enzyme-MOF Conjugates The significantly enhanced performance of enzyme-MOF conjugates in terms of activity, stability, and reusability increased their potential for numerous applications. However, large-scale implementation of such conjugates is still hampered by the loss of their individual (i.e., enzymatic or MOF) functionality and the loss of their combined catalytic power under high temperature and variable pH conditions, both of which are encountered in common manufacturing conditions. As such, enzyme-MOF conjugate applications are mostly limited to the bioreactor development stage for biomedicine or biosensors, where their chemical stability and thermodynamic compatibility are supported by the relatively mild operating conditions. Specific examples are shown in Figure 3. Figure 3a illustrates the proposed strategy for an enzyme-MOF bioreactor formation to be used for industrial textile manufacturing and environmental treatments. Shieh et al. (89) encapsulated catalase in ZIF-90 structures to build a protective shelter of sorts that prevented the proteolytic degradation of the enzyme; such proteolytic degradation is normally encountered during silk degumming. The ZIF-90 shelter was considered as a model due to its formation by coordinating Zn-ions with imidazole-2-carboxaldehyde, resulting in a zeolitic structure of a porous nature. The pore size of such structures is smaller than the size of the protease enzyme, but larger than that of the H2O2 catalase’s substrate. The authors showed that the shielding was effective at protecting against enzyme unfolding. Further, the shielding limited the denaturation of the enzyme at the interface with the MOF support while ensuring high enzyme stability in both high temperatures and even urea conditions (90). A complementary study (87) embedded lipase in ZIF-8 structures to be used as conjugates for ester hydrolysis. The analysis showed that the lipase-ZIF-8 conjugates exhibited high catalytic activity and stability in the ester hydrolysis conditions while also displaying favorable enantioselectivity and reusability in the kinetic resolution of secondary alcohols and in nonaqueous mediums. Figure 3b shows an example of a bioreactor to be used in biomedicine. Lian et al. (91) developed a tyrosinase‐MOF nanoreactor to activate the prodrug paracetamol for cancer therapy. The PCN-333 support used consisted of an atrimeric-oxo cluster and a planar triangular ligand (4,4′,4′′-s-triazine-2,4,6-triyltribenzoic acid (TATB)) that self-assembled into a supertetrahedron (STH) which exposed the metal clusters on the corners while the TATB formed the tetrahedron face. The synthesized MOF structure displayed two types of cavities with diameters of 4.2 nm and 5.5 nm, respectively. By generating ROS and depleting glutathione (GSH), the product of the enzymatic conversion of paracetamol was expected to be toxic to the cancer cells. Indeed, the analysis showed that the proposed tyrosinase‐PCN-333 nanoreactors caused significant cell death in the presence of paracetamol for up to 3 days after being internalized. The toxicity was measured by the 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide assay. Conversely, the free enzymes experienced a dramatic reduction of activity in only a few hours upon encountering the cellular platforms. 85

Figure 3. Schematic diagram of typical applications of enzyme-MOF conjugates. (a) A bioreactor for industrial manufacturing and environmental treatments. Reprinted with permission from ref. (89). Copyright 2015 American Chemical Society. (b) A bioreactor for biomedicine. Reprinted with permission from ref. (91). Copyright 2018 Wiley‐VCH. (c) A biosensor using enzyme-MOF conjugates. Reprinted with permission from ref. (93). Copyright 2013 American Chemical Society.

86

In a complementary study, Liang et al. (92) coated live eukaryotic cells with a bioactive shell; the coating was achieved by growing a ZIF-8 film directly onto the β-galactosidase-coated cell surface. The surface enzyme is known to be involved in the metabolization of nutrients. Analysis showed that the MOF film not only protected the enzymes, but also enabled the cells to survive in a simulated lactosebased environment for more than 7 days. The cells also displayed only a 30% decrease in their viability as compared to a 99% decrease for the control (naked, uncoated cells). Lastly, Figure 3c shows an example of enzyme-MOF conjugates used for the development of a biosensor. Ma et al. (93) employed ZIFs as matrices for the coimmobilization of methylene green (MG) and glucose dehydrogenase (GDH) enzymes and electrochemical biosensor formation. The combination of MG as an electrocatalyst and GDH as a detector was shown to allow for continuous measurement of a typical amperometric response. Further, it was shown that such a response depended on the conversion between glucose and gluconolactone, which were used as substrates. Sensitivity was evaluated under continuous-flow conditions using in vivo microdialysis and was shown to allow for glucose recordings in real time with a linear range of 0.1−2.0 mM. Moreover, the ZIF-based biosensor was highly selective to glucose more so than other endogenous, electroactive species such as sodium ascorbate (SA), 3, 4-dihydroxyphenylacetic acid (DOPAC), dopamine (DA), uric acid (UA), and 5-hydroxytryptamine (5-HT)—all commonly employed for measurements in the cerebral system. In a complementary study, Wang et al. (94) developed an efficient biosensor for ultrasensitive and rapid detection of bisphenol A (BPA) using a Cu-MOF support and a tyrosinase enzyme. The Cu-MOF ([Cu(bdc)(ted)0.5]·2DMF·0.2H2O) was formed from 1,4‐benzenedicarboxylic acid (H2bdc), triethylenediamine (ted), N, N‐dimethylformamide (DMF) and was formed via hydrothermal reaction. Tyrosinase is a copper-containing enzyme that can catalyze the oxidation of BPA using molecular oxygen (95). The resulting biosensor showed a high sensitivity of 0.2 A M-1, a wide linear range of 5.0×10-8 to 3.0 × 10-6 M, and a low detection limit of only 13 nM. The excellent biosensing performances were attributed to the large surface area and pore structure of the Cu-based MOF conjugates that not only allowed for high enzyme loading but also for increased catalytic yield of the immobilized enzyme. The Cu-MOF-based bisphenol biosensor had a response time for BPA detection of less than 11 s, and not only displayed selectivity for bisphenols but also for heavy metal ions (96). Patra et al. (97) also developed a sensitive biosensor for the detection of glucose by using MIL-100(Fe) and Pt-nanoparticle composites as supports for GOx immobilization. The resulting biosensor displayed a high sensitivity of 71 mA M-1 cm-2, a low limit of detection of only 5 µM, and a low response time of less than 5 s. The positive results were attributed to the improved performance of the enzyme due to the transfer of the Fe ions from the center of the MIL-100(Fe) structure.

87

Conclusions and Outlook MOFs diversity in structure and in physicochemical properties, user-guided design and functionality, large surface area, and uniform and controlled pore geometries have all led to their increased implementation as platforms for enzyme immobilization. However, even though the examples discussed in this chapter demonstrated increased enzyme performance due to such MOF interfaces, the field of enzyme-MOF conjugates and their implementation in industry is still in its infancy. The limited analysis of system robustness and stability not only restricts its large-scale implementation, but also caps the potential for extended applications. For instance, surface adsorption cannot effectively avoid enzyme leaching due to the weak enzyme-support interaction. Further, the proximity of the supports leads to enzyme denaturation. Covalent binding can overcome the drawback of the weak binding force imposed by a direct surface contact; however, such binding does not protect the immobilized enzymes from harsh environments, which further hinders implementation. Additionally, while encapsulation successfully avoids enzyme leaching and shields the immobilized enzymes from potentially harsh environments, such an immobilization strategy is limited by the controllability and user-specificity of the synthesized pore sizes that are required to match the size of the enzyme. Lastly, the in situ method limits the ability to synthesize a variety of enzyme-MOF conjugates as it requires mild conditions and water-based precursors. The future of enzyme-MOF conjugates will be dictated by the pace of exploration into the functionality and synthesis methods of MOFs and how such methods could be interfaced with the specific, limited shelf life and compatibility of the enzymatic system being implemented. The mechanisms that could potentially lead to the design of a functional, novel, and highly specific enzyme-MOF conjugate or hierarchical hybrid with increased enzyme performance and implementation capability need to be further investigated. Fully understanding these mechanisms would not only ensure practical applications in ex vivo environments, but would further develop robust, biocatalytic-based industrial processes. Moreover, efficiency and stability would be increased, while the overall cost and labor needed for design, user-preparation training (when large-scale implementation is being considered), and studying the real operational conditions of the enzyme would be reduced. These factors will, in part, define the overall market operational integration and capability. As such, future studies should explore (1) the development of novel, organic linkers that can be easily obtained and integrated at a low cost; (2) large scale production of biocatalysts that are stable in a variety of operational conditions; (3) the development of design strategies that allow for control of the interface between the enzyme and the MOF support to ensure high activity and stability; (4) the overall performance of the immobilized enzymes based on the presumed industrial settings (e.g., types of reactors) and conditions (e.g., temperature and chemical environments); and (5) an economic evaluation analysis of the entire cost of enzyme-based process applications in an industrial setting to enable high yield and adaptable integration of various platforms. 88

Acknowledgments This work was funded by the National Science Foundation (NSF) grant 1454230.

References 1. 2.

3.

4.

5. 6.

7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

Zhou, H. C.; Long, J. R.; Yaghi, O. M. Introduction to Metal-Organic Frameworks. Chem. Rev. 2012, 112, 673–674. Liang, K.; Ricco, R.; Doherty, C. M.; Styles, M. J.; Bell, S.; Kirby, N.; Mudie, S.; Haylock, D.; Hill, A. J.; Doonan, C. J.; Falcaro, P. Biomimetic mineralization of metal-organic frameworks as protective coatings for biomacromolecules. Nat.Commun. 2015, 6. Wang, X.; Lu, X. B.; Wu, L. D.; Chen, J. P. 3D metal-organic framework as highly efficient biosensing platform for ultrasensitive and rapid detection of bisphenol A. Biosens. Bioelectron. 2015, 65, 295–301. Carvalho, E. A.; Goes, L. M. D.; Uetanabaro, A. P. T.; da Silva, E. G. P.; Rodrigues, L. B.; Pirovani, C. P.; da Costa, A. M. Food Chem. 2017, 221, 1499–1506. Hasanbeigi, A.; Price, L. J. Cleaner Prod. 2015, 95, 30–44. Sodaneath, H.; Lee, J. I.; Yang, S. O.; Jung, H.; Ryu, H. W.; Cho, K. S. J. Environ. Sci. Health., A: Toxic/Hazard. Subst. Environ. Eng. 2017, 52, 1099–1111. Gao, A. Q.; Shen, H. W.; Zhang, H. J.; Feng, G. C.; Xie, K. L. J. Cleaner Prod. 2017, 164, 277–287. Hakala, T. K.; Liitia, T.; Suurnakki, A. Carbohydr. Polym. 2013, 93, 102–108. Kim, K. J.; Lee, J. M.; Ahn, E. B.; Eom, T. J. Cellulose. 2017, 24, 3503–3511. Kumar, N. V.; Rani, M. E.; Gunaseeli, R.; Kannan, N. D. Int. J. Biol. Macromol. 2018, 111, 289–295. Apetrei, I. M.; Rodriguez-Mendez, M. L.; Apetrei, C.; de Saja, J. A. Sens. Actuators, B 2013, 177, 138–144. Truppo, M. D. ACS Med. Chem. Lett. 2017, 8, 476–480. Kurek, W.; Koszelewski, D.; Ostaszewski, R.; Zadlo-Dobrowolska, A. Process Biochem. 2017, 60, 92–97. Yuan, Y. S.; Yang, L.; Liu, S. P.; Yang, J. D.; Zhang, H.; Yan, J. J.; Hu, X. L. Spectrochim. Acta, Part A 2017, 176, 183–188. Das, R.; Ghosh, S.; Bhattacharjee, C. Lwt-Food Sci. Technol. 2012, 47, 238–245. Gilman, A.; Laurens, L. M.; Puri, A. W.; Chu, F.; Pienkos, P. T.; Lidstrom, M. E. Microb. Cell Fact. 2015, 14, 182. Bernal, C.; Guzman, F.; Illanes, A.; Wilson, L. Food Chem. 2018, 239, 189–195. Bilal, M.; Iqbal, H. M. N.; Hu, H. B.; Wang, W.; Zhang, X. H. J. Environ. Manage. 2017, 188, 137–143. Bilal, M.; Asgher, M.; Iqbal, H. M. N.; Hu, H. B.; Zhang, X. H. Environ. Sci. Pollut. Control Ser. 2017, 24, 7035–7041. 89

20. Manai, I.; Miladi, B.; El Mselmi, A.; Hamdi, M.; Bouallagui, H. Environ. Technol. 2017, 38, 880–890. 21. Jemli, S.; Ayadi-Zouari, D.; Hlima, H. B.; Bejar, S. Crit. Rev. Biotechnol. 2016, 36, 246–258. 22. Porter, J. L.; Rusli, R. A.; Ollis, D. L. Chembiochem. 2016, 17, 197–203. 23. Es, I.; Vieira, J. D. G.; Amaral, A. C. Appl. Microbiol. Biotechnol. 2015, 99, 2065–2082. 24. Bilal, M.; Asgher, M.; Parra-Saldivar, R.; Hu, H. B.; Wang, W.; Zhang, X. H.; Iqbal, H. M. N. Sci. Total Environ. 2017, 576, 646–659. 25. Sahoo, P. C.; Sambudi, N. S.; Park, S. B.; Lee, J. H.; Han, J. I. Catal. Lett. 2015, 145, 519–526. 26. Zhu, X.; He, B.; Zhao, C. W.; Ma, Y. H.; Yang, W. T. Langmuir 2017, 33, 5577–5584. 27. VandeZande, G. R.; Olvany, J. M.; Rutherford, J. L.; Rasmussen, M. Methods Mol. Biol. 2017, 1504, 165–179. 28. Marschelke, C.; Raguzin, I.; Matura, A.; Fery, A.; Synytska, A. Soft Matter 2017, 13, 1074–1084. 29. Hata, Y.; Sawada, T.; Sakai, T.; Serizawa, T. Biomacromolecules 2018, 19, 1269–1275. 30. Vitola, G.; Buning, D.; Schumacher, J.; Mazzei, R.; Giorno, L.; Ulbricht, M. Macromol Biosci. 2017, 17 doi:10.1002/mabi.201600381. 31. Zhang, X. M.; Jing, L. Y.; Chang, F. F.; Chen, S.; Yang, H. Q.; Yang, Q. H. Chem. Commun. 2017, 53, 7780–7783. 32. Verma, N.; Kumar, N.; Upadhyay, L. S. B.; Sahu, R.; Dutt, A. Anal. Lett. 2017, 50, 1839–1850. 33. Masi, C.; Kumar, K. R. N.; Raja, N. C. G.; Umesh, R. Res. J. Pharm., Biol. Chem. Sci. 2017, 8, 153–161. 34. Khoshnevisan, K.; Vakhshiteh, F.; Barkhi, M.; Baharifar, H.; Poor-Akbar, E.; Zari, N.; Stamatis, H.; Bordbar, A. K. Mol. Catal. 2017, 442, 66–73. 35. Schroeder, M. M.; Wang, Q. M.; Badieyan, S.; Chen, Z.; Marsh, E. N. G. Langmuir 2017, 33, 7152–7159. 36. Fernandez-Lopez, L.; Pedrero, S. G.; Lopez-Carrobles, N.; Gorines, B. C.; Virgen-Ortiz, J. J.; Fernandez-Lafuente, R. Enzyme Microb. Technol. 2017, 98, 18–25. 37. Liese, A.; Hilterhaus, L. Chem. Soc. Rev. 2013, 42, 6236–6249. 38. Yong, J. K. J.; Cui, J. W.; Cho, K. L.; Stevens, G. W.; Caruso, F.; Kentish, S. E. Langmuir 2015, 31, 6211–6219. 39. Secundo, F. Chem. Soc. Rev. 2013, 42, 6250–6261. 40. Wang, Z. G.; Wan, L. S.; Liu, Z. M.; Huang, X. J.; Xu, Z. K. J. Mol. Catal. B: Enzym. 2009, 56, 189–195. 41. Jochems, P.; Satyawali, Y.; Diels, L.; Dejonghe, W. Green Chem. 2011, 13, 1609–1623. 42. Zhou, Z.; Hartmann, M. Chem. Soc. Rev. 2013, 42, 3894–3912. 43. Hartmann, M.; Kostrov, X. Chem. Soc. Rev. 2013, 42, 6277–6289. 44. Zhou, H. C.; Long, J. R.; Yaghi, O. M. Chem. Rev. 2012, 112, 673–674. 45. Yao, Z. Y.; Guo, J. H.; Wang, P.; Liu, Y.; Guo, F.; Sun, W. Y. Mater. Lett. 2018, 223, 174–177. 90

46. An, Y. Y.; Lu, L. P.; Zhu, M. L. Acta Crystallogr. Sect. C: Struct. Chem. 2018, 74, 418–423. 47. Alqadami, A. A.; Khan, M. A.; Siddiqui, M. R.; Alothman, Z. A. Microporous Mesoporous Mater. 2018, 261, 198–206. 48. Jin, H.; Wollbrink, A.; Yao, R.; Li, Y. S.; Caro, J.; Yang, W. S. J. Membr. Sci. 2016, 513, 40–46. 49. Pan, Y. C.; Liu, Y. Y.; Zeng, G. F.; Zhao, L.; Lai, Z. P. Chem. Commun. 2011, 47, 2071–2073. 50. Taddei, M.; Steitz, D. A.; van Bokhoven, J. A.; Ranocchiari, M. Chem.−Eur. J. 2016, 22, 3245–3249. 51. Babu, R.; Roshan, R.; Kathalikkattil, A. C.; Kim, D. W.; Park, D. W. ACS Appl. Mater. Interfaces 2016, 8, 33723–33731. 52. Tari, N. E.; Tadjarodi, A.; Tamnanloo, J.; Fatemi, S. Microporous Mesoporous Mater. 2016, 231, 154–162. 53. Cabello, C. P.; Arean, C. O.; Parra, J. B.; Ania, C. O.; Rumori, P.; Palomino, G. T. Dalton Trans. 2015, 44, 9955–9963. 54. Jiang, H. X.; Wang, Q. Y.; Wang, H. Q.; Chen, Y. F.; Zhang, M. H. Catal. Commun. 2016, 80, 24–27. 55. DeStefano, M. R.; Islamoglu, T.; Hupp, J. T.; Farha, O. K. Chem. Mater. 2017, 29, 1357–1361. 56. Silva, C. G.; Luz, I.; Xamena, F. X. L. I.; Corma, A.; Garcia, H. Chem.−Eur. J. 2010, 16, 11133–11138. 57. Hajek, J.; Vandichel, M.; Van de Voorde, B.; Bueken, B.; De Vos, D.; Waroquier, M.; Van Speybroeck, V. J. Catal. 2015, 331, 1–12. 58. Peterson, G. W.; Mahle, J. J.; DeCoste, J. B.; Gordon, W. O.; Rossin, J. A. Angew. Chem., Int. Ed. 2016, 55, 6235–6238. 59. Shearer, G. C.; Chavan, S.; Bordiga, S.; Svelle, S.; Olsbye, U.; Lillerud, K. P. Chem. Mater. 2016, 28, 3749–3761. 60. Liang, W. B.; Coghlan, C. J.; Ragon, F.; Rubio-Martinez, M.; D’Alessandro, D. M.; Babarao, R. Dalton Trans. 2016, 45, 4496–4500. 61. Jiang, Z. R.; Wang, H. W.; Hu, Y. L.; Lu, J. L.; Jiang, H. L. ChemSusChem. 2015, 8, 878–885. 62. Vermoortele, F.; Bueken, B.; Le Bars, G.; Van de Voorde, B.; Vandichel, M.; Houthoofd, K.; Vimont, A.; Daturi, M.; Waroquier, M.; Van Speybroeck, V.; Kirschhock, C.; De Vos, D. E. J. Am. Chem. Soc. 2013, 135, 11465–11468. 63. Zhang, Z. J.; Zhang, S. L.; Yao, Q. L.; Chen, X. S.; Lu, Z. H. Inorg. Chem. 2017, 56, 11938–11945. 64. Morris, W.; Briley, W. E.; Auyeung, E.; Cabezas, M. D.; Mirkin, C. A. J. Am. Chem. Soc. 2014, 136, 7261–7264. 65. Zhang, Z. H.; Duan, F. H.; Tian, J. Y.; He, J. Y.; Yang, L. Y.; Zhao, H.; Zhang, S.; Liu, C. S.; He, L. H.; Chen, M.; Chen, D. M.; Du, M. ACS Sens. 2017, 2, 982–989. 66. Chen, Y. J.; Li, P.; Modica, J. A.; Drout, R. J.; Farha, O. K. J. Am. Chem. Soc. 2018, 140, 5678–5681. 67. Zhou, H. C.; Kitagawa, S. Chem. Soc. Rev. 2014, 43, 5415–5418. 68. Gkaniatsou, E.; Sicard, C.; Ricoux, R.; Mahy, J. P.; Steunou, N.; Serre, C. Mater. Horiz. 2017, 4, 55–63. 91

69. Wu, X. L.; Hou, M.; Ge, J. Catal. Sci. Technol. 2015, 5, 5077–5085. 70. Liu, W. L.; Wu, C. Y.; Chen, C. Y.; Singco, B.; Lin, C. H.; Huang, H. Y. Chem.−Eur. J. 2014, 20, 8923–8928. 71. Liu, W. L.; Yang, N. S.; Chen, Y. T.; Lirio, S.; Wu, C. Y.; Lin, C. H.; Huang, H. Y. Chem.−Eur. J. 2015, 21, 115–119. 72. Lian, X. Z.; Fang, Y.; Joseph, E.; Wang, Q.; Li, J. L.; Banerjee, S.; Lollar, C.; Wang, X.; Zhou, H. C. Chem. Soc. Rev. 2017, 46, 3386–3401. 73. Mehta, J.; Bhardwaj, N.; Bhardwaj, S. K.; Kim, K. H.; Deep, A. Coord. Chem. Rev. 2016, 322, 30–40. 74. Zhao, M.; Zhang, X. M.; Deng, C. H. Chem. Commun. 2015, 51, 8116–8119. 75. Shih, Y. H.; Lo, S. H.; Yang, N. S.; Singco, B.; Cheng, Y. J.; Wu, C. Y.; Chang, I. H.; Huang, H. Y.; Lin, C. H. ChemPlusChem. 2012, 77, 982–986. 76. Jung, S.; Kim, Y.; Kim, S. J.; Kwon, T. H.; Huh, S.; Park, S. Chem. Commun. 2011, 47, 2904–2906. 77. Lykourinou, V.; Chen, Y.; Wang, X. S.; Meng, L.; Hoang, T.; Ming, L. J.; Musselman, R. L.; Ma, S. Q. J. Am. Chem. Soc. 2011, 133, 10382–10385. 78. Chen, Y.; Lykourinou, V.; Vetromile, C.; Hoang, T.; Ming, L. J.; Larsen, R. W.; Ma, S. Q. J. Am. Chem. Soc. 2012, 134, 13188–13191. 79. Feng, D. W.; Liu, T. F.; Su, J.; Bosch, M.; Wei, Z. W.; Wan, W.; Yuan, D. Q.; Chen, Y. P.; Wang, X.; Wang, K. C.; Lian, X. Z.; Gu, Z. Y.; Park, J.; Zou, X. D.; Zhou, H. C. Nat. Commun. 2015, 6, 5979. 80. Lian, X. Z.; Chen, Y. P.; Liu, T. F.; Zhou, H. C. Chem. Sci. 2016, 7, 6969–6973. 81. Capdevila, D. A.; Marmisolle, W. A.; Tomasina, F.; Demicheli, V.; Portela, M.; Radi, R.; Murgida, D. H. Chem. Sci. 2015, 6, 705–713. 82. Lyu, F. J.; Zhang, Y. F.; Zare, R. N.; Ge, J.; Liu, Z. Nano Lett. 2014, 14, 5761–5765. 83. Wu, X. L.; Ge, J.; Yang, C.; Hou, M.; Liu, Z. Chem. Commun. 2015, 51, 13408–13411. 84. Liang, K.; Ricco, R.; Doherty, C. M.; Styles, M. J.; Bell, S.; Kirby, N.; Mudie, S.; Haylock, D.; Hill, A. J.; Doonan, C. J.; Falcaro, P. Nat. Commun. 2015, 6, 7240. 85. Li, P.; Moon, S. Y.; Guelta, M. A.; Harvey, S. P.; Hupp, J. T.; Farha, O. K. J. Am. Chem. Soc. 2016, 138, 8052–8055. 86. Zhang, Q.; Su, J.; Feng, D. W.; Wei, Z. W.; Zou, X. D.; Zhou, H. C. J. Am. Chem. Soc. 2015, 137, 10064–10067. 87. He, H. M.; Han, H. B.; Shi, H.; Tian, Y. Y.; Sun, F. X.; Song, Y.; Li, Q. S.; Zhu, G. S. ACS Appl. Mater. Interfaces. 2016, 8, 24517–24524. 88. Tully, J.; Yendluri, R.; Lvov, Y. Biomacromolecules. 2016, 17, 615–621. 89. Shieh, F. K.; Wang, S. C.; Yen, C. I.; Wu, C. C.; Dutta, S.; Chou, L. Y.; Morabito, J. V.; Hu, P.; Hsu, M. H.; Wu, K. C. W.; Tsung, C. K. J. Am. Chem. Soc. 2015, 137, 4276–4279. 90. Liao, F. S.; Lo, W. S.; Hsu, Y. S.; Wu, C. C.; Wang, S. C.; Shieh, F. K.; Morabito, J. V.; Chou, L. Y.; Wu, K. C. W.; Tsung, C. K. J. Am. Chem. Soc. 2017, 139, 6530–6533. 91. Lian, X.; Huang, Y.; Zhu, Y.; Fang, Y.; Zhao, R.; Joseph, E.; Li, J.; Pellois, J. P.; Zhou, H. C. Angew. Chem. 2018, 57, 5725–5730. 92

92. Liang, K.; Richardson, J. J.; Doonan, C. J.; Mulet, X.; Ju, Y.; Cui, J. W.; Caruso, F.; Falcaro, P. Angew. Chem., Int. Ed. 2017, 56, 8510–8515. 93. Ma, W. J.; Jiang, Q.; Yu, P.; Yang, L. F.; Mao, L. Q. Anal. Chem. 2013, 85, 7550–7557. 94. Wang, X.; Lu, X. B.; Wu, L. D.; Chen, J. P. Biosens. Bioelectron. 2015, 65, 295–301. 95. Canbay, E.; Akyilmaz, E. Anal. Biochem. 2014, 444, 8–15. 96. Lu, X. B.; Wang, X.; Wu, L. D.; Wu, L. X.; Dhanjai; Fu, L.; Gao, Y.; Chen, J. P. ACS Appl. Mater. Interfaces. 2016, 8, 16533–16539. 97. Patra, S.; Crespo, T. H.; Permyakova, A.; Sicard, C.; Serre, C.; Chausse, A.; Steunou, N.; Legrand, L. J. Mater. Chem. B 2015, 3, 8983–8992.

93

Chapter 7

Protease-Catalyzed Polymerization of Tripeptide Esters Containing Unnatural Amino Acids: α,α-Disubstituted and N-Alkylated Amino Acids Kousuke Tsuchiya* and Keiji Numata* Biomacromolecules Research Team, RIKEN Center for Sustainable Resource Science, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan *E-mail: [email protected]. *E-mail: [email protected].

Chemoenzymatic polymerization utilizing the aminolysis reaction catalyzed by proteases in aqueous solution is a useful method to synthesize various types of polypeptides in an environmentally benign manner. Various ester derivatives of amino acids can be polymerized using proteases. However, the polymerization of unnatural amino acid esters generally results in no polypeptide formation because proteases exhibit high substrate specificity for certain amino acid residues. To overcome the poor affinity of unnatural amino acids for proteases, we prepared tripeptide esters containing unnatural amino acids, such as 2-aminoisobutyric acid (Aib) and sarcosine (Sar). The tripeptide esters can be polymerized by papain to obtain polypeptides with a periodic sequence containing these unnatural amino acids. The incorporation of unnatural amino acids in the periodic sequences of polypeptides caused drastic changes in their secondary structure and physical properties.

© 2018 American Chemical Society

Introduction Depending on their amino acid sequences, polypeptides are unique biopolymers with diverse functional and physical properties. Biological or chemical synthetic methods have been developed to synthesize artificial polypeptides and to achieve precise control of their amino acid sequences. A biological synthetic method uses living microbes as host-reaction fields via a genetic transformation technique. Solid-phase polypeptide synthesis is generally used to synthesize polypeptides by a chemical synthetic method. These methods offer sophisticated sequential control in target polypeptides but are costly and produce low-yields. Another synthetic approach, chemoenzymatic polymerization of amino acid ester derivatives using proteases has been developed to synthesize polypeptides in an environmentally benign manner (1, 2). This technique is applicable for the syntheses of various types of polypeptides, including homopolypeptides (3, 4), random/block copolypeptides (5–7), and polypeptides with special structures, such as star and telechelic shapes (8, 9). In addition to DNA-coding amino acids, unnatural amino acids can be incorporated into polypeptide sequences and confer novel functionalities and/or physical properties. A representative example in nature is a variety of antibiotic peptides containing N-alkylamino acid residues, such as sarcosine (N-methylglycine, Sar). The existence of N-alkylamino acids allows the antibiotic peptides to resist proteolytic degradation (10). This feature indicates that unnatural amino acids possess a poor affinity for proteases. Recently, it was demonstrated that the papain-catalyzed copolymerization of amino acid esters with ω-aminoalkanoates (nylon monomers) as an unnatural amino acid unit provided polypeptides containing nylon units in their sequence (Figure 1a) (11, 12). The obtained nylon-containing polypeptides exhibited a melting point. This result indicates that the introduction of nylon units enables polypeptides to be subjected to thermal processing greater than the melting temperature; polypeptides usually decompose before melting. However, the papain-catalyzed copolymerization of nylon monomers is impeded by the extremely poor reactivity of nylon, resulting in low-nylon introduction rates of up to 15%, even with a high-feed ratio of nylon to amino acids. Papain is an extracellular cysteine protease and shows a relatively broad substrate specificity. Therefore, we utilized papain for chemoenzymatic polymerization of various amino acid esters (1, 2). The substrate specificity of papain relies on specific interactions between the substrate amino acids and subsites in a substrate pocket around its catalytic center. Each subsite favors specific amino acid residues, and recognition is regulated by the combination of subsite interactions with the amino acid residues. Early fundamental research on substrate specificity using polyalanines revealed that the subsite next to the catalytic center in papain strictly recognizes l-amino acid residues in polypeptides; whereas the subsites far from the catalytic center are tolerant of d-amino acid residues (13). This feature suggests that an appropriate sequential design of target polypeptides enables polymerization of unfavorable substrates.

96

Figure 1. (a) Chemoenzymatic copolymerization of Leu and nylon ethyl esters using papain and (b) general strategy for introduction of unnatural amino acids into polypeptide backbone using tripeptide esters. In contrast to other solution-based polypeptide syntheses, such as the ring-opening polymerization of N-carboxy amino acid anhydrides (NCAs), protease-catalyzed polymerization can adopt oligopeptide esters as a monomer. Various periodic sequences, including alternating polypeptides, are attainable by polymerization of oligopeptide monomers. Dipeptide esters, such as GlyAla and LysLeu ethyl esters, were successfully polymerized using proteases to afford the corresponding polypeptides with alternating sequences (14, 15). Recently, we also performed chemoenzymatic copolymerization of ValProGly tripeptide and ValGly dipeptide ethyl esters in the presence of papain. The resulting polypeptide possessed a sequence similar to the periodic motif in elastin (i.e., ValProGlyValGly) (16). The resulting polypeptide showed an elastin-mimetic, reversible structural transition dependent on the temperature. This method of synthesizing specific periodic sequences can be applied to unnatural amino acids by sandwiching them between natural amino acids (Figure 1b). Modification with natural amino acids at both N- and C-terminals is expected to mitigate the poor affinity of papain according to the previous evaluation of the substrate specificity of subsites in the substrate pocket (13). In this context, two types of unnatural amino acids [namely, 2-aminoisobutyric acid (Aib) and Sar] were incorporated into tripeptide ester monomers (17). The chemoenzymatic polymerization of tripeptide esters using papain successfully afforded polypeptides that consist of periodic sequences containing unnatural amino acids. 97

General Procedure for Chemoenzymatic Polymerization The general procedure for the chemoenzymatic polymerization of amino acid esters is as follows. A solution of amino acid HCl salt in an aqueous buffer is placed in a glass tube equipped with a stir bar and stirred at 40°C until all monomers are dissolved. In most cases, the buffer solution concentration is 1 M and the pH is greater than 7, which activates the monomers. A solution of a protease in the aqueous buffer is then added to this solution in one portion. In the case of papain, the final concentration of papain is usually 50 mg/mL at optimal condition. The mixture is stirred at 800 rpm and 40°C for 2–24 h. As the polymerization proceeds, the precipitation of insoluble polypeptide occurs. After the mixture cools to room temperature, the precipitate is collected by centrifugation at 7000 rpm for 15 min at 4°C. The crude product is washed twice with deionized water and lyophilized to afford the polypeptide as a white solid. The obtained polypeptides are characterized by 1H NMR spectroscopy and matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOFMS). Dipeptide and tripeptide esters are used as well as a monomer for chemoenzymatic polymerization. An optimized monomer concentration is in a range of 0.1–2.0 M depending on the amino acid residues.

Synthesis of Polypeptides Containing Aib Chemoenzymatic polymerization of amino acid esters using papain has been intensively studied for the synthesis of various types of polypeptides. Hydrophobic amino acids, such as alanine and leucine, become water-insoluble polypeptides after papain-catalyzed polymerization, resulting in precipitation of the product during the reaction. The precipitated polypeptides can be easily isolated from papain by centrifugation of the reaction mixture. To investigate the ability to polymerize an unnatural amino acid, Aib, we attempted papain-catalyzed polymerization of an Aib ethyl ester in a phosphate buffer solution under various conditions (Figure 2). No precipitate appeared, even after a prolonged reaction time of 24 h, and only the starting material was obtained. A dipeptide ethyl ester consisting of Aib and Ala units, AibAla-OEt, was also used for polymerization with papain in a phosphate buffer solution. However, alanine modification at the C-terminal of Aib did not improve the reactivity, which resulted in no polymerization. This indicates that Aib has a poor affinity for papain and is less reactive than natural amino acid monomers. The copolymerization of AibAla-OEt with the Ala ethyl ester (1:1 molar ratio in feed) in the presence of papain afforded a polypeptide as a white precipitate in a very low yield (< 5%). The resulting polypeptide contained small amounts of the Aib unit (i.e., up to 5 mol%) revealing the low reactivity of AibAla-OEt. In contrast, the papain-catalyzed polymerization of the tripeptide ethyl ester containing Aib, AlaAibAla-OEt, provided a moderate yield of polypeptides of up to 30%. The incorporation of Aib units in the obtained polypeptides was confirmed by 1H NMR spectroscopy. The chemoenzymatic polymerization of AlaAibAla-OEt at different monomer concentrations was performed. The results are summarized in Figure 98

3. The polymer yield was maximized at a monomer concentration of 0.25 M and reduced with an increasing monomer concentration. The polypeptide was no longer obtained at a concentration of more than 0.5 M. The Aib content in the polypeptide, which was calculated from 1H NMR spectra, was almost comparable to the theoretical value (33 mol%) at the lower monomer feed concentration. However, the Aib content slightly decreased as the feed concentration increased due to the additional Ala insertion in poly(AlaAibAla) by transamidation, which was revealed by MALDI-TOFMS (17). The Ala insertion was suppressed at low-feed concentrations.

Figure 2. Chemoenzymatic polymerization of Aib-containing monomers in the presence of papain.

The MALDI-TOFMS of the products revealed that the tetramer (AlaAibAla)4 and pentamer (AlaAibAla)5 were mainly obtained with a series of small peaks derived from Ala-inserted poly(AlaAibAla). The conversion of the AlaAibAla-OEt monomer to poly(AlaAibAla) during chemoenzymatic polymerization was monitored by a time-course study, as shown in Figure 4. The polypeptide formation rapidly occurred within 30 min, and the conversion of the monomer was saturated at approximately 50%. The chemoenzymatic polymerization reaction was kinetically controlled by using moderately activated ester monomers. Therefore, the tandem aminolysis of amino acid esters catalyzed by an enzyme rapidly proceeds in the early stage of polymerization, whereas the aminolysis reaction competes with hydrolysis in the later stage. The polymerization behavior is similar to the chemoenzymatic polymerization of natural amino acids, revealing that the reaction is regulated by the alanine moiety in the tripeptide, which shows a high affinity for papain. 99

Figure 3. The effect of the monomer concentration on the yield for papain-catalyzed polymerization of AlaAibAla-OEt (square) and the Aib content in poly(AlaAibAla) (circle). (Reproduced with permission from reference (17). Copyright 2017 Royal Society of Chemistry.)

To investigate the effect of Aib units on the secondary structure of the polypeptide, IR and circular dichroism (CD) spectroscopic analyses were performed on the poly(AlaAibAla) prepared by chemoenzymatic polymerization of AlaAibAla-OEt. The IR and CD spectra are shown in Figure 5. In the IR spectra, a peak derived from the stretching vibration mode of the carbonyl group (1700–1600 cm−1) is defined as the amide I region. A peak shift in the amide I region reflects the specific environment of the amide bonds and depends on the secondary structures of the polypeptides (18, 19). Polyalanine (polyAla), which has the natural backbone structure of poly(AlaAibAla), showed a strong, sharp peak at 1630 cm−1, corresponding to a β-sheet structure. Poly(Ala-r-AibAla) containing only 5 mol% of the Aib unit exhibits a similar profile with a slight shoulder at 1660 cm−1, corresponding to an α-helical structure. In contrast, the IR spectrum of poly(AlaAibAla) showed a substantial peak shift to 1660 cm−1, indicating a structural transition from a β-sheet to α-helical conformation. This finding revealed that the periodic introduction of Aib units into the polypeptide backbone effectively induced a helical conformation. The Aib-containing triad, XaaXaaAib, in polypeptide sequences tends to assemble into a helical conformation because of the steric hinderance of the two methyl groups at the α-carbon (20–22). A 5 mol% introduction of Aib in poly(Ala-r-AibAla) was not adequate to induce the α-helix structure. 100

Figure 4. Time-course study of monomer conversion during chemoenzymatic polymerization of AlaAibAla-OEt (0.25 M) in the presence of papain at 40°C. (Reproduced with permission from reference (17). Copyright 2017 Royal Society of Chemistry.)

The secondary structure of poly(AlaAibAla) in a solution state was also investigated by CD spectroscopy in 2,2,2-trifluoroethanol. The polyAla sequence is most likely to adopt a β-strand/sheet structure, even in organic solvents that strongly induce helical conformations (23). The CD profile of polyAla showed a very weak negative peak at 218 nm and a positive peak at 193 nm, indicating the formation of a β-strand (sheet) structure. This β-strand propensity was unchanged with a small amount of Aib, as shown in the CD profile of poly(Ala-r-AibAla). On the other hand, poly(AlaAibAla) exhibited a drastic change in the CD profile compared to polyAla. A strong negative Cotton effect was observed with two negative peaks at 218 and 208 nm and a positive peak at 191 nm. This result revealed that poly(AlaAibAla) adopted an α-helix structure with a right-handed screw direction. 101

Figure 5. (a) IR spectra of Aib-containing polypeptides and (b) CD spectra of Aib-containing polypeptides in a 2,2,2-trifluoroethanol solution (100 mM). (Reproduced with permission from reference (17). Copyright 2017 Royal Society of Chemistry.)

Figure 6. Chemoenzymatic polymerization of Sar-containing tripeptide esters in the presence of papain.

102

Synthesis of Polypeptides Containing Sar This tripeptide ester can be applied to other types of unnatural amino acids. Polysarcosine (polySar), also referred to as polypeptoide, is a promising polypeptide-like material for bioapplications due to its biocompatible, hydrophilic nature, which is similar to that of poly(ethylene glycol) (24, 25). We synthesized two types of tripeptide esters containing the Sar unit, namely, GlySarGly and AlaSarAla ethyl esters, as the monomers for chemoenzymatic polymerization. Similar to the Aib-containing monomers, Sar and SarGly dipeptide ethyl esters were inactive in papain-catalyzed polymerization; no polypeptide was obtained by chemoenzymatic polymerization of these monomers. In contrast, GlySarGly-OEt was converted to poly(GlySarGly) using papain in a phosphate buffer solution (Figure 6). Papain-catalyzed polymerization of AlaSarAla-OEt in a phosphate buffer solution also afforded the corresponding polypeptide, poly(AlaSarAla). In the case of the Sar-containing sequences, the resulting polypeptides were water-soluble; therefore, no precipitate was obtained after the chemoenzymatic polymerization. The resulting mixture was subjected to ultrafiltration for the removal of papain followed by dialysis to isolate the polypeptides. Extraction using organic solvents, such as chloroform, was also effective in the case of the isolation of poly(AlaSarAla). Polyalanine and polyglycine are water-insoluble because of hydrophobic aggregates that form via intermolecular hydrogen-bonding. The Sar units in the polypeptide backbone can break the hydrogen bonds by the methyl group on the N atom. Therefore, the introduction of Sar units into polyGly and polyAla caused significant changes in the solubility in water by breaking hydrogen bonds. The resulting polypeptides were totally soluble in water and other organic solvents, such as chloroform and acetonitrile. The formation of poly(GlySarGly) and poly(AlaSarAla) was confirmed by MALDI-TOFMS, as shown in Figure 7. The molecular weights of poly(GlySarGly) and poly(AlaSarAla) were in the range from 400 to 2000, and the maximum degree of polymerization was 10 (30 residues) and 4 (12 residues), respectively. In the case of papain-catalyzed polymerization of GlySarGly-OEt, only poly(GlySarGly) with a free carboxylic acid at the C-terminal was obtained. Because the resulting poly(GlySarGly) was highly water-soluble, the ester group at C-terminal was assumed to be easily hydrolyzed in the presence of papain during the reaction.

103

Figure 7. MALDI-TOF mass spectra of (a) poly(GlySarGly) and (b) poly(AlaSarAla) synthesized by papain-catalyzed polymerization.

Conclusions Three types of tripeptide esters containing unnatural amino acids, namely, AlaAibAla-OEt, GlySarGly-OEt, and AlaSarAla-OEt, were rationally designed to mitigate the poor affinity of unnatural amino acids for papain. The papain-catalyzed polymerization of these tripeptide esters successfully provided polypeptides with periodic sequences containing unnatural amino acids. The introduction of Aib units into a polyalanine backbone induced a drastic structural transition from a β-sheet to an α-helix, whereas the introduction of Sar units changed the solubility of the polypeptides. The chemoenzymatic polymerization of tripeptide esters enhances the versatility of the substrates and can be applied to various types of unnatural amino acids. The incorporation of unnatural amino acids into polypeptides will open a way to finetune their material properties and confer novel functionalities.

Acknowledgments This work was financially supported by Impulsing Paradigm Change through Disruptive Technologies Program (ImPACT), JSPS KAKENHI Grant Number JP17K18361, and JST ERATO Grant Number JPMJER1602. 104

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

Tsuchiya, K.; Numata, K. Macromol. Biosci. 2017, 1700177. Numata, K. Polym. J. 2015, 47, 537–545. Ma, Y.; Sato, R.; Li, Z.; Numata, K. Macromol. Biosci. 2016, 16, 151–159. Baker, P. J.; Numata, K. Biomacromolecules 2012, 13, 947–951. Tsuchiya, K.; Numata, K. ACS Macro Lett. 2017, 6, 103–106. Numata, K.; Baker, P. J. Biomacromolecules 2014, 15, 3206–3212. Fagerland, J.; Finne-Wistrand, A.; Numata, K. Biomacromolecules 2014, 15, 735–743. Tsuchiya, K.; Numata, K. Macromol. Biosci. 2016, 16, 1001–1008. Ageitos, J. M.; Baker, P. J.; Sugahara, M.; Numata, K. Biomacromolecules 2013, 14, 3635–3642. Agrawal, S.; Adholeya, A.; Deshmukh, S. K. Front. Pharmacol. 2016, 7, 333. Yazawa, K.; Gimenez-Dejoz, J.; Masunaga, H.; Hikima, T.; Numata, K. Polym. Chem. 2017, 8, 4172–4176. Yazawa, K.; Numata, K. Polymers 2016, 8, 194. Schechter, I.; Berger, A. Biochem. Biophys. Res. Commun. 1967, 27, 157–162. Qin, X.; Xie, W.; Tian, S.; Cai, J.; Yuan, H.; Yu, Z.; Butterfoss, G. L.; Khuong, A. C.; Gross, R. A. Chem. Commun. 2013, 49, 4839–4841. Qin, X.; Khuong, A. C.; Yu, Z.; Du, W.; Decatur, J.; Gross, R. A. Chem. Commun. 2013, 49, 385–387. Gudeangadi, P. G.; Tsuchiya, K.; Sakai, T.; Numata, K. Polym. Chem. 2018, 9, 2336–2344. Tsuchiya, K.; Numata, K. Chem. Commun. 2017, 53, 7318–7321. Huang, W.; Krishnaji, S.; Tokareva, O. R.; Kaplan, D.; Cebe, P. Macromolecules 2014, 47, 8107–8114. Rabotyagova, O. S.; Cebe, P.; Kaplan, D. L. Macromol. Biosci. 2010, 10, 49–59. Demizu, Y.; Doi, M.; Sato, Y.; Tanaka, M.; Okuda, H.; Kurihara, M. Chem. Eur. J. 2011, 17, 11107–11109. Solà, J.; Helliwell, M.; Clayden, J. J. Am. Chem. Soc. 2010, 132, 4548–4549. Galoppini, E.; Fox, M. A. J. Am. Chem. Soc. 1996, 118, 2299–2300. Shinchuk, L. M.; Sharma, D.; Blondelle, S. E.; Reixach, N.; Inouye, H.; Kirschner, D. A. Proteins 2005, 61, 579–589. Robert, L.; Corinna, F.; Arlett, G. J. Polym. Sci., Part A: Polym. Chem. 2013, 51, 2731–2752. Ueda, M.; Makino, A.; Imai, T.; Sugiyama, J.; Kimura, S. Chem. Commun. 2011, 47, 3204–3206.

105

New Materials Based on Polysaccharides

Chapter 8

Sustainable Development of Polysaccharide Polyelectrolyte Complexes as Eco-Friendly Barrier Materials for Packaging Applications Kai Chi and Jeffrey M. Catchmark* Department of Agricultural and Biological Engineering, The Pennsylvania State University, 226 Agricultural Engineering Building, Shortlidge Road, University Park, Pennsylvania 16802, United States *E-mail: [email protected].

Interest in exploring high-performance eco-friendly barrier materials that could replace synthetic polymers whose composition and manufacturing processes present ecological challenges is growing. Polysaccharides are natural biopolymers already produced in large volumes for many industries including papermaking, textiles, and food production. Cellulose, starch, chitin, and their chemical derivatives, including carboxymethyl cellulose (CMC) and chitosan (CS) are among the highest volume, least expensive biopolymers produced. These polymers, however, are highly hydrophilic and do not possess adequate liquid barrier properties. Exceptional barrier behavior using these polymers has been achieved by combining them in polyelectrolyte complexation. Specifically, cationic CS and anionic CMC have been combined under high-shear homogenization to creat nanostructured particles that electrostatically coalesce during dehydration, forming a dense insoluble material. The current study demonstrates that this material is resistant to the penetration of grease (TAPPI T 559 cm-02, kit number 12), vegetable oil, and water. With the addition of rigid cellulose nanocrystals, the resulting materials exhibited improved mechanical and water vapor barrier properties. This work demonstrates that electrostatic complexation can be used to produce

© 2018 American Chemical Society

sustainable polysaccharide-based materials with unprecedented performance useful for replacing synthetics or higher cost alternatives in many high-volume applications including paper, food engineering, textiles, packaging, and construction.

Introduction Currently, the production of plastics is growing because of their widespread applications in every aspect of our everyday life. As of 2015, approximately 6.3 billion tonnes of produced plastic is estimated to have become waste, with 9% recycled, 12% incinerated, and 79% accumulated on our planet (1). It is predicted that roughly 12 billion tonnes of plastic waste will end up in landfills or the natural environment by 2050 (1), causing ecological threats such as ocean plastic pollution, microplastic contamination in tap water, and leachate infiltration. More unfortunately, approximately 99% of produced plastics are derived from petrochemicals, which are a limited and nonrenewable resource from economic and environmental perspectives (2, 3) and present several key negatives including price volatility, unreliable global supply, and greenhouse gas emissions (4, 5). As the largest plastics market, the packaging industry consumes roughly 44% of produced plastics (6). In the case of food packaging, traditional petroleum-based plastic packaging materials, such as polyethylene, polypropylene, polystyrene, and polyethylene terephthalate, are widely used because of their affordable cost and superior performance as compared to naturally occurring and biobased polymers. However, petroleum-derived polymer materials typically exhibit limited recyclability, poor biodegradability, and a large environmental footprint, all of which have posed significant waste disposal and environmental pollution issues. Therefore, increased interest in exploring sustainable, cost-effective, and ecologically compatible materials with excellent mechanical and barrier properties for the food packaging industry has driven research on the development of bioderived polymeric materials. Polysaccharide-based composites and blends have shown desirable mechanical performance, gas barrier properties, and antimicrobial properties, opening up new opportunities for scientists and engineers to develop novel packaging systems for food and other applications (7). Nonetheless, the hydrophilic nature of these polysaccharides has caused the resulting materials to have inferior mechanical, thermochemical, and barrier properties as compared to their petroleum-derived counterparts (8). In particular, the gas and water vapor barrier properties of some polysaccharide composite films are impaired at higher relative humidity (RH) conditions (9), which limits their further application as barrier coatings or films for certain food packaging applications. Many efforts have been devoted to developing polysaccharide polyelectrolyte complex materials with adequate barrier properties that could be promising alternatives to petrochemical polymers for food and other packaging applications (10–15). The oppositely charged polysaccharides could spontaneously form electrostatic crosslinking within the resulting materials, contributing to improvement in mechanical and barrier properties (16). Nanocellulose, a sustainable 110

bionanomaterial with many extraordinary features (such as robust mechanical properties, biodegradability, renewability, and high surface area) (17), is emerging as a barrier property enhancement filler or barrier coating for paper and paperboard packaging materials (18). Improved water vapor and oxygen barrier properties have been accomplished by incorporating nanocellulose into various biopolymer or synthetic polymer matrices (9). However, the incorporation of nanocellulose into a binary polyelectrolyte complex system has not been reported, and little is known about the role of nanocellulose in the structure−property relationship of ternary polyelectrolyte complex systems. Specifically, in the case of the paper and paperboard packaging industry, the uncoated lignocellulosic-based paper and paperboard exhibit desirable properties such as low cost, sustainability, recyclability, and biodegradability, yet their porous microstructure and hygroscopic properties have resulted in insufficient barrier properties against water, oxygen, and oil. Usually, the surfaces of paper and paperboard are coated with unsustainable materials such as fossil-based waxes, latex, plastics, and aluminum to develop high-barrier packaging materials at the expense of eco-friendliness and biodegradability. Replacing these coatings with compostable bioderived materials would make a significant impact on the disposable packaging problem and the environment. This chapter presents our recent research on developing polysaccharide polyelectrolyte complex materials as novel barrier coating materials for paper and paperboard packaging applications. Exceptional barrier performance has been achieved by subjecting multipolysaccharide systems to polyelectrolyte complexation coupled with a high-shear homogenization process, opening up new opportunities for high-performance eco-friendly biomaterials in paper, packaging, food engineering, and many other applications.

Experimental Section Materials High purity chitosan (CS) (ChitoClear, Primex Inc, Iceland) was obtained with an average molecular weight (MW) of 214 kDa and a degree of deacylation of 90%. The carboxymethyl cellulose (CMC) sodium salt (lot # 419273) had an average MW of 90 kDa and a degree of substitution of 0.7 according to the supplier (Sigma-Aldrich, St. Louis, MO). Lab-made cellulose nanocrystals (CNC) were extracted from microcrystalline cellulose (Avicel PH-101, Sigma-Aldrich) by a sulfuric acid hydrolysis process. The conditions of acid hydrolysis and subsequent purification processes were detailed in previous studies (19, 20). All chemical reagents including toluene, n-heptane, formic acid, and sodium hydroxide were used as received. Pristine paperboard substrate was kindly provided by Southern Champion Tray (Chattanooga, TN). It was coated on one side with clay, and the uncoated side was used for the coating experiments. High quality nanopure water with a resistivity of 18.2 MΩ/cm (Millipore Milli-Q UF Plus) was used to prepare sample solutions or dispersions. 111

Preparation of Polysaccharide Polyelectrolyte Complex Coating Materials The polysaccharide polyelectrolyte complex (PPC) coating materials were prepared by a high-shear homogenization process. Initially, a 4.5% (w/v) polysaccharide solution was prepared by dispersing CS (or CMC) powder into nanopure water, followed by the addition of formic acid to adjust the solution pH to different values (pH = 4 and pH = 5.7). The polysaccharide solution was then magnetically stirred at 60°C overnight to ensure complete dissolution. For binary PPC coating materials, CS and CMC solutions were combined and subjected to immediate high-shear homogenization by a homogenizer (T25 Ultra-Turrax, IKA) at a speed of 20,000 rpm for 20 min. For ternary PPC coating materials, the desired amount of 1% (w/v) CNC stock suspension was added into CMC solution, followed by stirring at 50°C for at least 2 h. The CMC/CNC mixture was subsequently added into the pre-prepared CS solution and homogenized under exactly the same conditions as the binary PPC materials. The CNC contents in ternary PPC materials were designed to be 5 and 10% (w/w) of the total solid content of the coating materials. In addition, the pH of ternary PPC materials was fixed at 5.7. Thus, the resulting PPC coating materials included CS/CMC_pH=4, CS/CMC_pH=5.7, CS/CMC/CNC5% and CS/CMC/CNC10%. All PPC materials were degassed for 5 min to remove the air bubble formed during the homogenization process before either casting to form free-standing films or coating on the paperboard substrates.

Preparation of PPC Films Various binary and ternary PPC films were fabricated using a casting/ evaporation method. Generally, the degassed PPC dispersion was casted into a glass petri dish. After a 2-day evaporation at 35°C and 20% RH, the PPC film was peeled from the glass petri dish and further dried in a vacuum oven (25 psi) at 60°C for 8 h to remove the residual water. All film samples were conditioned in a desiccator at ambient temperature for at least 3 days before any further analysis.

Preparation of PPC-Coated Paperboards PPC-coated paperboard samples were prepared through a dip coating process according to our previous study (11). The uncoated, porous paperboard substrate was immersed in a freshly prepared PPC dispersion for 5 min. Afterward, the coated sample was removed, allowed to drain, and dried in an oven at 140°C for 15–20 min. The individual polysaccharide–coated paperboards were prepared using the same protocol and were considered control samples. The coating weight was defined as the weight difference per square meter between the initial (before coating) and final weight (after coating and oven drying) of the paperboard substrate. The pristine and coated paperboard samples were stored in a desiccator at room temperature and 20% RH before the morphology and liquid barrier performance analysis. 112

Characterization The surface charge content of individual polysaccharides (amino group of CS, carboxylate group of CMC, and sulfate group of CNC) was determined by conductometric titration. Conductivity and pH values were simultaneously monitored using a SevenExcellence pH/conductivity meter (Mettler Toledo, Columbus, OH) with an InLab 730 conductivity probe. Sample preparation, titration conditions and processes, and the calculation of surface group content have been described in previous studies (11, 17). The particle size of PPC coating materials was measured by a laser diffractometer (Mastersizer 2000, Malvern Panalytical) at 25°C. At least three measurements were taken for each sample. The viscosity of PPC coating materials was determined using a Zahn cup (type 2). Zeta potential measurement of various PPC coating materials was conducted using a Zetasizer Nano ZS (Malvern Panalytical) at 25°C. Five runs were performed for each sample, and the data was fitted using the Smoluchowski model. Scanning electron microscopy (SEM) morphology of PPC coating materials was carried out with a Nova NanoSEM 630 field emission scanning electron microscope (FEI) operating at an accelerating voltage of 3 kV, a beam current of 42 pA, and a working distance of 4 mm. Before SEM imaging, PPC coating dispersions were diluted (~0.001 wt %), lyophilized, and sputter coated with a thin layer (~3–5 nm) of iridium. For PPC film samples, the surface and cryofractured cross section morphologies were observed using the same SEM instrument and operation conditions as previously described. Tensile tests of PPC films were performed using a dynamic mechanical analyzer (Q800, TA Instruments Inc., New Castle, DE). Rectangular strips with the dimensions of 50 mm × 5 mm × 0.05 mm (length × width × thickness) were cut from films and tested at a strain ramp rate of 5%/min and a gauge length of 10 mm until break at ambient conditions. For each sample, at least five specimens were used for tensile testing. The swelling of PPC films was determined by immersing the film strips (20 × 50 mm2) in water. The films were removed from water at various times ranging from 1 to 24 h, blotted between filter papers to remove excess surface liquid, and reweighed at various times up to 24 h. The degree of swelling (DS) was calculated using the following formula: DS (%) = (weight of swollen film − weight of initial dry film)/(weight of initial dry film) × 100. At least three replicates were measured for each sample. The water vapor transmission rate (WVTR) of PPC films was determined gravimetrically following the ASTM Standard E 96 procedure with slight modification, according to our previous study (11). For PPC-coated paperboard samples, the surface and cross section morphologies were characterized by SEM imaging. The morphology of individual polysaccharide (CS and CMC)–coated paperboard was also observed. In addition, the liquid (grease, water, and vegetable oil) barrier properties of individual polysaccharide– and PPC-coated paperboards were analyzed at room temperature. The grease resistance property was evaluated by the standard method for paper and paperboard known as the Kit test (TAPPI T 559 cm-02), as described in a previous study (11). The water and vegetable oil resistance tests were performed by depositing a small volume of liquid (~10 μL) on the surface 113

of coated paperboard samples. Different locations on the coated paperboard were selected, and the results were reported as the average of five measurements.

Results and Discussion PPC Coating Materials In this study, various binary and ternary polysaccharide-based coating materials were developed by combining them in polyelectrolyte complexation. During the complexation process, strong electrostatic association via Coulombic attraction is believed to be the major interaction between cationic CS and anionic cellulose materials (CMC and CNC), forming a physically crosslinked network structure. Other intermolecular interactions such as hydrogen bonding and van der Waals forces could also exist because of the similar saccharide structure. Many internal (polysaccharide backbone structure, MW, degree of substitution of ionic groups, etc.) and external parameters (pH, ionic strength, charge molar ratio, etc.) can impact the formation, structures, and properties of PPC materials (21). Previous studies have demonstrated that the charge molar ratio between polycations and polyanions is the most important factor that can determine the structure and properties of PPCs (21, 22). The 1:1 charge molar ratio can ensure the majority of available surface ionic groups of polysaccharides electrostatically interact with their counterparts, contributing to a high degree of crosslinking within the resulting materials. The evolution of surface charge content of individual polysaccharides as a function of solution pH is shown in Figure 1a. CS has glucosamine and acetyl glucosamine randomly distributed on its backbone. The pKa of CS is ~6.0–6.3, resulting in the protonation of amine groups in acidic conditions with cationic charges. CMC, on the other hand, is a weak polyanion and can be deprotonated at a pH higher than its pKa (~4.5–5). At different pH values, the charge molar ratio between CS and CMC was different. Therefore, two different solution pH values (pH = 4 and 5.7) were chosen for binary PPC material composed of CS and CMC in order to understand the influence of the degree of ionic crosslinking on the structure and properties of the resulting materials. More cationic charges derived from protonated amine groups of CS existed at pH = 4, while the 1:1 charge stoichiometry was achieved when the pH reached 5.7. CNC had anionic sulfate groups on its surface because of the esterification of surface hydroxyl groups in the acid hydrolysis process. The amount of sulfate groups, 0.23 mmol/g, was stable from pH = 3 to 10, as the pKa of sulfate is ~1.5–2. Considering such a large difference in the density of ionizable groups, CNC was incorporated into the CS/CMC blend by weight fraction to fabricate ternary PPC materials. Properties such as particle size, viscosity, and zeta potential of various PPC coating materials were characterized (Figure 1b) before applying the coating on paperboard substrates. These parameters are important, as they are the main factors that determine the formation, surface coverage, and properties of the resulting coating layer on the paperboard surface. For binary PPC materials, an increase of pH from 4 to 5.7 induced larger particle size (49.5 μm vs 93.7 μm), lower viscosity (93.2 cSt vs 70.8 cSt), and smaller zeta potential value (20.4 114

mV vs 8.1 mV), implying increased electrostatic interaction between oppositely charged CS and CMC and therefore a high degree of crosslinking within the resulting material. The particle size distribution (PSD), determined by using the formula PSD = [Dv (90) − Dv (10)]/Dv (50), exhibited an increase from 2.81 to 3.23 (data not shown in Figure 1), indicating a wider PSD in the CS/CMC blend at pH = 5.7. PPC coating material with a wider PSD is believed to be beneficial to the effective coverage of porous paperboard substrates, as smaller particles could fill up spaces between large particles. With the addition of CNC, the ternary PPC materials showed continuously increased particle size and lowered viscosity and zeta potential. At 10% CNC addition, large particles (101 μm) were formed, causing a decrease in the viscosity (61.2 cSt) of coating materials. It could be explained that these large complex aggregates displayed reduced surface area of the hydrophilic polysaccharide and thus reduced the viscosity in water. Therefore, the inherent high viscosity of individual polysaccharide (especially for CS) was lost when polyelectrolyte complexation occurred. This particle size–viscosity relationship has been depicted in a previous study (10), and the current results are in agreement. The smaller zeta potential indicated that some CNC might participate in the ionic complexation with CS. However, it should be noted that other intermolecular interactions (like hydrogen bonding and van der Waals forces) between CNC and CMC or CS might also exist because of their similar saccharide structures. The morphology of PPC materials is shown in Figure 1c,d. Fiber-shaped particles were found for binary PPC prepared at pH = 4 (Figure 1c), whereas a mixture of fiber-shaped and platelet-like particles was observed for ternary PPC with 10% CNC content (Figure 1d). In addition, larger particles can be seen in Figure 1d, which supports the particle size measurement. The morphology of PPC particles is more dependent on the homogenization and lyophilization processes. Under the same production process and sample preparation procedure, ternary PPC material with 10% CNC displayed larger, platelet-like particles, suggesting the participation of CNC in the interaction with CS and CMC to form larger particles. The formation of fiber-shaped particles in this study is hypothesized to originate from the homogenization process. The application of immediate high-shear blending opens up a new process element in which the PPC materials were sheared during the ionic complexation process, yielding isolated fiber-like particles. PPC Films The properties of solution-casted PPC films were characterized as represented in Figure 2. Figure 2a–c depicts the surface and cross section morphologies of various PPC films analyzed by SEM. All PPC films were smooth and homogeneous as observed at the surface and cross section, independent of the preparation pH and addition of CNC. No pores or voids could be detected in the films, suggesting that the films were dense and compact. The addition of CNC, even at 10% weight fraction (Figure 2c), did not cause any discernable phase separation, indicating good compatibility between CNC and the CS/CMC blend. The distribution of CNC in the CS/CMC blend is difficult to observe 115

without the help of energy dispersive electron spectroscopy analysis with sulfur mapping of CNC. An evenly distributed sulfur element was observed in PPC film with the addition of 10% CNC as reported in our previous study (11), suggesting the uniform distribution of CNC in PPC film. Such uniform distribution of rigid CNC contributes to the nanoreinforcing effect of CNC, resulting in an improvement in the mechanical and water vapor barrier properties of PPC films. It is also worth mentioning that the ternary PPC films were prepared by initially mixing anionic CNC and CMC, followed first by homogenizing with CS under high-shear conditions and finally by solution casting into films. Both CMC and CNC have negative charges and similar backbone structures, yet they possess a large difference in charge density and backbone rigidity. This work hypothesizes that the high surface charge (almost 10 times larger than CNC based on conductometric titration results) (11) and flexible CMC polymer chain segments might be more prone to electrostatic interaction with the cationic CS polymer chains, thus forming a densely entangled polymer network. By contrast, the rigid and much less charged CNC was unable to wrap around the CS polymer chains and form the crosslinked network structure. Therefore, the ternary PPC films developed in this study might be based on the individualized distribution of CNC in a matrix of CS/CMC with a crosslinked structure.

Figure 1. Characterization of PPC coating materials: (a) surface charge content of individual polysaccharides (CMC, CS, and CNC) as a function of pH; (b) particle size, viscosity, and zeta potential of PPC coating materials prepared at various conditions, including binary PPC prepared at pH = 4 (CS/CMC_pH = 4) and 5.7 (CS/CMC_pH = 5.7) and ternary PPC prepared at pH = 5.7 and different CNC loading levels (5 and 10%); SEM images of freeze dried CS/CMC_pH = 4 (c) and CS/CMC/CNC10% (d). (a) and (b) reproduced with permission from reference (11). Copyright 2018 Elsevier. 116

Figure 2. Characterization of free-standing PPC films: SEM surface and cross section (inset) morphologies of CS/CMC_pH = 4 (a), CS/CMC_pH = 5.7 (b), and CS/CMC/CNC10% (c); tensile (d), water swelling (e), and water vapor barrier (f) properties of various PPC films. (a–d) and (f) reproduced with permission from reference (11). Copyright 2018 Elsevier.

The tensile properties, including tensile strength (TS), Young’s modulus (E), and elongation at break, of various PPC films are depicted in Figure 2d. A high level of ionic crosslinking within the CS/CMC blend should improve the strength and stiffness of the film. This assumption is supported by the tensile data obtained for binary PPC films with varied degrees of electrostatic crosslinking. PPC film prepared at pH = 5.7 had a TS and E of 43.2 MPa and 3.1 GPa, respectively, which were 51 and 57% higher than those of PPC film prepared at pH = 4. Several previous studies have reported a similar beneficial influence of ionic complexation on the mechanical properties of PPC materials (13, 15). With the incorporation of CNC, both the TS and E values of the resulting PPC films significantly increased, which is primarily attributable to the uniform distribution and interface compatibility of rigid, highly crystalline CNC within the CS/CMC matrix, as evidenced by the energy dispersive electron spectroscopy data (11). Nanocellulosic materials (including CNC, cellulose nanofibrils, and bacterial nanocellulose) with high crystallinity (~70–90%), a high Young’s modulus (~100–130 GPa), and a large surface area (several hundred m2/g) are well known for their exceptional nanoreinforcing efficiency in many polymer nanocomposites (17, 23, 24). The distribution and interface compatibility of CNC within the polymer matrix are two key elements of its reinforcing efficiency. A high loading level of CNC (>10 wt %) in the PPC materials could cause CNC self-aggregation, impairing both TS and E, as evidenced in our recent study (11). CS/CNC film was also prepared as another control sample to elucidate the role of CMC in the 117

ternary PPC films. The TS and E of CNC/CS film were 45.2 ± 3 MPa and 3.3 ± 0.3 GPa, respectively, and were inferior to those of ternary PPC films with the same CNC content. This result supports our hypothesis that the high level of electrostatic crosslinking between CS and CMC might play a more significant role in improving the stiffness and strength of PPC films. The swelling kinetics of PPC films is shown in Figure 2e. All films showed rapid swelling within 1 h, and no significant increase in DS was observed after 3 h. Binary PPC film with pH = 4 represented the maximum DS at any time over other PPC films. The CS and CMC control films were also tested for DS at the same conditions (data not shown here). Both individual polysaccharides were completely dissolved within 2 h. The lowest DS was observed for binary PPC films prepared at pH = 5.7, possibly because of the high level of crosslinking within the films. Films with stronger crosslinking could result in fewer available hydrophilic groups, such as hydroxyl groups, that can bind water. For ternary PPC films, the incorporation of CNC negatively impacted the DS. Indeed, CNC is an extremely hydrophilic material and can absorb a large amount of water. Overall, it seems that PPC material can become resistant to water swelling after polyelectrolyte complexation occurs. The WVTR of PPC film samples is shown in Figure 2f. WVTR is a critical parameter for materials used in food packaging, as the transportation of water vapor from the surrounding environment (especially at higher moisture conditions) into food products can significantly impact the shelf life of those products. Generally, electrostatic crosslinking and the addition of evenly distributed crystalline CNC were found to be beneficial to the improvement of WVTR in our study. Binary PPC film with a high level of crosslinking (pH = 5.7) exhibited significantly decreased WVTR (from 32,000 to 13,220 g μm m-2 d-1). On the other hand, with incorporation of 10 wt % CNC, the WVTR of PPC film was further decreased by 40% (from 13,220 to 7982 g μm m-2 d-1). Many factors, including temperature, pressure, crystallinity, wettability, film density, and structure and pore size, can impact the WVTR of a film sample. The homogeneous, defect-free, and densely packed PPC films, as observed from SEM analysis, could block effective travel for the diffusion of water vapor. The role of electrostatic crosslinking in the improvement of WVTR has been previously reported (11, 15). It is speculated that ionic interaction could enhance the cohesive energy density of a PPC film, therefore largely inhibiting the diffusion of water vapor (18). Improved water vapor barrier properties of PPC films after the incorporation of highly crystalline CNC could possibly be attributed to the synergetic roles of impermeable crystalline regions of CNC and a dense percolating nanoparticle network, which together increase the tortuousness of the diffusion path water vapor must take to permeate the film. Furthermore, other intermolecular interactions between CNC and a CS/CMC matrix, possibly including ionic bonding, hydrogen bonding, and van der Waals forces, could lead to the formation of a densely packed network structure, higher cohesive energy density, and lower free volumes with lower permeability to water vapor. High levels of CNC content (>10 wt %) in PPC films were found to impair the water vapor barrier property in our recent study (11) because of the aggregation behavior of CNC. 118

PPC-Coated Paperboards The surface and cross section morphologies of uncoated and PPC-coated paperboards were analyzed by SEM as shown in Figures 3 and 4. As a control, the morphology of paperboard coated with individual polysaccharide (CS and CMC) was also evaluated. The pristine, uncoated paperboard (Figure 3a,a1) exhibited a porous structure with pore size ranging from 10 to 100 μm. Many holes and cavities were seen on the surface of the pristine paperboard. In addition, small pores (~5 μm) were present (not shown here) at the cellulose fiber surface. The porous characteristic of pristine paperboard leads to poor barrier performance against water, oxygen, oil, and grease. Individual polysaccharides were deposited on the paperboard substrate at a coating weight of ~8 g/m2, which could be considered a starting point for well-coated paperboard intended for barrier applications (25). In the case of CS-coated paperboard (Figure 3b,b1), partial coverage of the substrate was observed. The number of holes was largely decreased, yet some small pores (~10 μm) still existed. Further, the covered surface was not smooth and showed a dented feature that can be seen in the inset of Figure 3b. It seems that the CS solution penetrated the paper sheet and thus could not form a continuous film on the paperboard substrate. This result is unexpected, as the CS solution used for coating had a solid content of 3.5% and high viscosity. The CMC-coated paperboard showed homogeneous and void-free surface characteristics. The CMC coating’s cross section image displayed a layered structure (~10–15 μm in thickness), implying the formation of a continuous film on the paperboard.

Figure 3. SEM images of surface (a–c) and cross section (a1–c1) morphologies of pristine paperboards (a and a1) and CS- (b and b1) and CMC-coated (c and c1) paperboards with the coating weight of 8 g/m2. 119

Deposition of the PPC coating materials onto cellulosic paperboard substrate led to different surface features as seen in Figure 4. A lower coating weight (4 g/m2) was applied for CS/CMC prepared at pH = 4. The pores within the paperboard were filled by the coating material, yet the covered surface showed roughness and dented features (see white arrows in the inset of Figure 4a). Thus, a higher coating weight is needed for a smooth and fully covered surface, which could ensure the formation of continuous film on the substrate. With the coating weight of 8 g/m2, all PPC-coated paperboard samples exhibited cavity-free and smooth surface features, indicating the formation of a continuous layer of coating material on the paperboard substrates. The appropriated particle size and viscosity of PPC materials are believed to be of great importance in forming homogeneous, continuous, and compact barrier layers on the paperboard substrate (11). In the case of ternary PPC with a 10% CNC-coated paperboard sample, larger particles (white arrows in Figure 4d) were observed, in agreement with the previous particle size data.

Figure 4. SEM images of surface morphology of PPC-coated paperboards with magnifications of 200 and 600 times (inset): CS/CMC_pH = 4 with the coating weight of 4 (a) and 8 g/m2 (b); CS/CMC_pH=5.7 (c) and CS/CMC/CNC10% (d) with the coating weight of 8 g/m2. 120

The liquid barrier properties of individual polysaccharide– and PPCcoated paperboards are represented in Figure 5. Generally, the individual polysaccharide–coated paperboard samples exhibited very poor barrier performance against grease, water, and vegetable oil, as compared to paperboards coated with PPC materials. The hydrophilic nature, uneven coverage on paperboard, and high wettability may contribute to the penetration of these liquids. In comparison, the PPC materials developed in this study show significantly enhanced liquid barrier properties. In the case of paperboard coated with binary PPC materials prepared at pH = 4, improvements in grease (kit number of 6), water penetration (2 days), and vegetable oil penetration (2 days) resistance were observed. With a high degree of ionic crosslinking, paperboard coated with binary PPC materials prepared at pH = 5.7 showed the best barrier performance against the liquids studied here, exhibiting a kit number of 12 (comparable to a polyethylene–coated paper or paperboard) and water and vegetable oil penetration resistance up to 7 days. The addition of CNC into this binary PPC system could affect the liquid barrier performance of the resulting materials, especially at high CNC content. The addition of 10% CNC notably deteriorated the grease, water, and vegetable oil resistance performance of coated paperboard. Possible reasons could be the increased wettability and hydrophilicity and the existence of CNC-aggregated domains that introduce porosity after incorporating a high content of CNC. A recent study by Gicquel et al. (25) has revealed the poor resistance of CNC coating against grease, as evidenced by the quick penetration (5 s) of grease into the pure CNC-coated paper substrate. CNC is a highly hydrophilic material, and the increased content of CNC in the PPC coating material contributes to the water wettability of the resulting PPC barrier layer. Additionally, as depicted in our previous study (11), the larger particle size and decreased viscosity of PPC coatings induced by the incorporation of CNC could affect the microstructure of the coatings, possibly leading to the existence of mesopores or micropores within the coating layer. The aforementioned defects are vulnerable to liquid penetration and especially to water penetration, as water could act as a plasticizer to split up the hydrogen-bonded CNC aggregated domains that provide the space and pathway for liquid water. All ternary PPC material–coated paperboard was considered grease resistant, as the kit numbers were larger than 8. SEM surface and cross section morphologies of paperboards coated with CMC (Figure 3c,c1) and CS/CMC prepared at pH = 4 (Figure 4b) displayed defect-free, uniform surface coverage of the paperboard substrate with a continuous film having formed on the surface. However, the liquid barrier data from Figure 5 suggest that achieving a homogeneous and continuous layer on top of the paperboard substrate is not sufficient to resist liquid penetration. Therefore, this work hypothesizes that surface wettability and electrostatic crosslinking are key parameters that could dominate the barrier performance of PPC-coated paperboard. Theories of capillary penetration indicate that there is a link between the surface wettability and the permeation of liquid. Assuming that there are no severe defects, a liquid is unable to penetrate a barrier layer if it is difficult for the liquid to wet it (18). For CMC- and ternary PPC material–coated paperboards, the surface water wettability is high because of the hydrophilic nature of CMC 121

and CNC, thereby accelerating the penetration and migration of water through the paperboard. On the other hand, electrostatic crosslinking within the PPC materials could ensure a continuous and densely packed structure within the barrier layer.

Figure 5. Liquid (grease, water, and vegetable oil) barrier properties of polysaccharide control and PPC-coated paperboards. Reproduced with permission from reference (11). Copyright 2018 Elsevier.

Conclusions Sustainable, cost-effective, and ecologically compatible materials derived from renewable natural resources have attracted a tremendous level of attention as replacements for a broad array of high-volume commercial materials based on petroleum-derived compounds. In this work, novel binary and ternary PPC materials composed of CS, CMC, and CNC were successfully fabricated through a high-shear blending approach. The developed PPC film materials exhibited homogeneous, densely packed morphological characteristics and improved mechanical and water vapor barrier properties, which were ascribed to the uniform distribution and good interfacial compatibility of CNC within the electrostatically crosslinked CS/CMC matrix. PPC films with 10 wt % CNC content showed TS and a Young’s modulus of 60.6 MPa and 4.7 GPa, respectively, and a WVTR of 7982 g μm m-2 d-1. When applied as barrier coatings on porous paperboard substrate, the PPC materials functioned as efficient barrier layers that resisted the penetration of water, oil, and grease. A high level of electrostatically crosslinked PPC coating without added CNC could make the resulting coated paperboard 122

grease resistant (kit number of 12), as well as water and vegetable oil resistant up to 7 days. It is expected that such sustainable and ecologically compatible PPC materials may be competitive barrier alternatives to petroleum-derived polymers for food packaging and handling as well as many other product applications. Furthermore, such edible barriers may offer new approaches for creating high-performance foods where control over the transport of aqueous solutions and oils is important.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

Geyer, R.; Jambeck, J. R.; Law, K. L. Sci. Adv. 2017, 3, e1700782. Chen, M.; Smith, P. M. Biomass Bioenergy 2017, 102, 52–61. Chen, M.; Smith, P. M.; Thomchick, E. Renewable Energy Focus 2017, 22, 1–9. Chen, M.; Smith, P. M.; Wolcott, M. P. BioProducts Business 2016, 1, 42–59. Chen, M.; Smith, P. M. BioProducts Business 2018, 3, 51–62. Wyman, I.; Auras, R.; Cheng, S. Green Chem. 2017, 19, 4737–4753. Cazón, P.; Velazquez, G.; Ramírez, J. A.; Vázquez, M. Food Hydrocolloids 2017, 68, 136–148. Rhim, J.-W.; Park, H.-M.; Ha, C.-S. Prog. Polym. Sci. 2013, 38, 1629–1652. Wang, J.; Gardner, D. J.; Stark, N. M.; Bousfield, D. W.; Tajvidi, M.; Cai, Z. ACS Sustainable Chem. Eng. 2017, 6, 49–70. Basu, S.; Plucinski, A.; Catchmark, J. M. Green Chem. 2017, 19, 4080–4092. Chi, K.; Catchmark, J. M. Food Hydrocolloids 2018, 80, 195–205. Dai, L.; Long, Z.; Chen, J.; An, X.; Cheng, D.; Khan, A.; Ni, Y. ACS Appl. Mater. Interfaces 2017, 9, 5477–5485. Schnell, C. N.; Galván, M. V.; Peresin, M. S.; Inalbon, M. C.; Vartiainen, J.; Zanuttini, M. A.; Mocchiutti, P. Cellulose 2017, 24, 4393–4403. Soni, B.; Schilling, M. W.; Mahmoud, B. Carbohydr. Polym. 2016, 151, 779–789. Yaich, A. I.; Edlund, U.; Albertsson, A.-C. Cellulose 2015, 22, 1977–1991. Ho, T. T. T.; Zimmermann, T.; Ohr, S.; Caseri, W. R. ACS Appl. Mater. Interfaces 2012, 4, 4832–4840. Chi, K.; Catchmark, J. M. Nanoscale 2017, 9, 15144–15158. Hubbe, M. A.; Ferrer, A.; Tyagi, P.; Yin, Y.; Salas, C.; Pal, L.; Rojas, O. J. BioResources 2017, 12, 2143–2233. Chi, K.; Catchmark, J. M. Carbohydr. Polym. 2017, 175, 320–329. Chi, K.; Catchmark, J. M. Cellulose 2017, 24, 4845–4860. Sæther, H. V.; Holme, H. K.; Maurstad, G.; Smidsrød, O.; Stokke, B. T. Carbohydr. Polym. 2008, 74, 813–821. Rodrigues, M. N.; Oliveira, M. B.; Costa, R. R.; Mano, J. F. Biomacromolecules 2016, 17, 2178–2188. Dufresne, A. Curr. Opin. Colloid Interface Sci. 2017, 29, 1–8. Liu, K.; Catchmark, J. M. Cellulose 2018, 25, 2273–2287. Gicquel, E.; Martin, C.; Yanez, J. G.; Bras, J. J. Mater. Sci. 2017, 52, 3048–3061. 123

Chapter 9

Effects of Monomer Compositions and Molecular Weight on Physical Properties of Alginic Acid Esters Yusuke Matsumoto,1 Daisuke Ishii,2 and Tadahisa Iwata*,1 1Science of Polymeric Materials, Department of Biomaterial Sciences, Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan 2Department of Chemistry for Life Sciences and Agriculture, Faculty of Life Sciences, Tokyo University of Agriculture, 1-1-1 Sakuragaoka, Setagaya-ku, Tokyo 156-8502, Japan *E-mail: [email protected].

Effects of monomer compositions and molecular weights on physical properties were investigated among alginic acid (Alg) esters. Tensile tests of Alg ester films clearly revealed that the molecular weight rather than monomeric composition affects the tensile properties. Furthermore, Alg esters with the shorter side chain showed the higher tensile strength and higher glass transition temperature (Tg) while elongation at break was almost maintained. In particular, Alg propionate prepared from Alg with the mannuronic acid content of 56 mol % and viscosity grade of 1000 cp had a tensile strength of 63 MPa and a Tg of 192 °C. These properties are comparable with or even higher than those of conventional plastics such as polycarbonate. The results suggest the potential of Alg as a starting material for plastics with high thermostability and favorable mechanical properties.

© 2018 American Chemical Society

Alginic acid (Alg) is a natural polysaccharide that can be extracted from brown algae. Alg consists of (1→4)-linked β-D-mannuronic acid (M) and α-L-gluronic acid (G), the C5 epimer of mannuronic acid (1–3). Because of a carboxyl group linked to the C5 carbon in the sugar backbone, Alg has high hydrophilicity and pH-dependent solubility and is capable of gelation in the presence of multivalent metal cations (4, 5). It is also biocompatible and biodegradable (6, 7). Because of these properties, Alg has been utilized in medicine and the food industry (8, 9). However, to date, there have been no reports of the development of thermoplastic and/or organosoluble materials from Alg. The major obstacle to utilizing polysaccharides as plastic materials is the presence of hydroxy groups in the sugar backbones. Because the hydroxy groups form inter- and intramolecular hydrogen bonds, most polysaccharides are not thermoplastic or soluble in organic solvents. Various derivatization methods have been applied to loosen the interactions within and between polysaccharide backbones. The most popular derivatization method is esterification. Esterification has several advantages over other derivatization methods for the following reasons. First, the thermal, mechanical, and other (e.g., optical) properties of polysaccharide derivatives can be widely controlled by the type, composition, degree of substitution (DS), and regiospecificity of the substituents. Second, the reagents used in the esterification reactions (carboxylic acids and reaction catalysts) are readily available. Third, many polysaccharide esterification reactions proceed at ambient or mildly elevated temperatures. Because of these advantages, polysaccharide ester derivatives such as cellulose acetate have been widely utilized in various industrial fields (10–12). The synthesis and properties of nonionic polysaccharides—namely, cellulose, curdlan, glucomannan, pullulan, and α-(1→3)-glucans—have been reported previously (12–18). However, there have been fewer reports on the synthesis of ester derivatives of ionic polysaccharides, such as chitin, chitosan, and Alg (19, 20). We have succeeded in preparing Alg esters with fully substituted hydroxyl groups (21). By esterifying all the hydroxyl groups, we obtained thermoplastic Alg materials that were soluble in organic solvents, and consequently, we were able to manufacture Alg ester films by hot pressing and solvent casting. The monomer composition and molecular weight have a significant influence on the physical properties of copolymers (22). The M/G composition of underivatized Alg affects its gelation ability and the strength of the gel (23). However, there has been no detailed investigation of the relationship between the M/G composition and the physical properties of derivatized Alg. In the present chapter, we report the effects of M/G composition and molecular weight on the physical properties of Alg ester derivatives. We also report the effect of the length of the alkyl side chain of the Alg ester derivatives.

126

Experimental Section Materials Four sodium alginates (M/G_viscosity = 39/61_100cp, 56/44_100cp, 69/31_100cp, and 56/44_1000cp) were obtained from KIMIKA CO., Ltd., Tokyo, Japan. The carboxylic acids (acetic acid, propionic acid, hexanoic acid, and octanoic acid) and their anhydrides (except for n-octanoic anhydride) were purchased from FUJIFILM Wako Pure Chemical Corporation, Tokyo, Japan; n-octanoic anhydride was purchased from Tokyo Chemical Industry Co., Ltd., Tokyo, Japan. All other reagents were obtained from commercial sources and used without further purification.

Pretreatment of Alg Each sodium alginate (5.0 g) was dissolved in 300 mL of distilled water, and pH of the solution was adjusted to 2.1 by adding 6 M HCl. Nine hundred milliliters of acetone was added to the mixture and stirred for 30 min to remove water and NaCl, and the precipitate was recovered by filtration. The precipitate was stirred in 400 mL of a 3:1 mixed solution of acetone and water and recovered by filtration. The washing process was repeated three times and acid-form Alg were obtained.

Esterification of Alg As shown in Table 1, Alg esters (Alg propionate, hexanoate, and octanoate) were prepared by using four types of Alg, according to a previously reported method (21). First, Alg was completely swollen in water. The water absorbed by the Alg was then removed by stirring in acetone. The acetone-exchanged Alg was recovered by filtration and stirred into 100 mL of carboxylic acid to obtain solvent-exchanged Alg. After solvent exchange, the pretreated Alg was stirred into carboxylic anhydride (50 mL) at 40 °C. The mixture was reacted at 40 °C for 3 h by adding 60% aqueous perchloric acid (0.5 mL) as the catalyst and carboxylic acid (50 mL) as the reaction medium. The reaction mixture was poured into 1 L of distilled water. The precipitate was recovered by filtration, washed several times with distilled water, and dried in vacuo overnight. Four and three-tenths grams of Alg propionate (AlgPr) (56/44_1000cp) were obtained. During the preparation of Alg hexanoate (AlgHe) and Alg octanoate (AlgOc), the reaction mixture was poured into a sodium bicarbonate (100 g)/distilled water (200 mL) solution. The precipitate was filtered, washed with distilled water, and dried in vacuo to obtain the sodium salts of AlgHe and AlgOc. The sodium-salts forms were transformed into their carboxylic-acid forms by dissolving them in propionic acid (100 mL) and precipitating them in distilled water (800 mL). The precipitates were recovered by filtration, washed in water, and dried in vacuo to obtain AlgHe (39/61_100cp, 5.2 g; 56/44_100cp, 4.4 g; 69/31_100cp, 2.9 g; 56/ 44_1000cp, 4.5 g) and AlgOc (56/44_1000cp, 4.6 g) (Scheme 1). 127

Scheme 1. Preparation of Alg ester derivatives.

Proton Nuclear Magnetic Resonance Measurement Proton nuclear magnetic resonance (1H NMR) spectra were recorded on a JEOL JNM-A500 FT-NMR (500 MHz) spectrometer (JEOL Ltd., Tokyo, Japan), using trifluoroacetic acid-d (TFA-d) (99.5% atom D, Sigma-Aldrich, Co., Ltd., St. Louis, MO, USA). The chemical shifts (δ) are reported as parts per million (ppm). 1H NMR [δ (ppm), in trifluoroacetic acid-d]:AlgPr [1.22 (-OOCCH2CH3), 2.5 (-OOCCH2CH3)], AlgHe [0.9 (-OOC(CH2)4CH3), 1.3 (-OOC(CH2)2(CH2)2CH3), 1.6, 1.7 (-OOCCH2CH2(CH2)2CH3), 2.5 (-OOCCH2(CH2)3CH3)], AlgOc [0.8 (OOC(CH6)CH3), 1.3 (-OOC(CH2)2(CH2)4CH3), 1.6 (-OOCCH2CH2(CH2)4CH3), 2.4 (-OOCCH2(CH2)5CH3)]. DS values were calculated by DS = (CH3/3)/(Ring-H/5). Gel Permeation Chromatography The number- and weight-average molecular weights (Mn and Mw) and polydispersity indices (Mw/Mn) were estimated by gel permeation chromatography (GPC) (LC-10ADVP system, Shimadzu, Co., Ltd., Kyoto, Japan, equipped with an RID-10A reflective index detector) in N,N-dimethylacetamide (DMAc) at 40 °C. The Alg esters were eluted at 0.6 mL/min and permeated through a polystyrene column (KD-804, 7.8 mm i.d. × 300 mm, Shodex Co., Ltd., Japan). Calibration curves were obtained by using pullulan standards (Shodex Co., Ltd.). Preparation of Solvent-Cast Films AlgPr, AlgHe, and AlgOc (250 mg) were dissolved in acetone (10 mL) and cast on a Teflon petri dish (5 cm diameter). The solvent was then evaporated in air at room temperature. The obtained films were dried in vacuo for 1 day to remove the solvent. Tensile Tests Tensile tests were carried out on the Alg ester cast films at room temperature using an EZ-test machine (Shimadzu, Co., Ltd.). The Alg esters films [4 mm (W) × 30 mm (L), initial gauge length of 10 mm] were stretched at a crosshead speed of 20 mm/min. 128

Thermogravimetric Analysis Thermogravimetric analysis (TGA) was carried out using an STA6000 (Perkin Elmer Inc., Waltham, MA, USA) in a nitrogen atmosphere. Each sample was heated from 30 to 500 °C at 10 °C/min.

Differential Scanning Calorimetry Differential scanning calorimetry (DSC) thermograms were recorded using a DSC8500 system (Perkin Elmer Inc., Waltham, MA, USA) in a nitrogen atmosphere. The Alg esters were initially heated from 30 to 210 °C at 20 °C/min and held at 200 °C for 1 min. Then, the samples were immediately cooled to -30 °C and held at that temperature for 5 min. The second heating scan was run from -30 °C to 200 °C at 20 °C/min. The temperature was calibrated using indium as an external standard.

Dynamic Mechanical Analysis Hot-pressed films were obtained by using a Mini Test Press 10 (Tokyo Seiki Co. Ltd., Tokyo, Japan). The powdered Alg esters were sanded using Kapton films and pressed at 200 °C (AlgPr), 180 °C (AlgHe), and 160 °C (AlgOc). The thickness of the films was adjusted to 500–700 μm using a 0.5-mm-thick stainlesssteel spacer. Dynamic mechanical analysis (DMA) measurements were performed on a DVA-2200S dynamic mechanical analyzer (IT Keisoku Seigyo Co., Ltd., Kyoto, Japan). The temperature sweep scans were performed at 100 Hz from 30 to 250 °C at a heating rate of 2 °C/min in a nitrogen atmosphere. The shearing strain applied to the sample was 0.05%.

Table 1. Characteristics of Alg Ester Derivatives DS

Yield(%) AlgHe

AlgOc

Mw/Mn

39/61_100cp

71.6

2

17.2

1.6

56/44_100cp

47.2

2

14.4

1.6

69/31_100cp

43.1

2

14.1

1.5

60.0

2

—

—

47.2

2

32.9

3.7

42.5

2

29.1

3.6

AlgPr AlgHe

Mw (104)

56/44_1000cp

129

Results and Discussion Preparation of Alg Esters Solvent exchange is an indispensable process for the complete esterification of the hydroxyl groups in Alg. The swellability of Alg in water depends on the M/G ratio and the molecular weight. In the present study, Alg 39/61_100cp, Alg 56/44_100cp, and Alg 69/31_100cp required 200 mL of water and swelling. Alg 56/44_1000cp required 500 mL of water to achieve maximum swelling Figure 1 shows side chain 1H NMR spectra of AlgHe, that was prepared from Alg with a different M/G ratio and molecular weight. The peaks of the representative side chain protons and ring protons were observed. DS of all the Alg esters calculated from the ratio of peaks area of methyl proton and sugar backbone was 2.0. This indicates that all the hydroxyl groups in the Alg were esterified. The averaged molecular weights are summarized in Table 1. The GPC analysis revealed that the AlgHe samples (39/61_100cp, 56/44_100cp, and 69/ 31_100cp) had similar molecular weights, namely, 17.2 × 104, 14.4 × 104, and 14.1 × 104, respectively. Because of the similar molecular weights of these AlgHe, the influence of molecular weight can be ignored when considering the effect of the M/G ratio. In contrast, AlgHe (56/44_1000cp) had a Mw of 32.9 × 104, which was much larger than the Mw values of the other AlgHe. The effect of molecular weight can be considered by comparing AlgHe (56/44_1000cp) with the other AlgHe samples.

Figure 1. 1H NMR spectra of four different AlgHe. 130

Solvent-Cast Films and Their Tensile Properties All the AlgHe samples were dissolved in acetone, irrespective of their M/G ratio or molecular weight. The AlgHe solvent-cast films were prepared from the acetone solutions. Figure 2 shows the translucent and self-standing cast films of AlgHe and their stress–strain curves. Whereas the AlgHe films (all 100cp) had similar tensile strengths (16.8–18.2 MPa) and Young’s moduli (0.26–0.30 GPa), the AlgHe (56/44_1000cp) film had a much higher tensile strength (26.8 MPa) and Young’s modulus (0.41 GPa). These results suggest that the M/G ratio does not affect the film properties. However, the molecular weight of the Alg esters has a significant effect on the film properties.

Figure 2. Cast films and their stress–strain curves of AlgHe sample of (a) 39/61_100cp, (b) 56/44_100cp, (c) 69/31_100cp, and (d) 56/44_1000cp.

Figure 3 shows the solvent-cast films of the Alg (56/44_1000cp) esters with different side chain lengths [(a), (b), and (c)] and their stress–strain curves. All the Alg esters were self-standing and transparent. Comparing among the Alg esters (56/44_1000cp), when the length of alkyl side chains decreased, the tensile strength and Young’s moduli of the Alg esters increased, whereas the elongations at break were almost the same. These results suggest that the strength of the Alg ester films increases with a decrease in the number of carbon atoms in the alkyl side chain, whereas the M/G ratio does not affect the film properties. The tensile properties of the Alg esters and conventional plastic materials—namely, polystyrene (PS), poly(methyl methacrylate) (PMMA), and polycarbonate (PC)—are summarized in Table 2 (24). The tensile strength of AlgPr (56/44_1000cp), which was 63 MPa, was higher than the tensile strengths of polystyrene and PMMA and was comparable to the tensile strength of PC. Furthermore, the 56/44_1000cp Alg ester had a larger elongation at break than did PS and PMMA. These results indicate that high-molecular-weight Alg esters have tensile properties that are comparable or even superior to those of conventional plastic materials. 131

Figure 3. Cast films and their stress–strain curves of Alg ester derivatives (56/44_1000cp): (a) AlgPr, (b) AlgHe, and (c) AlgOc.

Thermal Properties The thermal behaviors of the Alg esters were analyzed by TGA and DSC. The TGA thermograms are shown in Figure 4. The decomposition temperature at 5% weight loss (Td5%) increased from 170 °C to 200–250 °C following esterification of the Alg hydroxyl groups. However, there was no clear dependence on the M/G ratio or molecular weight. Except for AlgHe (56/44_1000cp), the AlgHe samples did not exhibit any characteristic peaks in the first run of DSC thermogram (data not shown). Although there was a small peak in the first DSC run thermogram of AlgHe (56/44_1000cp), it was too small to be considered a melting point. These results confirm that AlgHe is an amorphous polymer. However, there were no significant glass transitions during the second runs of DSC analysis. A similar absence of glass transition has also been reported based on the DSC results for ester derivatives of xylan, glucomannan, and pullulan (14–17, 25).

Figure 4. TGA thermograms of (1) (a) Alg and AlgHe samples of (b) 39/61_100cp, (c) 56/44_100cp, (d) 69/31_100cp, and (e) 56/44_1000cp and (2) Alg ester derivatives (56/44_1000cp): (A) AlgPr, (B) AlgHe, and (C). AlgOc. 132

DMA Results for Alg Ester Hot-Pressed Films Hot-pressed films of the Alg esters were prepared to estimate their glass transition temperature (Tg) values by DMA. Temperature dependence of the dynamic loss tangent (tan δ) is shown in Figure 5. According to the peaks of the loss tangent thermograms, the Tg values of the AlgHe were determined to be within the range of 160–182 °C [Figure 5(1)]. AlgHe (69/31_100cp) had a much higher tan δ value than the other Alg esters. Figure 5(2) shows the loss tangent curves of the (56/44_1000cp) Alg esters. According to the loss tangent curves, the Tg values of the Alg esters were 176, 182, and 192 °C for AlgOc, AlgHe, and AlgPr, respectively. The thermal properties of the polysaccharides are summarized in Table 2. The Tg values of the polysaccharides were much higher than those of conventional lplastics (polystyrene, PMMA, and polycarbonate) (24). These results indicate that the Alg esters are highly thermostable amorphous plastics.

Figure 5. DMA curves of (1) AlgHe samples of (a) 39/61_100cp, (b) 56/44_100cp, (c) 69/31_100cp, and (d) 56/44_1000cp and (2) Alg ester derivatives (56/44_1000cp): (A) AlgPr, (B) AlgHe, and (C) AlgOc. 133

Table 2. Tensile and Thermal Properties of Alg Ester Derivatives

AlgHe

Tensile Strength (MPa)

Elongation at break (%)

Young modulus (GPa)

Td5% (°C)

Tga (°C)

39/61_100cp

17 ± 1.1

20 ± 8.8

0.26 ± 0.03

253

160

56/44_100cp

17 ± 1.2

12 ± 2.0

0.27 ± 0.06

221

176

69/31_100cp

18 ± 1.8

10 ± 1.7

0.30 ± 0.05

250

171

64 ± 5

35 ± 12

0.74 ± 0.15

237

192

27 ± 2

35 ± 17

0.41 ± 0.13

204

182

18 ± 2

30 ± 4

0.25 ± 0.04

210

176

AlgPr AlgHe

56/44_1000cp

AlgOc

50

Polystylene

134

PMMA

50 60

Polycarbonate Glass transition temperature were measured by DMA. Onset temperature of weight decrease.

a

b

3

2

3

4

100

(26) Onset temperature of weight decrease.

2 c

ca.

300b

213c ca.

450d

(27) 10% weight decrease temperature.

100 105 145 d

(28)

Conclusions In this research, effects of monomer composition and molecular weight as well as side chain length of Alg esters to their physical properties were investigated. AlgHe were prepared from four kinds of Alg with different ratios of mannuronic acid (M) to gluronic acid (G) and different molecular weights. All the Alg esters formed self-standing film, which enabled us to investigate the effect of M/G ratio and Mw on the tensile properties by solvent-casting method. From tensile test of AlgHe samples, it was revealed that tensile properties were clearly affected by Mw but not by M/G ratio. The effect of side chain length was further investigated by Alg esters prepared from high Mw (1000cp) Alg. It was revealed that the Alg esters with a shorter side chain showed higher tensile strength, while the elongation at break was almost unaffected. In particular, AlgPr (56/44_1000cp) gave the highest tensile strength (63 MPa) and glass transition temperature (192 °C), which are higher than those of most amorphous plastics. The results of the present study indicate that Alg esters have great potential as plastic materials.

Acknowledgments This work was carried out as part of the “Innovative Synthesis of High-Performance Bioplastics from Polysaccharides” project supported by JST ALCA Grant Number JPMJAL1502, Japan.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

Draget, K. I.; Smidsrød, O.; Skjåk-Bræk, G. Carbohydr. Polym. 1994, 25 (1), 31–38. Trujillo-Roldán, M. A.; Moreno, S.; Segura, D.; Galindo, E.; Espín, G. Appl. Microbiol. Biotechnol. 2003, 60, 733–737. Venegas, M.; Matsuhiro, B.; Edding, M. Bot. Mar. 1993, 36, 47–51. Matsumoto, T.; Mashiko, K. Biopolymers 1990, 29, 1707–1713. Grant, G. T.; Morris, E. R.; Rees, D. A.; Smith, P. J. C.; Thom, D. FEBS Lett. 1973, 32, 195–198. Remminghorst, U.; Rehm, B. H. A. Biotechnol. Lett. 2006, 28, 1701–1712. Andersen, T.; Strand, B. L.; Formo, K.; Alsberg, E.; Christensen, B. E. Carbohydr. Chem. 2012, 37, 227–258. Vos, P. D.; Haan, B. D.; Schilfgaarde, R. V. Biomaterials 1997, 18, 273–278. Ertesvåg, H.; Valla, S. Polym. Degrad. Stabil. 1998, 59, 85–91. Edgar, K. J.; Buchanan, C. M.; Debenham, J. S.; Rundquist, P. A.; Seiler, B. D.; Shelton, M. C.; Tindal, D. Prog. Polym. Sci. 2001, 26, 1605–1688. Morooka, T.; Norimoto, M.; Yamada, T.; Shiraishi, N. J. Appl. Polym. Sci. 1984, 29, 3981–3990. Crepy, L.; Miri, V.; Joly, N.; Martin, P.; Lefebvre, J. M. Carbohydr. Polym. 2011, 83, 1812–1820. Marubayashi, H.; Yukinaka, K.; Enomoto-Rogers, Y.; Takemura, A.; Iwata, T. Carbohydr. Polym. 2014, 103, 427–433. 135

14. Enomoto-Rogers, Y.; Ohmomo, Y.; Takemura, A.; Iwata, T. Carbohydr. Polym. 2014, 101, 529–599. 15. Enomoto-Rogers, Y.; Ohmomo, Y.; Iwata, T. Carbohydr. Polym. 2013, 92, 1827–1834. 16. Danjo, T.; Enomoto-Rogers, Y.; Takemura, A.; Iwata, T. Polym. Degrad. Stabil. 2014, 109, 373–378. 17. Enomoto-Rogers, Y.; Iio, N.; Takemura, A.; Iwata, T. Eur. Polym. J. 2015, 66, 470–477. 18. Puanglek, S.; Kimura, S.; Enomoto-Rogers, Y.; Kabe, T.; Yoshida, M.; Wada, M.; Iwata, T. Sci. Rep. 2016, 6, 30479. 19. Schweiger, R. G. J. Org. Chem. 1962a, 27, 1786–1789. 20. Schweiger, R. G. J. Org. Chem. 1962b, 27, 1789–1791. 21. Matsumoto, Y.; Ishii, D.; Iwata, T. Carbohydr. Polym. 2017, 171, 229–235. 22. Martinez-Gomez, F.; Encinas, V. M.; Matsuhiro, B.; Pavez, J. J. Appl. Polym. Sci. 2015, 132, 42398–42408. 23. Andriamanantoaninaa, H.; Rinaudob, M. Polym. Int. 2010, 59, 1531–1541. 24. Billmeyer, F. W. J. Textbook of polymer science; J. Wiley: New York, 1984. 25. Fundador, N. G. V.; Enomoto-Rogers, Y.; Takemura, A.; Iwata, T. Polymer 2012, 53, 3885–3893. 26. Mehta, S.; Biederman, S.; Shivkumar, S. J. Mater. Sci. 1995, 30, 2944–2949. 27. Gałka, P.; Kowalonek, J.; Kaczmarek, H. J. Therm. Anal. Calorim. 2014, 115, 1387–1394. 28. Zhou, W.; Yang, H.; Zhou, J. J. Anal. Appl. Pyrol. 2007, 78 (2), 413–418.

136

Chapter 10

Preparation of Hydrophobically Modified Cashew Gum Through Reaction with Alkyl Ketene Dimer Atanu Biswas,*,1 Sanghoon Kim,1 Megan Buttrum,1 Roselayne F. Furtado,2 Carlucio R. Alves,3 and H. N. Cheng*,4 1National Center for Agricultural Utilization Research, USDA Agricultural Research Services, 1815 N. University Street, Peoria, Illinois 61604, United States 2Embrapa Agroindústria Tropical, Rua Dra. Sara Mesquita 2270, CEP 60511-110, Fortaleza, CE, Brazil 3State University of Ceará, Chemistry Department, Silas Munguba Av. 1.700, 60740-020, Fortaleza, CE, Brazil 4Southern Regional Research Center, USDA Agricultural Research Service, 1100 Robert E. Lee Boulevard, New Orleans, Louisiana 70124, United States *E-mail: [email protected]. *E-mail: [email protected].

Hydrophobic modification of polysaccharides is useful for imparting special properties, such as viscosity improvement and additive compatibility. An interesting reagent for hydrophobic modification is alkyl ketene dimer (AKD). A brief review is given here of AKD-modified polysaccharides. In addition, new data relating to AKD-modified cashew gum (CG) are presented. Reactions have been conducted at 90 ° C, using dimethyl sulfoxide as a solvent and 4-dimethylaminopyridine as a promoter. Samples with degrees of substitution (DS) up to 0.016 have been made; the higher DS samples tend to form insoluble dispersions. This new polymer shows some of the properties of other AKD-adducted polysaccharides and may be a useful addition to the family of hydrophobically modified polysaccharides.

© 2018 American Chemical Society

Introduction Hydrophobically modified water-soluble polymers constitute a useful class of industrial additives that can impart surfactant-like behavior and desirable rheological properties to aqueous polymer solutions and dispersions (1, 2). These are usually hydrophilic polymers with hydrophobic side chains or terminal groups. The hydrophobic groups often exhibit associative behavior in water, thereby providing shear-thinning rheological characteristics. Some well-known examples of these polymers are hydrophobically modified hydroxyethyl celluloses (HMHECs) (3, 4), hydrophobically modified ethoxylated urethanes (HEURs) (5, 6), and hydrophobically modified alkali-swellable emulsions (HASEs) (7–9). Polysaccharides are especially amenable to hydrophobic modification, and many polysaccharide derivatives have been reported (3, 4, 10–12). These are frequently studied for their rheological behavior (13, 14). More recently, some hydrophobically modified polysaccharides have been used for enzyme immobilization (15) and as drug carriers (16, 17). Alkyl ketene dimer (AKD, Scheme 1) is a relatively inexpensive, commercially available sizing agent for paper (18, 19). It is synthesized from saturated or unsaturated fatty acids, with comparatively few byproducts after reactions. Its β-lactone functionality permits the reaction with hydroxy groups on polysaccharides. Indeed, a fair amount of papers have appeared on its reaction with polysaccharides; a summary of the recent AKD/polysaccharide derivatives is given in Table 1.

Scheme 1. Chemical structure of AKD.

In Table 1, we have excluded the use of AKD for paper sizing where many publications exist (18, 19). It is understood that in paper sizing at least some of the AKD reacts with the cellulose chains on paper to form covalent bonds (20, 21). Recently, there has been renewed interest in reacting AKD with different types of cellulose, such as microcrystalline cellulose (22), microfibrillated cellulose (23, 24), cellulose nanofibers (25, 26), and bacterial cellulose (27). In addition, several publications have appeared on the reaction of AKD with water-soluble polysaccharides. These include AKD adducts of carboxymethyl cellulose (28), hydroxyethyl cellulose (29), and starch (30, 31). The reaction procedures can involve no solvent (25, 28), solvents that dissolve both AKD and polysaccharides (22, 24, 29, 30), or solvents that dissolve only AKD (23, 26, 31). 138

Table 1. Derivatives of AKD on Polysaccharides Solventa used (catalyst)

Substrate

Refs.

Cellulose, microcrystalline

1,3-dimethyl-2-imidazolinone/LiCl

(22)

Cellulose, microfibrillated

EtAc (1-methylimidazole)

(23)

Cellulose, microfibrillated

DMAc/LiCl

(24)

Cellulose nanofibers

No solvent, aqueous AKD dispersion

(25)

Cellulose nanofibers

Chloroform

(26)

Cellulose, bacterial

Supercritical CO2

(27)

Carboxymethyl cellulose

No solvent, aqueous AKD dispersion

(28)

Hydroxyethyl cellulose

DMF, DMAc, DMSO (enzyme)

(29)

Starch

DMF, DMAc, DMSO (enzyme)

(30)

Starch, microporous

Ethanol

(31)

Cashew gum (CG)

DMSO

This work

a DMAc = N,N-dimethylacetamide; DMF = N,N-dimethylformamide; DMSO = dimethyl sulfoxide; EtAc = ethyl acetate.

In this work, we have used AKD to hydrophobically modify cashew gum (CG). CG is a bark exudate from the cashew tree (Anacardium occidentale L.) (32, 33). It is a hydrophilic, branched high-molecular-weight polysaccharide comprising 72% β-D-galactopyranose, 14% α−D-glucopyranose, 4.6% α-L-arabinofuranose, 3.2% α-L-rhamnopyranose, and 4.5% β-D-glucuronic acid. The presence of glucuronic acid in CG imparts an anionic character to this polymer. In cold water, it swells into a gel but dissolves rapidly when heated. It has been used for food, drug, and cosmetic applications as an encapsulant, binder, emulsion former, and rheology modifier (32, 33). Although many derivatives of CG are known, thus far no report has appeared on the AKD derivatives of CG.

Experimental Materials Gum exudate from the cashew tree was collected from Embrapa Experimental Station at Pacajus (Fortaleza-Ceará). It was ground to 100 mesh particle size, dissolved in water, centrifuged at 10,000 rpm at 4 °C for 20 min, and filtered to remove insoluble materials. Water/ethanol mixture at 1:4 (v/v) ratio was added to the solution for 24 h to precipitate the polysaccharides. The precipitate was filtered, washed exhaustively with acetone, and dried in a hot air circulation oven. This final material is called CG in this work. AKD (AquapelTM 364) was a gift from Solenis, LLC (Wilmington, DE, USA). Other chemicals were acquired from Sigma-Aldrich (St. Louis, MO, USA). 139

Synthesis of AKD-CG The CG was dried overnight in a vacuum oven at 70–80 °C before use. Typically, 1 g of CG, 0.02–0.5 g of AKD, and 0.125–0.15 g of 4-dimethylaminopyridine (DMAP) were dissolved in 4 mL DMSO and heated at 90 °C for 3–18 h in a Reacti-ThermTM (Fisher Scientific, Pittsburgh, PA, USA) with constant stirring. The reaction mixture was then cooled down, and isopropanol was added dropwise with stirring to prevent any large clumps from forming. The precipitate was washed with methylene chloride and then dried under a vacuum at 50–75 °C overnight.

NMR Analysis For NMR analysis, each AKD-CG sample was dissolved in d6-DMSO in an NMR tube (at a concentration of 10% or higher). The 1H NMR spectra were acquired on a Bruker DRX 400 spectrometer (Karlsruhe, Germany) at ambient temperature using standard operating conditions. The 1H chemical shifts were referenced to tetramethylsilane at 0 ppm.

Solution Viscosity The viscosity of AKD-CG solutions was measured with a Vibro viscometer (Model SV-10, A&D Company, Tokyo, Japan). Each AKD-CG sample was dissolved in deionized water to make a 4% solution and transferred to a glass sample cup (capacity of 13 mL) mounted in the water jacket. The temperature of the sample cup was maintained at 20 °C during the measurement by using a water circulator (Polystat Model 106, Cole Parmer, Vernon Hills, IL, USA). Viscometer reading was calibrated against water at 20 °C.

Results and Discussion The reaction of AKD is shown in Scheme 2. In the literature, the reaction of AKD with a polysaccharide is usually carried out in a polar aprotic solvent such as N,N-dimethylformamide (DMF) (29), N,N-dimethylacetamide (DMAc) (29, 30), dimethyl sulfoxide (DMSO) (30), 1,3-dimethyl-2-imidazolidinone (22), or t-butyl methyl ether (29). We chose DMSO as the solvent. An initial study was made to zoom in on the experimental parameters needed (Table 2). For samples A1–A4, the reaction time changed from 1 to 18 h. After 1 h of reaction, the degree of substitution (DS) was 0.0022, but at 3, 6, and 18 h, the DS was relatively constant at 0.0027–0.0030. The samples (A1, A5, and A6) reflected the temperature dependence of the reaction. Interestingly, the DS stayed roughly the same at 90 °C and 70 °C but decreased at 120 °C. The samples A1, A7, and A8 were meant to show the effect of DMAP. No observed difference in DS was observed at different DMAP levels. In view of the data obtained, we proceeded to use the reaction conditions for sample A1 for most of this work. 140

Scheme 2. Reaction of AKD with CG.

Table 2. Degree of Substitution of AKD-CG Products as a Function of Reaction Time and Temperature CG (g)

Sample #

AKD (g)

DMAP (g)

DMSO (mL)

Temp (° C)

Time (h)

DS

A1

1

0.064

0.15

4

90

3

0.0030

A2

1

0.064

0.15

4

90

1

0.0022

A3

1

0.064

0.15

4

90

6

0.0029

A4

1

0.064

0.15

4

90

18

0.0027

A5

1

0.064

0.15

4

120

3

0.0015

A6

1

0.064

0.15

4

70

3

0.0033

A7

1

0.064

0

4

90

3

0.0032

A8

1

0.064

0.3

4

90

3

0.0032

The 1H NMR spectrum of AKD is shown in Figure 1. The CG contains multiple sugar residues, and their peaks can be found under the broad spectral pattern at ca. 3.0–5.5 ppm. The protons attached to C1’s of the sugars occur at the downfield region, together with the hydroxy groups (4.2–5.5 ppm) (34–36). The peak at 1.1 ppm is due to the methyl peak in rhamnose, one of the sugars present in CG. The peaks at ca. 1.25 and 0.8 ppm correspond to the CH2 and the CH3 of AKD. The peak area for the CG at 3.0–5.5 ppm (A1) corresponds approximately to 10 protons on the anhydroglucose basis. Since each AKD molecule contains two methyl groups and each methyl has three hydrogens, the methyl area (A2) needs to be divided by six. Thus, the DS can be calculated via the following expression: 10*A2/(6*A1). We then proceeded to vary the amount of AKD used in the reaction. As it turned out, the DS of the products was almost linearly proportional to the amount of AKD used (up to about DS 0.006) but then started to level off at higher DS values (Figure 2). Also, near DS 0.006, a precipitate could be visibly observed; the amount of precipitate increased as more AKD was added. Since AKD is highly hydrophobic, it is not surprising that higher substitution levels of AKD caused 141

increasing aggregation of the hydrophobic groups and increasing insolubility of the product. This phenomenon was also observed for AKD derivatives of starch and hydroxyethyl cellulose (29, 30).

Figure 1. 1H NMR spectrum of sample A1 dissolved in d6-DMSO.

Figure 2. The degree of substitution as a function of weight of AKD used in the reaction. 142

The aggregative property of AKD-CG samples suggested that these samples might also show increased viscosity. The solution viscosities of 4% solutions of the AKD-CG products are shown in Table 3 and Figure 3. The viscosity increased up to DS 0.006, but at higher DS values, the viscosity dropped again due to precipitate formation.

Table 3. Viscosity of 4% Aqueous Solutions of AKD-CG Derivatives as a Function of DS and Other Experimental Parameters Sample

AKD (g)

DMAP (g)

B1

0

B2

0.02

0.15

B3

0.064

B4

DS

Time (h)

Viscosity (cP)

0

1.32

3

0.0013

1.35

0.15

3

0.0029

1.37

0.084

0.15

3

0.0053

1.58

B5

0.104

0.15

3

0.0059

1.45

B6

0.25

0.125

19

0.0112

1.4

B7

0.5

0.125

19

0.0160

1.39

Figure 3. Viscosity of 4% aqueous solutions of AKD-CG derivatives as a function of DS. 143

Thus, the properties of the AKD–CG adducts depend on the degree of substitution. For a soluble product, the DS needs to be less than 0.006. For a DS 0.006 or higher, a dispersion is obtained. In the literature, an anionically and hydrophobically modified starch has been used for the formation of emulsions (37). A hydrophobically modified anionic polymer has been employed successfully to disperse graphite particles in an aqueous solution (38). An anionically and hydrophobically modified cellulose derivative has also been helpful in the suspension polymerization of a vinyl monomer (39). In addition, aqueous dispersions of hydrophobically modified polymers have been used as flocculants for treatment of oily wastewater (40). Perhaps this new hydrophobically modified CG can be considered for some of these applications as well.

Conclusions This work is part of our efforts to derivatize agro-based raw materials to diversify their structures and to assess their properties. Since CG is widely available, we have modified it with a hydrophobic reagent (AKD) which can impart a surfactant-like property to this polymer. AKD-CG samples with degrees of substitution up to 0.016 have been made, and the chemical structures have been verified with NMR. The samples with degrees of substitution higher than 0.006 tend to form dispersions. Further work will be needed to explore possible applications for this new polymer.

Acknowledgments The authors thank Karl Vermillion for supplying the NMR spectra. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. USDA is an equal opportunity provider and employer.

References 1. 2. 3. 4. 5. 6. 7. 8.

Associative Polymers in Aqueous Media; Glass, J. E., Ed.; ACS Symposium Series 765; American Chemical Society: Washington, DC, 2000. Winnik, M. A.; Yekta, A. Curr. Opin. Colloid Interface Sci. 1997, 2, 424–436. Landoll, L. M. J. Polym. Sci., Polym. Chem. Ed. 1982, 20, 443. Tanaka, R.; Meadows, J.; Williams, P. A.; Phillips, G. O. Macromolecules 1992, 25, 1304–1310. Emmons, W. D.; Stevens, T. E. U.S. Patent 4,079,028, March 14, 1978. Glass, J. E. Adv. Chem. Ser. 1986, 213, 391. Shay, G. D.; Eldridge, E.; Kail, E. U.S. Patent 4,514,552, April 30, 1985. Shay, G. D.; Kravitz, F. K.; Brizgys, P. V.; Kersten, M. A. U.S. Patent 4,801,671, January 31, 1989. 144

9. 10. 11. 12. 13. 14. 15.

16. 17. 18. 19. 20. 21. 22. 23.

24. 25. 26. 27. 28. 29. 30. 31. 32. 33.

34. 35.

36. 37.

Shay, G. D.; Kravitz, F. K.; Brizgys, P. V. ACS Symp. Ser. 1991, 462, 121. Cunha, A. G.; Gandini, A. Cellulose 2010, 17, 875–889. Cunha, A. G.; Gandini, A. Cellulose 2010, 17, 1045–1065. Cheng, H. N.; Gu, Q. M. Polymers 2012, 4, 1311–1330. Miller, D.; Löffler, M. Coll. Surf., A 2006, 288, 165–169. Henni, W.; Deyme, M.; Stchakovsky, M.; LeCerf, D.; Picton, L.; Rosilio, V. J. Coll. Interf. Sci. 2005, 281, 316–24. Mohan, T.; Rathner, R.; Reishofer, D.; Koller, M.; Elschner, T.; Spirk, S.; Heinze, T.; Stana-Kleinschek, K.; Kargl, R. Biomacromolecules 2015, 16, 2403–2411. Jung, B.; Shim, M.-K.; Park, M. J.; Jang, E. H.; Yoon, H. Y.; Kim, K.; Kim, J.H. Int. J. Pharmaceutics 2017, 520, 111–118. Na, K.; Lee, T. B.; Park, K.-H.; Shin, E.-K.; Lee, Y. B.; Choi, H.-K. Eur. J. Pharm. Sci. 2003, 18, 165–173. Brander, J.; Thorn, I. Surface Application of Paper Chemicals; Blackie Academic: London, 1997. Reynolds, W. F. The Sizing of Paper, 2nd ed.; TAPPI: Atlanta, 1989. Bottorff, K. J.; Sullivan, M. J. Nordic Pulp Paper Res. J. 1993, 8, 86–95. Nahm, S. H. J. Wood Chem. Technol. 1986, 6, 89–112. Yoshida, Y.; Isogai, A. Cellulose 2007, 14, 481–488. Yan, Y.; Amer, H.; Rosenau, T.; Zollfrank, C.; Dorrstein, J.; Jobst, C.; Zimmermann, T.; Keckes, J.; Veigel, S.; Gindl-Altmutter, W.; Li, J. Cellulose 2016, 23, 1189–1197. Song, X.; Chen, F.; Liu, F. Carbohydr. Polym. 2012, 88, 417–421. Delgado-Aguilar, M.; Gonzalez, I.; Jimenez, A. M.; Tarres, Q.; Quintana, G.; Mutje, P. Cellulose Chem. Technol. 2016, 50, 369–375. Missoum, K.; Julien, B.; Naceur, B. Patent Applications, U.S. 2016168696 (A1); W.O. 2015011364 (A2). Russler, A.; Wieland, M.; Bacher, M.; Henniges, U.; Miethe, P.; Liebner, F.; Potthast, A.; Rosenau, T. Cellulose 2012, 19, 1337–1349. Lähteenmäki, M.; Känköhen, H.; Kloow, G.; Ruppert, O.; Leupin, J. A.; Gosselink, E. P. U.S. Patent 6,600,033, July 29, 2003. Gu, Q.-M.; Cheng, H. N. ACS Polym. Prepr. 2005, 46, 30–31. Qiao, L.; Gu, Q.-M.; Cheng, H. N. Carbohydr. Polym. 2006, 66, 135–140. Dang, C.; Xu, M.; Yin, Y.; Pu, J. BioResources 2017, 12, 5775–5789. Kumar, A.; Moin, A.; Shruthi, R.; Ahmed, A.; Shivakumar, H. G. Curr. Drug Ther. 2012, 7, 2–12. Ribeiro, A. J.; Souza, F. R. L.; Bezerra, J. M. N. A.; Oliveira, C.; Nadvorny, D.; Soares, M. F. R.; Nunes, L. C. C.; Silva-Filho, E. C.; Veiga, F.; Soares-Sobrinho, J. L. Carbohydr. Polym. 2016, 147, 188–200. Cheng, H. N.; Neiss, T. G. Polym. Rev. 2012, 52, 81–114. Quelemes, P. V.; Araújo, A. R.; Plácido, A.; Delerue-Matos, C.; Maciel, J. S.; Bessa, L. J.; Ombredane, A. S.; Joanitti, G. A.; Soares, M. J. S.; Eaton, P.; Silva, D. A.; Leite, J. R. S. A. Carbohydr. Polym. 2017, 157, 567–575. Pitombeira, N. A. O.; Neto, J. G. V.; Silva, D. A.; Feitosa, J. P. A.; Paula, H. C. B.; Paula, R. C. M. Carbohydr. Polym. 2015, 117, 610–615. Nilsson, L.; Bergenståhl, B. J. Colloid Interface Sci. 2007, 308, 508–513. 145

38. Xu, Q.; Somasundaran, P. In Adsorption of Nonionic Surfactants, Anionic/Nonionic Surfactant Mixtures, and Hydrophobically Modified Polymers on Minerals and its Effect on Their Flotation and Dispersion; Proceedings of the XVIII International Mineral Processing Congress, Sydney, Australia, March 23-28, 1993. https://pdfs.semanticscholar.org/ cc3e/05df7c872502f1a26124bf98026232a7e13e.pdf (accessed July 16, 2018) 39. Craig D. H. U.S. Patent 4,868,238, September 19, 1989. 40. Lü, T.; Qi, D.; Zhao, H.; Cheng, Y. Polym. Sci. Eng. 2015, 55, 1–7.

146

Bio-Related Polyesters, Polyamides, and Polyurethanes

Chapter 11

Salicylic Acid-Based Poly(anhydride-esters): Synthesis, Properties, and Applications Yue Cao and Kathryn E. Uhrich* Department of Chemistry, University of California, Riverside, California 92354, United States *E-mail: [email protected].

Salicylic acid is known for its various therapeutic effects, including antipyretic, anti-inflammatory, and antimicrobial capabilities. To address the issue of short in vivo half-life, this bioactive molecule was incorporated into the polymer backbone with linker molecules to yield the biocompatible and biodegradable salicylic acid-based poly(anhydride-esters) (SAPAE). Localized sustained release of salicylic acid was realized upon the hydrolytic degradation. Polymer properties such as degradation were modulated by the linker molecule. In this review, the influence of synthetic strategies, structural property relationships, and biomedical applications are highlighted.

Introduction Salicylic acid (SA) is a naturally occurring phenolic acid and an important active metabolite of aspirin. It has many therapeutic effects, including antipyretic, anti-inflammatory, and antimicrobial properties (1, 2). The earliest evidence of SA’s medicinal use goes back over 3500 years (3–5). Ancient Egyptians, early Romans, ancient Chinese, early Native Americans, and Southern African Khoikhoi peoples used willow leaf extracts containing SA to treat various inflammatory disorders (4). SA still plays a key role in modern medicine and is on the World Health Organization’s (WHO) Model List of Essential Medicines, which covers the most important medications needed in a basic health system (6). © 2018 American Chemical Society

Similar to other small molecule drugs, the pharmaceutical use of SA is limited owing to its short half-life in vivo (about 2 h) (7, 8). Repeated administration, often used to maintain an effective dose, may result in side effects and potential toxicity (9, 10). To achieve controlled sustained release of SA, a polymer delivery system of SA-based poly(anhydride-esters) (SAPAE) was developed, as shown in Figure 1. The SA is chemically incorporated into the polymer backbone with the “linker” molecule. Upon hydrolysis of the anhydride and ester bonds, SA and the linker molecule are released.

Figure 1. SAPAE degradation releases SA and adipic acid upon anhydride–ester bond hydrolysis.

Compared to other polymer delivery systems, the PAE exhibits some significant advantages. First, the high loading of SA: it makes up 71 wt % of the polymer 1, which contains the adipic acid linker (Figure 1). Second, the release of SA can be modified by varying the linker groups (11). In contrast, the homopolymer of SA does not have tunability in releasing bioactives (12, 13). Third, PAE generally undergoes surface erosion, resulting in a sustained, near zero-order bioactive release. In this review, we will introduce different aspects of SAPAE, including the polymer chemistry, structural property relationships, and applications as a biomaterial.

Polymerization Methods Several polymerization methods have been successfully utilized in the synthesis of polyanhydrides, including melt condensation, solution polymerization, interfacial polymerization, and ring-opening polymerization (14–18). In our SAPAE system, melt condensation and solution polymerization are applicable. The comparison and investigation of the two synthetic strategies 150

provide an in-depth understanding of the chemistry involved in the polymerization, as well as the influence of polymerization method on the polymer properties. Melt condensation polymerization is widely used in polyanhydride synthesis owing to its ease and reproducibility, as well as its great potential for scale-up (19–21). However, due to high temperatures (over 150 °C) utilized in the process, melt condensation has limitations, including reversible thermal depolymerization, the formation of cyclic molecules, and a thermal stability requirement for monomers (15, 20, 21). In contrast, the formation of polyanhydrides at or even below room temperature could be realized through solution polymerization (14, 16–18). The main limitations of solution polymerization are the strict, required, exact stoichiometry and its frequent yields of polymers with low molecular weights (14, 15). As illustrated in Figure 2, the polymer 4 can be obtained via two synthetic routes: melt condensation and solution polymerization (22). The two polymerization strategies share the same intermediate 2, which is obtained through a modified esterification reaction (23). For the melt condensation method, the diacid 2 reacts with excess acetic anhydride to form the mixed anhydride 3. The monomer 3 is heated under a vacuum to yield the SAPAE, with acetic anhydride as a byproduct. For solution polymerization, the triphosgene is used as a coupling agent for diacid 2 in the presence of triethylamine at 0 °C.

Figure 2. Modified synthetic route to salicylic diced 2. SAPAE 4 prepared via solution polymerization and melt condensation.

As shown in Table 1, the melt condensation polymer exhibited a higher molecular weight than the solution-made polymer, which contributed to a higher glass transition temperature, degradation temperature, and Young’s modulus (24). A similar trend was observed in substituted SAPAE (24). 151

Table 1. Properties of Polymer 1 Synthesized via Melt Condensation and Solution Polymerization Polymer

Mw

Polydispersity index (PDI)

Tg (° C)c

Td (° C)d

Young’s modulus (kPa)

1a

31,800

1.8

46

290

1500

1b

10,000

1.1

23

240

330

a

b

Obtained via melt condensation; data from reference (23). Obtained via solution polymerization; data from reference (23). c Tg is the glass transition temperature. d Td is the thermal decomposition temperature.

Figure 3. 13C NMR (DMSO-d6) spectra of polymer made via solution polymerization 4a (top) and melt condensation. Adapted with permission from reference (20). Copyright 2008 Taylor & Francis. 152

Spectroscopy analysis provided more detailed structural information on the differences of polymers obtained through the two synthetic methods (22). As shown in Figure 3, 13C NMR spectra of the solution-made polymer 4a showed 13 distinct carbon peaks that clearly correlated to the expected structure. In contrast, the melt condensation polymer exhibited more than 20 additional carbon peaks, indicating a more complex composition that is different from the solution-made polymer. Differences in chemical structure were also confirmed by the IR spectra: only one type of anhydride–ester bond was observed for the solution-made polymer 4a, whereas multiple anhydride–ester bonds were noted for the melt condensation polymer. The formation of a complex composition in the melt condensation polymer was considered relevant to the salicylate–anhydride rearrangement. The proposed mechanism is shown in Figure 4. The ortho-ester groups of the salicylate 5 react with the adjacent aromatic anhydrides to form aromatic ester–anhydride 6 at a high temperature during melt condensation. The driving force of rearrangement may be the formation of thermodynamically stable aromatic ester. The thermodynamic rearrangement was further confirmed by the control experiment: the solution-made polymer exhibited additional 13C NMR and IR signals after it was exposed to melt condensation conditions. Besides rearrangement, the anhydride exchange at high temperature is also attributed to a complex composition, which yields a mix of polymers with high and low molecular weights, as well as cyclic polymers (15). Recently, a one-pot melt condensation strategy was developed for the synthesis of SAPAE (25). As shown in Figure 5, the mix anhydride monomer was obtained through the solvent-free multicomponent reaction. Compared with the traditional melt condensation method shown in Figure 2, the one-pot polymerization strategy significantly reduced raw material consumption and increased synthetic efficiency, as well as minimized waste generation. The SAPAE obtained from one-pot polymerization had similar molecular weight, thermal properties, and cytocompatibility. The development of a one-pot strategy not only increases the synthetic efficiency and reduces the use of chemicals, but also enables industrial-level scale-up and the low-cost manufacturing of SAPAE.

Structural Property Relationships As mentioned above, a great advantage of PAE over other controlled release systems is its tunable release property. PAEs with different linkers have different thermal properties, hydrophobicity, and degradation rates. Figure 6 illustrates the chemical structures of different linkers incorporated into the PAE backbones, including linear, branched, and oxygen-containing aliphatic as well as aromatic blocks (11, 26).

153

Figure 4. Proposed mechanism of the salicylate–anhydride rearrangement during melt condensation.

Figure 5. One-pot polymerization of SAPAE with adipic linker.

Figure 6. Chemical structures of different linkers incorporated into the backbone of SAPAE. 154

PAE with the linear aliphatic linkers exhibits a monotonic relationship between the length of the linker and the physical properties. As shown in Figure 7, molecular weight and thermal decomposition temperature (Td) increase as the linker length increases. Decreased glass transition temperature (Tg) is owing to the increased flexibility. Polymers with aromatic linkers have the highest Td and Tg due to the relatively stable electronic structure and rigid skeleton of aromatic systems (Table 2). For polymers 13 and 14, steric hindrance that originates from the bulk branch chain decreases the dense packing of polymer chains, resulting in greater free volume and lowered Tg. The diglycolic linker in 15 and glutaric linker in 7 are similar in length and size but different in hydrophobicity, so the comparison of 15 and 7 would be interesting (26). The lower Tg of 15 compared to 7 is due to the higher flexibility of the C–O bond than the C–C bond (27, 28).

Figure 7. Influence of linear aliphatic chain length on SAPAE properties. Adapted with permission from reference (9). Copyright 2005 American Chemical Society.

Degradation tests were performed at 37 °C in the phosphate buffered saline (PBS) buffer solution (pH = 7.4). Polymers with linear aliphatic linkers exhibited similar release behaviors: they fully degraded in 8–16 days with a short lag time (0–3 days) (Figure 8). Polymer 7 was fully degraded in 10 days. For polymers 11 and 12 with aromatic linkers, a longer lag time and degradation time (20 days) were needed. For polymers 13 and 14 with branched linkers, less than 20% was degraded in 30 days (11). Polymer 15, with the oxygen-containing linker, exhibited the fastest degradation rate, which fully degraded in 24 hours (26). 155

Table 2. Properties of SAPAEs with Different Linkers Mw

Polydispersity index (PDI)

Tg (° C)

Td (° C)

Contact angle (degree)

12,600

1.8

50

234

75

11a

17,100

1.8

118

403

81

12a

8800

1.7

106

405

77

Polymer

Linear aliphatic 7a Aromatic

Branched aliphatic 13a

39,800

1.5

44

261

95

14a

16,500

1.6

23

276

93

1.5

35

326

49.5a

Oxygen-containing aliphatic 15b a

14,500

Data from reference (9).

b

Data from reference (24).

Here, water contact angles were measured to evaluate the relative hydrophobicity, which may influence hydrolytic degradation rate. Polymers with branched linkers have the highest contact angle over 90 degrees, which indicates their hydrophobic nature, as well as the lowest degradation rates (11). The polymer exhibiting the fastest degradation, polymer 15 with the hydrophobic diglycolic linker, showed the smallest contact angle of 49.5 degrees, whereas that of polymer 7 was 75 degrees (26). Copolymerization is a widely used strategy to modify the polymer properties by chemically incorporating another repeating unit into the copolymer backbone. In SAPAE systems, the mechanical properties, thermal properties, and degradation behaviors could be tuned to meet the needs of various applications (26, 29). Copolymer 16 was obtained by incorporating the para-carboxyphenoxyhexane (pCPH) repeat unit (Figure 9). Compared to that of the homopolymer, increased Tg (33 °C versus 27 °C) greatly enhanced the processability of 16 at room temperature. A stronger influence on the copolymer’s mechanical properties was observed. Owing to the brittle nature of the pCPH monomer, the ductility of the copolymer decreased with the increase of pCPH content (29). Copolymers with higher pCPH content exhibited lower tensile strength, toughness, ultimate elongation, as well as a higher Young’s modulus. Meanwhile, a higher content of hydrophobic pCPH in copolymers resulted in a slower degradation rate, and therefore increased the stability of the degradable copolymer (29). The better processability, stronger mechanical property, and enhanced stability make these copolymers great candidates for bioactive-releasing implantable materials. Similarly, incorporating the hydrophilic monomer, diglycolic acid, led to a relatively hydrophilic copolymer 17, which has a smaller contact angle of 65.4 degrees compared to that of homopolymer 1 (83 degrees) (26). 156

157

Figure 8. In vitro degradation profile of SAPAE with (a) linear aliphatic linkers, (b) aromatic linkers, (c) branched aliphatic linkers, and (d) oxygen-containing aliphatic linkers. (a), (b), and (c) are adapted with permission from reference (9). Copyright 2005 American Chemical Society. (d) is adapted with permission from reference (24). Copyright 2009 John Wiley and Sons.

Figure 9. Chemical structures of copolymers 16 and 17.

Biomedical Applications As described above, the bioactive molecule SA has several therapeutic properties, including antipyretic, anti-inflammatory, and antimicrobial capabilities. Owing to the ability of a localized and sustained release of SA, SAPAE has great potential in biomedical applications. The antibacterial property of SAPAE was evaluated through an in vitro bacterial adhesion test. As shown in Figure 10, the Pseudomonas aeruginosa biofilm formation on an SAPAE surface was hindered by the released SA, which prevents cell accumulation by five orders of magnitude compared to an inactive control (30). The in vivo implantation experiment further confirmed the anti-inflammatory effect of SAPAE. By modulating the inflammatory response, the protection of the implant was realized. Similarly, SAPAE also effectively hindered the formation of Salmonella biofilms (31). The anti-inflammatory property of SA also benefits bone regeneration by reducing the production of pro-inflammatory cytokines that impair bone healing (32–36). SAPAE/bone graft mixtures significantly enhance bone regeneration in diabetic rats and accelerate bone regeneration in normoglycemic animals (32). As shown in Figure 11, SAPAE/bone graft-treated rats, diabetic and normoglycemic, had a significantly higher bone fill percentage than the controls after four weeks. After 12 weeks, SAPAE/bone graft-treated diabetic rats had 43% more bone fill than the control, whereas there was no statistical difference in normoglycemic rats. 158

Figure 10. Time course photographs of biofilm formation on active SAPAE and inactive control polymer pellets. Original magnification 100 ×. Adapted with permission from reference (28). Copyright 2006 Elsevier.

Figure 11. Microcomputed tomography (micro-CT) images of bone formation within bone defect regions in diabetic (A, C, E, G) and normoglycemic (B, D, F, H) rats implanted with SAPAE/bone graft mixture (E, F, G, H) or with bone graft control alone (A, B, C, D). Scale bar = 1 mm. Adapted with permission from reference (30). Copyright 2013 Elsevier.

159

Summary SAPAE was designed to realize the tunable, localized sustained release of salicylic acid, a bioactive molecule with a variety of therapeutic effects, including antipyretic, anti-inflammatory, and antimicrobial capabilities. The biocompatible and biodegradable polymer was obtained through different synthetic methods, melt condensation and solution polymerization, in two to three steps. The success of a one-pot polymerization strategy further enables industrial-level scale-up and the low-cost manufacturing of SAPAE. To meet the requirement of various applications, the thermal and physical properties of SAPAE could be tuned through the structural modification of the linker molecule. Degradation rate is closely related to the hydrophobicity of linkers. With different linkers, the degradation period of the polymer ranges from hours to months. SAPAE has been successfully used in biomedical applications such as bone regeneration and the inhibition of biofilm formation. Therefore, the ease of synthesis, tunable properties, and controlled releasing behavior together make the SAPAE an excellent biocompatible and biodegradable material showing great potential in various biomedical applications.

References 1.

Mahdi, J. G.; Mahdi, A. J.; Mahdi, A. J.; Bowen, I. D. The Historical Analysis of Aspirin Discovery, Its Relation to the Willow Tree and Antiproliferative and Anticancer Potential. Cell Proliferation 2006, 39, 147–155. 2. Madan, R. K.; Levitt, J. A Review of Toxicity from Topical Salicylic Acid Preparations. J. Am. Acad. Dermatol. 2014, 70, 788–792. 3. Fuster, V.; Sweeny, J. M. Aspirin: A Historical and Contemporary Therapeutic Overview. Circulation 2011, 123, 768–778. 4. Mackowiak, P. A. Brief History of Antipyretic Therapy. Clin. Infect. Dis. 2000, 31, S154–S156. 5. Dawn, C. A History of Aspirin. Clin. Pharm. 2014, 6. 6. WHO Model Lists of Essential Medicines. http://www.who.int/medicines/ publications/essentialmedicines/en/ (accessed May 20, 2018). 7. Ping, M.; Zizhen, L.; Meng, X.; Rui, J.; Weirui, L.; Xiaohong, W.; Shen, M.; Gaimei, S. Naturally Occurring Methyl Salicylate Glycosides. Mini-Rev. Med. Chem. 2014, 14, 56–63. 8. Rowland, M.; Riegelman, S. Pharmacokinetics of Acetylsalicylic Acid and Salicylic Acid after Intravenous Administration in Man. J. Pharm. Sci. 1968, 57, 1313–1319. 9. Wu, P.; Grainger, D. W. Drug/Device Combinations for Local Drug Therapies and Infection Prophylaxis. Biomaterials 2006, 27, 2450–2467. 10. Kost, J.; Langer, R. Responsive Polymeric Delivery Systems. Adv. Drug Delivery Rev. 2012, 64, 327–341. 11. Prudencio, A.; Schmeltzer, R. C.; Uhrich, K. E. Effect of Linker Structure on Salicylic Acid-Derived Poly(anhydride−esters). Macromolecules 2005, 38, 6895–6901. 160

12. Liming, T.; Rabenstein, M.; Kricheldorf Hans, R. Macrocycles 14. Polycondensations of Aspirin and Other Monomers of Salicylic Acid. Macromol. Chem. Phys. 2001, 202, 1497–1504. 13. Shalaby, W. S.; Donald, F. K.; Steven, A. Homopolymers and Copolymers of Salicylate Lactones. U.S. Patent 5082925, Jan 21, 1992. 14. Leong, K. W.; Simonte, V.; Langer, R. Synthesis of Polyanhydrides: Melt-polycondensation, Dehydrochlorination, and Dehydrative Coupling. Macromolecules 1987, 20, 705–712. 15. Domb, A. J.; Langer, R. Polyanhydrides. I. Preparation of High Molecular Weight Polyanhydrides. J. Polym. Sci., Part A: Polym. Chem. 1987, 25, 3373–3386. 16. Domb, A. J.; Ron, E.; Langer, R. Poly(anhydrides). 2. One-Step Polymerization Using Phosgene or Diphosgene as Coupling Agents. Macromolecules 1988, 21, 1925–1929. 17. Yoda, N.; Miyake, A. Synthesis of Polyanhydride. I. Mixed Anhydride of Aromatic and Aliphatic Dibasic Acids. Bull. Chem. Soc. Jpn. 1959, 32, 1120–1126. 18. Subramanyam, R.; Pinkus, A. G. Synthesis of Poly(terephtha1ic Anhydride) by Hydrolysis of Terephthaloyl Chloride/Triethylamine Intermediate Adduct. Characterization of Intermediate Adduct. J. Macromol. Sci., Pure Appl.Chem. 1985, 22, 23–31. 19. Hill, J. W. Studies on Polymerization and Ring Formation. VI. Adipic Anhydride. J. Am. Chem. Soc. 1930, 52, 4110–4114. 20. Hill, J. W.; Carothers, W. H. Studies of Polymerization and Ring Formation. XIV. A Linear Superpolyanhydride and a Cyclic Dimeric Anhydride from Sebacic Acid. J. Am. Chem. Soc. 1932, 54, 1569–1579. 21. Hill, J. W.; Carothers, W. H. Studies of Polymerization and Ring Formation. XIX.1 Many-Membered Cyclic Anhydrides. J. Am. Chem. Soc. 1933, 55, 5023–5031. 22. Schmeltzer, R. C.; Johnson, M.; Griffin, J.; Uhrich, K. Comparison of Salicylate-Based Poly(anhydride-esters) Formed via Melt-Condensation versus Solution Polymerization. J. Biomater. Sci., Polym. Ed. 2008, 19, 1295–1306. 23. Schmeltzer, R. C.; Anastasiou, T. J.; Uhrich, K. E. Optimized Synthesis of Salicylate-Based Poly( anhydride-esters). Polym. Bull. 2003, 49, 441–448. 24. Carbone, A. L.; Song, M.; Uhrich, K. E. Iodinated Salicylate-Based Poly(anhydride-esters) as Radiopaque Biomaterials. Biomacromolecules 2008, 9, 1604–1612. 25. Faig, J. J.; Smith, K.; Moretti, A.; Yu, W.; Uhrich, K. E. One‐Pot Polymerization Syntheses: Incorporating Bioactives into Poly(anhydride‐esters). Macromol. Chem. Phys. 2016, 217, 1842–1850. 26. Carbone, A. L.; Uhrich, K. E. Design and Synthesis of Fast‐Degrading Poly(anhydride‐esters). Macromol. Rapid Commun. 2009, 30, 1021–1026. 27. Sengwa, R. J. Solvent Effects on Microwave Dielectric Relaxation in Poly(ethylene glycols). Polym. Int. 1999, 45, 43–46. 28. Torabi, S.; Jahani, F.; Van Severen, I.; Kanimozhi, C.; Patil, S.; Havenith Remco, W. A.; Chiechi Ryan, C.; Lutsen, L.; Vanderzande Dirk, J. M.; Cleij 161

29.

30.

31.

32.

33.

34.

35.

36.

Thomas, J.; Hummelen Jan, C.; Koster, L. J. A. Strategy for Enhancing the Dielectric Constant of Organic Semiconductors Without Sacrificing Charge Carrier Mobility and Solubility. Adv. Funct. Mater. 2014, 25, 150–157. Whitaker‐Brothers, K.; Uhrich, K. Poly(anhydride‐ester) Fibers: Role of Copolymer Composition on Hydrolytic Degradation and Mechanical Properties. J. Biomed. Mater. Res., Part A 2004, 70A, 309–318. Bryers, J. D.; Jarvis, R. A.; Lebo, J.; Prudencio, A.; Kyriakides, T. R.; Uhrich, K. Biodegradation of Poly(anhydride-esters) into Non-steroidal Anti-inflammatory Drugs and Their Effect on Pseudomonas aeruginosa Biofilms In Vitro and on the Foreign-Body Response In Vivo. Biomaterials 2006, 27, 5039–5048. Rosenberg, L. E.; Carbone, A. L.; Römling, U.; Uhrich, K. E.; Chikindas, M. L. Salicylic Acid‐Based oly(anhydride esters) for Control of Biofilm Formation in Salmonella enterica serovar Typhimurium. Lett. Appl. Microbiol. 2008, 46, 593–599. Wada, K.; Yu, W.; Elazizi, M.; Barakat, S.; Ouimet, M. A.; RosarioMeléndez, R.; Fiorellini, J. P.; Graves, D. T.; Uhrich, K. E. Locally Delivered Salicylic Acid from a Poly(anhydride-ester): Impact on Diabetic Bone Regeneration. J. Controlled Release 2013, 171, 33–37. Mitchell, A.; Kim, B.; Snyder, S.; Subramanian, S.; Uhrich, K.; O’Connor, J. P. Use of Salicylic Acid Polymers and Bone Morphogenetic Protein-2 to Promote Bone Regeneration in Rabbit Parietal Bone Defects. J. Bioact. Compat. Polym. 2015, 31, 140–151. Subramanian, S.; Mitchell, A.; Yu, W.; Snyder, S.; Uhrich, K.; O’Connor, J. P. Salicylic Acid-Based Polymers for Guided Bone Regeneration Using Bone Morphogenetic Protein-2. Tissue Eng., Part A 2015, 21, 2013–2024. Mitchell, A.; Kim, B.; Cottrell, J.; Snyder, S.; Witek, L.; Ricci, J.; Uhrich Kathryn, E.; Patrick O’Connor, J. Development of a Guided Bone Regeneration Device Using Salicylic Acid‐poly(anhydride‐ester) Polymers and Osteoconductive Scaffolds. J. Biomed. Mater. Res., Part A 2013, 102, 655–664. Yu, W.; Bien‐Aime, S.; Mattos, M.; Alsadun, S.; Wada, K.; Rogado, S.; Fiorellini, J.; Graves, D.; Uhrich, K. Sustained, Localized Salicylic Acid Delivery Enhances Diabetic Bone Regeneration via Prolonged Mitigation of Inflammation. J. Biomed. Mater. Res., Part A 2016, 104, 2595–2603.

162

Chapter 12

L-Lysine-Modified

Poly(ester-urethane) Based on Polycaprolactone for Controlled Release of Hydrocortisone Karla A. Barrera-Rivera* and Antonio Martínez-Richa* Departamento de Química, División de Ciencias Naturales y Exactas, Universidad de Guanajuato, Noria Alta S/N, Guanajuato, Guanajuato 36050, México *E-mail: [email protected]. *E-mail: [email protected].

A linear L-lysine poly(ester-urethane) (PUR) was successfully prepared from polycaprolactone diol (DEG1), hexamethylenediisocyanate (HDI), and L-lysine ethyl ester dihydrochloride. Evidence from Fourier transform infrared (FT-IR) spectra indicates that intermolecular hydrogen-bonded species of PUR are notably influenced by the presence of L-lysine. Thermal and morphological properties of polyurethanes (PUs) were characterized by differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA). It was found that L-lysine provokes profound changes in the thermal and crystalline properties of the polymer. The presence of L-lysine has a strong influence in the water-uptake behavior of the polymer. Films of PU containing 1% (w/w) hydrocortisone were placed in three different buffer systems: (1) citrate buffer (pH 4.0), (2) phosphate buffer (pH 7.0), and (3) borate buffer (pH 10.0) at room temperature and monitored in vitro up to 48 h for their release behavior. Amino acid-containing degradable PUR was tested for delivery of hydrocortisone under different pH conditions. The release profiles of hydrocortisone occurred by a two-stage process and followed a non-Fickian behavior.

© 2018 American Chemical Society

Introduction In the field of biomedical research, there is great interest in the synthesis of degradable polymers with potential applications in drug delivery and tissue engineering. Based on their good mechanical properties, flexibility, and biocompatibility, some polyurethanes (PUs) have been used to fabricate biomedical devices. In particular, biodegradable PUs have been mainly employed in the design of drug-delivery systems (1). Enzyme-catalyzed polymerization may become a versatile method for the production of sustainable PUs: Lipases, for example, are renewable catalysts with high catalytic activities. The most prominent advantage of using hydrolytic enzymes for the production of polymers is the reversible polymerization–degradation reaction that allows chemical recycling. Polymer chains that contain enzymatically hydrolyzable moieties can be specifically cleaved by a hydrolytic enzyme to produce potentially re-polymerizable low-molecular-weight fragments that can be recycled (2). Yarrowia lipolytica lipases (YLL) are attracting the interest of scientists and industrial researchers due to their ability to catalyze important high-value applications in the food, pharmaceutical, fine-chemicals, and environmental industries (3). Hydrocortisone [11,17,21-trihydroxy-(11b)-pregn-4-ene-3,20-dione] is a hydrophobic corticosteroid drug widely used in the treatment of various skin condition symptoms, such as redness, swelling, itching, as well as certain types of arthritis, allergies, and asthma. It relieves symptoms related to certain hormone deficiencies and has immunosuppressive action. Hydrocortisone has also been used as a supportive-care medication for cancer (4). In this study, hydrocortisone was used as a model for drug-controlled release. We recently reported a study of the degradation behavior of two different PURs, each synthesized from (1) polycaprolactone (PCL) diol obtained by way of enzymatic polymerization and (2) hexamethylenediisocyanate (HDI). One sample is the neat linear PU; the other contains L-lysine as amino acid chain extender. Under composting conditions, half-maximal degradation time (t50) of both samples is approximately 25 days (5). In this work, we report a study to investigate the potential use as a drug-delivery system of these biodegradable, nontoxic PURs.

Experimental Materials Before use, 97% ε-polactone (ε-CL) 97% was dried over calcium hydride and distilled under reduced pressure. Diethylene glycol (DEG) 99%, Lewatit VP OC 1026 beads, stannous 2-ethylhexanoate ~95%, HDI ≥98%, L-lysine ethyl ester dihydrochloride (L-Lys; purity ≥99%), triethylamine 99%, 1,2-dichloroethane anhydrous 99.8%, and hydrocortisone 98% were purchased from Sigma Aldrich and used as received. Citrate buffer (pH 4.0), phosphate buffer (pH 7.0), borate buffer (pH 10.0), and ethanol were purchased from Karal and used as received. YLL was obtained and immobilized according to a procedure described in the literature (6). 164

Lipase Isolation and Immobilization Lipase production by YLL was made as previously reported by Barrera et al. (6). Before immobilization, Lewatit VPOC 1026 beads were activated with ethanol (1:10 beads:ethanol) for 5 h, washed with distilled water, and dried under vacuum for 24 h at room temperature. The beads (1 g) were shaken in a rotatory shaker in 15 mL of YLL-lipase solution with 0.2 mg/mL at 4 °C for 14 h. After incubation, the carrier was filtered off, washed with distilled water, and then dried under vacuum for 24 h at room temperature. Synthesis of α,ω-Telechelic poly(ε-caprolactone)diol PCL diol (DEG1) was prepared as described in (7), from 10 mmol of ε-CL and 1 mmol of DEG in the presence of 12 mg YLL-1026 (immobilized lipase). Vials were stoppered with a Teflon silicon septum and placed in a thermostated bath at 120 ºC for 6 h. No inert atmosphere was used. After the reaction was stopped, the enzyme was filtered off, and the PCL diol was dried at room temperature for 12 h and stored at ambient temperature in a desiccator until used. Synthesis of PURs Bearing Amino Acids PURs were prepared according to the literature (8). Dry PCL diol (2.5 g), HDI in the appropriate amount (OH:NCO ratio = 1:1), and 15 mL of 1,2-dichloroethane were charged in a round bottom flask. The catalyst, stannous 2-ethylhexanoate (1% mol by PCL diol moles) was added and stirred for 1 h and 30 min at 80 ºC; after that time, 0.3498 mmol of amino acid and 1.046 mmol of triethylamine were added to the reaction mixture and allowed 5 h for reaction. The resulting solution was poured over a Teflon petri dish (d-10 cm). The solution was allowed to stand at ambient temperature for 24 h for solvent evaporation. The film was then released and dried at room temperature (Scheme 1). Using 1H-NMR, it was determined that approximately 65% of PU chains contain L-lysine as end-group (5).

Scheme 1. Chemo-enzymatic synthesis of PUR bearing L-lys. 165

Preparation of Films Containing 1% Hydrocortisone (w/w) PUR films were mixed in solution (1,2-dichloroethane as solvent) with 1% (w/ w) of hydrocortisone and stirred at room temperature until they were completely dissolved; the solution was then poured over a Teflon dish (d-10 cm) and allowed to stand at room temperature for solvent evaporation. The film was then released and dried at room temperature. In Vitro Hydrocortisone-Release Studies Hydrocortisone release was evaluated in vitro by ultraviolet–visible light (UV-VIS) spectroscopy. A PUR film (1 × 1 cm) was placed in 3 mL of buffer at different pH values in a quartz cell. The hydrocortisone concentration was determined using a Perkin-Elmer UV-VIS spectrometer lamba 25 at 248 nm during 48 h. Samples were analyzed by duplicate. Swelling Degree Film specimen with a dimension of 3 × 3 cm was dried to a constant weight (Wd) at room temperature. Subsequently, the dried specimen was placed in a flask with 50 mL of the different pH solutions (pH 4, 7, and 10) for 24 h. The swollen specimen was then taken out of the flask, wiped with a filter paper, and weighed (Ww). The swelling degree is calculated according to Equation 1.

where Wd is the weight of the dried film, and Ww is the weight of the swollen film.

Characterization of the PURs Fourier Transform Infrared Spectra were obtained at room temperature using attenuated total reflection technique on films deposited over a diamond crystal on a Perkin-Elmer Spectrum Two™ spectrometer with an average of four scans at 4 cm-1 resolution. Peak analysis was performed using the Origin 6.1 computer program for PC. Thermal Analysis PU samples were conducted on a Mettler Toledo differential scanning calorimeter (DSC)-820e using heating and cooling rates of 10 °C/min. Thermal scans were performed from 25 to 80 ºC, 80 to −90 °C, and −90 to 80 °C in air atmosphere. 166

Thermogravimetric Analysis Thermogravimetric analysis (TGA) scans were recorded on a TA Instruments simultaneous TGA-DSC SDT Q600 at 10 °C/min with dynamic atmosphere of high-purity air at 100 mL/min.

Results and Discussion Two different PUs were synthesized from the enzymatically synthesized PCLdiol: the neat PURDEG1HDI and the PUR modified with L-lysine amino acid as end-group (PURDEG1 L-lys). The reaction was followed by FT-IR by checking the disappearance of the stretching vibration of the isocyanate group (-N=C=O at 2270 cm−1), which confirms the completion of the reaction between the hydroxyl group from the enzymatically synthesized PCL-diol and the isocyanate group from HDI after 90 min. The FTIR analysis of PURDEG1 and the PURDEG1 L-lys films (containing hydrocortisone) revealed the typical bands of PCL-based PUs: (1) a peak in the region of 3350 cm-1 due to N-H stretching; (2) -CH2- stretching (from 3100 to 2700 cm-1); (3) C=O stretching (1724 cm-1); (4) N-H bending and C-N stretching from the urethane group (1530 and 1242 cm-1); and (5) C-O-C stretching (1166 cm-1). After analyzing the peak pattern of both spectra, some differences were observed between them. A sharper peak in the PURDEG1 HDI spectrum centered at 1530 (urethane group) was observed. In addition, small differences in shape for the broad peak in the 1200−1120 region [due to C-O and C-N(C=O) stretching band] were apparent. The main difference observed in the spectra was for the carbonyl stretching band (C=O), which contains information on intermolecular and self-associated species through hydrogen bonding (Figure 1).

Figure 1. FT-IR spectra (carbonyl zone) of the synthesized PURs. To estimate the content of the inter-associated and the self-associated carbonyl bonds by way of hydrogen bonds, peak deconvolution analysis was performed. The degree of the carbonyl groups participating in hydrogen bonding 167

can be recorded by the carbonyl hydrogen bonding index, R (9), derived by using Equation 2:

The results are reported in Table 1.

Table 1. FT-IR Data for the Synthesized PURs

It is evident that the content of inter-associated species in the PUR decreases in the presence of L-lys. This is expected because L-lys moieties compete with the N-H bonds of the urethane functional group to create associated species. In the hydroxyl zone (Figure 2), a shoulder at 3520 cm-1 (“free” OH) is more evident in the FT-IR spectrum of PURDEG1 L-lys (see arrow). In addition, the peak shape between the two spectra differs slightly in this zone.

Figure 2. FT-IR spectra (hydroxyl zone) of the synthesized PURs:PURDEG1 L-lys (top line) and PURDEG1 (bottom line). 168

Thermal Properties Thermal properties for PURs containing hydrocortisone were registered from 25–80 °C followed by cooling and reheating from −90 to 80 °C. The results from DSC measurements are listed in Table 2. In the first heating (25−80 °C), PURDEG1HDI L-lys showed a flat baseline, whereas PURDEG1HDI displayed a transition due to pure PCL-phase melting (Figure 3). This behavior indicates that melt transition for PCL is inhibited by the presence of amino acid in the polymeric matrix of PURDEG1 HDI. In the second heating, a difference in peak positions (by 3.4 °C) and the heat of fusion are apparent.

Figure 3. DSC curves of the synthesized PURs. A) First heating and B) second heating.

Figure 4. Crystallization DSC curves of the synthesized PURs. Figure 4 shows the thermograms for PCL crystallization (cooling from 80 to −90 °C) for synthesized PURs. Whereas PURDEG1 HDI L-Lys shows a crystallization peak at 3.9 °C, the other PUR has a Tc of −10.7 °C. In addition, greater enthalpy is recorded for the L-lysine containing PUR. Differences in crystalization temperatures and the heat of fusion indicates that more crystalline zones are induced during the second heating in the amino acid-containing sample. 169

Table 2. Thermal Properties Obtained by DSC for the Synthesized PURS-Containing Hydrocortisone Sample

PURDEG1HDI DEG1 L-LYS HDI

Tm (°C) First heating

Tm (°C) Second heating

ΔHm (J/mol) First heating

Tc (°C)

ΔHm (J/mol) Second heating

44.8

18.93

36.9

25.2

−10.7

−

−

40.3

29.9

3.9

TGA results for PURDEG1 and PURDEG1 L-lys containing hydrocortisone are shown in Figure 5 and Table 3. The onset decomposition temperatures (Tonset) for PURDEG1HDI, and PURDEG1 L-lys are 292 and 291 °C, respectively. Similarly, the decomposition temperature peaks (Tmax) for these compounds are at 325 and 329 °C, respectively. Between these two compounds, L-lysine chain extender unit is only present in PURDEG1 L-lys. Because the hard-segment content of the compound increases when chain-extender content is greater, the decomposition temperature or thermal resistance of the compound slightly increases.

Figure 5. TGA curves of the synthesized PURs. 170

Table 3. Thermal Properties Obtained by TGA of the PURs PUR

Tonset (°C)

Decomposition (%)

Tmax (°C)

Decomposition (%)

PURDEG1HDI

292

12

325

85

PURDEG1 L-Lys HDI

291

15

329

83

Swelling Degree L-Lysine is an α-amino acid and is an essential building block for all proteins in the body. It contains an ε-amino group in the aliphatic side chain (R group), which gives the molecule a hydrophilic character (10). Swelling degrees of the synthesized PURs containing hydrocortisone PURDEG1HDI in different buffers at pH 4, pH 7, and pH 10 for 24 h were 0.3%, 0.3%, and 1.2%, respectively, and for PURDEG1 L-lys were 2.5%, 7.1%, and 1.1% (see Table 4), respectively. The swelling degree depends on the presence of L-lysine and the amount on PCL content in the PUR. Three factors affect swelling behavior. One factor is the presence of L-lysine because this moiety is hydrophilic and favors the absorption of water. Another important factor is the relative amount of PCL in the PURs, which influences swelling behavior because of its hydrophobic nature. Lower content of PCL in the PUR reflects in a higher swelling degree. The third factor is the physical cross-link effect of the hard segment. The presence of hard segment acts as physical cross-links in the PUs and subsequently enhances their physical properties as mechanical strength, hardness, and solvent resistance. In that regard, the presence of L-lysine and the lower relative amount of PCL chains in the PUR explain the higher values of water uptake recorded for the PURDEG1 L-lys sample at pH = 4 and 7. At greater pH (more basic) values, a decrease in the swelling degree is observed for the PURDEG1 L-lys at pH = 10 after 24 h of being in contact with the buffer solution; values for both polymers are approximately the same. After 48 h, degradation of PURDEG1 L-lys occurred. The release profiles for hydrocortisone at three different pH values for the two samples are shown in Figures 6 and 7. The hydrocortisone release consists of a two-stage process: an initial rapid-release stage followed by a second slowerrelease stage. After 48 h, the amount of hydrocortisone released from PURDEG1HDI without L-lysine at pH 4 was approximately 16%; at pH 7 it was 25%; whereas at pH 10 a maximum at 15% was observed after 35 h. For this sample, after 35 h a sharp decrease in the release was observed, which can be attributed to the degradation of the PU under basic conditions at pH 10.

171

Table 4. Swelling Degree of the Synthesized PUs at Room Temperature PUR DEG1 HDI

DEG1 L-lys

pH

Swelling degree (%) at 24 h

Swelling degree (%) at 48 h

4

0.3

−0.2

7

0.3

0.2

10

1.2

0.6

4

2.5

2.1

7

7.1

1.8

10

1.1

−12.6

Figure 6. In vitro drug-release profile for hydrocortisone under different hydrolytic conditions for PURDEG1HDI.

The amount of hydrocortisone released from the PUR containing L-lysine after 48 h at pH 4 was approximately 14%; at pH 7 it was 18%; and at pH 10 it was 50%. Under the conditions studied, PURDEG1 L-lys shows a better performance as a delivery system than PURDEG1 because more amount of drug is delivered if the same amount of time is considered. There is no evidence of decomposition for this sample under basic conditions at this stage. 172

Figure 7. In vitro drug-release profile for hydrocortisone under different hydrolytic conditions for PURDEG1 L-lys.

Table 5. Values k and n (±95% Confidence Intervals) Obtained by Plotting the Logarithm of the Hydrocortisone Fraction Release Versus the Logarithm of Time at pH Values 4, 7, and 10a Matrix

k, (h)-n

n

R2

PURDEG1HDI (pH = 4)

2.5 × 10-2

0.51

0.98

PURDEG1HDI (pH = 7)

1.7 ×

10-2

0.70

0.97

3.2 ×

10-2

0.60

0.99

0.59

0.99

1.0

1.0

0.62

0.99

PURDEG1HDI (pH = 10) PURDEG1 L-Lys HDI (pH = 4) PURDEG1 L-Lys HDI (pH =7) PURDEG1 L-Lys HDI (pH =10) a

R2

2.13 × 1×

10-2

10-2

6.0 ×

10-2

values are also reported.

Kinetics of Drug Release The drug-transport mechanism was modeled using Equation 3:

where Mt/M∞ describes the portion of drug released at time t (M∞ is considered the same as the amount total drug loaded in each polymer); k is the constant of release rate and n corresponds to an important exponent value, which can be used to define the release mechanism. Normal Fickian diffusion for a thin polymer membrane is defined by n < 0.5, whereas case II transport is characterized by n = 1.0. 173

Using linear regression, the intercept and slope of the plot log (Mt/M∞) against time were determined. From these values, k and n were determined. All release profiles presented n values between 0.5 and 1 as listed in Table 5, which suggests that the release of hydrocortisone follows a non-Fickian behavior (11, 12).

Conclusions Biodegradable PURs have been successfully synthesized from PCL-diol, DEG, and HDI; one of the samples contains L-lys as chain extender/end-group. Chemical structure and molecular features were determined using FT-IR. DSC analysis reveals striking differences in crystallinity and transition temperatures. High temperature stability (up to ~290 °C) for the PURs was observed from the TGA thermograms. As expected, the hydrophilicity of PURs is increased with the presence of L-lysine as revealed by water-swelling analysis at different pH values. Hydrocortisone-release curves depend on pH, and release is more effective for PURDEG1 L-lys sample. Drug release occurs by a non-Fickian mechanism.

Acknowledgments The authors acknowledge financial support from Consejo Nacional de Ciencia y Tecnología (CONACyT), Grant 153922.

References 1.

2. 3. 4. 5. 6. 7.

8.

9.

Basu, A.; Farah, S.; Kunduru, K. R.; Doppalapudi, S.; Khan, W.; Domb A. J. Advances in Polyurethane Biomaterials. Series in Biomaterials 108; Woodhead Publishing: Cambridge, MA, 2016; pp 217–246. Yanagishita, Y.; Kato, M.; Toshima, K.; Matsumura, S. ChemSusChem. 2008, 1, 133–142. Brígida, A. I. S.; Amaral, P. F. F.; Coelho, M. A. Z.; Gonçalves, L. R. B. J. Mol. Catal. B: Enzym. 2014, 101, 148–158. Golbert-Gist, A. Chem. Eng. News 2005, 83, 25. Arrieta, M. P.; Barrera-Rivera, K. A.; Salgado, C.; Martínez Richa, A.; López, D.; Peponi, L. Poly. Degrad. Stabil. 2018, 152, 139–146. Barrera-Rivera, K. A.; Flores-Carreón, A.; Martínez-Richa, A. J. Appl. Polym. Sci. 2008, 2, 708–719. Barrera-Rivera, K. A.; Marcos-Fernández, A.; Martínez-Richa, A. Green Polymer Chemistry: Biocatalysis and Biomaterials;Cheng, H. N.,Gross, R. A., Eds.; ACS Symposium Series 1043; American Chemical Society: Washington, DC, 2010; pp 227–235. Barrera-Rivera, K. A.; Martínez-Richa, A. Green Polymer Chemistry: Biobased Materials and Biocatalysis; ACS Symposium Series 1192; American Chemical Society: Washington, DC, 2015; pp 27–40. Seymour, R. W.; Estes, G. M.; Cooper, S. L. Macromolecules 1970, 3, 579–583. 174

10. Nelson, D. L.; Cox, M. M. Lehninger Principles of Biochemistry, 4th ed.; Worth Publishers: Conway, AR, 2017; pp 76–79. 11. Ritger, P. L.; Peppas, N. A. J. Controlled Release 1987, 5, 23–36. 12. Ritger, P. L.; Peppas, N. A. J. Controlled Release 1987, 5, 37–42.

175

Chapter 13

Rational Synthesis of Biobased Hyperbranched Poly(ester)s for Sustained Delivery Tracy Zhang,1 Bob A. Howell,2 Steven J. Martin,1 Brandon Zhu,1 Daniel Zhang,1 and Patrick B. Smith*,1 1Michigan

State University, 1910 West St. Andrews Road, Midland, Michigan 48640, United States 2Central Michigan University, Department of Chemistry, Dow Science Complex 263, Mount Pleasant, Michigan 48859, United States *E-mail: [email protected].

Probability models have been used to direct the preparation of hyperbranched poly(ester)s (HBPEs) from glycerin and either adipic acid (AA) or succinic acid (SA) possessing predictable structures using monomer stoichiometry as the only independent variable for polymerizations taken to high extents of reaction. Quantitative 13C nuclear magnetic resonance (13C NMR) was used to determine the glycerol substitution patterns and the extent of reaction of both glycerol and AA, which were incorporated into a Miller–Macosko conditional probability model to predict the HPBE MW and dispersity. The model accommodates the difference in primary and secondary hydroxyl reactivity and substituent effects. In all cases, the predicted Mw values were in excellent agreement with the absolute molecular weights determined by size-exclusion chromatography (SEC) with light-scattering detection. Several active agents were covalently bonded to these and similar HBPEs and were shown to release effectively through either chemical or enzymatic hydrolysis. Moreover, time-release profiles were shown to be a function of composition, catalyst type and level, and physical form. The release kinetics were nearly linear with time.

© 2018 American Chemical Society

Introduction Hyperbranched poly(ester)s (HBPEs) from renewable sources have a number of interesting applications due to the fact that they are composed of nontoxic building blocks and are biodegradable. HBPEs have been synthesized from many different biobased monomers—including those from multifunctional alcohols, such as glycerol, pentaerythritol, xylitol, and other sugar alcohols—as well as multifunctional acids, such as sebacic, adipic, aconitic, succinic, and azelaic (1–11). Many of the applications targeted for these bio-based HBPEs include the delivery of actives, such as pharmaceuticals, pesticides, and antimicrobials (12–21). HB polymers are synthesized from mutually reactive multifunctional monomers. The use of bimolecular nonlinear polymerization (BMNLP) methodology to obtain soluble HBPEs has been described elsewhere (22). An idealized structure of an HBPE from glycerol and adipic acid (AA) is given in Scheme 1. The BMNLP model can be quantitatively applied only for monomers of equal functional-group reactivity. For this reason, a different model is required to predict the structure of glycerin-based HBPEs because the primary hydroxyl units were shown to be 3.3 times more reactive towards esterification than the secondary hydroxyl unit (11). The Miller–Macosko conditional probability model has been used to model the structures produced from these syntheses. The end-group composition of this type of HBPE was shown to be dependent on the monomer stoichiometry, primarily being that of the excess monomer. This reaction strategy enables the design of HBPEs with either hydroxyl or acid end-groups from glycerol and diprotic acids. The ability to easily control end-group functionality is a valuable attribute for HB systems because many polymer properties depend on it, including solubility, solution and melt viscosity, and thermal properties. The end-groups are also available for subsequent reaction, such as covalently bonding actives (15, 23–25). A large variety of bonding chemistries—including hydroxyl, carboxylic acid, olefin, amine, ketone, aldehyde, and several others—is accessible using HBPEs as a platform. The subject of polymer conjugates for controlled release was recently reviewed (26). HBPEs composed of biobased polyfunctional alcohols and acids are degradable to their monomeric building blocks either by chemical or enzymatic hydrolysis. Glycerol, AA, and succinic acid (SA) are on the Food and Drug Administration’s generally-recognized-as-safe (GRAS) list (27), such that the hydrolytic degradation of these types of HBPEs to monomers does not yield toxic products. The rate of hydrolysis has been shown to be a function of poly(ester) composition (12), decreasing as the hydrophobicity increases, presumably due to the lower water solubility of the material. The rate of biodegradation under physiological conditions was shown to depend on the HBPE composition and to range from several days to weeks. Based on these findings, the results presented here are focused on the use of HBPEs from glycerol and adipic or SuA as time-release platforms because they are expected to provide an excellent combination of polymer molecular dimensions, compositions, and structure for the preparation of conjugates from 178

a variety of active agents. Moreover, these building blocks are very inexpensive and benign and can be obtained from renewable sources. Several applications will be presented: the binding of (1) 2-undecanone (2-UD), a natural mosquito repellant; (2) salicylic acid, a therapeutic drug used for the treatment of acne and inflammation; and (3) naproxen, a nonsteroidal anti-inflammatory drug.

Scheme 1. Structure of HPBE from glycerol and AA.

Materials and Methods Glycerol, AA, and dibutyltin oxide were obtained from Sigma Aldrich and used without further purification. Common solvents and reagents were obtained from Thermo Fisher Scientific or Sigma Aldrich. Glycerol, AA, salicylic acid, naproxen, and 2-UD were obtained from Sigma Aldrich. Nuclear magnetic resonance (NMR) spectra were obtained using a Bruker Avance 300 MHz spectrometer. Proton and carbon chemical shifts are reported in parts-per-million (δ) with respect to tetramethylsilane as an internal reference (δ = 0.00). Quantitative 13C NMR spectra were obtained using a 90° pulse width, a pulse-repetition time of 10 s, gated decoupling without nuclear Overhauser effect, a sweep width of 31 KHz, 13.1 K points, and 3.0-Hz line broadening. The 1H NMR spectra were obtained using a 10° pulse width and a 5-s repetition rate. Size-exclusion chromatography (SEC) was performed for HBPE during synthesis or after synthesis to follow its formation. SEC was performed using a Malvern GPC Max instrument with low-angle light scattering and refractive-index detectors to determine the absolute weight–average molecular weight (Mw). The instrument was operated with two Agilent PL gel 3-μm MIXED-E columns in series. The eluent was tetrahydrofuran at a flow rate of 1 mL min-1. Typical concentration of the samples was approximately 1.0 mg mL-1. High-performance liquid chromatography (HPLC) was conducted using a Waters Breeze 2 system comprised of a 2707 autosampler, a 1525 gradient pump, 179

a 5 CH column oven at 35 °C, a Waters Symmetry C18, 3µ column (3.0 × 150 mm), and a 2489 ultraviolet light/visible light detector. The flow rate was 1 mL/min, and 10-µL injections were made with the following gradient program: 90/10 water/acetonitrile for 9 min to 50/50 for 1 min to 0/100 (pH = 2.0 was maintained using H3PO4/acetonitrile). General methods of synthesis and characterization have been described (28). As an example, the glycerol-AA poly(ester) containing hydroxyl end-groups was prepared using melt polymerization. Into a dry 100-mL, three-necked, round-bottomed flask fitted with a magnetic stirring bar and a condenser bearing a gas-inlet tube was placed 5.02 g (54.3 mmol) of glycerol, 5.95 g (40.7 mmol) of AA, and 0.15 wt.% dibutyltin oxide. The flask was mounted in an oil bath maintained at 150° C, and the mixture was stirred under a blanket of nitrogen for approximately 4 h and then for 9 h at decreased pressure (15 torr). The product (8.77 g) was obtained as a clear, colorless, viscous liquid. Several lots of the HBPE were synthesized. The MW was typically 2,000–3,000 g mol-1: 1H NMR (δ, DMSO-d6), 1.51–1.52 (m, OCOCH2CH2), 2.08–2.30 (m, OCOCH2CH2), 3.24–5.22 (H from glycerol [substituted and unsubstituted]); 13C NMR (δ, DMSO-d6) 23.9 (OCOCH2CH2), 33.1 (OCOCH2CH2), 59.6–75.5 (carbon atoms from glycerol), 172.2–172.86 (ester carbonyl); IR (ATR, cm-1) 3462 (m, broad, O-H stretch), 2952 (s), 2874 (m, C-H aliphatic), 1737 (vs, ester C=O), and 1173 (s, C-O stretch).

Bonding of 2-UD to the Glycerol-AA HBPEs Containing Hydroxyl End-Groups Into a dry, 100-mL, two-necked, round-bottomed flask fitted with a magnetic stirring bar and take-off adapter was placed 10.46 g of glycerol-AA HBPE (89.85 mmol of hydroxyl groups;), 7.65 g (89.85 mmol) of 2-UD, and 4 mg of toluenesulfonic acid (TsOH). The flask was mounted in an oil bath maintained at 120 °C, and the mixture was stirred for 30 min. The pressure in the flask was reduced to 4–5 torr and stirred continuously at 20 °C for 4 h. A viscous, light yellow liquid (13.94 g) was obtained. The MW of the polymer increased dramatically from 2900 g/mol for the starting HBPE to 5100 g/mol for the ketalized HBPE. The 13C NMR spectrum of the product contained resonances from the ketal structure at δ 110.8–110.9.

Release of the 2-UD Release of the 2-UD was conducted under acidic conditions, typically using phytic acid as a catalyst at 37 °C. The release of 2-UD, as well as the hydrolysis of the HBPE, was monitored using SEC, NMR, and gas chromatography–mass spectrometry (GC-MS). These degradation studies were carried out using 100 mg of glycerol-AA-UD, 100 mg of glycerol, and 2 and 15 weight percent (wt %) phytic acid added as a 50% solution in water. 180

Salicylic-Acid Attachment (Glycerol-AA-SA) Into a dry 250-ml, three-necked, round-bottomed flask fitted with a magnetic stirring bar and a Soxhlet extractor containing 40 g of 4 Å molecular sieves and bearing a condenser equipped with a gas-inlet tube was placed a solution of 10.01g (85.9 mmol of hydroxyl end-groups) of the hydroxyl-terminal HBPE, 11.86 g (85.9 mmol) of salicylic acid, and 20 mg of p- TsOH in 140 ml of triglyme. The solution was stirred at 150 °C (oil bath) under nitrogen for 12 h. The mixture was poured into excess diethyl ether to remove triglyme and free salicylic acid and precipitate the conjugate. The conjugate was obtained as a viscous, clear, slightly yellow liquid with an MW of 3200 g/mol (vs. 500 g/mol for the starting polymer). Based on NMR analysis, approximately 20 wt % of the available hydroxyl endgroups had been converted to salicylate esters. Naproxen was also bonded through esterification. Salicylic-Acid and Naproxen Release Hydrolysis of the HBPE conjugate was determined as a function of time in order to measure the rate of drug release. The polymer conjugate (8.0 mg) was loaded into a vial, and then an additional 7.6 mL of pH 7.0 aqueous buffer was injected into the vial to create a polymer dispersion of 1 mg/mL (an additional 0.4 mL of rat liver microsomes was later added). The vial was then inserted into a water bath maintained at 37°C and vigorously stirred to facilitate a fine suspension of the polymer in the buffer solution. For the naproxen conjugate, 8 mg of the polymer was heated in an oven with 18 µL of dimethylsulfoxide (DMSO) because the HBPE-conjugate was a solid at room temperature and was not readily dispersible in the buffer solution. Simultaneously, at least 400 µL of rat liver microsomes was defrosted in an ice bath (the microsomes were stored at liquid-nitrogen temperatures before hydrolysis). Time (15 min) was allowed for both the polymer to be suspended and the microsomes to defrost. Afterwards, 400 µL of rat liver microsomes were directly injected into the polymer-solution vial. At this time, the stirrer was turned down to the lowest setting, so as not to degrade the enzymes. Samples (1.0 mL) were taken directly from this vial after 0, 15, 30, and 45 min and at 1, 1.5, and 2 h. These samples were placed in Agilent Technologies 2-mL vials along with 200 µL of tetrahydrocortisone F to denature the enzymes. The samples were then analyzed by HPLC to determine the amount of active released as a function of time.

Results The glycerol-SuA and glycerol-AA HBPEs were fully characterized using 13C NMR spectroscopy and SEC (11). These techniques, together with Miller–Macosko models, were shown to thoroughly define the structure of HBPEs from glycerol and AA. Five glycerol-substitution patterns are possible for the esterification with SuA as listed in Table 1. They include the primary monoester, which in previously reported HBP terminology, is a terminal unit labeled TG; the secondary monoester, 181

which is also a terminal unit labeled T1,3; the primary–primary diester linear unit labeled L1,3; the primary–secondary diester linear unit labeled L1,2; and the trisubstituted dendritic unit labeled D (29). The 13C NMR resonances for all these structures have been assigned (11) and were used to quantitatively define the glycerol substitution pattern as well as the extent of reaction for the minor (carboxyl) species, Pb, and the mole ratio of the monomer building blocks, both those incorporated into the HBPE and those unreacted. These values were entered into the Miller–Macosko conditional probability model to predict the structure and molecular weight of HBPEs prepared with varying monomer stoichiometry (11, 30–32) . The model used a series of “super-species”—G00 through G21—which are defined in Figure 1. These super-species corresponded to the glycerol-substitution patterns listed Table 1. G00 represents unsubstituted glycerol; the primary and secondary monoesters, G10 and G01, correspond to Tg and T1,3, respectively; the two diesters, primary–primary, G20 (L1,3) and primary–secondary, G11 (L1,2); and the triester, G21 (D). The distribution of glycerol super-species was quantitatively determined using NMR. The MW values and dispersities of these species can then be predicted by using conditional probability to randomly combine the glycerol super-species with AA. Because the distribution of glycerol ester species was analytically determined, the model accommodates the difference in primary and secondary hydroxyl reactivity and substituent effects for the glycerol hydroxyl reactivity on substitution of one of the other three hydroxyl groups. The model assumes that the adipic-acid carboxyl groups had equal reactivity and react independently and that there were no intramolecular reactions (loops). The mole fraction of glycerol, f0, is equal to the moles of glycerol divided by the total number of moles of all glycerol species, N3A. The mole fractions f1, f2, and f3 are those for the monoesters, diesters, and triesters, respectively. The mole fraction of the ester species is given by g1 though g3. The molecular weight values for each glycerol species are shown in the final column of Figure 1. As a housekeeping tool, because a molecule of water is lost for each esterification reaction, the molecular weight of the glycerol unit is assumed to lose 1 mass unit and the acid group to lose 17 mass units. Thus, the glycerol masses given in the far right column of Figure 1 are m0 for unreacted glycerol, which is equal to the mass of glycerol; m3A, that of glycerol with one ester unit is m1, which was defined as m3A – 1; and so on. The value of the Miller–Macosko conditional probability model is that it is possible to accurately predict the molecular weight and dispersity of HB polymers from the initial monomer stoichiometry and the extent of reaction for monomers with functional groups of different reactivity such as those in this study. Accurate prediction of molecular weight is not possible with the more conventional Flory–Stockmayer (F-S) model (on which the BMNLP model is based), which assumes equal reactivity of functional groups (33). The Miller–Macosko conditional probability model makes it possible to use a targeted synthesis for a given molecular weight and structure for HB polymers simply by selecting the initial monomer stoichiometry for polymerizations taken to high conversion. It gives control and precision to the synthesis of these materials.

182

Table 1. Chemical Shift Assignments for the Glycerol-SuA HBPE

To demonstrate the power of the conditional probability model, a series of HBPEs of varying monomer stoichiometry or, more precisely, functional group stoichiometry, [–OH]/[–COOH] varying from 1.52 to 2.16 were produced. 13C NMR spectroscopy was used to determine the composition of substituted glycerol species and the degree of branching (DB %) as a function of stoichiometry for this series (Table 2). 183

Figure 1. Definition of glycerol and glycerol ester species, their mole fractions, and that of the adipate ester species used in the conditional probability model.

The degree of branching was calculated from Equation (1) (34):

The values for D, L1,2, and L1,3 as a function of reaction time, were obtained directly from the areas of the 13C NMR spectra and are listed in Table 2. Table 3 lists the extent of reaction of the carboxyl functionality (Pb), the Mw as determined by SEC–low-angle light scattering (LALS), and the Mw predicted using the Miller–Macosko model using the NMR-determined super-species distribution. The agreement between predicted molecular weight values (Miller–Macosko) and those determined by SEC-MALLS is excellent. The F-S model fails to correctly predict the MW values and gel points for this series of HBPEs due to the unequal reactivity of the glycerol hydroxyl units. Because the rate of esterification for the secondary hydroxyl is approximately one third that of the primary hydroxyls, the number of trifunctional species formed at a particular extent of reaction is decreased relative to that predicted by the F-S model. Therefore, the reaction can be pushed to greater extents of reaction without gel formation. 184

Table 2. The [–OH]/[–COOH] Reactant Stoichiometries, Mole Ratio of Substituents, and Degrees of Branching for the Series of HBPEs DB%

Stoichiometry

Gly

TG

T13

L13

L12

D

2.16

24

0.15

0.37

0.05

0.24

0.13

0.06

2.00

23

0.12

0.37

0.05

0.26

0.14

0.06

1.93

32

0.14

0.34

0.05

0.26

0.12

0.09

1.74

30

0.08

0.31

0.05

0.31

0.15

0.10

1.69

32

0.08

0.30

0.04

0.32

0.15

0.11

1.52

36

0.04

0.26

0.03

0.34

0.18

0.15

Table 3 lists the gel condition predicted for each syntheses stoichiometry using the standard F-S model. Pb,gel is the F-S predicted extent of reaction for the carboxyl functionality at the gel point. For example, the F-S model predicts the gel point for a stoichiometry of 2.0 at Pb,gel = 1.0. According to the F-S model, the HBPE samples of Table 3 with stoichiometry 1.74 and 1.69 having Pb = 0.92 and 0.91, respectively, should be very near the gel point and that of stoichiometry 1.52 with Pb = 0.90 should be well past the gel point. In fact, all of the samples have finite weight- Mw and do not exhibit gelation. The gel point predicted by the Miller–Macosko super-species model is given by Equation (2):

Table 3. The Determined Absolute Mwa and Predicted Mwb for Syntheses with Varying Stoichiometry and the Extent-of-Reaction Gel-Point Predicted by the F-S Model Pb (NMR)

Stoichiometry

Determined Mw (SEC-LALS)

Predicted Mw (Miller–Macosko)

Pb, gel (F-S)

2.16

0.96

1100

1569

no gel

2.00

0.93

1250

1454

1.00

1.93

0.91

2100

2015

0.98

1.74

0.92

2400

2938

0.93

1.69

0.91

4900

3545

0.92

1.52

0.90

14000

14492

0.54

a SEC-LALS. concentration.

b

Miller–Macosko model using NMR-determined super-species

All of the samples in Table 3 would be predicted to be well outside the gel window from this equation. This is indeed what was observed.

185

Controlled Release from the HBPE Platform The following section describes the use of these HBPEs as a platform to covalently bond active agents for delivery to an area of need and to release them in a controlled way.

2-UD as the Active Agent 2-UD has been shown to be an effective insect repellant, but it is quite volatile and must be reapplied every 15 to 30 min to maintain its effectiveness (35–39). Therefore, the effectiveness of 2-UD could be improved by time-release. 2-UD can be covalently bonded to an HBPE with hydroxyl end-groups through a ketalization reaction. However, this reaction generates several different products, which include intramolecular dioxane (six-membered ring) and dioxolane (five-membered ring) structures. Intermolecular ketals can also be formed, which would link two HBPE molecules together, leading to an increase in molecular weight. Examples of the structures that may be formed from intramolecular ketalization are presented in Scheme 2 and those from intermolecular reactions in Scheme 3. The terminal groups of the glycerol HBPE are predominantly composed of one primary and one secondary hydroxyl unit rather than two primary hydroxyl units. This results from the fact that the primary hydroxyls have a greater probability of being converted to an ester unit during the initial polyesterification because they possess greater reactivity and are present in 2-fold concentration relative to the secondary hydroxyl. Therefore, the intramolecular HBPE–2-UD conjugate would be expected to possess predominantly dioxolane structures. HBPEs of glycerol-AA and glycerol-SuA were successfully capped with 2-UD, with the resulting products being identified as glycerol-AA-UD and glycerol-SuA-UD, respectively. The hydrolytic degradation of both products was investigated. SEC chromatograms of glycerol-AA and the corresponding 2-UD conjugate are shown in Figure 2. The chromatogram of the glycerol-AA-UD conjugate demonstrates a significant increase in Mw compared with that of the original HBPE. Approximately 11 wt % 2-UD was covalently bonded to the HBPE, which corresponds to about approximately 2-UD molecules/HBPE on average. This would give rise to a small increase in molecular weight from 2900 to approximately 3200. The fact that the Mw increased to 5100 g/mol suggests that significant intermolecular ketal formation occurred. Because the molecular weight of the HBPE almost doubled upon ketalization, each HBPE contained approximately one intermolecular ketal. Therefore, the distribution of ketal structures was approximately 1.5 intramolecular ketals/HBPE and 0.5 intermolecular (one 2-UD links two HBPEs) ketals/HBPE or 3 times as many intramolecular as intermolecular ketal units. The reason for this unexpectedly large amount of intermolecular ketal formation was most likely because the ketalization was performed without solvent. If the mixture were diluted significantly, a smaller amount of intermolecular ketal formation would be expected. 186

Figure 2. SEC chromatograms of glycerol-AA (bottom) and the crude product formed by the reaction with 2-UD (top).

Scheme 2. Dioxolane (left) and dioxane (right) structures formed from the intramolecular reaction of hydroxyl-terminal HBPEs with 2-UD.

187

Scheme 3. Intermolecular ketal formation from the reaction of hydroxyl-terminal HBPEs with 2-UD.

Figure 3.

13C

NMR spectrum of glycerol-AA-UD. 188

Surprisingly, the addition of the 2-UD to the HBPE generated a less viscous conjugate than the starting polymer. This decrease in viscosity was presumably due to a disruption of the hydrogen-bonding network of the initial HBPE with hydroxyl end-groups as they were converted to 2-UD ketals. The ketalization reaction was carried out in the presence of excess 2-UD. Therefore, the crude product contained free 2-UD. A resonance for the quaternary carbon atom of the ketal is clearly evident at δ 110.1 in the 13C NMR spectrum of the product mixture (see Figure 3). A resonance for a ketone carbonyl at δ 205 demonstrates the presence of a small amount of free 2-UD. The other resonances are assigned on the spectrum. The “R” groups attached to glycerol are either ester or hydroxyl, and those attached to AA are either ester or acid units. When strong acids—such as hydrochloric acid (pKa= –8), TsOH (pKa= –2.8), or phytic acid (pKa = 1.8)—were used to catalyze the hydrolytic degradation of the HBPE-UD conjugates, very effective release of 2-UD was observed. Tosylate and phosphate esters—such as ethyl tosylate, isopropyl tosylate, and triethyl phosphate— were also very effective catalysts for the hydrolytic release of UD. The release rate could be varied from hours to weeks by varying the level of acid in the formulation. For example, in the absence of acid in the formulation, UD was released over a period of months. The release of 2-UD was monitored using NMR spectroscopy, SEC, and GC-MS. SEC chromatograms, demonstrating the degradation of glycerol-AA-UD and release of 2-UD catalyzed by 15% aqueous phytic acid solution at 37 °C after 2 and 7 days, are shown in Figure 4. The glycerol-AA-UD HBPE conjugate was completely hydrolyzed during this period of time. The solutions in which degradation was occurring were periodically analyzed by GC-MS. The concentration of free 2-UD increased rapidly in the first two days and leveled off thereafter, indicating that all of the 2-UD was released within the first 2 days. Degradation of the glycerol-AA HBPE continued during the remaining period. SEC and GC data for the degradation of glycerol-AA-UD in the presence of 15% aqueous phytic acid solution are listed in Table 4.

Figure 4. SEC chromatograms reflecting the degradation of glycerol-AA-UD in the presence of phytic acid: Initial mixture (top) and mixture after 2 (middle) and 7 (bottom) days. 189

Table 4. SEC and GC Data for the Hydrolytic Degradation of Glycerol-AA-UD and the Release of 2-UD in the Presence of Phytic Acid Material Glycerol-AA-UD

Time (days)

Mw (kDa)

GC relative peak area of 2-UD

0

6.6

49

2

3.4

541

7

1.9

487

Figure 5. 13C NMR spectra reflecting the hydrolytic degradation of glycerol-AA-UD and release of 2-UD in the presence of phytic acid: Initial mixture (top) and mixture after 24 h at 37 ºc (bottom). A more quantitative study of the degradation of the glycerol-AA-UD HBPE and release of 2-UD was conducted using a mixture of glycerol-AA-UD and glycerol in a 1:1 mole ratio with either a 15% or a 2% aqueous phytic acid solution at 37°C. The degradation solutions were periodically analyzed using NMR spectroscopy and SEC. The 13C NMR spectra for an initial degradation solution containing 15 wt % (based on polymer) phytic acid and the same solution after 24 h at 37 ºC are shown in Figure 5. The ketal resonances observed in the spectrum of 190

the starting material are not present in the spectrum of the sample after 24 h in the presence of phytic acid. Peaks due to free 2-UD are observed in the spectrum as is the acid carbonyl of AA at δ176 due to hydrolysis of the HBPE. This illustrates that all the 2-UD is released in fewer than 24 h under these conditions. A detailed summary of the rate of hydrolytic release of 2-UD as a function of phytic-acid concentration and temperature is listed in Table 5. A plot of the percentage of 2-UD released as a function of time is shown in Figure 6 showing that the release of UD was complete after 24 h in 15% phytic-acid solution at physiological temperature (37 °C) but released much more slowly in a solution containing 2% phytic acid. These kinetic experiments indicate that both temperature and acid concentration affect the degradation rate. The degradation was much faster at physiological temperature than at room temperature (approximately 21°C).

Table 5. Release of 2-UD from Glycerol-AA-UD as a Function of Acid Concentration and Temperature Phytic acid present (wt %)

Degradation temp (°C)

Control experiment (no phytic acid present) 15 37

21

2

37

21

Time (h)

13C

SEC Mw (kDa)

NMR data Free UD (wt %)

Bound UD (wt %)

0

4.7

11.1

0.0

2

3.9

9.6

2.9

4

3.9

8.5

3.8

6

2.3

7.4

5.3

24

2.1

0.0

12.2

48

1.4

0.0

11.3

24

3.3

8

3

2

4.1

7

1.2

4

4

6.5

1.6

6

2.7

N/A

N/A

24

2.4

3.6

4.5

48

2.1

1.8

5.4

24

3.7

N/A

N/A

These observations clearly demonstrate that 2-UD is released in its active form from the glycerol-based HBPE conjugate even under mild conditions. Timerelease formulations of targeted release rates can therefore be prepared by control of the level of catalyst in the formulation. 191

Figure 6. Release of 2-undecanone from glycerol-AA-UD in the presence of phytic-acid solution at 37 ºC.

Salicylic Acid as an Active Agent Salicylic acid is a common therapeutic drug used for the treatment of acne and inflammation (40–42). It is administered at relatively high doses in the treatment of acne, which often gives rise to skin irritation due its acidic nature. Time-release of salicylic acid would enhance the duration of its effectiveness while minimizing skin irritation. Salicylic acid was covalently bonded to the glycerol-AA HBPE with hydroxyl end-groups by esterification through its carboxylic acid functionality as shown in Scheme 4. The resulting polymer possessed a molecular weight of approximately 3200 g/mol (SEC–multi-angle light scattering). The quantitative 13C NMR spectrum of the polymer, shown in Figure 7, contained resonances for the ester carbonyl at δ 172.6, whereas that of the salicylic-acid conjugate contained a resonance for the ester carbonyl of the polymer at δ 172.7, one for the ester carbonyl from the conjugate at δ 168.3, as well as the aromatic carbon atoms of salicylic acid, which are labeled on the spectrum. The small acid carbonyl of AA just downfield from carbon “c” was due to a slight amount of hydrolysis of the HBPE during attachment of the salicylic acid. The small peak just downfield of carbon “2” of salicylic acid was due to the presence of a small amount of free salicylic acid. 192

Scheme 4. Synthesis of salicylic-acid conjugate of the glycerol-AA HBPE.

An HPLC method was developed to quantify the release of salicylic acid by enzymatic hydrolysis as a function of time in the presence of rat liver microsomes. Calibration plots were constructed using salicylic-acid standards of known concentration such that the percent of drug released from the conjugate as a function of time could be determined. The glycerol-AA-salicylic acid HBPE conjugate was not water soluble; rather, it was a fine suspension created by stirring the solution. Care was taken to stir each solution at a constant rate and amount of time to achieve reproducible release kinetics. The enzymatic hydrolysis of the conjugate was performed in duplicate to evaluate the reproducibility of the experiment. The results, given in Figure 8, show that the hydrolysis rates were quite reproducible. The release rates of salicylic acid for both samples were nearly linear within the 2-h testing window, indicating a fairly constant release rate that is consistent with zero order–release kinetics. 193

Figure 7. Quantitative 13C NMR spectrum of a glycerol-AA HBPE containing hydroxyl end-groups (top) and the salicylic-acid conjugate (bottom).

Figure 8. Rate of salicylic-acid release from the glycerol-AA-salicylic acid conjugate from duplicate degradation experiments. 194

An HBPE conjugate of glycerol-SuA-salicylic acid was shown to release salicylic acid much more slowly than that from the glycerol-AA-salicylic acid conjugate (Figure 9). However, the salicylic-acid release from the glycerol-SA-salicylic acid conjugate was also nearly linear with time.

Figure 9. Salicylic-acid release from glycerol-AA-SA and glycerol-SuA-SA HBPE conjugates.

Figure 10 shows the results of a control study of salicylic-acid rate of release from the glycerol-AA HBPE conjugate with and without the presence of rat liver microsomes. The circles give the rate in the presence of microsomes and the square gives the rate without microsomes. These data clearly show that the rate of salicylic-acid release from the conjugate is orders of magnitude slower without the catalytic effect of rat liver microsomes. The release rate of salicylic acid from HBPEs of glycerol-AA and of glycerolSuA conjugates was shown to be quite different, indicating the ability to control the rate of release based on HBPE composition. The release rates from both HBPEs were also linear with time, suggesting zero-order kinetics for the release. This was probably due to the fact that the conjugates were not water soluble but were rather suspensions in the degradation solution. The suspended conjugate particles were probably eroded from the outer surface in the presence of enzymes, giving rise to zero-order kinetics for the release of salicylic acid. This type of kinetics has been observed for similar systems (43). 195

Figure 10. Release of salicylic acid from the glycerol–AA–salicylic acid conjugate in the presence and absence of rat liver microsomes.

The probable cause for the release of salicylic acid from the glycerol-AA HBPE conjugate being faster than that from the glycerol-SuA HBPE conjugate was the relative particle size of the two dispersions. The glycerol-SuA HBPE conjugate was a stiffer material and more difficult to disperse and therefore probably had a larger particle size.

Naproxen as an Active Agent Naproxen is a nonsteroidal anti-inflammatory drug used for the treatment of pain, stiffness, and swelling. A conjugate was prepared by bonding naproxen (its structure is shown in Figure 11) by esterification to a hydroxyl-terminated HBPE of glycerol-AA. This conjugate also effectively released naproxen by enzymatic hydrolysis with rat liver microsomes. Duplicate kinetic runs for the glycerol-AA-naproxen conjugate are shown in Figure 11. The release of naproxen was significantly slower than that for the salicylic-acid conjugates under the same conditions, but linear release rates were again observed, suggesting the same mechanism for release. 196

Figure 11. Release of naproxen from capped glycerol-AA HBPE in duplicate experiments.

Conclusions A series of HBPEs from glycerol and AA were prepared as a function of initial monomer stoichiometry, thus providing an excellent sample set to demonstrate the power of concerted characterization and modeling tools. The 13C NMR spectra were completely assigned in terms of glycerol, and the five glycerol ester species as a function of initial monomer stoichiometry and the Mw was determined for each. The Mw values were predicted from the measured distribution of glycerol ester species using the Miller–Macosko conditional probability model for this set of HBPEs of different stoichiometry. The model accounted for the unequal reactivity, as well as possible substituent effects, of the glycerol primary and secondary alcohol groups. The predicted and measured molecular weights were in excellent agreement. These tools facilitate the preparation HBPEs of well-defined and reproducible structure. HBPEs from glycerol-AA and glycerol-SuA conjugates were then converted to conjugates with a number of active agents, namely 2-UD, salicylic acid, and naproxen. These HBPEs were shown to provide a useful bio-derived platform for the controlled release of bioactive agents. This delivery platform was shown to be non-toxic, inexpensive, and versatile, thus allowing for tunable time-delivery formulations. The active agents—2-UD, salicylic acid, and naproxen—were covalently bonded to the HBPE, one by ketalization and the others by esterification, and then released through either chemical or enzymatic hydrolysis. The loading level and rate of release were controlled by proper choice of HBPE structure, molecular weight and catalyst level. 197

Acknowledgments The authors thank Dr. Thomas Chamberlain and Dendritech, Inc. for assistance with HPLC analysis of the release rates of the salicylic-acid conjugates and Michigan State University and Central Michigan University for funding.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

18. 19. 20. 21. 22.

Yang, Y.; Lu, W.; Cai, J.; Hou, Y.; Ouyang, S.; Xie, W.; Gross, R. A. Macromolecules 2011, 44, 1977–1985. Jang, Y.-I.; An, B.-K. Polymer 2015, 78, 193–201. Stumbé, J.-F.; Bruchmann, B. Macromol. Rapid Commun. 2004, 25, 921–924. Jena, K. K.; Raju, K. V. S. N.; Prathab, B.; Aminabhavi, T. M. J. Phys. Chem. B 2007, 111, 8801–8811. Kumari, S.; Mishra, A. K.; Krishna, A. V. R.; Raju, K. V. S. N. Prog. Org. Coat. 2007, 60, 54–62. Lin, Q.; Long, T. E. Macromolecules 2003, 36, 9809–9816. Brioude, M. de M.; Guimarães, D. H.; Fiúza, R. da P.; Prado, L. A. S. de A.; Boaventura, J. S.; José, N. M. Mater. Res. 2007, 10, 335–339. Kumari, S.; Mishra, A. K.; Chattopadhyay, D. K.; Raju, K. V. S. N. J. Polym. Sci. Part A., Polym. Chem. 2007, 45, 2673–2688. Somisetti, V.; Allauddin, S.; Narayan, R.; Raju, K. V. S. N. RSC Adv. 2015, 5, 74003–74011. Saeed, H. A. M.; Eltahir, Y. A.; Xia, Y.; Wang, Y. Adv. Mater. Res. 2014, 937, 80–85. Zhang, T.; Howell, B. A.; Dumitrascu, A.; Martin, S. J.; Smith, P. B. Polymer 2014, 55, 5065–5072. Coneski, P. N.; Rao, K. S.; Schoenfisch, M. H. Biomacromolecules 2010, 11, 3208–3215. Cao, W.; Zhou, Z.; Wang, Y.; Zhu, L. Biomacromolecules 2010, 11, 3680–3687. Ifran, M.; Seiler, M. Ind. Eng. Chem. Res. 2010, 49, 1169–1196. Lin, C.; Gitsov, I. Macromolecules 2010, 43, 10017–10030. Shi, X.; Wang, S. H.; Lee, I.; Shen, M.; Baker, J. R., Jr. Biopolymers 2009, 91, 936–942. Ye, L.; Letchford, K.; Heller, M.; Liggins, R.; Guan, D.; Kizhakkedathu, J. N.; Brooks, D. E.; Jackson, J. K.; Burt, H. M. Biomacromolecules 2011, 12, 145–155. Chaterjee, S.; Ramakrishnan, S. Macromolecules 2011, 44, 4658–4664. Gao, C.; Yan, D. Prog. Polym. Sci. 2004, 29, 183–275. Mishra, M. K., Kobayashi, S. Star and Hyperbranched Polymers. CRC Press: Boca Raton, FL, 1999, p 53. Coullerez, G.; Lundmark, S.; Malmstrom, E.; Hult, A.; Mathieu, H. J. Surf. Interface Anal. 2003, 35, 693–708. Dvornic, P. US Patent 6,812,298. 198

23. Weimer, M. W.; Fréchet, J. M. J.; Gitsov, I. J. Polym. Sci., Part A: Polym. Chem. 1998, 36, 955–970. 24. Bartels, J. W.; Cheng, C.; Powell, K. T.; Xu, J.; Wooley, K. L. J. Polym. Sci. Part A: Polym. Chem. 2006, 44, 4782–4794. 25. Bartels, J. W.; Cheng, C.; Powell, K. T.; Xu, J.; Wooley, K. L. Macromol. Chem. Phys. 2007, 208, 1676–1687. 26. Seidi, F.; Jenjob, R.; Crespy, D. Chem Rev. 2018, 118, 3965–4036. 27. https://www.fda.gov/Food/IngredientsPackagingLabeling/ FoodAdditivesIngredients/ucm091048.htm#ftnA (accessed September 14, 2018). 28. Zhang, T.; Howell, B. A.; Smith, P. B. Ind. Eng. Chem. Res. 2017, 56, 1661–1670. 29. Kulshrestha, A. S.; Gao, W.; Gross, R. A. Macromolecules 2005, 38, 3193–3204. 30. Macosko, C. W.; Miller, D. R. Macromolecules 1976, 9, 199–206. 31. Miller, D. R.; Macosko, C. W. Macromolecules 1980, 13, 1063–1069. 32. Sarmoria, C.; Miller, D. R. Macromolecules 1991, 24, 1833–1845. 33. Flory, P. J. Principles of Polymer Chemistry; Cornell University Press: Ithaca, NY, 1953; Chapter 9. 34. Wyatt, V. T.; Strahan, G. D. Polymers 2012, 4, 396–407. 35. Witting-Bissinger, B. E.; Stumpf, C. F.; Donohue, K. V.; Roe, R. M. J. Med. Entomol. 2008, 45, 891–898. 36. Roe, R. M. Method for repelling insects; US Patent 2003/0083387 A1, May 1, 2003. 37. Bissinger, B. W.; Zhu, J.; Apperson, C. S.; Sonenshine, D. E.; Watson, D. W.; Roe, R. M. Am. J. Trop. Med. Hyg. 2009, 81, 685–690. 38. Witting-Bissinger, B. E.; Stumpf, C. F.; Donohue, K. V.; Roe, R. M. Method of repelling insects; US Patent 6,437,001 B1, August 20, 2002. 39. Bissinger, B. W.; Apperson, C. S.; Watson, D. W.; Arellano, C.; Sonenshine, D. E.; Roe, R. M. Med. Vet. Entomol. 2011, 25, 217–226. 40. Amann, R.; Peskar, B. A. Eur. J. Pharmacol. 2002, 447, 1–9. 41. Amborabé, B.-E.; Fleurat-Lessard, P.; Chollet, J.-F.; Roblin, G. Plant Physiol. Biochem. 2002, 40, 1051–1060. 42. Titus, S.; Hodge, J. Am. Fam. Physician 2012, 86, 734–740. 43. Faig, J. F.; Smith, K.; Moretti, A.; Yu, W.; Uhrich, K. E. Macromol. Chem. Phys. 2016, 217, 1842–1850.

199

Chapter 14

Aromatic Bioplastics with Heterocycles Sumant Dwivedi and Tatsuo Kaneko* Graduate School of Advanced Science and Technology, Energy and Environment Area, Japan Advanced Institute of Science and Technology, 1-1 Asahidai, Nomi, Ishikawa 923-1292, Japan *E-mail: [email protected].

In the interest of establishing sustainable societies, the production of polymers from bio-based renewable materials has gained momentum owing to the availability of raw materials at low costs from fermented materials. White-biotechnology has catalyzed the production of bio-based raw materials and plays a significant role in lowering final product costs. Various kinds of bio-based aliphatic plastics have been developed for several decades, and high-performance polymers having heterocyclic and/or aromatic structures have been commercialized. The wide expansion of bioplastics is driven by outstanding progress in the processes for refining biomass feedstocks. These feedstocks produce the building blocks that allow versatile and adaptable polymer chemical structures to achieve tailored properties and functionalities. In this chapter, the recent progress in bioplastics composed of heterocyclic and/or aromatic structures, such as polyamides, polyureas, and polyimides, is described and their molecular structure-performance relationships are reviewed. Additionally, the recent scientific achievements regarding high-spec engineering-grade bio-based polymers are discussed.

Introduction Prior to the 1960s, the polymer industry relied heavily on petroleum-derived chemistry, refinery, and engineering processes. The industrial perspectives on the high-economy generation processes had not changed until the negative impact on the environment reached a critical level by the late 20th century. The 21st century © 2018 American Chemical Society

began with two serious challenges for the polymer industry: global warming and dwindling fossil resources. One of the promising methodologies for combating these problems is to use sustainable resources rather than fossil-based resources. Biomass feedstocks are a promising resource because of their continuous availability and sustainability. Biomass feedstocks can be converted into raw materials for polymer production, and the resulting polymers are called “bio-based polymers.” Currently, several kinds of bio-based polymers such as polylactides (PLA), poly(hydroxy alkanoates) (PHAs), succinate derived polymers, and others are gaining pace through rapid development and commercialization as a response to waste accumulation problems largely encountered in the agricultural, marine fishery, and construction industries (1–6). The progress made in bio-derived polymers is recognized as one of the most successful innovations in the polymer industry for addressing environmental issues. In terms of market development, it has been projected that in the case of bio-based polymers, their production capacity will triple from 5.1 metric tonnes in 2013 to 17 million tonnes in 2020 (7). Bio-based drop-in substitutes for PET and the new polymers PLA and PHA show the fastest rates of market growth. The lion’s share of capital investment is expected to take place in Asia (8–11). The bio-based production capacity of 5.1 million tonnes represented a 2% share of the overall structural polymer production of 256 million tonnes in 2013 (12). The bio-based polymer turnover was about €10 billion worldwide in 2013. There are several successful examples of commercialization of these polymers, including the pilot-scale production of polylactide (PLA) at NatureWorks and Corbion/Total; poly(trimethylene terephthalate) (PTT) at DuPont; poly(isosorbide carbonate) at Mitsubishi Chemicals; bio-based polyamides at Arkema, Toray, BASF, DSM, and others; and poly(ethylene 2,5-furandicarboxylate) (PEF) at Synvina (12, 13). Bio-based polymers are being applied to general and engineering situations. For example, because of the improvements in the physical durability and processability of PLA, it has been used in the packaging industry (1, 14, 15). In addition, owing to the superior gas barrier properties of PEF, it is being used in bottles, films, and other packaging materials in the food and beverage industry. Furthermore, bio-based PTT is analogous to petroleum-derived PTT, and its bio-based, sustainable nature and intrinsically flexible chain properties afford comfortable stretching and shape recovery properties that make it attractive and promising. The current general approach for bio-based plastic processability entails physical modification and optimization of polymer processing, including the optimization of processing parameters, extruder screw design, selection of appropriate additives, and post-orientation for strain-induced crystallization (16–21). These developments in the processing conditions have made bio-based polymers analogous to certain petroleum-derived polymers and have helped to establish a sizeable market. The aforementioned bio-based polymers have been remarkably well developed because of their high mechanical strengths. However, it was estimated that these polyesters will only replace a small percentage of the nondegradable plastics currently in use owing to their poor thermal resistance (22). As a result, high-performance environmentally friendly polymers from bio-based materials that are degradable after usage into natural molecules are urgently desired in 202

order to improve human life. Many nondegradable engineering plastics have rigid conjugated rings, such as benzene, benzimide, benzoxazole, benzimidazole, or benzothiazole (23). The introduction of a heterocyclic and/or aromatic component into a thermoplastic polymer backbone is an efficient method for intrinsically improving material performance. Additionally, the continuous sequence of aromatic rings can be a mesogenic group. Molding in the thermotropic liquid crystalline (LC) state can induce molecular orientation, imparting anisotropy to mechanical performance, which sometimes dramatically increases mechanical strength and Young’s modulus (24). Outstanding reviews and textbook knowledge are available for bio-based polyesters and other kinds of aliphatic bio-derived polymers (25–29); however, a systematic review for the aromatic and heterocyclic bio-based engineering plastics is lacking. This chapter reviews the recent important developments in the preparation of various rigid polymers derived from phytochemical monomers and other fermented products. Moreover, an exhaustive discussion on the significant progress in heterocyclic/aromatic bio-based polyamides (PAs), polyureas (PU), and polyimides (PIs), as well as their (nano)composites is included for a better understanding of structure-performance relationships and value addition through material hybridization.

Bio-Based Polyamides Polyamides (PAs or Nylon) are widely used as engineering thermoplastics because of their unique physicochemical properties, and their high thermomechanical properties attract many researchers in polymer industries (30–32). The amide linkages exhibit hydrogen bonding, and thereby good thermal stability and mechanical properties. Conventional PAs account for a significant fraction of the thermoplastics field. However, based on the previously mentioned background of environmental issues, several PA scientists in both industry and academia have trialed the process conversion from petro-chemistry to white-biotechnology by incorporating cost-performance tradeoffs. Most of the development began with the syntheses of aliphatic bio-based monomers for PAs by DSM, Evonik, Arkema, and others (12, 13). Figure 1 illustrates the structures of various bio-based PAs with minimized carbon footprint, such as PA11 derived from castor oil and other PAs. PA6,10 is synthesized by the interfacial polymerization of hexamethylenediamine and sebacoyl chloride, which was also derived from castor oil to afford a wide variety of molecular designs involving PAs. Furthermore, the properties of the bio-derived PAs were benchmarked with the commercial PA6 and PA6,6. Several aliphatic bio-polyamides such as PA4 and PA4,6 showed quite good thermomechanical properties. These PAs exhibit a wide range of applications, from flexible tubes to high-end electronic devices.

203

Figure 1. Syntheses of bio-derived polyamides.

The introduction of aromatic/heterocyclic moieties in the bio-based PA backbone improved their thermomechanical properties. Several industries attempted to prepare “partially” bio-derived PAs, in which only the aliphatic component was bio-derived. Mitsubishi Chemical Corporation prepared poly(m-xylylene adipamide) (MXD-6) by polycondensation of m-xylylene diamine with bio-derived adipic acid. MXD-6 exhibits a glass transition in the 204

range 85–100 °C and a melting point in the range 235–240 °C (1, 3). QianaTM developed by DuPont is an aliphatic-aromatic nylon fiber with a melting point of 275 °C (1–3, 33–36). In general, PAs absorb moisture ranging from 2 to 10% owing to the amide linkage interacting with the water molecules via hydrogen bonding, which has a plasticizing effect and reduces their melting point, glass transition temperature, and mechanical properties (37–40), although the impact strength and resilience increase. Since PAs with longer interconnecting alkyl chains have lower contents of amide linkages, they reveal lower water absorption. However, the potential of bio-PAs to provide superior material performance has yet to be exploited by incorporating heterocyclic and/or aromatic structures in the PA backbone. Reactive alicyclic or aromatic structures do not widely exist as biomolecules for exploitation in monomer production. Furthermore, the toxicity of these superior moieties for bio-PAs renders their biological availability quite low. Itaconic acid (IA, Figure 1), which is widely produced by the fermentation of Aspergillus terreus, possesses two carboxylic groups separated by a vinylidene group adjacent to a carboxylic group as a α,β-unsaturated compound (41). The conventional polycondensation reaction of IA with a diamine in the solution phase was extremely difficult to accomplish due to the three-way branching resulting from the Michael addition and amidation of three amines to the double bond and two carboxylic acids of IA, respectively. Therefore, the polymerization of the complex monomer was conducted using 1:1 organic salts of IA and diamine followed by melt condensation, leading to the formation of rigid five-membered pyrrolidone groups along the bio-PA backbone with superior thermomechanical properties (41, 42). The heterocyclic ring enhances the fatigue resistance and wearing durability, and promotes strong intermolecular hydrogen bonding. Its glass transition temperature can be attributed to the molecular motion of the amorphous regions, such as translation motion, the alkyl chain length, atomic vibrations, movability of the crystal lattice, net amide content, and chain uncoiling. The amide content plays a critical role in deciding the Tg as a result of hydrogen bonding, which leads to better lattice arrangement and a greater degree of crystallization. The Tg of PA increases with increasing crystallization. Greater crystallinity impacts the major properties of PAs such as high moisture absorption, abrasion resistance, high stiffness, high density, high moduli, high chemical resistance, and better dimensional stability, but decreases elongation, impact resistance, thermal expansion, and gas permeability (43, 44). The melting temperatures of aliphatic PAs are greatly affected by the alternation of odd or even alkyl chain length in the diamine moiety. It was observed that PAs with odd numbers of methylene groups have higher melting temperatures than those with even numbers of methylenic groups, owing to the varying levels of inter/intra-molecular hydrogen bonding. For instance, the melting temperatures of PA4,6, PA6,6, PA6,10, and PA6,12 are 300 °C, 260 °C, 225 °C, and 210 °C, respectively (45). In the case of PAs and coPAs with varying degrees of crystallinity, the polymer molecular weight greatly affects the melting point. Similar trends were also observed with the heterocyclic IA-based bio-PA. One of the motivations for the development of bio-based polymers was their biodegradability, which is becoming increasingly important due to strong public 205

concerns about waste management. It is worth mentioning that the pyrrolidone ring was opened by ultraviolet (UV) irradiation and through composting in a landfill for extensive periods of time, leading to a reduction in the molecular weight of the polymer, thereby demonstrating the degradation ability of bio-based engineering thermoplastics (41). Aromatic PAs with very high glass transition temperatures (156–242 °C) have also been synthesized from bio-based monomers that introduce a heterocyclic ring (42). IA reacted with aromatic diamines such as 4,4′-diaminodiphenylether, p-phenylenediamine, m-xylenediamine, and p-xylenediamine, leading to the syntheses of aromatic bio-PAs. Out of these structures, the p-phenylenediamine-based bio-PA reveals the highest thermal stability due to the absence of ether and methylene linkages among the aromatic groups. Owing to their rigid backbones, especially the entirely p-substituted derivative, aromatic PAs have high thermomechanical properties (tensile and impact resistance), are chemically resistant to alkali, and are hydrolytically stable. The two most famous commercial aramids are NomexTM and KevlarTM. Morgan (DuPont) discovered NomexTM in 1958 and it was later commercialized in 1961 (1, (46, 47)). On the other hand, KevlarTM was discovered by Kwolek (DuPont) in 1965 and it was later commercialized in 1971. NomexTM was synthesized by the polymerization of m-phenylene diamine and isophthaloyl chloride. NomexTM has a high melting point (400 °C) and thermal stability, and allows easy processing in the form of fibers from aprotic polymer solutions (1–3). However, owing to the poor solubility of aromatic bio-PAs in the commonly used organic solvents, the processing becomes difficult, which limits their applications. The behavior of hybrid composites is a balance of the advantages and disadvantages of each component, in which the advantages of one type of fiber could compensate for the lack of it in another fiber. A small content of carbon fibers was found to be capable of significantly improving the mechanical properties of PA11, including stiffness, elastic deformation before yield, and creep properties (43–45). However, carbon fiber is not renewable and quite expensive compared to other fiber types because of the high consumption of energy during its manufacturing. Compared to wood fiber composites (30% wood fiber), hybridization with carbon fiber (10% wood fiber and 20% carbon fiber) increased the tensile and flexural moduli by 168% and 142%, respectively (46). The Izod impact strength of the hybrid composites exhibited a good improvement compared to that of the wood fiber composites. Hence, hybridization of a small amount of carbon fibers with natural fiber could be another cost‐effective alternative for developing high-performance biocomposites. The addition of inorganic fillers in the PA matrix that act as nucleating sites enhances the net crystallinity and abrasion resistance of the resulting composite material. Nanohybridization of the IA-based bio-PA with montmorillonite (MMT) clays enhances not only the moduli through improved molecular orientation from cryogenic treatment but also the elongation degree for Na-MMT. The increment in the strain energy density observed from improved elasticity was attributed to the greater flexibility of the polymer backbone due to the catalytic hydrolysis of the pyrrolidone moiety through interactions with the silicate layers (42). 206

Transparent high-performance PAs possess largely amorphous structures, which are controlled by a precise molecular design of monomers having aromatic units interlinked with (cyclo)aliphatic units (47). Path-breaking advancement requires connecting the dots along the multidisciplinary scientific progress. Producing a bio-based diamine with an aromatic structure in the backbone requires precise molecular design. Takaya et al. developed an artificial biosynthetic route to 4-aminocinnamic acid (4ACA) in Escherichia coli (E. coli) cells to obtain 4-aminophenyl alanine (4APhe; Figure 2) from glucose followed by deamination (48). However, the yield of 4ACA was too low for realistic commercialization of the technology. Adopting white-biotechnology led to the improved yield of 4ACA through genetically engineered E. coli that was at least 30 times more than that obtained using the conventional deamination techniques with 90% purity (47). A microbial catalyst was employed to reduce 4ACA to 4-aminohydrocinnamic acid (4AHCA), and this molecule was utilized for polycondensation to yield a PA with a very high glass transition temperature of more than 240 °C and 10 % weight loss temperature above 390 °C. On the other hand, the synthesis of diamine through the [2+2] photocycloaddition reaction of the 4ACA salt yielded 4,4′-diaminotruxillic acid-based monomers (4ATA-acid/methyl/ethyl ester). The polycondensation of 4ATA-acid and 4ATA-methyl ester led to the formation of a bio-PA with transparency over 95% at 550 nm and a tensile strength of over 400 MPa (the highest among transparent plastics and borosilicate glass). The ultrahigh mechanical strength of the bio-PA was attributed to the prospective spring-like flexible function of the phenylenecyclobutanyl backbone. However, it is interesting to note that the spring function of the phenylenecyclobutanyl backbone is also responsible for degradation under UV irradiation, due to the ease of cleavage along the strained ring. This suggests photo-degradation, which is an indispensable aspect of the biopolymers development. These bio-PAs may be utilized for versatile applications in the manufacturing of automobile parts, flexible electronic devices, sensors, and other optical materials as well.

Bio-Based Polyureas Polyurea (PU) is an elastomeric polymer suitable for application in ballistic protection owing to its characteristic set of properties: light weight, high elasticity, high flexibility, heat resistance, impact resistance, and high energy absorbing capacity (49, 50). It is also one of the most successful materials used in the coating industry, having multiple applications in many fields of activity owing to its fast curing, chemical resistance, low flammability, good stability and durability, and excellent bonding properties to all types of surfaces, especially metals (51). The synthesis of PU elastomeric films consists of a rapid polyaddition reaction between two oligomeric species: a diamine and a diisocyanate. The PU films described in the literature are generally obtained by simply mixing two commercial components: an isocyanate component (Isonate 143L, Basonat HI-100, Vestanat IPDI) and an oligomeric diamine (Versalink P-1000, Jeffamine D-2000 and D-400, Jefflink 754, Clearlink 1000) (52–54). 207

Figure 2. Biosyntheses of 4-aminocinnamic acid using genetically-manipulated Escherichia coli (studied by Prof. N. Takaya of Tsukuba University, Japan). Reproduced from ref. (67). Copyright © 2014 American Chemical Society.

The aforementioned PUs have an interesting microstructure consisting of two distinct microphases with different characteristics: aromatic ring regions (hard domains) that are chemically bonded and homogenously distributed throughout the aliphatic polymer chain matrix (soft domains) (55). Depending on the composition, the properties of a PU coating change dramatically. Thus, a higher aliphatic chain content increases the flexibility, but decreases the strength, whereas a higher content of aromatic rings leads to increased strength but lower elasticity. Another important aspect related to the PU coatings is their transition from the glassy to a rubbery state during the deformation process, which occurs earlier in the case of high impact forces (49, 56). The amount of energy dissipated through the PU film is comparable to its loading-unloading behavior, which is related to the hysteresis area under the stress–strain curve. The PU properties for a certain type of application can be designed by choosing the right components and the adequate proportions between them. Thus, the behavior of these materials can be controlled by the synthesis parameters. Several kinds of bio-based PUs based on 4-aminophenylalanine (4-APhe, available from genetically modified E. coli., as presented in Figure 2) and 4-APhe methyl ester along with several kinds of diisocyanates (varying degrees of aromaticity: MDI, MMDI, 1,3 PDI, and 1,4 PDI) were synthesized (57, 58), as shown in Figure 3. Overall, almost all of these bio-based PUs have higher thermal degradation temperatures of up to 200 °C, whereas the carboxylic acid containing 4-APhe shows superior thermomechanical properties than the PU obtained from methyl-ester-based 4-APhe, with unified diisocyanate structures. This observation was attributed to the interchain hydrogen bonding of carboxylic acid with the amide and carbonyl units. Furthermore, bio-based PUs with a greater degree of hydrogen bonding and rigid aromatic backbone showed superior tensile properties. On the contrary, flexible interlinked aromatic structures containing bio-based PU exhibited better ductilities and higher strain energy densities (above 208

10 Jcm–3). In general, these bio-based PUs showed better thermal stability and mechanical properties than conventional aliphatic PUs or the PU spray elastomer (Spandex) (49, 57).

Figure 3. Syntheses of bio-based polyureas.

On the other hand, introducing ionic interactions among the repetitive units (4ATA-acid and aromatic diisocyanate MDI, Figure 3) in PU results in a significant transition from a highly transparent film to a colored one due to the formation of a metal-ligand charge transfer complex (47). Furthermore, the 4ATA-acid based PU reveals significant swelling characteristics in alkaline solutions that lead to highly improved mechanical properties, especially for multivalent cationic interactions with the 4ATA-carboxylate repetitive unit. The high solubility of these bio-based PUs in aprotic solvents ensures easy processability and is comparable with those of conventional PUs. During the past decades, great progress has been made in the development of segmented thermoplastic polyurethanes/polyureas (TPUs), which are applied in many areas such as the automotive industry, clothing, sportswear, and medical applications (59, 60). Conventionally, TPUs are prepared from polymeric diols (which act as the soft segment (SS)), diisocyanates, and low molecular diol or diamine chain extenders (the low molecular compounds together form the hard segment (HS)). However, the toxicity of the diisocyanates should not be ignored, especially for the TPUs employed in biomedical applications. Several 209

isocyanate-free routes have been developed for preparing TPUs, starting from carbonylbiscaprolactam, di-tert-butyltricarbonate, cyclocarbonates, transition metal catalyzed transurethanization, etc. (61). A metal-free organic catalyst was utilized as an isocyanate-free route for synthesizing a series of segmented PUs from renewable resources: dimethylcarbonate (DMC, metabolic engineered E. coli.), diamino-terminated polypropylene glycol (PPGda, potentially bio-based and originating from propylene oxide), and 1,4-diaminobutane (DAB or putrescine) (62). The renewable segmented PUs contain monodisperse HSs. Dynamic mechanical analysis of the PUs reveals a sharp glass transition, a sharp flow transition, and a flat rubbery plateau. The flow and maximum use temperature (FL) of PUs increases with the increasing number of urea groups in the corresponding dicarbamates. In addition, for a constant HS length, varying the length of the SS allows the modulation of the thermomechanical properties of PUs, enabling their application as adhesives, soft elastomers, or rigid plastics.

Bio-Based Polyimides The development of PAs and PUs as engineering plastics triggered the evolution of highly chemically and thermally resistant and mechanically strong polymers through the incorporation of aromatic/heterocyclic rings in the polymer backbone. In the 1950s, DuPont employed pyromellitic dianhydride and 4, 4′-oxydianiline as monomers and invented a new class of polymers with repetitive imide units, polyimides, and commercialized the material as Kapton® (63). Since then, many scientific reports have been published on the development of more advanced polyimides, though scarcely any report has mentioned the synthesis of biopolyimides. The recent progress in the preparation of a bio-based polyimide from the exotic amino acid 4ACA (Figure 2) has drawn special attention due to the advanced optical, thermo-mechanical, and memory characteristics of the polyimide (64–66). The various kinds of diamine synthesis as monomers from 4ACA have been prepared through photo-cycloaddition reactions. The low thermal stability of cyclobutane in the truxillate backbone motivated researchers to dimerize 4ACA through alternative routes (67–71). Grubb’s olefin metathesis was utilized to prepare 4, 4′-diaminostilbene (DAS) and its reduced counterpart 4, 4′-(ethane-1,2-diyl)dianiline (EDDA) for polyimide synthesis (68). Furthermore, a salt of 4, 4′-diaminotruxillate was utilized for the preparation for polyimide-hybrid materials (70). Figure 4 represents the preparation of a polyimide using the 4ACA-based diamine monomer, and mostly comprised thermal imidization, except in the case of polyimide-metal oxide as a hybrid material. Among all the products of polycondensation, CBDA exhibits the highest inherent viscosity, irrespective of the diamine type, which can be attributed to the higher reactivity of the anhydride unit attached to the aliphatic cyclobutane ring that produces a positive inductive effect (67, 72). Furthermore, the biopolyimide possesses high chemical resistance, especially the DAS-based polyimides, which are insoluble in non-polar/protic-polar/aprotic solvents; this may be attributed to 210

the conjugated trans-stilbene unit of the diamine, which provides structural rigidity. However, EDDA-based diamine and ATA-methyl/ethyl ester are soluble only in concentrated sulfuric acid, probably due to the coplanar rings of the aromatic-imide structures (68). These synthesized biopolyimides have an amorphous nature, except the polyimide from CBDA, which may be ascribed to the cyclobutane rings, phenylenes, and imide rings that induce partial crystallization. It is important to mention that a slight change in the monomer structure can significantly affect the polymer crystallinity, e.g., replacing the methyl with ethyl in the ATA-ester in CBDA-based polyimide decreases the crystallinity by 20% (73). The copolymers also show high chemical resistance to non-polar/polar-protic/aprotic solvents; however, all the co-polyimides exhibit solubility in concentrated sulfuric acid and trifluoro acetic acid. Nevertheless, it was found that reducing the torsional energy of the dianhydride ring significantly improves the solubility of the biopolyimide in aprotic solvents (69). Moreover, a water-soluble biopolyimide has been introduced through the neutralization of the carboxylic acid moieties in each repetitive unit (72). One-step chemical imidization of the ATA salt and BCDA was performed in the presence of isoquinoline, and thereafter, the ATA salt neutralized with an acid. The polyimide resulting from chemical imidization reveals less chemical resistance, as it showed solubility in a series of aprotic solvents, which in turn offers an opportunity for the sol-gel condensation reaction. The sol-gel reaction comprises the hydrolysis and condensation of metal alkoxide units. Titanium and zirconium alkoxide precursors were added to a solution of polyimide in DMAc with acidic catalysts to yield a polyimide-metal oxide composite film as a hybrid material. Furthermore, the content of the metal oxide was varied 10–50 wt% and the average particle size of the oxide was around 5 nm (for 50 wt% TiO2 or ZrO2) (70). The transmittance of the 4ATA-derived homo/co-polyimide film was measured for a normalized thickness of 10–20 μm with a cutoff wavelength of 450 nm (T450). In general, it was observed that poly(amic acid) (PAA) from all 4ATA-derived monomers exhibit an almost transparent to pale-yellow color (T450 > 88%) due to the variation in dianhydride. Aromatic dianhydrides exhibit more intense colors than their aliphatic counterparts due to the electron-rich benzenes of ATA and dianhydrides that induce strong charge-transfer (CT) interactions (67). However, it is important to mention that all the 4ATA-derived polyimides were more transparent than Kapton®. Usually, transmittance increases with the loss of conjugation in the polymer backbone (67–74). The optical properties of the 4ATA salt and BCDA-based polyimide show outstanding transmittance, presumably due to the alicyclic dianhydride, which induces weak intra and intermolecular CT interactions. It is important to mention that metal oxide particle dispersion in the polymer matrix and the band gaps of ZrO2 (5.0–5.85 eV) and TiO2 (3.2 eV) critically affect film transparency, with a greater band gap resulting in a higher transparency (75, 76). The refractive index and Abbe number of the polyimide hybrids increase with metal oxide content, and indicate the sol-gel reaction between the 4ATA salt and M-OH (M = Ti, Zr) to form M-O-M structures (70, 77). In other words, controlling the amount of metal oxide 211

in the polyimide matrix provides an opportunity to tune the refractive index of the composite, as well as its Abbe number, for advanced optical applications.

Figure 4. Syntheses of bio-based polyimides. The DAS-based polyimide exhibits exceptionally high thermal resistance compared to the EDDA-based counterpart due to the extended conjugation in the polymer backbone structure. The aromaticity in the polymer backbone is responsible for the greater thermal resistance. In other words, CBDA-based polyimides with DAS or EDDA show lower thermal resistances than the other aromatic dianhydrides due to the absence of π-electron conjugation. The 10% weight loss temperatures (Td10) of the DAS- and EDDA-based polyimides were in the range 400 °C to 600 °C. In particular, the OPDA-DAS-based polyimides show an exceptionally high Td10 value of 600 °C, which is the highest reported for any polyimide (68). On the other hand, the 4ATA-based homopolymer exhibits satisfactory thermal resistance, with Td10 values ranging from 399 °C to 425 °C, which are lower than DAS/EDDA due to the presence of the strained and non-conjugated cyclobutane ring of truxillic acid along the diamine unit, compared to DAS. Along similar lines, 4ATA-based copolymers show a Td10 value around the arithmetic mean of the corresponding values of the individual dianhydride-based homopolymers. The glass transition temperature was sufficiently high for several polyimides (> 250 °C) and difficult to determine for most polyimides. However, the polymer backbone with lower torsion energy exhibits glass transition behavior and promotes easy processability in aprotic solvents (67, 68). Furthermore, the 212

incorporation of metal oxides in polyimides increases the Tg, as well as Td10, of the polymer hybrids due to the uniform dispersion of rigid inorganic components. Contact angle measurements of the 4ATA-based polyimide revealed the film surface to be hydrophobic. The cell compatibility of biopolyimide was tested based on its adherence to L929 fibroblast cells. All the polyimides exhibit similar results with good cell compatibility with increasing incubation time, which is similar to Kapton®. The mechanical properties of super-engineering plastics have always been the center of attention and are mostly expressed in terms of Young’s modulus (E), elongation at break (ε), tensile strength (σ), and strain energy (U). Monomer structure and polymer molecular weight are among the significant factors that affect the mechanical properties. For instance, the kinked sulfone unit of DSDA increases polymer flexibility (good ε, high σ, and E), whereas the rigid cyclobutane/a-single benzene of PMDA makes the polymer brittle (low ε). In general, the DAS/EDDA-based polymers reveal high tensile strengths (σ = 54–132 MPa), moduli (E = 2.3–4.3 GPa), and a large window of polymers with varying degrees of elongation (ε = 1.5–7.8%) (67, 74). In general, most of the 4ACA-derived polymers show higher moduli than Kapton®. The 4ATA-based polyimide exemplifies a large window of modulation of the mechanical properties modulation (σ = 28–114 MPa, E = 3–13 GPa, and ε = 0.9–9.4%) by varying the structure of the dianhydride(s). This observation can be envisaged to the structural disordering that renders interchain stacking difficult and improves the ability of the carbonyl-connecting groups to rotate the benzene rings of the tetra-acid (74). Furthermore, the copolyimide ductility of the copolymer substantially increased compared to the corresponding homopolymer, which can be suitably used in foldable devices. Recently, it was observed that the greater surface hydrophilicity of the biopolyimide induces stronger interfacial interactions with the ITO nanolayer, resulting in high transparency, smoothness, and high robustness against mechanical deformation (78). Hybridization of polyimide with graphite is important for application in electronic devices (79). The adhesion strength of the ATA-based homopolyimide on a carbon substrate was found to be equivalent to that of a conventional instant super glue prepared from α-cyanoacrylate polymers. The HOMO and LUMO levels calculated from cyclic voltammetry, and the amalgamation of the information with molecular simulations based on the density functional theory provides a detailed insight of the memory characteristics of these 4ATA salt/BCDA polyimide hybrids. It was found that the introduction of TiO2/ZrO2 as electron acceptors into the neat polyimide matrix results in a lower LUMO level, which is one of the basic requirements of memory devices, and facilitates as well as stabilizes the CT interactions. In other words, the type of memory greatly depends on the content of the metal-oxide (0–50 wt%). The neat polyimide showed no memory properties, but introducing TiO2/ZrO2 up to 5 wt% yields dynamic random access memory (DRAM); up to 10 wt%, the polyimide-hybrids behave as static random access memory (SRAM), while beyond 10 wt%, the materials act as write once read many (WORM). In other words, the memory device revealed longer retention time and lower threshold voltage with increasing TiO2/ZrO2 and exhibited bi-switchable characteristics (70, 78, 80). The introduction of 213

SiO2 in the 4ATA-ester/CBDA polyimide leads to improved optical, dielectric, and dielectric breakdown strengths. On the other hand, the 4ATA-ester/CBDA polyimide hybrid with silica showed lower transparency and electrical properties due to prospective localized interactions with carboxylic acid and silanols (71). These observations clearly exemplify the opportunity of these hybrid materials in transparent memory, display, and insulating devices with high robustness.

Future Outlook In the pursuit of sustainable development and reduction of environmental impact, bio-based polymers are among the best materials to tackle global environmental concerns. Undoubtedly, the production of various kinds of monomers from bio-based precursors presents an alternative to the conventional dependence on fossil-based monomers for biobased polymer production.

Figure 5. Transparent sheets of high-performance bioplastics. Reproduced from ref. (47). Copyright © 2016 American Chemical Society. The bioderived engineering plastics show either equivalent or better thermomechanical characteristics with good transparency (Figure 5) compared to the conventional engineering plastics. Many biomolecules are multifunctional, which is advantageous in the molecular design of heterocyclic polymers having high thermal/mechanical performances. Furthermore, hybridization of biopolymers with various fillers leads to superior properties. In particular, the application of biopolyimides (hybridized with metal oxides) offers an opportunity for the development of advanced materials such as transparent memory devices or as an ophthalmological material. Further strategic exploration of bio-based monomers, as well as the design of new monomers, presents an outstanding prospect for new and expanded horizons for bio-based polymer chemistry.

References 1. 2.

3.

Steinbuchel, A. Biopolymers; Wiley-VCH: Weinheim, Germany, 2001. Domb, A. J.; Kost, J.; Wiseman, D. M; Handbook of Biodegradable Polymers; Harwood Academic Publishers: London, 1997; ISBN 90-5702-153-6. Klass, D. L. Biomass for Renewable Energy, Fuels, and Chemicals; Academic Press: San Diego, CA, 1998; ISBN 0-12-410950-0. 214

4. 5.

6. 7. 8. 9. 10. 11. 12.

13.

14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.

Ladmiral, V.; Jeannin, R.; Fernandes Lizarazu, K.; Lai-Kee-Him, J.; Bron, P.; Lacroix-Desmazes, P.; Caillol, S. Eur. Polym. J. 2017, 93, 785–794. Toledo, P. V.; Limeira, D. P.; Petri, D. Abstracts of Papers, 255th ACS National Meeting & Exposition, New Orleans, LA, United States 2018, PMSE-271. Watanabe, H.; Fujimoto, A.; Nishida, J.; Ohishi, T.; Takahara, A. Langmuir 2016, 32, 4619–4623. Llevot, A.; Grau, E.; Carlotti, S.; Grelier, S.; Cramail, H. Macromol. R. Commun. 2016, 37, 9–28. Krishnaswamy, R. K.; Van Walsem, J.; Peoples, O. P.; Shabtai, Y.; Padwa, A. R. US Pat. Appl. 20150132512 A1 20150514, 2015. Kaneko, T.; Tran, H. T.; Matsusaki, M.; Akashi, M. Chem. Mater. 2006, 18, 6220–6226. Wei, L.; McDonald, A. G. Mater. 2016, 9, 1–23. Babu, R. P.; O’Connor, K.; Seeram, R. Prog. Biomater. 2013, 2, 1–16. Avantium Report. Renewable Chemicals into Bio-Based Materials: From Lignocellulose to PEF. http://biobasedperformancematerials.nl/ upload_mm/3/5/7/651bed82-390b-4435-a006-7909570de736_BPM %202017%20-%20Speaker%2006%20-%20Ed%20de%20Jong %20-%20Renewable%20chemicals%20into%20biobased%20materials%20-%20from%20lignocellulose%20to%20PEF.pdf (accessed May 19, 2018). Avantium Report. PEF, a 100% Bio-Based Polyester: Synthesis, Properties & Sustainability. http://euronanoforum2015.eu/wp-content/ uploads/2015/06/PlenaryII_PEF_a_100_bio-based_polyester_GertJanGruter_11062015_final.pdf (accessed May 19, 2018). Fukushima, K.; Yoshiharu, K. Polym. Int. 2006, 55, 626–642. Ikada, Y.; Jamshidi, K.; Tsuji, H.; Hyon, S. H. Macromolecules 1987, 20, 904–906. Duan, Y.; Liu, J.; Sato, H.; Zhang, J.; Tsuji, H.; Ozaki, Y.; Yan, S. Biomacromolecules 2006, 7, 2728–2735. Serizawa, T.; Yamashita, H.; Fujiwara, T.; Kimura, Y.; Akashi, M. Macromolecules 2001, 34, 1996–2001. Cicala, G.; Latteri, A.; Saccullo, G. J. Polym. Environ. 2017, 25, 750–758. Yonezawa, N.; Okamoto, A. Polym. J. 2009, 41, 899–928. Zachariades, A. E.; Economy, J. Polym. Eng. Sci. 1983, 23, 266–270. Saulnier, B.; Ponsart, S.; Coudane, J.; Garreau, H.; Vert, M. Macromol. Biosci. 2004, 4, 232–237. Nagarajan, V.; Singh, M.; Kane, H.; Khalili, M.; Bramucci, M. J. Polym. Environ. 2016, 14, 281–287. Imai, Y. High Perform. Polym. 1995, 7, 337–345. Kaneko, T.; Tran, H. T.; Shi, D. J.; Akashi, M. Nature Mater. 2006, 5, 966–970. Ricarda, N.; Anthony, J. M.; Cathie, M. Nature Biotechnol. 2004, 22, 746–754. Mathews, A. S.; Kim, I.; Ha, C. S. Macromol. Res. 2007, 15, 114–128. 215

27. Lu, Y.; Hu, Z.; Wang, Y.; Fang, Q. X. J. Appl. Polym. Sci. 2012, 125, 1371–1376. 28. Garrison, T. F.; Murawski, A.; Quirino, R. L. Polymers 2016, 8, 262. 29. Nakajima, H.; Dijkstra, P.; Loos, K. Polymers 2017, 9, 523. 30. Kawasaki, N.; Nakayama, A.; Yamano, N.; Takeda, S.; Kawata, Y.; Yamamoto, N.; Aiba, S. Polymer 2005, 46, 9987–9993. 31. Winnacker, M.; Rieger, B. Macromol. Rapid Commun. 2016, 37, 1391–1413. 32. Moran, C. S.; Barthelon, A. B.; Pearsall, A.; Mittal, V.; Dorgan, J. R. J. Appl. Polym. Sci. 2016, 133, 43626. 33. Schouwer, F. D.; Claes, L.; Claes, N.; Bals, S.; Degrèvec, J.; Vos, D. E. D. Green Chem. 2015, 17, 2263–2270. 34. Winnacker, M.; Vagin, S.; Auer, V.; Rieger, B. Macromol. Chem. Phys. 2014, 215, 1654–1660. 35. Jasinska, L.; Villani, M.; Wu, J.; van Es, D.; Klop, E.; Rastogi, S.; Koning, C. E. Macromolecules 2011, 44, 3458–3466. 36. Vanhaecht, B.; Rimez, B.; Willem, R.; Biesemans, M.; Koning, C. E. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, 1962–1971. 37. Koning, C. E.; Teuwen, L.; DeJong, R.; Janssen, G.; Coussens, B. High Perform. Polym. 1999, 11, 387–394. 38. Gaymans, R. J.; van Utteren, T. E. C.; van Berg, J. W. A.; Schuyer, J. J. Polym. Sci., Part B: Polym. Chem. 1977, 15, 537–545. 39. Steeman, P.; Nijenhuis, A. Polymer 2010, 51, 2700–2707. 40. Hasan, M. M.; Zhou, Y.; Mahfuz, H.; Jeelani, S. Mater. Sci. Eng. 2006, 429, 181–188. 41. Ali, M. A.; Tateyama, S.; Kaneko, T. Polym. Degrad. Stab. 2014, 109, 367–372. 42. Ali, M. A.; Kaneko, T. Microbial Applications; Springer-Verlag: Berlin Heidelberg, 2017; Vol. 2, pp 279−289. 43. Wang, Z.; Wei, T.; Xue, X.; He, M.; Xue, J.; Song, M.; Wu, S.; Kang, H.; Zhang, L.; Jia, Q. Polymer 2014, 55, 4846–4856. 44. Coiai, S.; Passaglia, E.; Pucci, A.; Ruggeri, G. Materials 2015, 8, 3377–3427. 45. Nikiforov, A. A.; Vol’fson, S. I.; Okhotina, N. A.; Rinberg, R.; Hartmann, T.; Kroll, L. Russ. Metallurgy 2017, 4, 279–282. 46. Zhao, M.; Yi, D.; Camino, G.; Frache, A.; Yang, R. RSC Adv. 2017, 7, 861–869. 47. Tateyama, S.; Msuo, S.; Suvannasara, P.; Oka, Y.; Miyazato, A.; Yasaki, K.; Teerawatananond, T.; Muangsin, N.; Zhou, S.; Kawasaki, Y.; Zhu, L.; Zhou, Z.; Takaya, N.; Kaneko, T. Macromolecules 2016, 49, 3336–3342. 48. Kawasaki, Y.; Nag, A.; Minakwa, H.; Masuo, S.; Kaneko, T. Appl. Microbiol. Biotechnol. 2018, 102, 631–639. 49. Toader, G.; Rusen, E.; Teodorescu, M.; Diacon, A.; Stanescu, O. P.; Rotariu, T.; Rotaiu, A. J. Appl. Polym. Sci. 2016, 133, 43967. 50. Sasaki, K.; Crich, D. Org. Lett. 2011, 13, 2256–2259. 51. Grujicic, M.; Pandurangan, B.; He, T.; Cheeseman, B. A.; Yen, C. F.; Randow, C. L. Mater. Sci. Eng. A 2010, 527, 7741–7751. 216

52. Holzworth, K.; Jia, Z.; Amirkhizi, A. V.; Qiao, J.; Nemat-Nasser, S. Polymer 2013, 54, 3079–3085. 53. Leventis, N.; Chidambareswarapattar, C.; Mohite, D. P.; Larimore, Z. J.; Lu, H.; Sotiriou-Leventis, C. J. Mater. Chem. 2011, 21, 11981–11986. 54. Aries, R. S. US Patent No. 2,945,841 (7-19-1960). 55. Reinecker, M.; Soprunyuk, V.; Fally, M.; Sanchez-Ferrer, A.; Schranz, W. Soft Matter 2014, 10, 5729–5738. 56. Qian, X.; Song, L.; Yu, B.; Yang, W.; Wang, B.; Hu, Y.; Yuen, R. K. K. Chem. Eng. J. 2014, 236, 233–241. 57. Grewal, M. S.; Taya, K.; Tateyama, S.; Kaneko, T. Macromol. Symp. 2017, 375, 1600194. 58. Shin, H.; Tateyama, S. J. Mater. Life Soc. 2014, 26, 25–31. 59. Tan, Y.; Liu, Y.; Chen, W.; Liu, Y.; Wang, Q.; Li, J.; Yu, H. ACS Sustainable Chem. Eng. 2016, 4, 3766–3772. 60. Jin, X.; Tateyama, S.; Kaneko, T. Polym. J. 2015, 47, 727–732. 61. Versteegen, R. M.; Kleppinger, R.; Sijbesma, R. P.; Meijer, E. W. Macromolecules 2006, 39, 772–783. 62. Tang, D.; Mulder, D. J.; Bart, A.; Noordover, J. Macromol. Rapid Commun. 2011, 32, 1379–1385. 63. Andrady, A. L. Plastics and Environment; Wiley-Intersciences, 2003; Vol. 1, pp 418−466. 64. Kitazume, T.; Tanaka, A.; Takaya, N.; Nakamura, A.; Matsuyama, S.; Suzuki, T.; Shoun, H. Eur. J. Biochem. 2002, 269, 2075–2082. 65. Kane, K. F.; Fiske, M. J. J. Bacteriol. 1985, 161, 963–966. 66. Tribe, E. D. Novel microorganism 4 method, U.S. Patent Number 4,681,852, 1987. 67. Suvannasara, P.; Tateyama, S.; Miyasato, A.; Matsumura, K.; Shimoda, T.; Ito, T.; Yamagata, Y.; Fujita, T.; Takaya, N.; Kaneko, T. Macromolecules 2014, 47, 1586–1593. 68. Kumar, A.; Tateyama, S.; Yasaki, K.; Ali, M. A.; Takaya, N.; Singh, R.; Kaneko, T. Polymer 2016, 83, 182–189. 69. Dwivedi, S.; Kaneko, T. Polymer 2018, 10, 368. 70. Huang, T. T.; Tsai, C. L.; Tateyama, S.; Kaneko, T.; Liou, G. S. Nanoscale 2016, 8, 12793–12802. 71. Dwivedi, S.; Sakamoto, S.; Kato, S.; Mitsumata, T.; Kaneko, T. RSC Adv. 2018, 8, 14009. 72. Kaneko, T.; Sakamoto, S.; Kobayashi, T.; Dwivedi, S. 255th ACS National Meeting & Exposition POLY-634, New Orleans, LA, United States, March 18–22, 2018. 73. Althues, H.; Henle, J.; Kaskel, S. Chem. Soc. Rev. 2007, 36, 1454–1465. 74. Shin, H.; Wang, S.; Tateyama, S.; Kaneko, D.; Kaneko, T. Ind. Eng. Chem. Res. 2016, 55, 8761–8766. 75. Zhang, S.; Li, Y.; Lin, D.; Wang, X.; Zhao, X.; Shao, Y.; Yang, S. Eur. Polym. J. 2005, 41, 1097–1107. 76. Chen, C. J.; Hu, Y. C.; Liou, G. S. Polym. Chem. 2013, 4, 4162–4171. 77. Tsai, C. L.; Chen, C. J.; Wang, P. H.; Lin, J. J.; Liou, G. S. Polym. Chem. 2013, 4, 4570–4573. 217

78. Dwivedi, S.; Kaneko, T. J. Appl. Polym. Sci. 2018 DOI:10.1002/ app.20180278. 79. Fang, Y.; Hester, J. G. D.; Su, W.; Chow, J. H.; Sitaraman, S. K.; Tentzeris, M. M. Sci. Rep. 2016, 6, 39909. 80. Harnett, M. E.; Alderman, J.; Wood, T. J. Colloid Surf. B: Biointerfaces 2007, 55, 90–97.

218

Bio-Based Phenolics and Composites

Chapter 15

Ferulic Acid- and Sinapic Acid-Based Bisphenols: Promising Renewable and Safer Alternatives to Bisphenol A for the Production of Bio-Based Polymers and Resins Louis Hollande and Florent Allais* AgroParisTech, Chaire Agro-Biotechnologies Industrielles (ABI), CEBB 3 rue des Rouges Terres, F-51110 Pomacle, France *E-mail: [email protected].

In the context of fossil resources depletion and ecological concerns, many bio-based aliphatic polymers and materials exhibiting smooth biodegradability have been developed. Nevertheless, the applications of these polymers are limited because of their poor thermomechanical properties and durability. To alleviate these drawbacks, an efficient method has been developed consisting of inserting aromatic subunits into the polymer chain. However many key commercial aromatics are derived from petrochemical feedstocks and are toxic. Much focus has been placed on bio-based aromatic compounds to replace monomers and polymers, such as bisphenol A, methylene diisocyanate, and phenolic resins. In this context, a great deal of effort was dedicated to the development of new aromatic chemical platforms which include ferulic and sinapic acids, two naturally occurring p-hydroxycinnamic acids. Obtained through chemoenzymatic synthetic processes, these bis- and trisphenol platforms have been used for the preparation of renewable aromatic epoxy resins and cyclocarbonates. The curing of the latter with di- and triamines resulted in the corresponding epoxy-amine resins and non-isocyanate polyurethane (NIPU) oligomers, respectively. The structure, thermomechanical properties, and chemical resistance of these novel bio-based materials derived from ferulic and sinapic acids were studied in order to evaluate their potential industrial applications.

© 2018 American Chemical Society

Introduction Epoxy resins were studied because of their remarkable mechanical properties, thermal and chemical resistance, low density, and adhesion. These properties offer a wide range of applications (e.g., coatings, adhesives, or composites). Epoxy resins constitute one of the most important fields of application in the thermosets market (1). The polyurethanes (PU) market has reached about 5% of the global polymer market, with an estimated global market of 14 Million tonnes in 2010 (2). The versatile, adaptable properties of PU materials find applications in various fields, such as thermoplastics, thermosets, or elastomers for foams, coatings, adhesives, and insulation (3, 4). Nevertheless, despite their very interesting properties, epoxy resins and PUs have several major drawbacks. As epoxy resins are derived from petrochemicals and the latter have become increasingly scarce and expensive, and because they are not recyclable due to their crosslinked structures, the interest in bio-based epoxy monomers has increased. Moreover, much concern has been recently raised about the toxicity of bisphenol A (BPA), the precursor of the diglycidyl ether of BPA (DGEBA), the most widespread epoxy monomer. Indeed, because its structure is similar to estrogen, BPA is recognized as an endocrine disruptor, contributing to reduced fertility and potential cancers (5, 6). Recently, the United States Department of Agriculture prohibited the use of BPA-based packaging for children’s products (7). Despite the excellent properties of DGEBA-based materials, their toxicity has increased the demand for BPA-free products and thus promoted research for safer and greener alternatives (8). Therefore, to restrict the use of DGEBA (9, 10), many works have focused on the valorization of renewable resources (e.g., terpenes (11, 12), linseed oil derivatives (13), sucrose soyate (14), rosin acid oligomers (15), and isosorbide (16–18)) versus epoxy resins. However, the major drawback of these aliphatic bioresources is the lack of rigidity compared to aromatic DGEBA, inducing lower mechanical properties. In the case of PUs, the isocyanates are toxic compounds, and their synthesis involves the use of the very harmful phosgene. Therefore, several isocyanates are now on the restricted list of annex XVII of REACH (Registration, Evaluation, Authorisation and Restriction of Chemicals) (19). The need for safer alternatives increased interest in NIPUs, particularly with respect to the ring-opening polymerization of polyfunctionnal cyclocarbonates by diamines (20–25). Cyclocarbonate monomers avoid both the use of isocyanate and the production of harmful by-products. In this alternative, poly(β-hydroxyurethanes) are obtained, and the pendant hydroxyl moieties, differing from classical PUs, offer a new range of properties and resulting NIPUs generally display an increased polarity (20), higher glass transition temperature (Tg) due to interchain hydrogen bonds (26), better thermal stability, lower solubility in organic media, and higher water uptake (27). In addition to the replacement of isocyanate, the substitution of petro-based monomers with renewable resources, including soybean oil (28), linseed oil (29), cyclic terpene (limonene) (30), and sebacic acid (31, 32), has also been investigated. Many bio-based NIPUs developed to date have been based on vegetable oils, with high functionality and varied structures of triglycerides, 222

leading to low Tg thermosets or thermoplastics with low . Higher mechanical properties could be obtained by using more rigid aromatic bio-based compounds. In view of the considerations above, naturally occurring phenolic compounds appear to be promising building blocks for the production of epoxy resins and NIPUs. The literature highlighted interest in the potential of phenolic moieties to compete with DGEBA. Recently, several epoxy-amine resins, based on phenolic extracts from green tea tannins (33), vanillin and vanillic acid (34, 35), quercetin (36), or phloroglucinol (37) (an aromatic triol contained in the bark of fruit trees), have been reported. In contrast to the numerous alternatives proposed for DGEBA-free resins, the design of new bio-based genuine bisphenols was scarcely described. A diphenolic ester obtained via the condensation of levulinic acid (bio-based) and phenol (petro-sourced) was studied as a BPA alternative for epoxy thermoset applicatons (38). Hernandez et al. described fully bio-based bisphenols obtained from guaiacaol and vanillyl alcohol and their application toward epoxy resins (39). Vanillin-based bisphenols have been described as a platform of high interest for the design of novel rigid bio-based polymers (40). A eugenol-based bisphenol was studied for high-performance thermosetting resins (41). While phenols have been extensively investigated for the production of epoxy resins, to the best of our knowledge, publications on biophenol-based NIPUs are very scarce. Creosol-based bisphenols were glycidylated, carbonated, and cured with various diamines (42). Vanillin was also studied as a biophenol precursor for cyclocarbonate synthesis but has not been engaged yet in the synthesis of NIPUs (43). In this context, an interesting alternative consisted of the valorization of naturally occurring p-hydroxycinnamic acids, and more particularly of ferulic and sinapic acids. Ferulic acid is naturally present at relatively high concentrations in the cell walls of several plants (e.g., rice, wheat, and corn). Many researchers have engaged in large-scale production of this acid using techniques such as extraction from agricultural wastes (44), enzymatic fractionation (45, 46), or more recently by fermentation (47). Sinapic acid, offering an extra methoxy group at the ortho position, can be extracted directly from biomasses such as canola cake or mustard bran. This chapter describes the chemoenzymatic syntheses of ferulic acid- and sinapic acid-based bis- and trisphenols as well as their corresponding epoxy resins and cyclocarbonates. The preparation of epoxy-amine resins and NIPU oligomers from the curing of the latter epoxy resins and cyclocarbonates, respectively, with di- and triamines will also be described. The thermomechanical properties and the degradability of the novel bio-based materials will be discussed.

Experimental Materials and Methods All reagents and enzymes (Candida antarctica lipase type B, Trametes versicolor laccase aka CAL-B) were purchased from Sigma-Aldrich; epichlorhydrin and decane diamine were purchased from Acros Organics; isophorone diamine was purchased from Chemical Industry Co.; furfurylamine, isosorbide, and diglycidyl ether of bisphenol A were purchased from Alfa Aesar. 223

All reagents were used as received. CO2 was purchased from Linde. Solvents were purchased from ThermoFisher Scientific, and dimethylformamide was dried on an mBraun SPS 800 system. Deuterochloroform (CDCl3) was purchased from Euriso-top. Evaporations were conducted under reduced pressure at a temperature below 40 °C for usual solvents and at 60 °C for dimethylformamide. Column chromatographies were carried out with an automated flash chromatography (PuriFlash 4100, Interchim) and prepacked INTERCHIM PF-30SI-HP (30 µm silica gel) columns using a gradient of cyclohexane and ethyl acetate for the elution. Carbonations were made in a 100 mL Paar autoclave equipped with magnetic stirring. FT-IR analyses were performed on a Cary 630 FT-IR spectrometer with ATR. UV analyses were performed on a Cary 60 UV-Vis instrument from Agilent Technologies by dissolving the samples in acetonitrile. Melting points were measured on a Mettler Toledo MP50 Melting Point System at 2 °C min-1. NMR analyses were recorded on a Bruker Fourier 300 instrument. 1H-NMR spectra of the samples were recorded in CDCl3 at 300 MHz; chemical shifts were reported in parts per million (CDCl3 residual signal at δ = 7.26 ppm). 13C-NMR spectra of the samples were recorded at 75 MHz (CDCl3 signal at δ = 77.16 ppm). HRMS were recorded by the PLANET platform at URCA (Université de Reims-Champagne Ardenne) on a Micromass GC-TOF spectrometer. Thermogravimetric analyses (TGA) were recorded on a Q500 analyzer from TA Instrument. About 10 mg of each sample was heated at 10 °C min-1 from 30 to 800 °C under nitrogen flow (60 mL min-1). Differential scanning calorimetry (DSC) thermograms were obtained using a TA Q20 instrument, under inert atmosphere (N2), with calibration obtained with indium, n-octadecane, and n-octane standards. For each sample, about 10 mg was weighed in a pan, which was then sealed and submitted to three heat/cool/heat cycles: heating from 30 to 250 °C at 10 °C min-1 and cooling from 250 to -50 °C at 20 °C min-1. Glass transition temperatures (Tg) were determined at the inflexion value in the heat capacity jump. High pressure size exclusion chromatography (HPLC-SEC) was performed at 70 °C on an Infinity 1260 system from Agilent Technologies with quadruple detection (IR, UV, MALS, viscosymetry) and two PL-gel 5-mm-mixed D columns (300 mm x 7.5 mm) in (1) DMF (flow rate 1 m .min-1) using polyethylene glycol/polyethylene oxide (PEG/PEO) calibration and toluene as internal standard for syringaresinol (SYR)-based NIPUs, and (2) THF (flow rate 1 mL min-1) using polystyrene calibration and toluene as an internal standard for ferulic acid-based NIPUs. Dynamic mechanical analyses (DMA) were performed on a DMA Q800 analyzer from TA Instrument. The DMA samples had a rectangular geometry (length: 20 mm; width: 10 mm; thickness: 1 mm). Uniaxial stretching of samples was performed while heating at a rate of 3 °C min-1 from 25 to 250 °C, keeping frequency at 1 Hz. Deformation was kept at 0.1% (amplitude of 6 μm) to stay in the linear viscoelastic region. The storage modulus (E′) and tan δ curves as a function of the temperature were recorded and analyzed using the TA Instruments Universal Analysis 2000 software. E′ is the elastic response and is related to the mechanical energy per cycle when the sample is deformed. E″ is the viscous response and is related to the dissipated energy per cycle upon deformation. The loss factor corresponds to the angle 224

between in-phase and out-of-phase components of the modulus in the cyclic motion and is defined as tan δ = E″/E′. The temperature Tα of the relaxation process, corresponding to the relaxation of the networks starting to coordinate large-scale motions, was determined as the temperature at the peak maximum of the tan δ curve. To study degradation, samples were freeze-dried for 20 h at -20 °C with an Alpha 1-2 LD Plus system from CHRIST©, supplied with a Vacuubrand RZ2.5 vacuum pump.

Enzymatic Preparation in Bulk of Bis-O-dihydroferuloyl Isosorbide Ferulic acid (250 g, 1.29 mol, 1 equiv) was dissolved in ethanol (900 mL, 250 g L-1) in the presence of a few drops of concentrated hydrochloric acid and heated at reflux for 2 days. The reaction mixture was then cooled to room temperature (RT) and put under argon before adding Pd/C 10 wt % (4.5 g, 5 wt % ). The mixture was stirred for 18 h at RT under H2 flow, and the reaction was monitored by NMR until the complete conversion of the starting material. The solution was filtered on Celite, and the solvent was removed under vacuum. The resulting crude product was filtrated over a pad of Celite and silica gel to afford ethyl dihydroferulate as a white powder (277 g, 96%, melting point at 42 °C). Isosorbide (1 equiv) and ethyl dihydroferulate (3 equiv) were melted and magnetically stirred at 75 °C before adding CAL-B (10% by weight relative to the total weight of isosorbide and ethyl dihydroferulate). The reaction mixture was kept under reduced pressure for three days. The reaction mixture was then dissolved in ethyl acetate (or acetone) and filtered to remove CAL-B beads. The solvent was evaporated under vacuum, and the crude product was purified by flash chromatography on silica gel eluted with a ratio of cyclohexane:ethyl acetate 70:30 to collect unreacted ethyl dihydroferulate, then 45:55 to gather bis-O-dihydroferuloyl isosorbide (85%, IDF).

Enzymatic Preparation of Syringaresinol Concentrated sulphuric acid (0.3 mL) was added dropwise to a solution of sinapic acid (1 g, 4.5 mmol, 1 equiv) in 15 mL of ethanol (0.3 M) at RT. The mixture was stirred with a magnetic stirrer under reflux at 80 °C for 23 h. The reaction mixture was allowed to cool, and the solvent was evaporated under reduced pressure. The product was then dissolved in 400 mL of ethyl acetate and washed with a saturated solution of NaHCO3 (2 × 60 mL) and brine (50 mL). The organic phase was dried over anhydrous magnesium sulphate and filtered, and the solvent was evaporated under reduced pressure. Ninety-three percent of crude ethyl sinapate was obtained as a white powder and used without further purification. Diisobutylaluminium hydride (1M in dichloromethane, 66 mL, 66 mmol, 3.3 equiv) of were added dropwise at 0 °C to a solution of ethyl sinapate (5 g, 225

20 mmol, 1 equiv) in 100 mL of dichloromethane (0.2 M) under nitrogen. The mixture was then stirred with a magnetic stirrer at RT for 3 h. A saturated solution of sodium potassium tartrate was added dropwise to the mixture at 0 °C until precipitation occurred. Dichloromethane (100 mL) and water (100 mL) were added for dissolution. The reaction media was acidified with a solution of HCl (2 M) until pH 4. The mixture was then stirred vigorously for 12 h. The organic phase was separated, washed, dried over anhydrous magnesium sulphate, and filtered, and the solvent was evaporated under reduced pressure. Crude sinapyl alcohol was then purified on silica gel using cyclohexane:ethyl acetate as an eluant (50:50). Seventy percent (70%) of pure product (viscous oil) was obtained. Sinapyl alcohol (4 g, 19 mmol, 1 equiv) was dissolved in acetonitrile (80 mL) in a triple-neck round bottom flask equipped with a cooling system, and a citrate/phosphate buffer (320 mL) at pH 5 was added (C = 0.05 M). A solution containing 35.2 mg of laccase from Trametes versicolor (0.1 U/mg of substrate) in 50 mL of citrate/phosphate buffer was added dropwise to the sinapyl alcohol solution at a rate of 9 mL/h with a syringe pump. The reaction mixture was stirred vigorously with a magnetic stirrer in the presence of O2 (air) at 50 °C for 470 min in the dark. At the end of the reaction, the product was extracted with 2 × 150 mL of dichloromethane and 2 × 150 mL of ethyl acetate. The organic phases were combined, dried over anhydrous magnesium sulfate, and filtered, and the solvent was evaporated under reduced pressure. Then 3.7 g of pure SYR was obtained as a white powder (93% yield), and no further purification was needed. Higher purity may be obtained through recrystallization (48) but did not enhance SYR’s reactivity in further transformations (49).

General Procedure for the Synthesis of Diglycidyl Compounds SYR (3 g, 11.9 mmol, 1 equiv) was dissolved in epichlorohydrin (11.3 mL, 239 mmol, 20 equiv). Triethyl benzyl chloride (TEBAC) (327 mg, 2.39 mmol, 0.2 equiv) was added, and the suspension was stirred for 2 h at 80 °C. The reaction medium was cooled down, and NaOH (5 M) (11.4 mL, 95.6 mmol, 8 equiv) was added. The biphasic solution was vigorously stirred for 4 h and then extracted with ethyl acetate (3 × 100 mL). The organic layers were washed with brine (80 mL), dried over anhydrous magnesium sulfate, filtered, and concentrated. The crude product was purified on silica gel using cyclohexane:ethyl acetate (40:60) as an eluent to provide diglycidyl ether of SYR (SYR2EP) as a white solid (3.3 g, 88%).

Epoxy-Amine Thermosets Formulations As one primary amine of the curing agent can react with two epoxy moieties, the resins were formulated with a 2:1 epoxy-amine ratio. SYR2EP (800 mg, 2 equiv) was molten at 120 °C, then diamine (1 equiv) was quickly added, and the mixture was vigorously stirred for 1 min at 120 °C. In the case of the solid diamine 226

DA10, an additional stirring time at 120 °C was necessary to get a complete dissolution of the amine in the epoxy precursor. The liquid and homogeneous mixture was then poured into a silicone mold. Finally, the formulation was cured in an oven, following this thermal curing program: 2 h at 120 °C, 18 h at 160 °C, 1 h at 180 °C. For DGEBA and IDF precursors, the temperature curing program was as follows: 2 h at 80 °C, 18 h at 110 °C, 1 h at 140 °C. The different compositions of epoxy-amine resins are depicted in Table 1. The theoretical crosslink density mentioned herein is defined as the molar amount of crosslinks (i.e., the amount of nitrogen atoms from the curing agent) for a formulation of 100 g of resin.

General Procedure for the Synthesis of Cyclocarbonates (48, 49) A solution of epoxy precursor SYR2EP (5.05 g, 9.52 mmol) and lithium bromide (41 mg, 0.476 mmol, 0.05 equiv) in DMF (20 mL) were sealed in a 100 mL autoclave. The system was stirred at 80 °C under 20 bar of carbon dioxide for 24 h. The solvent was removed by distillation, and the crude product was solubilized in ethyl acetate (150 mL) and washed with water (3 × 60 mL) to remove DMF traces. The organic phase was dried over anhydrous magnesium sulfate, filtered, and concentrated to provide the SYR-derived biscyclocarbonate (SYR2Cy) as a white amorphous solid (4.75 g, 81% crude yield).

General Procedure for the Synthesis of NIPUs The cyclocarbonate precursor was melted around 80 °C, and the adequate amount of diamine (DA10, DIFFA, and IPDA) was added. A ratio of 1:1 was chosen, considering that one five-membered ring cyclocarbonate reacts only once with a primary amine. The system was then manually homogenized, transferred to a rubbery mold, and cured. In order to avoid freezing of the system and to ensure optimal conversion, temperature programs were established: -

-

Ferulic acid-based NIPUs: Materials were first cured at 80 °C (a temperature inferior to the temperature of crosslinking as determined by DSC) and then at 100 °C (prior to the temperature of crosslinking as determined by DSC). For uniformity in the study, the same curing process was applied to all NIPU materials: 5 h at 80 °C followed by 10 h at 100 °C. Sinapic acid-based NIPUs: materials were first cured from 80 to 160 °C (at 10 °C•h-1) and then maintained at 160 °C for 18 h (prior to the exothermic reaction peak observed in the DSC scan).

The different compositions are reported in references (50) and (51) for ferulic acid- and sinapic acid-based NIPUs, respectively. 227

Table 1. Theoretical crosslink densities for ferulic acid- and sinapic acid-based epoxy precursors Epoxy precursor (wt %)

Curing agent (wt %)

Theoretical crosslink density (mol of crosslink x10²/100 g)

DGEBA (79.8)

DA10 (20.2)

23.5a

DGEBA (74.4)

DIFFA (25.6)

21.9a

DGEBA (80.0)

IPDA (20.0)

23.5a

IDF2EP (88.3)

DA10 (11.7)

13.6a

IDF2EP (84.7)

DIFFA (15.3)

13.1a

IDF2EP (88.4)

IPDA (11.6)

13.7a

BDP2EP (87.7)

DA10 (12.3)

14.3a

BDP2EP (84.0)

DIFFA (16.0)

13.7a

BDP2EP (87.8)

IPDA (12.2)

14.4a

BDF(amine)2EP (86.9)

DA10 (13.1)

15.4a

BDF(amine)2EP (83.0)

DIFFA (17.0)

14.5a

BDF(amine)2EP (87.0)

IPDA (13.0)

15.3a

GTF3EPc (87.0)

DA10 (13.0)

(15.1a + 11.0b) = 26.1

GTF3EP (83.1)

DIFFA (16.9)

(14.4a + 10.5b) = 24.9

GTF3EP (87.1)

IPDA (12.9)

(15.2a + 11.0b) = 26.2

SYR2EP (86.0)

DA10 (14.0)

16.2a

SYR2EP (81.9)

DIFFA (18.0)

15.4a

SYR2EP (86.2)

IPDA (13.8)

16.2a

a Calculated by considering only the amount of nitrogen atoms. b Calculated by considering the amount of GTF3EP units. c GTF3EP: triepoxide deriving from GTF.

Degradation Study Cured IDF-, SYR-, and DGEBA-based epoxy-amine resin samples with dimensions of about 7 × 7 × 1 mm with weights ranging from 75 to 95 mg were placed in a 10 mL sealed tube. A solution of NaOH (3 M) or HCl (3 M) was added. The vials were heated at 60 °C for 40 min for equilibration. Incubations were then continued for 5, 24, 30, 48, and 96 h. The residual solid was then washed with deionized water, freeze-dried, and then weighed to determine the mass loss. For each data point, the degradation was reproduced in triplicate to determine a mean mass loss.

228

Estrogen Binding Affinity Testing The agonistic potential of bisphenols was analyzed following a literature method (47) in which ERα transcriptional activity is monitored by using ERα reporter cells (HELN ERα cell line). Activity was measured in relative light units (RLU), and 100% activity was assigned to the RLU value obtained with 10 nM estradiol. The vehicle (Dimethyl sulfoxide, aka DMSO) was tested as a control without any compound.

Results and Discussions Synthesis of the Bio-Based Epoxy Precursors Synthesis of the Ferulic Acid- and Sinapic Acid-Based Bis- and Trisphenols The chemical esterification of p-hydroxycinnamic acids with alcohol is unselective, and side reactions generally lead to unwanted product that needs to be removed by purification steps, generating waste to dispose of. In opposition to the chemical-catalyzed reactions, the lipase-mediated enzymatic synthesis offers some advantages, such as a higher selectivity (fewer co-products, easier purification) and milder reaction conditions (52). Thus, enzymatic catalysis represents an environmentally friendly process for synthesizing esters. However, the enzymatic esterification of p-hydroxycinnamic acid with alcohols suffers from some drawbacks such as slow kinetic reaction and/or low conversion. It is possible to improve the efficiency of this reaction by reducing the conjugated unsaturations via a palladium-catalyzed hydrogenation (53–56) or by using the corresponding alkyl p-hydroxycinnamate (56–58). To obtain the needed polyfunctional phenolic architecture, ferulic acid is thus first transformed into ethyl ferulate through a two-step, one-pot process involving a Fisher esterification followed by a palladium-catalyzed hydrogenation (Scheme 1). Ethyl dihydroferulate is then (di/tri)merized by transesterification with bio-based di/triol following the sustainable solvent-free synthesis pathway developed by our team (59) using reagents in bulk at 75 °C under high vacuum and catalyzed by 10 wt % of immobilized Candida antarctica Lipase B (CAL-B) (Scheme 1). We chose three different bio-based diols with various structures: the aliphatic 1,4-butanediol, the rigid cycloaliphatic isosorbide, and the trifunctional glycerol with the aim to tailor the final thermoset properties by playing with the nature of the linker between the phenolic units. The syntheses using 1,4-butanediol, isosorbide, and glycerol lead to the butanediol diferulate (BDF), the isosorbide diferulate (IDF), and the glycerol triferulate (GTF), respectively. In addition, apart from transesterification, CAL-B is also efficient for catalyzing a wide range of reactions, including amidation (60) and transamidation (60, 61). Thus, in order to introduce an amide function in the bisphenol structure—expecting higher network properties due to more hydrogen bonds from the amide function—an amide-containing ferulic acid derivative was synthesized from 1,4-butanediamine and dihydroferulate, leading to the butandiamidediferulate [BDF(amide)] (63). 229

Scheme 1. Bis- and trisphenols syntheses from ferulic and sinapic acids. Adapted with permission from ref. (57). Copyright 2013 Royal Society of Chemistry.

From sinapic acid, we decided to focus our efforts toward the synthesis of one of its naturally occurring dimers, SYR. Indeed, this dimer possesses not only two bisphenol moieties but also a very rigid bisfuranic linker with ether bonds that are more resistant to acids/bases than the ester/amide bonds present in ferulic acidbased bis- and trisphenols IDF, BDF, BDF(amide), and GTF. SYR was prepared from sinapic acid via a three-step chemo-enzymatic process (Scheme 1). Sinapic acid (3) is first converted into the corresponding ester (4) through a classical Fisher esterification, 4 is then reduced into the corresponding allylic alcohol (5), and the latter undergoes a highly regioselective laccase-mediated oxidative dimerization to provide SYR (Scheme 1) (64). We were glad to observe that the purity of the resulting crude product was sufficient to engage it in the functionalization step without further purification.

Endocrine Binding Affinity Test According to the recognized toxicity of BPA due to its ability to disrupt endocrine activity, estrogenic activity evaluation was first conducted to ensure the elaboration of viable alternatives. Bisphenol derivatives (SYR, IDF, BDF, and BPA) and their bisepoxy derivatives (SYR2EP, IDF2EP, and DGEBA) were evaluated. BPA is known to bind to and activate estrogen receptor ERα, as a 230

partial agonist of ERα. ERα is a member of the nuclear hormone receptor family, whose activity is regulated by the steroid and estrogen sex hormone 17β-estradiol (E2). For this test, the estrogenic activities were measured with ERα and benchmarked against E2 (Figure 1). The sex hormone E2 shows a 100% activity at about 5 × 10-10 M. For BPA, estrogenic activity is observed for concentrations higher than 10-7 M, and about 60% of activity is observed at 10-5 M. In contrast, all ferulic acid- and sinapic acid-based macrobisphenols, IDF, BDF, and SYR, exhibited a low activity, comparable to the control (~10%), even with increased concentration. The lack of estrogenic disruption for these compounds is coherent with their structure: all bisphenols display a longer spacer between the two phenolic moieties in comparison with BPA, this distance having a key role in the disruption mechanism (41). The bisepoxy derivatives SYR2EP, BDF2EP, IDF2EP, and DGEBA, showed no endocrine disruption, in accordance with a previous study showing no endocrine disruption for DGEBA (65), and thus confirming the importance of the free phenols for the toxicity. Based on these results demonstrating ferulic acid- and sinapic acid-based bisphenols innocuousness, SYR, IDF, and BDF can be considered genuine safer BPA replacements (66).

Figure 1. Estrogenic activity of the bisphenols and their corresponding epoxides at 10-5 M and E2 at 10-8 M.

Epoxies and Epoxy-Amine Resins The functionalization of bisphenols toward the corresponding epoxy precursors was attempted through glycidylation using epichlorohydrin. It is noteworthy to mention that, even if it can now be bio-based from glycerol through 231

the Epicerol® process from Solvay, epichlorohydrin is a rather toxic chemical that needs to be handled with caution. This method consists of forming a phenolate that is able to attack each one of the three carbons of epichlorohydrin, two of them leading to the epoxy ring opening (Scheme 2). However, Routes 1 and 3 are predominant. Thus, at the end of the first step, a second one is necessary to close the open forms from attacks 2 and 3 by a basic treatment. The crude mixture is therefore treated with a sodium hydroxide solution to form the oxirane rings from chlorinated intermediates. For the glycidylation step of ferulic acid-based bis- and trisphenols, in order to avoid the partial hydrolysis/transesterification/transamidation of the internal ester moieties (65) and that of the diamide, the processes that use NaOH at high temperature and long reaction time were ruled out (67, 68). Thus, another synthetic route for the glycidylation of phenolic compounds involving a catalytic amount of triethylbenzyl ammonium chloride (TEBAC) in the presence of a large excess of epichlorohydrin (35, 37, 69) was used (Scheme 1). The BDFamide that presents a lower solubility in epichlorhydrin has to be more diluted and required more equivalent of epichlorohydrin. Diglycidylated BDF, diglycidylated IDF, triglycidylated GTF, and diglycidylated BDF(amide) are named BDF2EP, IDF2EP, GTF3EP, and BDF(amide)2EP, respectively. The four epoxy precursors are characterized by 1H-NMR, which allow for the determination of the epoxy functionality of each epoxy precursor by integrating the signal at 3.30 ppm corresponding to the proton on the tertiary carbon of the oxirane ring. We obtained epoxy functionalities for BDF2EP, IDF2EP, GTF3EP, and BDF(amide)2EP of 1.83, 1.90, 2.80, and 1.95, respectively.

Scheme 2. Glycidylation mechanism by SN2 (Route 1) or epoxy ring opening followed by a basic treatment (Route 3) and oxetane formation (Route 2). Reproduced with permission from ref. (62). Copyright 2017 Elsevier.

As there is no ester/amide bond in SYR, its glycidylation was first tested following a classical procedure (excess of epichlorohydrin under alkaline conditions) (67, 68) but led to less than 2 epoxy functions per SYR moiety. The 232

use of a catalytic amount of TEBAC (35, 37, 69) allowed access to SYR2EP as a white solid with an 88% yield and with epoxy functionality of 2.0.

Thermosets Synthesis: Formulation and Curing Amines are common hardeners for epoxy resins and can react in a two-step mechanism to form a crosslinked network (Scheme 2). According to this mechanism, we supposed that the secondary amines formed would react contrary to less reactive hydroxyls, formed upon epoxy ring opening. Consequently the theoretical epoxy-amine ratio is 2:1, and this stoichiometric ratio was chosen for all formulations in order to obtain high crosslinking rates. Two bio-based amines (DA10 and DIFFA) and one extensively used fossil-based amine (IPDA), presenting different rigidities, were chosen as hardeners. To ensure that ferulic acid- and sinapic acid-based epoxy precursors synthesized present a good reactivity toward aliphatic diamines, BDF2EP/IPDA (1/1) and SY2EP/IPDA (1/1) were analyzed by DSC (62, 64). The results showed an exothermal peak centered around 100 to 110 °C corresponding to the epoxy-amine reaction for BDF2EP/IPDA and SYR2EP/IPDA. These data proved a reactivity of SYR2EP and BDF2EP comparable to DGEBA (exothermal peak centered around 107 °C). The delayed onset of cure with SYR2EP is more likely due to its higher melting point. For the formulations, BDF2EP, IDF2EP, and SYR2EP had to be molten and rapidly mixed with the corresponding amine hardeners. All resins were rapidly blended in a molten state at 120 °C to warrant formulation homogeneity (in the case of solid DA10) and prevent bubbles. In order to ensure chain mobility and to obtain optimal crosslinking content, curing temperature programs were defined to slow down the crosslinking reaction: -

-

BDF2EP and IDF2EP: The formulations will be cured at T < TDSC crosslinking peak then at T > TDSC crosslinking peak. The curing program used is thus 2 h at 80 °C, then 10 h at 110 °C, and finally 1 h at 140 °C. It is noteworthy to mention that, under these curing conditions set for the BDF2EP/IPDA formulation, each monomer may lead to different extents of cure. Therefore, any differences in polymer properties, although they may be a result of different extents of cure, are ultimately due to the structural differences of the monomers. SYR2EP: The formulations were cured first at Tfus < T ≈ TDSC crosslinking peak (2 h at 120 °C) then at T > TDSC crosslinking peak (16 h at 160 °C, and then 1 h at 180 °C).

DSC thermograms for all the formulations do not exhibit transitions that would correspond to further curing reactions. These results indicate that the degree of thermoset polymer cure is high and that further curing reactions that might occur due to remaining oxirane moieties are low since they were below the detection limit of DSC experiments. This processing guaranteed the homogeneity of materials. 233

One can observe that the thermosets formulated with the bio-based epoxy precursors possess good optical properties; they are translucent like the DGEBA-based corresponding references. Nevertheless, the resins cured with DIFFA displayed an intense coloration and low translucence, as the starting amine used is dark brown (Figure 2).

Figure 2. Photographs of some thermosets, IDF2EP- and BDF2EP-based epoxy-amines (top), and SYR2EP-based epoxy-amines (bottom). Adapted with permission from ref. (62), Copyright 2017 Elsevier; and ref. (64), Copyright 2017 Wiley.

Figure 3. Td5% of the epoxy-amine resins under nitrogen (10 °C min-1).

234

Table 2. Characterization of ferulic acid- and sinapic acid-based epoxy-amine resins

235

Thermoset

Td5% (°C)a

wt %chara

Tg (°C)b

Tα (°C)b

E′glassy (GPa) at 0 °Cc

E′glassy (GPa) at 50 °Cc

E′elastic (MPa) at 200 °Cc

DGEBA-IPDA

331

10.2

150

174

1.7

X

21

DGEBA-DA10

337

5.2

98

97

1.0

X

23

DGEBA-DIFFA

301

14.8

92

110

2.1

X

21

BDF2EP-IPDA

314

10.3

51

61

2.4

X

11

BDF2EP-DA10

314

9.5

33

33

1.0

X

11

BDF2EP-DIFFA

306

15.6

32

41

1.8

X

8

BDF(amide)2EP-IPDA

282

18.2

75

X

X

X

X

BDF(amide)2EP-DA10

297

18.2

69

X

X

X

X

BDF(amide)2EP-DIFFA

271

23.7

74

X

X

X

X

IDF2EP-IPDA

295

12.7

85

99

2.0

X

9

IDF2EP-DA10

292

11.8

54

68

1.5

X

13

IDF2EP-DIFFA

304

18.9

63

85

2.0

X

12

GTF3EP-IPDA

294

16.4

73

92

1.7

X

16

GTF3EP-DA10

293

14.0

54

70

2.4

X

24

GTF3EP-DIFFA

284

24.7

58

67

1.7

X

8

SYR2EP-IPDA

298

24.0

126

157

X

0.8

12 Continued on next page.

Table 2. (Continued). Characterization of ferulic acid- and sinapic acid-based epoxy-amine resins

a

Thermoset

Td5% (°C)a

wt %chara

Tg (°C)b

Tα (°C)b

E′glassy (GPa) at 0 °Cc

E′glassy (GPa) at 50 °Cc

E′elastic (MPa) at 200 °Cc

SYR2EP-DA10

303

23.1

73

102

X

0.9

14

SYR2EP-DIFFA

288

29.4

102

108

X

1.1

10

Values determined by thermogravimetric analyses (TGA) under nitrogen at 10 °C values determined by DMA (frequency 1 Hz, amplitude 6 μm, 3 °C min-1).

min-1;

b

values determined by DSC under nitrogen at 10 °C min-1;

c

236

Characterization of the Ferulic Acid- and Sinapic Acid-Based Epoxy Thermosets Thermal Characterizations Thermogravimetric Analysis Thermal stability of cured thermosets was studied by thermogravimetric analysis (TGA) under nitrogen flow (Figure 3). Td5% was defined as the temperature at which the thermoset lost 5 wt % of its initial mass; wt %char corresponds to the relative amount of stable residue at a high temperature (700 °C). Table 2 sums up the values of Td5% and wt %char of the thermosets prepared. Generally, it is observed that ferulic acid- and sinapic acid-based thermosets showed slightly lower thermal stabilities than the corresponding DGEBA-containing thermosets, except for the DIFFA-containing thermoset series. It may be due to the thermolysis of the ester and ether functions of the bio-based epoxy precursors and/or the presence of electro-donating methoxy substituents on the phenol as previously discussed (70). Indeed, for the IPDA and DA10 series, the Td5% of the BDF2EP-containing thermoset is around 15 °C lower than the DGEBA-containing one, and IDF2EP-, GTF3EP-, and SYR2EP-containing thermosets present lower thermal stabilities than the BDF2EP (Table 2). The Td5% of the BDF2EP-IPDA and BDF2EP-DA10 thermosets are around 314 °C, while the IDF2EP-(IPDA/DA10), GTF3EP-(IPDA/DA10), and SYR2EP-(IPDA/DA10) ones range from 293 to 303 °C. This result may be correlated with the structure of the linker between the two phenol moieties (64, 66). In the DIFFA series, TGA shows lower gaps between the thermal stabilities of all the thermosets. This could mean that the thermal degradation of this network series is initiated by the rupture of the curing agent segment, which is the weak point of the DIFFA-containing thermoset structures. By comparing BDF2EP- and BDF(amide)2EP-containing thermosets, we observed that the amide-containing thermosets present thermal stabilities 15 to 30 °C lower than the corresponding ester-containing ones. This result was unexpected because the amide bond usually presents a higher stability than the ester one (71); however, in some cases, this difference is not systematically observed (72). Therefore, the lower thermal stabilities of amide-containing thermosets may be due to an incomplete curing of the network caused by a partial evaporation of the curing agent during the solvent casting operation and the curing under a 200 °C air flow. As for the high-temperature (700 °C) stable char content (Table 2) for the four thermoset series, it is noteworthy that ferulic acid-based thermosets, despite their aliphatic structure, lead to equivalent or superior char content compared to their DGEBA-containing equivalent. Moreover, sinapic acid-based materials are the ones that have the highest char content, probably because of the very rigid bisfuranic linker. The char content is a parameter of importance for the thermal behavior of epoxy thermosets, especially in fire retardation applications. For example, in the IPDA-series, the wt %char increased from 10.2 for DGEBA, to 12.7 wt % for IDF2EP, to 16.4 wt % for wt % GTF3EP, and to 24.0 wt % for SYR2EP. Thermograms also show that ferulic acid- and sinapic acid-based 237

thermosets produce different char contents that can be classified according to the linker structure. Indeed, in the three series, one can observe that wt %char BDF2EP < wt %char IDF2EP < wt %char GTF3EP < wt %char BDF(amine)2EP < wt %char SYR2EP.

Differential Scanning Calorimetry

Figure 4. Tg of the epoxy-amine resins.

Physical properties of thermosets can be assessed by measuring their glass transition temperature (Tg) by DSC. Since applications of epoxy resins usually require high thermal and mechanical properties, high glass transition temperatures (Tg) are targeted. DSC analyses were thus performed on all the crosslinked thermosets. Tg values obtained from the DSC curves are presented in Table 2 and Figure 4. Compared to their DGEBA equivalents, all ferulic acid- and sinapic acid-based thermosets lead to lower Tg values. Their curing with hardeners presenting various rigidities led to a wide range of Tg (from 32 to 126 °C). The most flexible aliphatic amines DA10 generally led to the lowest Tg. The aromatic rings of DIFFA conferred higher rigidity and induced higher Tg (102 °C). Finally, the cyclic configuration of IPDA reduces the mobility of the curing agent, and the resulting more rigid network showed the highest Tg (126 °C). One can therefore conclude that Tg IPDA > Tg DIFFA > Tg DA10. When comparing the curing of the six resins (BDF2EP, BDF(amide)2EP, IDF2EP, GTF3EP, SYR2EP, and DGEBA) with IPDA, the following trend is observed for the Tg: DGEBA (150 °C) > SYR2EP (126 °C) > IDF2EP (85 °C) > BDF(amide)2EP (75 °C) > GTF3EP (73 °C) > BDF2EP (51 °C). Different parameters might explain the differences observed between the five epoxies. For instance, the molar mass of ferulic acid238

and sinapic acid-based epoxies being higher to that of DGEBA, the final curing agent concentration and thus the theoretical crosslinking density are reduced. Secondly, the flexibility of the linker and the distance between the phenol moieties also impact the Tg. Indeed, BDF2EP leads to the lower Tg values due to its aliphatic structure, which allows an important mobility of the segment between the two ferulic units. With GTF3EP, the Tg value increases due to the higher crosslink concentration caused by the epoxy functionality of GTF3EP of 2.8. One can also observe that, for the three amines series, BDF(amide)2EP-containing thermosets lead to higher Tg than corresponding BDF2EP-containing thermosets. For example, Tg is increased from 51 °C for BDF2EP-IPDA thermoset to 75 °C for BDF(amide)2EP-IPDA. The same trend is observed for the other DA10- and DIFFA-containing thermosets. The higher Tg observed for the amide-containing thermoset is most probably related to the hydrogen bonds formed in the network by the amide functions as they increase interchain interactions and greatly reduce the mobility of the network. Thus, the energy needed to break these weak interactions is higher than in the case of BDF2EP. The incorporation of amide functions in the precursor epoxy therefore allows conferring better properties than when using a trifunctional or cyclic linker. Finally, while both IDF2EP and SYR2EP present a rigid bicyclic core and two aromatic rings, the aliphatic segment of IDF2EP provides higher flexibility; whereas in the case of SYR2EP, the rigid fused rings are directly linked to the aromatic rings, resulting in higher rigidity. Interestingly, with a Tg of 126 °C, SYR2EP-IPDA resin comes closer to the commercial DGEBA-IPDA reference (Tg = 150 °C) with nearly competitive mechanical properties. Thermomechanical Characterization of Ferulic Acid- and Sinapic Acid-Based Thermosets Mechanical properties of the cured resins were evaluated by DMA. DMA temperature scans gave values of the glassy modulus at 50 °C (E′glassy), rubbery storage modulus at 200 °C (E′rubbery), and alpha transition temperatures (Tα) related to glass transition, and determined from the peak of loss modulus versus temperature (Table 2). Because of the difficulty of preparing the samples linked to the high fusion temperature of the amide-containing epoxy precursor, BDF(amide)2EP-containing thermosets were not analyzed by DMA. According to the Tg determined by DSC, the mechanical transition temperatures (Tα) of the ferulic acid- and sinapic acid-based thermosets are lower than their DGEBA-based counterparts. For the IPDA-containing series, Tα is reduced from 174 °C for the DGEBA-IPDA thermoset to 157, 61, 99, and 92 °C for SYR2EP-IPDA, BDF2EP-IPDA, IDF2EP-IPDA, and GTF3EP-IPDA, respectively. The same trend is observed in the two other series, except in the DA10-containing one, where IDF2EP-DA10 and GTF3EP-DA10 lead to equivalent Tα. As discussed in the previous section (DSC analyses), the chemical structure of the linker strongly influences the mobility of the network and its mechanical transition. Among the ferulic acid-based thermosets, IDF2EP leads to the higher Tα due to its bicyclic linker, making it more rigid than the other ferulic acid-based epoxy precursors. Concerning GTF3EP-containing thermosets, the 239

glycerol linker provides intermediate Tα. The lower Tα of each series are obtained with the BDF2EP-containing thermosets, whose aliphatic linker structure confers an important mobility to the molecule and reduces the energy needed to coordinate large-scale motions. The lower Tα obtained with the ferulic acid- and sinapic acid-based thermosets are partially due to the lower theoretical crosslink caused by the higher molar mass of the bio-based epoxy precursors compared to DGEBA (Table 1). This phenomenon is also highlighted by the lower elastic modulus values observed in the rubbery plateau for the bio-based thermosets. Indeed, the elastic modulus value at the rubbery state increases with the crosslink rate of the thermoset network. Concerning the elastic modulus values in the glassy state (E′(0 °C) for ferulic acid-based epoxies and E′(50 °C) for SYR2EP), it is interesting to note that the four bio-based epoxy precursors lead to equivalent or higher values than the corresponding DGEBA-containing thermoset (Table 2), proving the good mechanical properties and the viability of the ferulic acid-based thermosets at the glassy state.

Chemical Degradation Study of IDF2EP-IPDA- and SYR2EP-IPDA-Cured Resins Besides thermomechanical properties, chemical resistance is one of the most valuable features of epoxy-amine resins. To evaluate the degradability of our resins toward acid and basic conditions, a chemical degradation study was performed on IPDA-cured resins based on SYR2EP, IDF2EP, and DGEBA, as the two first resins are those that come close to DGEBA in terms of thermomechanical properties. Samples were immersed in alkali (NaOH 3 M) or acidic (HCl 3 M) solutions at 60 °C. Mass losses were determined as a function of the immersion time (Figure 5). As described previously (65), IDF2P-based resin showed a rapid decrease in mass upon immersion due to the readily hydrolyzable esters moieties. The hydrolysis of the ester bonds of IDF2EP during the degradation allows a potential chemical recycling and thus appears as an attractive end-of-life option. The outcome proved totally different with SYR2EP as its IPDA resin showed hydrolysis resistance comparable to that of DGEBA-IPDA, with no weight loss. Synthesis of the Bio-Based Cyclocarbonates The addition of carbon dioxide to oxirane rings constitutes an easy and green access to cyclocarbonates (73, 74). Indeed, the use of carbon dioxide, an abundant inexpensive renewable resource, is very appealing in this method. Nevertheless, the use of catalysts is necessary to achieve good conversions (e.g., metal salts (73, 75) or complexes (76, 77), silica-supported amines (76)). Recently, an efficient synthesis of cyclic carbonates under mild conditions in bulk under atmospheric pressure of CO2 was described, using a bifunctional quaternary phosphonium iodide catalyst (79). The lithium bromide-catalyzed addition of carbon dioxide onto oxirane rings seems to be the simplest way to access five-membered cyclocarbonates. To determine structure–NIPU thermomechanical properties relationships, the ferulic acid- and sinapic acid-based epoxies were used as raw 240

materials and carbonated using a catalytic amount of lithium bromide (5 mol %/oxirane) under high carbon dioxide pressure (20 bar) at 80 °C (Scheme 3). The conversion of the oxirane rings into cyclocarbonate rings is quantitative after 24 h (monitoring by 1H- or 13C-NMR spectroscopy). The final C5-cyclocarbonate precursor functionality is determined by integrating the signal at 5.15 ppm corresponding to the proton on the tertiary carbon of the cyclocarbonate ring (1.95 for BDF2Cy, 1.85 for IDF2Cy, 2.40 for GTF3Cy, and 2.0 for SYR2Cy). It is noteworthy to mention that the integration for the final C5 functionality is close to the initial epoxy one, demonstrating the almost quantitative conversion of the oxiranes.

Figure 5. Chemical degradation in NaOH 3 M at 50 °C (left) and chemical degradation in HCl 3 M at 50 °C (right) (▴ IDF2EP-IPDA, ♦ DGEBA-IPDA, and ● SYR2EP-IPDA). Adapted with permission from ref. (64). Copyright 2017 John Wiley & Sons.

Scheme 3. Lithium bromide-mediated carbonation of BDF2EP, IDF2EP, GTF3EP, and SYR2EP epoxies. 241

Synthesis of the NIPUs In the case of NIPUs, one cyclocarbonate only reacts once with a primary amine. Thus, trifunctional C5-precursor GTF3Cy allows access to crosslinked NIPU networks thanks to its C5-functionality higher than two, whereas difunctional C5-precursors BDF2Cy, IDF2Cy, and SYR2Cy lead to linear NIPU chains. With the successful preparation of BDF2Cy, IDF2Cy, and GTF3Cy, three NIPU series were prepared via their polyaddition with diamine. As ferulic acid- and sinapic acid-based epoxy precursors exhibit low melting temperature, and various studies (80) having shown that the reaction of cyclic carbonate and diamine occurs faster in bulk than in solvent, the reactants were mixed without any solvent, and polyadditions were performed in bulk without adding any catalyst. Polyadditions were carried out following a temperature program adapted to each cyclocarbonate. BDF2Cy, IDF2Cy, and GTF3Cy were cured at the lower temperature (100 °C), to both ensure the homogeneity of reaction media and avoid degradation of the fragile ester bond. Although a higher temperature curing would reduce the viscosity of the system and enhance the conversion, this was not attempted as transamidation and/or chain scission may occur (3). Inversely, for both SYR-Cy and BPA-Cy, which are more robust, a higher temperature program was chosen (160 and 180 °C, respectively). NIPU thermosets and thermoplastics were prepared by mixing the C5-precursor with the adequate amount of diamine, considering that a primary amine reacts with only one C5 cyclocarbonate (i.e., the C5 cyclocarbonate/amine ratio was fixed to 1:1 for all the formulations). Four diamines with various structures were chosen as curing agents: the oil-based aliphatic and cycloaliphatic amines, EDR148 from Huntsman, and IPDA, respectively, but also two bio-based diamines, aliphatic DA10 and the aromatic DIFFA; a triamine, tris(2-aminoethyl)amine (TREN), was also tested with SYR2Cy. As shown in Figure 6, which pictures the BDF2Cy-containing thermoplastic series, materials are translucent, except for the IPDA-containing one. The same trend is observed for the four series. The various diamine structures led to NIPUs with different physical properties, from flexible and elastic materials with low Tg, to rigid and brittle ones with higher Tg.

Figure 6. Photographs of BDF2Cy-based NIPUs. Reproduced with permission from ref. (48). Copyright 2016 American Chemical Society. 242

Characterization of the Ferulic Acid- and Sinapic Acid-Based NIPUs Table 3. Experimental data from TGA, DSC, SEC, and 1H-NMR

FT-IR and 1H/13C-NMR Spectroscopy

Figure 7. Mechanisms of the two C5-ring-opening routes. Reproduced with permission from ref. (48). Copyright 2016 American Chemical Society. The structures of the linear and crosslinked prepared NIPUs were first confirmed through FT-IR. The conversion was monitored by observing the disappearance of the carbonate signal (νC=O,carbonate ≈ 1800 cm-1) and the appearance of that of the carbamate (νC=O,carbamate ≈ 1700 cm-1, νNH ≈ 3400 243

cm-1). This simple analysis allows a qualitative estimation of the conversion, which is a very useful way to adjust the polymerization temperature. For all the linear NIPU synthesized, NMR analyses were carried out. The formation of a urethane link was confirmed by 13C NMR spectroscopy, with the appearance of new peaks for the quaternary carbon of the carbonyl moiety (e.g., at 156.0 and 156.4 ppm for SYR2Cy-based NIPU, when resonance of the carbonyl group of the starting carbonates was observed at ~. 155.0 ppm). In 1H-NMR spectra, the appearance of peaks between 7.1 and 7.7 ppm corresponds to the NH of the urethane moiety obtained by ring opening. 1H-NMR provides an estimated conversion (Table 3). NMR analyses also provide information on the regioselectivity of the ring-opening polyaddition. As depicted in Figure 7, the nucleophilic attack of the primary amine on cyclic carbonate moiety gives access to both primary and secondary alcohols. The signal of primary alcohol at 4.8–5.0 ppm and that of secondary OH at 3.38–3.57 ppm were attributed through 2D NMR and hydrogen-deuterium exchange. The ratio of primary and secondary hydroxyl groups (RI/II) can thus be defined by the alcohols/signals/integrations ratio, as described in the literature, and results are summed up in Table 3. RI/II values show the preferential formation of secondary alcohol, due to the electro-withdrawing group (PhOCH2-) that promotes a selectivity toward secondary alcohol. Finally, it is noteworthy to mention that FT-IR and 13C-NMR spectra show no evidence of transamidification of ferulic acid-based cyclocarbonates or unreacted amine. The latter information allows us to assess that residual cyclic carbonates are end-chain groups and calculate the degree of polymerization (Table 3).

Gel Permeation Chromatography The linear NIPUs were solubilized, into THF for BDF2Cy and IDF2Cy, and into DMF for SYR2Cy, in the presence of toluene as a flow marker and analyzed by SEC using PS standards. For all the NIPU, high conversions were obtained (between 75 and 95%). Molecular weights ranged between 3.5 and 10.0 kg/mol, which correspond to low DPn between 4 and 11. These low molar masses may be due to the incomplete conversion of the cyclocarbonates during the polyaddition reaction as previously determined by 1H-NMR. One can assume that these low conversions may be due to the increasing viscosity of the NIPU system that does not allow enough molecular motions for a quantitative reaction. Although the viscosity of the system may be decreased by performing the curing at a higher temperature, this was not attempted as transamidation and/or chain scission may occur. The discrepancies between the DPn values obtained via SEC and determined by 1H-NMR (Table 3) are due to the presence of hydrogen bonds, which minimize the hydrodynamic radius value of the solubilized NIPU chain compared to the PS standard ones. Moreover, the presence of a minor part of the NIPU oligomers of lower DPn induces a DPn error and minimizes the polydispersity (e.g., Đ of 1.1 for BDF2Cy-IPDA), whereas the other NIPUs, which show a more homogeneous distribution of the molar mass, present a higher Đ as expected for a polyaddition. 244

Thermal Characterization

Figure 8. Td5% of the NIPUs under nitrogen (10 °C min-1).

Thermogravimetric Analysis Thermal degradations of all NIPU materials were characterized by TGA under nitrogen flow to assess the influence of the monomer structure on the thermal stability and the charring propensity of the polymer (Figure 8). Table 3 shows the Td5% and the wt%char for all the thermosets and thermoplastics prepared. In terms of thermal degradation, the NIPU materials exhibit good thermal stabilities as their Td5% range between 225 and 280 °C. It is generally observed that, for the same diamine, the SYR2Cy- and BDF2Cy-containing NIPUs exhibit the highest thermal stability. As the BDF2Cy-containing NIPU presents equivalent thermal stabilities, it is assumed that the thermal stability of the NIPU system is mainly correlated with the stability of the internal ester of the C5-precursor in the ferulic acid-based NIPUs. IDF2Cy and GTF3Cy possess at least one secondary carbon in the α position of the ester, while BDF2Cy only possesses a primary one. As the alkyl groups are inductive donator groups, our assumption is that the higher inductive effect in the case of IDF2Cy and GTF3Cy destabilized further their ester compared to the BDF2Cy one, which is subjected to a lower inductive effect. One can observe that the crosslinked structure of the GTF3Cy-based NIPU does not seem to influence the thermal stability of the NIPU. It is assumed that the thermal degradation of all the ferulic acid-based NIPU starts by the thermolysis of the internal ester functions that are the more sensitive functions in these NIPUs. Concerning the high temperature char content, in the four series, DIFFA leads systematically to the highest char content of an 245

NIPU series due to the aromaticity brought by its two furanic rings that favors the charring mechanism during the NIPUs’ thermal degradation. The linear DA10 and TREN provide the lowest char content since the aliphatic chains present a very low charring propensity. IPDA and EDR148 lead to intermediate char content. We assume that the cyclic structure of IPDA and the EDR148 oxygen content may favor the cyclization and dihydroxylation mechanism, respectively. One can also notice that for the same diamine the SYR2Cy-containing NIPUs present a higher char content at 900 °C (Table 3). This is due to the higher stability of the bisfuranic moiety (i.e., cyclic ethers) compared to that of the esters of the ferulic acid-based cyclocarbonates, which favor the charring mechanism, a potent feature for fire retardation applications. Finally, it is noteworthy to mention that ferulic acid- and sinapic acid-based NIPUs have similar Td5% to that of the DGEBA-IPDA NIPU (i.e., 276 °C).

Differential Scanning Calorimetry (DSC) To assess the Tg of the prepared NIPU, DSC analyses were carried out (Figure 9). For the same cyclocarbonate precursor, the choice of diamines offers a wide range of Tg, from 18 °C with flexible aliphatic DA10 to 98 °C for more rigid cyclic IPDA. For the same amine used, BDF2Cy-based NIPUs offered the lowest Tg (62 °C) due to the flexibility of the aliphatic chain of the esters. Interestingly, in the case of NIPU thermoplastics, the rigidity of SYR offered the highest Tg, 20 °C higher than BPA. For NIPU thermosets (i.e., curing with TREN), IDF2Cy offered the lower Tg (47 °C) while SYR2Cy reached DGEBA2Cy value due to an increase of conformational barriers for chain motion brought by the extra methoxy substituent in SYR2Cy in comparison with ferulic acid-based cyclocarbonates.

Figure 9. Tg of the NIPUs. 246

Thermomechanical Characterization of Ferulic Acid- and Sinapic Acid-Based Thermosets To assess the mechanical properties of the crosslinked NIPU materials, the thermosets prepared from GTF3Cy were characterized by DMA. Unfortunately, due to the slight foaming of the GTF3Cy-IPDA thermosets, and the brittleness of SYR2Cy-based thermosets, these NIPU oligomers could not be mechanically characterized. GTF3Cy-containing NIPUs exhibit thermoset thermomechanical behaviors. Indeed, below their transition temperature, the three NIPUs show a glassy plateau. At temperatures higher than their Tg, the value of the elastic modulus is reduced until a rubbery plateau showing a particular structure of the polymer due to crosslinking points. Furthermore, the presence of a rubbery plateau for each NIPU sample reveals an elastomeric behavior for these materials. It is also observed that the two aliphatic-based NIPUs show equivalent and low Tα, while the DIFFA-containing NIPU exhibits a higher transition as observed by measuring the Tg values. Concerning the NIPU thermal transition behavior, one can observe a two-time mechanical relaxation phenomenon, not visible by DSC. Indeed, the EDR148- and DA10-containing NIPUs show two distinct tan δ peaks; whereas GTF3Cy-DIFFA presents a clear shoulder on the trace, allowing the determination of two mechanical transitions for each sample (Table 3). Interestingly, the Tg obtained by DSC corresponds to one of the two Tα obtained by DMA. Indeed, the NIPU thermosets present a tan δ peak corresponding to the Tg value previously determined for DA10 and EDR148 (i.e., ~ 40 °C) and for DIFFA (i.e., ~ 60 °C). The two aliphatic diamines lead to a second mechanical transition at a temperature lower than the Tg of the material, around 20 °C; whereas the DIFFA-containing NIPU presents a second mechanical transition at temperatures higher than the Tg of the material (~ 100 °C). It is assumed that these additional mechanical transitions are linked to the diamine structure. The 16 to 19 °C mechanical transitions may correspond to the relaxation of the linear aliphatic segment of DA10 and EDR148 diamines and the one at 102 °C to that of the diaromatic segment of DIFFA; whereas the 39 to 61 °C transitions are attributed to the GTF3Cy network segments. These results tend toward structured thermosets with soft segments (fully reacted diamines, but also potential partially reacted diamines that act as plasticizers) and hard segments (GTF3Cy), the latter being impacted by the soft ones as suggested by the rather large range observed for the transitions attributed to GTF3Cy. Such two-time mechanical transitions are of great interest as they allow NIPUs to exhibit intermediate mechanical behavior, between the glassy and the rubbery state, in a short temperature range. Even if DMA is not ideal to determine precisely the elastic modulus (E′) of the NIPU thermosets, it gives an order of magnitude for these values (Table 3), showing the strong mechanical behavior of these materials with E′ > 1 GPa below their Tg. At the rubbery state, DIFFA leads to a lower elastic modulus than those of DA10 or EDR148. This is representative of the lower crosslink density of the DIFFA-containing thermoset linked to its higher molar mass.

247

Conclusion With the low availability and fluctuating prices of fossil resources, and new regulations that restrict the use of harmful precursors, finding renewable and safe alternatives to the current petro-based building blocks used in polymers and materials remains a challenge. In this chapter, ferulic and sinapic acids, two naturally occurring p-hydroxycinnamic acids, have been used for the preparation of renewable and non-endocrine disruptive epoxies using efficient chemo-enzymatic synthetic routes. The latter were then cured with diamines to provide epoxy-amine resins whose thermal properties can be finely tuned by playing with the chemical structure of the epoxies. Among the novel epoxies, sinapic acid-derived SYR provided epoxy-amine resins exhibiting thermal properties close to that of DGEBA ones, making them less toxic and bio-based alternatives. Moreover, the ferulic acid-based epoxies bring a very interesting degradability character to the corresponding epoxy-amine resins in both acidic and basic media, thus allowing the recyclability of the materials. The ferulic acid- and sinapic acid-based epoxies were also readily transformed into the corresponding cyclocarbonates, which were then cured with di- and triamines to provide thermoplastic and thermoset NIPU oligomers. Similarly to the epoxy-amine resins, the thermomechanical properties of the materials can be easily tailored by judiciously playing with the structure of the cyclocarbonates. This chapter thus demonstrates the high potential of ferulic and sinapic acids as not only safer and sustainable BPA alternatives in epoxy-amine resins but also as aromatic building blocks for the production of thermoplastic and thermoset NIPU oligomers.

References Hamerton, I. Polym. Int. 1996, 41, 101–102. Nohra, B.; Candy, L.; Blanco, J.-F.; Guerin, C.; Raoul, Y.; Mouloungui, Z. Macromolecules 2013, 46, 3771–3792. 3. Maisonneuve, L.; Lamarzelle, O.; Rix, E.; Grau, E.; Cramail, H. Chem. Rev. 2015, 115, 12407–12439. 4. Sardon, H.; Pascual, A.; Mecerreyes, D.; Taton, D.; Cramail, H.; Hedrick, J. L. Macromolecules 2015, 48, 3153–3165. 5. Maffini, M. V.; Rubin, B. S.; Sonnenschein, C.; Soto, A. M. Mol. Cell. Endocrinol. 2006, 254–255, 179–186. 6. Chen, M.-Y.; Ike, M.; Fujita, M. Environ. Toxicol. 2002, 17, 80–86. 7. Food Additives & Ingredients; https://www.fda.gov/food/ ingredientspackaginglabeling/foodadditivesingredients/default.htm (accessed: September 4, 2016). 8. Nelson, A. M.; Long, T. E. Polym. Int. 2012, 61, 1485–1491. 9. Raquez, J.-M.; Deléglise, M.; Lacrampe, M.-F.; Krawczak, P. Prog. Polym. Sci. 2010, 35, 487–509. 10. Auvergne, R.; Caillol, S.; David, G.; Boutevin, B.; Pascault, J.-P. Chem. Rev. 2014, 114, 1082–1115. 1. 2.

248

11. Wu, G.-M.; Liu, D.; Liu, G.-F.; Chen, J.; Huo, S.-P.; Kong, Z.-W. Carbohydr. Polym. 2015, 127, 229–235. 12. Garrison, M. D.; Harvey, B. G. J. Appl. Polym. Sci. 2016, 43621, 1–12. 13. Supanchaiyamat, N.; Hunt, A. J.; Shuttleworth, P. S.; Ding, C.; Clark, J. H.; Matharu, A. S. RSC Adv. 2014, 4, 23304–23313. 14. Kovash, C. S. J.; Pavlacky, E.; Selvakumar, S.; Sibi, M. P.; Webster, D. C. ChemSusChem 2014, 7, 2289–2294. 15. Mantzaridis, C.; Brocas, A.-L.; Llevot, A.; Cendejas, G.; Auvergne, R.; Caillol, S.; Carlotti, S.; Cramail, H. Green Chem. 2013, 15, 3091–3098. 16. Lukaszczyk, J.; Janicki, B.; Kaczmarek, M. Eur. Polym. J. 2011, 47, 1601–1606. 17. Chrysanthos, M.; Galy, J.; Pascault, J.-P. Polymer (Guildf). 2011, 52, 3611–3620. 18. Hong, J.; Radojcic, D.; Ionescu, M.; Petrovic, Z. S.; Eastwood, E. Polym. Chem. 2014, 5, 5360–5368. 19. European Chemicals Agency Substance Information; https://echa.europa.eu/ substance-information/-/substanceinfo/100.239.193 (accessed: May 29, 2018). 20. Kihara, N.; Endo, T. J. Polym. Sci., Part A: Polym. Chem. 1993, 31, 2765–2773. 21. Kihara, N.; Kushida, Y.; Endo, T. J. Polym. Sci., Part A: Polym. Chem. 1996, 34, 2173–2179. 22. Tomita, H.; Sanda, F.; Endo, T. J. Polym. Sci., Part A: Polym. Chem. 2001, 39, 851–859. 23. Tomita, H.; Sanda, F.; Endo, T. J. Polym. Sci., Part A: Polym. Chem. 2001, 39, 860–867. 24. Tomita, H.; Sanda, F.; Endo, T. J. Polym. Sci., Part A: Polym. Chem. 2001, 39, 3678–3685. 25. Oestreich, S.; Struck, S. Macromol. Symp. 2002, 187, 325–332. 26. Ochiai, B.; Satoh, Y.; Endo, T. Green Chem. 2005, 7, 765–767. 27. Ochiai, B.; Kojima, H.; Endo, T. J. Polym. Sci., Part A: Polym. Chem. 2014, 52, 1113–1118. 28. Javni, I.; Hong, D. P.; Petrovic, Z. S. J. Appl. Polym. Sci. 2008, 108, 3867–3875. 29. Bähr, M.; Mülhaupt, R. Green Chem. 2012, 14, 483. 30. Bähr, M.; Bitto, A.; Mülhaupt, R. Green Chem. 2012, 14, 1447–1454. 31. Carré, C.; Bonnet, L.; Avérous, L. RSC Adv. 2014, 4, 54018–54025. 32. Carré, C.; Bonnet, L.; Avérous, L. RSC Adv. 2015, 5, 100390–100400. 33. Benyahya, S.; Aouf, C.; Caillol, S.; Boutevin, B.; Pascault, J.-P.; Fulcrand, H. Ind. Crops Prod. 2014, 53, 296–307. 34. Fache, M.; Auvergne, R.; Boutevin, B.; Caillol, S. Eur. Polym. J. 2015, 67, 527–538. 35. Fache, M.; Viola, A.; Auvergne, R.; Boutevin, B.; Caillol, S. Eur. Polym. J. 2015, 68, 526–535. 36. Kristufek, S. L.; Yang, G.; Link, L. A.; Rohde, J. B.; Robertson, M.; Wooley, K. ChemSusChem 2016, 9, 2135–2142. 249

37. Ménard, R.; Negrell, C.; Fache, M.; Ferry, L.; Sonnier, R.; David, G. RSC Adv. 2015, 5, 70856–70867. 38. Maiorana, A.; Spinella, S.; Gross, R. A. Biomacromolecules 2015, 16, 1021–1031. 39. Hernandez, E. D.; Bassett, W. A.; Sadler, J. M.; La Scala, J. J.; Stanzione, J. F. ACS Sustain. Chem. Eng. 2016, 4, 4328–4339. 40. Llevot, A.; Grau, E.; Carlotti, S.; Grelier, S.; Cramail, H. J. Mol. Catal. B: Enzym. 2016, 125, 34–41. 41. Harvey, G.; Sahagun, C. M.; Guenthner, A. J.; Groshens, T. J.; Cambrea, L. R.; Reams, J. T.; Mabry, J. M. ChemSusChem 2014, 7, 1964–1969. 42. Chen, Q.; Gao, K.; Peng, C.; Xie, H.; Zhao, Z. K.; Bao, M. Green Chem. 2015, 17, 4546–4551. 43. Fache, M.; Darroman, E.; Besse, V.; Auvergne, R.; Caillol, S.; Boutevin, B. Green Chem. 2014, 16, 1987. 44. Tilay, A.; Bule, M.; Kishenkumar, J.; Annapure, U. J. Agric. Food Chem. 2008, 56, 7644–7648. 45. Dupoiron, S.; Lameloise, M. L.; Pommet, M.; Bennaceur, O.; Lewandowski, R.; Allais, F.; Teixeira, A. R. S.; Rémond, C.; Rakotoarivonina, H. Ind. Crops Prod. 2017, 105, 148–155. 46. Dupoiron, S.; Lameloise, M. L.; Bedu, M.; Lewandowski, R.; Allais, F.; Teixeira, A. R. S.; Rakotoarivonina, H.; Rémond, C. Sep. Purif. Technol. 2018, 200, 75–83. 47. Overhage, J.; Steinbüchel, A.; Priefert, H.; Steinbu, A. Appl. Environ. Microbiol. 2003, 69, 6569–6576. 48. Vermes, B.; Seligmann, O.; Wagner, H. Phytochemistry 1991, 30, 3087–3089. 49. Hollande, L.; Jaufurally, A. S.; Ducrot, P.-H.; Allais, F. RSC Adv. 2016, 6, 44297–44304. 50. Ménard, R.; Caillol, S.; Allais, F. ACS Sustainable Chem. Eng. 2017, 5, 1446–1456. 51. Janvier, M.; Ducrot, P.-H.; Alais, F. ACS Sustainable Chem. Eng. 2017, 5, 8648–8656. 52. Lue, B.-M.; Karboune, S.; Yeboah, F. K.; Kermasha, S. J. Chem. Technol. Biotechnol. 2005, 8, 462–468. 53. Sabally, K.; Karboune, S.; Yebaoh, S.; Kermasha, S. Biocatal. Biotransform. 2005, 23, 37–44. 54. Sabally, K.; Karboune, S.; St-Louis, R.; Kermasha, S. J. Am. Oil Chem. Soc. 2006, 83, 101–107. 55. Sabally, K.; Karboune, S.; St-Louis, R.; Kermasha, S. Biotransformation 2007, 25, 211–218. 56. Feddern, V.; Yang, Z.; Xu, X.; Badiale-Furlong, E.; Almeida de Souza Soares, L. Ind. Eng. Chem. Res. 2011, 50, 7183–7190. 57. Figueroa-Espinoza, M.-C.; Villeuneuve, P. Food Chem. 2005, 53, 2779–2787. 58. Yang, Z.; Feddern, V.; Glasius, M.; Guo, Z.; Xu, X. Biotechnol. Lett. 2011, 33, 673–679. 59. Pion, F.; Reano, A. F.; Ducrot, P.-H.; Allais, F. RSC Adv. 2013, 3, 8988–8998. 250

60. Prasad, A. K.; Husain, M.; Singh, B. K.; Gupta, R. K.; Manchanda, V. K.; Olsen, C. E.; Parmar, V. S. Tetrahedron Lett. 2005, 46, 4511–4514. 61. Dhake, K. P.; Qureshi, Z. S.; Singhal, R. S.; Bhanage, B. M. Tetrahedron Letters 2009, 50, 2811–2814. 62. Poulhes, F.; Vanthuyne, N.; Bertrand, M. P.; Gastaldi, S.; Gil, G. J. Org. Chem. 2011, 76, 7281–7286. 63. Jaufurally, A. S.; Teixeira, A. R. S.; Hollande, L.; Allais, F.; Ducrot, P.-H. ChemistrySelect 2016, 1, 5165–5171. 64. Ménard, R.; Caillol, S.; Allais, F. Ind. Crops Prod. 2017, 95, 83–95. 65. Maiorana, A.; Reano, F.; Centore, R.; Grimaldi, M.; Balaguer, P.; Allais, F.; Gross, R. A. Green Chem. 2016, 18, 4961–4973. 66. Janvier, M.; Hollande, L.; Jaufurally, A. S.; Pernes, M.; Ménard, R.; Grimaldi, M.; Beaugrand, J.; Balaguer, P.; Ducrot, P.-H.; Allais, F. ChemSusChem 2017, 10, 738–746. 67. Kishi, H.; Akamatsu, Y.; Nogushi, M.; Fujita, A.; Matsuda, S.; Nishida, H. J. Appl. Polym. Sci. 2011, 120, 745–751. 68. Ferdosian, U. F.; Yuan, Z.; Anderson, M.; Xu, C. RSC Adv. 2014, 4, 31745–31753. 69. Nouailhas, H.; Aouf, C.; Le Guerneve, C.; Caillol, S.; Boutevin, B.; Fulcrand, H. J. Polym. Sci., Part A: Polym. Chem. 2011, 49, 2261–2270. 70. Harvey, G.; Guenthner, A. J.; Lai, W. W.; Meylemans, H. A.; Davis, M. C.; Cambrea, L. R.; Reams, J. T.; Lamison, K. R. Macromolecules 2015, 48, 3173–3179. 71. Soleimani, A.; Drappel, S.; Carlini, R.; Goredema, A.; Gillies, E. R. Ind. Eng. Chem. Res. 2014, 53, 1452–1460. 72. Zuo, J.; Li, S.; Bouzidi, L.; Narinz, S. S. Polymer (Guildf). 2011, 52, 4503–4516. 73. Rokicki, G.; Kuran, W. Bull. Chem. Soc. Jpn. 1984, 57, 1662–1666. 74. Reithofer, M. R.; Sum, Y. N.; Zhang, Y. Green Chem. 2013, 15, 2086–2090. 75. Liang, S.; Liu, H.; Jiang, T.; Song, J.; Yang, G.; Han, B. Chem. Commun. 2011, 47, 2131–2133. 76. Paddock, R. L.; Nguyen, S. T. J. Am. Chem. Soc. 2001, 123, 11498–11499. 77. Buchard, A.; Kember, M. R.; Sandeman, G.; Williams, C. K. Chem. Commun. 2011, 47, 212–214. 78. Yu, K. M. K.; Curcic, I.; Gabriel, J.; Morganstewart, H.; Tsang, S. C. J. Phys. Chem. A 2010, 114, 3863–3872. 79. Liu, S.; Suematsu, N.; Maruoka, K.; Shirakawa, S. Green Chem. 2016, 18, 4611–4615. 80. Couvret, D.; Brosse, J.-C.; Chevalier, S.; Senet, J.-P. Die Makromol. Chem. 1990, 191, 1311–1319.

251

Chapter 16

Application of Bio-Based Epoxy Resin as the Matrix for Composites Liang Yue* Department of Macromolecular Science and Engineering, Case Western Reserve University, Cleveland, Ohio 44106, United States *E-mail: [email protected].

The potential use of bio-based epoxy resins as sustainable alternatives to the currently used petroleum-derived epoxy is discussed. A new bio-based formulation with 85 wt % of diglycidyl ethers of either ethyl or pentyl diphenolate esters (DGEDP) mixed with 15 wt % glycidyl ether of eugenol (GE) demonstrates the feasibility of replacing petroleum-based epoxy resins in the vacuum infusion of fiberglass mats. The processability of bio-based epoxy resins being used has been rarely studied previously. Here, the processability of the bio-based DGEDP/GE formulation was studied by chemorheology. The bio-based formulation shows suitable viscosity and gelation time for composite vacuum-infusion processing while maintaining, and in some cases exceeding, the mechanical and thermal properties of the petroleum-based systems. The bio-based DGEDP/GE resin system is a good example of an application of bio-based epoxy resin as the matrix for composites, which may have great potential in the future.

Introduction The traditional epoxy resin is derived from petroleum product Bisphenol A (BPA). However, in recent years, decreasing petroleum reserves and increasing environmental problems have pushed toward more sustainable development. Additionally, the potential health impact of using BPA, which is believed to increase the risk of cancer, has received a lot of attention (1, 2). Therefore, the © 2018 American Chemical Society

demand for alternative epoxy monomers from more sustainable and BPA-free sources has increased tremendously. There has been an increasing interest in academia and the industry in recent years in investigating the synthesis of epoxy monomers from bio-based feedstock. Many reports on the use of various renewable bio-based resources to synthesize epoxy monomers have been published. Some raw materials used include plant oils (3), saccharides (4), polyphenols (5), lignin, and lignin derivatives (6). Some of those reported bio-based epoxy resin systems achieve comparable mechanical properties with the Bisphenol A diglycidyl ether (DGEBA) resin. For example, Robertson et al. reported on epoxy resins derived from plant-based phenolic acids, which showed comparable mechanical properties with the DGEBA resin (7). After an allylation step of salicylic acid and 4-hydroxybenzoic acid, the epoxidation is shown in Figure 1. The yield is 70% to salicylic acid and 48% to 4-hydroxybenzoic acid. The bio-based epoxy resins derived from phenolic acids exhibit comparable glass transition temperature, tensile strength, and modulus with the petroleum-based DGEBA resin as reported.

Figure 1. Epoxidation of allylated (a) salicylic acid and (b) 4-hydroxybenzoic acid. Reproduced from reference (7). Copyright 2016 American Chemical Society.

However, although many bio-based epoxy resin systems have been reported, there is little discussion about the processability of these resins. A lack of information on processing may limit their commercialization.

Bio-Based Diglycidyl Ethers of Alkyl Diphenolates Epoxy Resin More recently, Gross et al. reported on a series of bio-based epoxy resin systems synthesized from esters of diphenolic acid (8). The diphenolic acid has a similar molecular structure with BPA, which may result in similar physical properties. The presence of an extra carboxylic acid group could allow chemical modification to tune the properties further. The bio-based epoxy resins based on diphenolic acid are synthesized as shown in Figure 2. By variation of the ester side-chain group (methyl, ethyl, butyl, and pentyl), the properties of those diglycidyl ethers of alkyl diphenolates (DGEDP)-based resins vary. The yields of the synthesized epoxy resins are 85–97%. 254

Figure 2. Two-step synthetic pathway to synthesize n-alkyl diphenolate diglycidyl ethers that differ in the chain length of the n-alkanol ester moiety. The diglycidyl ethers of alkyl diphenolates (DGEDP-methyl ester, DGEDP-ethyl ester, DGEDP-n-butyl ester, and DGEDP-n- pentyl ester) were prepared. Reproduced from reference (8). Copyright 2015 American Chemical Society.

These DGEDP-based resins exhibit Newtonian behavior under rotational shearing (Figure 3), with viscosity values dependent on the ester chain length. Table 1 compares the viscosity of these DGEDP resins and DGEBA resin. As the ester chain length increases, the viscosity decreases. This is expected since a longer side chain length increases the flexibility of the monomers. DGEDP-methyl has a very high viscosity at room temperature (792 Pa·s). Such high viscosity is definitely not suitable for composite processing, but it may be ideal for adhesion applications. By increasing the ester chain length to DGEDP-pentyl, the viscosity dramatically decreases (12 Pa·s), close to DGEBA resin (4 Pa·s).

Figure 3. (a) Rotational rheology at 25 °C of synthesized epoxy resins compared to the DGEBA. (b) Viscosity as a function of ester length for the prepared resins with a trend line for the diamond-shaped symbols and excluding the square symbol. Reproduced from reference (8). Copyright 2015 American Chemical Society. 255

Table 1. Comparison of Viscosities, Experimental, and EEW of DGEDP (8)a

a

Viscosity (Pa·s)

Measured EEW

Theoretical EEW

DGEBA

4

175

170

DGEDP-methyl ester

792

215 ± 18

206

DGEDP-ethyl ester

63

222 ± 8.6

213

DGEDP-n-butyl ester

18

243 ± 7.4

227

DGEDP-n-pentyl ester

12

265 ± 3.7

234

DGEDP, diglycidyl ethers of alkyl diphenolates; EEW, Epoxide Equivalent. Weight.

The synthesized DGEDP resins were cured with stoichiometric amounts of isophorone diamine. The mechanical properties of these resins were compared with DGEBA resin cured under similar conditions. Tensile test results are shown in Figure 4. There are very small differences in tensile strength and Young’s modulus between DGEDP resins and DGEBA resins. DGEDP-methyl and DGEDP-ethyl exhibit an even higher modulus than the DGEBA resin.

Figure 4. (a) Stress–strain curves of the cured DGEDP–ester epoxy resins compared to cured DGEBA. (b) Storage modulus (DMA) and Young’s modulus (tensile testing) of the cured materials. Reproduced from reference (8). Copyright 2015 American Chemical Society. The glass transition temperatures of those DGEDP resins show a trend of decreasing as the ester chain length increases (Figure 5). The DGEDP-methyl has the highest glass transition temperature, which is comparable with the DGEBA resin. For many applications, higher glass transition temperature is preferred, but from the processing standpoint, lower viscosity is preferred, especially for vacuum-infusion processing. These bio-based DGEDP resins could be chosen for different applications such as adhesives, coatings, or matrices for composites, based on processability and working conditions. 256

Figure 5. (a) Alpha transition temperature (peak of loss modulus) and glass transition temperature as a function of ester length. (b) Tan(δ) as a function of temperature. Reproduced from reference (8). Copyright 2015 American Chemical Society.

The DGEDP epoxy resins from renewable resources exhibit the potential to replace the petroleum-based DGEBA resin. However, the disadvantages related to viscosity should also be further addressed.

A Bio-Based Reactive Diluent for Diglycidyl Ethers of Alkyl Diphenolate Resins The development of epoxy resins for infusion processing has been of great importance for the industry. Large pieces of engineering structural composites that are fabricated with high-fiber content, such as wind turbine blades, require epoxy resins with very low viscosity. Commercial resin products for the infusion application are usually based on the mixture of DGEBA with low-viscosity reactive diluents, such as monofunctional aliphatic glycidyl esters and aliphatic or aromatic glycidyl ethers (9, 10). In addition to decreasing the viscosity of the system, the reactive diluent should also extend the gelation time of the resin to ensure long enough processing time to finish the infusion. Besides, the reactive diluent should also react with the hardener but be nonreactive with the resin under storage conditions. A new kind of bio-based reactive diluent has been recently developed by Maiorana and colleagues (11). It is a monofunctional glycidyl ether reactive diluent synthesized from eugenol, which is extracted from plant-based essential oils, such as clove and nutmeg (12, 13). The bio-based reactive diluent glycidyl ether of eugenol (GE) is synthesized in a single step, as shown in Figure 6, with a yield ranging from 85 to 90%. 257

Figure 6. Synthesis of glycidyl ether of eugeno (GE)l. Reproduced from reference (11). Copyright 2016 Wiley.

The main target use of GE is to decrease the viscosity of the bio-based DGEDP epoxy resin. Since the monofunctional glycidyl ether will react with the diamine hardener, the cross-linking density, as well as the physical properties of the cured resin, will possibly be different from the neat resin. To study how the reactive diluent will impact the resin properties, the GE was formulated with DGEDP-pentyl at different ratios. DGEDP-pentyl was chosen because it has the lowest viscosity (around 12 Pa·s) of the DGEDP-based epoxy resins. From the processing viewpoint, it may be the most suitable resin for vacuum-infusion processing. Figure 7 shows the viscosity of the formulated resin as a function of GE loading (5, 10, 15, 20, 30, and 50 wt % to DGEDP-pentyl resin). GE could dramatically decrease the viscosity of DGEDP-pentyl resin. It was found that 5 wt % GE could reduce the viscosity by 38%, 10 wt % could reduce it by 70%, whereas 20 wt % could reduce the viscosity by 100% to about 1 Pa·s. The initial viscosities of these formulated resin systems are comparable with the commercial infusion resin system.

Figure 7. Average Newtonian viscosity as a function of percent GE in DGEDP-pentyl mixtures at room temperature. Reproduced from reference (11). Copyright 2016 Wiley. 258

Gelation time is another important parameter for infusion processing. The hardener used is isophorone diamine, and the stoichiometric ratio is based on the epoxied equivalent weights from both DGEDP-pentyl resin and the reactive diluent. Figure 8 shows the chemorheology study for the formulated resin systems with different GE loading. To speed up gelation, tests were conducted at 80 °C. The gel point is determined as the time at which Tan(δ) is independent of frequency as shown in Figure 8b.

Figure 8. (a) Gel time as a function of the reactive diluent concentration at 80 °C, and (b) determining the time of frequency independence of Tan(δ) for the 5 wt % GE composition. Reproduced from reference (11). Copyright 2016 Wiley.

Figure 9. (a) Storage modulus (25 °C), and (b) peak of the loss modulus (alpha transition) being related to the glass transition temperature as a function of increasing glycidyl eugenol. Reproduced from reference (11). Copyright 2016 Wiley. Figure 8a shows the gel time is extended significantly as the GE loading increases allowing for better processability. Processability of the bio-based DGEDP resin is improved by decreasing the system’s initial viscosity and extending the gelation time. The effect of GE on the mechanical and thermal properties of the cured formulated resin systems are studied with dynamic 259

mechanical analysis. The storage modulus of formulated resin systems with different GE loadings at 25 °C is shown in Figure 9a. The storage modulus remains similar (between 2 and 3 GPa) up to 30 wt % GE. The glass transition temperatures in Figure 9b are taken from the peak of the loss modulus correlated to the alpha transition temperature Tα, which is closely related to the glass transition temperature Tg. The formulated resin systems exhibit a clear trend of decreasing Tg with increasing GE loading. As shown in Figure 9b, the relation of Tg as the function of GE loading is nearly linear (slope = -1.36, R2 = 0.995). Unlike the storage modulus, the Tg of formulated resin systems is highly sensitive to the GE loading. Even with just 5 wt % GE, the Tg decreases by 6 °C.

Bio-Based Epoxy Resin Formula for Vacuum Infusion Molding To evaluate the processability and mechanical performance of the bio-based DGEDP resins for the application of fiber-reinforced composites, DGEDP-ethyl and DGEDP-pentyl utilizing 15 wt % GE were compared with two petroleumderived benchmark systems (DGEBA with 15 wt % GE and a Hexion infusion resin) (14). As previously discussed, resin viscosity is one of the most critical parameters for composites processing. Low viscosity of the resin results in good flow properties and fiber wetting, and the gelation time of resin represents the processable window.

Figure 10. Complex viscosity of epoxy resins and cross-linker as a function of time at 25 °C. Reproduced from reference (14). Copyright 2017 Elsevier. The chemorheology of curing could provide the complex viscosity as a function of time as shown in Figure 10. Both initial viscosity and gel time were important for processing. The Hexion infusion resin displays an initial viscosity of 2.5 Pa·s only, while the bio-based DGEDP-ethyl and DGEDP-pentyl have initial viscosities of 20.4 and 10.0 Pa·s, respectively. As the curing reaction happens, the viscosity of the resin system will gradually increase. Once the system reaches the gel point, the resin can no longer flow and loses its processability. The time from the curing start to the time when the system gelled can be considered as the processing window for the thermoset system. By adding 15 wt % GE to the bio-based resins, the initial viscosity of both resins reduced significantly. 260

Meanwhile, the gel time of the bio-based resins extend to more than 100 min, a processability comparable to industrial products. Glass fiber-reinforced epoxy composites with a fiber content of 75 wt % were prepared with all formulated resins by a vacuum infusion process, and the thermomechanical properties were compared by dynamic mechanical analysis as shown in Figure 11. At 25 °C, the DGEDP-ethyl/15 wt % GE composite displays an 11% higher storage modulus than the Hexion composites, and both bio-based resin systems have higher storage moduli than DGEBA. Besides, composites fabricated with DGEDP-based resin systems also exhibit better flexural modulus and strength than Hexion and DGEBA/15 wt % GE composites as listed in Table 2. The better performance of DGEDP-based systems on mechanical performance compared with BPA-based systems may be due to the interactions of the extra ester moieties with glass fibers rendering better interfacial adhesion between the matrix and fibers.

Figure 11. Storage modulus (left) and Tan(δ) as a function of temperature (right) for Hexion, DGEBA/15 wt % GE and DGEDP/15 wt % GE epoxy/glass fiber composites. Reproduced from reference (14). Copyright 2017 Elsevier.

Table 2. Flexural Properties of Epoxy/Glass Fiber Composites (14) Sample

Strength at break (MPa)

Strain at break (%)

Modulus (GPa)

Hexion

324.8 ± 23.3

2.5 ± 0.2

14.76 ± 0.32

DGEBA + GE

334.8 ± 18.1

2.7 ± 0.3

14.18 ± 0.41

DGEDP-ethyl + GE

356.9 ±11.8

2.6 ± 0.3

15.47 ± 0.29

DGEDP-pentyl + GE

350.0 ± 17.4

2.5 ± 0.1

15.78 ± 0.36

SEM images (Figure 12) of those glass fiber-reinforced composite fracture surfaces confirmed the good adhesion between DGEDP-based resins and glass fibers. There is no fiber pullout from the matrix and delamination between fibers. 261

Figure 12. SEM images of fracture surfaces for epoxy/glass fiber composites. Reproduced from reference (14). Copyright 2017 Elsevier.

Conclusions The engineering application of bio-based epoxy resins is hindered by a very narrow processing window due to the system’s high viscosity. The bio-based reactive diluent GE demonstrated good performance on improving the processability of bio-based DGEDP epoxy resins. The new formulation displays suitable viscosity and gelation time for composite vacuum-infusion processing while maintaining a comparable mechanical performance with petroleum-based systems. The successful example of DGEDP-based epoxy resin systems demonstrates the feasibility of replacing petroleum-based epoxy resin systems to apply as a matrix for fiber-reinforced polymer composites.

References 1. 2. 3. 4. 5.

Chen, M. Y.; Ike, M.; Fujita, M. Environ Toxicol. 2002, 17, 80–86. Maffini, M. V.; Rubin, B. S.; Sonnenschein, C.; Soto, A. M. Mol. Cell. 2006, 254–255, 179–186. Sudheer, K.; Sushanta, K. S.; Smita, M.; Sanjay, K. N. Ind. Eng. Chem. Res. 2017, 56, 687–698. Pan, X.; Sengupta, P.; Webster, D. C. Green Chem. 2011, 13, 965–975. Aouf, C.; Nouailhas, H.; Fache, M.; Caillol, S.; Boutevin, B.; Fulcrand, H. Eur. Polym. J. 2013, 49, 1185–1195. 262

6. 7. 8. 9. 10. 11. 12. 13. 14.

Auvergne, R.; Caillol, S.; David, G.; Boutevin, B.; Pascault, J. P. Chem. Rev. 2014, 114, 1082–1115. Yang, G.; Rohde, B. J.; Tesefay, H.; Robertson, M. L. ACS Sustainable Chem. Eng. 2016, 4, 6524–6533. Maiorana, A.; Spinella, S.; Gross, R. A. Biomacromolecules. 2015, 16, 1021–1031. Parker, P. H. U.S. Patent 3488404, 1967. Hove, C. L. F. D.; Heymans, D.; Steinbrecher, C.; Kotlweska, A.; Sand, R. V. U.S. Patent 2014/0316030 A1, 2014. Maiorana, A.; Yue, L.; Manas-Zloczower, I.; Gross, R. J. Appl. Polym. Sci. 2016, 133, 43635. Lugemwa, F. N. Molecules 2012, 17, 9274–9282. Mehmet, A.; Ertas, M.; Nitz, S.; Kollmannsberger, H. BioResources 2007, 2, 265–269. Yue, L.; Maiorana, A.; Patel, A.; Gross, R.; Manas-Zloczower, I. Composites,. Part A: Appl. Sci. Manuf. 2017, 100, 269–274.

263

Chapter 17

Strategic Assemblies of Modified Xylochemicals for New Bio-Based Polymers and Composites Joseph F. Stanzione III,* Elyse A. Baroncini, Alexander W. Bassett, and Silvio Curia Department of Chemical Engineering, Rowan University, 201 Mullica Hill Road, Glassboro, New Jersey 08028, United States *E-mail: [email protected].

In this chapter, we present an overview of the ongoing efforts in the Sustainable Materials Research Laboratory at Rowan University related to providing society with alternative (poly)phenolics that are derived from xylochemicals (wood-derived building blocks). The xylochemicals utilized are molecules that we currently obtain, or have the potential to obtain, from lignocellulosic biorefineries, including pulp and paper mills and fermenters. As lignin is known as Earth’s most prevalent aromatic biopolymer, our approach has strategically focused on designing, synthesizing, and testing monomers, oligomers, and resins containing lignin-derived units (such as guaiacyls, coniferyls, and courmaryls), as well as other bio-based molecules (e.g., cardanol and furandicarboxylic acid). Similar to nature, we strive to engineer materials that aid in sustaining life, withstand environmental stress factors, and serve multiple needs. Testing has included spectroscopic, rheological, thermomechanical, thermogravimetric, mechanical, and toxicity measurements to fundamentally understand the structure–processing–property–toxicity relationships of our novel systems. As a result of this work, we are generating a valuable library of xylochemical-based (poly)phenolics and their functionalized derivatives for use in high-performing polymer applications, including coatings, adhesives, composites, and membranes.

© 2018 American Chemical Society

Introduction The replacement of nonrenewable polymers with bio-based alternatives is necessary to avoid undesirable environmental and social costs that are connected to the use of fossil-derived chemicals (1). Historically, this call was initiated by the acknowledgment of the rapid depletion of petroleum reserves and the direct price increases of oil-derived chemicals (2, 3). Currently, these concerns are accompanied by customers and regulatory agencies demanding the development of materials that are environmentally benign, drastically less toxic to human health, and ultimately more sustainable (4, 5). As a result, interest in the sustainable development, manufacture, and commercialization of renewable, bio-based polymers and composites is expected to grow dramatically (1, 4, 5). Industrial analyses and predictions show that by 2030 both biodegradable and nonbiodegradable bio-based plastics will be widely used and that the renewable plastics market will account for about $6.2 billion (USD) (6, 7). In the plant kingdom, nature synthesizes and utilizes cellulose, hemiceullose, lignin, triglycerides, and other natural chemicals in complex combinations to sustain life. These materials are characterized by distinct building blocks arranged and connected with appropriate chemical functionality, such that systems they are a part of serve multiple needs related to chemical and microbial resistances, transport, energy, growth, strength, and toughness. Current polymers and polymer composites are derived from nonrenewable resources, and one can argue that these are constrained in their utility due to a limited palette of building blocks, as petroleum does not easily afford a variety of chemical structures or functionality. Additionally, compared to the petrochemical industry, biomass feedstock processing (especially environmentally benign refining and manufacturing techniques) is considered to be in its infancy. Thus, the Sustainable Materials Research Laboratory at Rowan University is dedicated to advancing fundamental and applied science and engineering technology related to polymers and composites research with the goal to utilize nature’s chemistries, both those renewed on an annual basis by the biosphere and those nature has provided us in the form of fossil reserves—preferably renewable resources—to enhance material performance and improve our global sustainability. With our academic, government, and industrial partners, we are achieving this by gaining a comprehensive understanding of the structure–processing–property relationships of our novel polymers and composites. This chapter is dedicated to highlighting our work related to lignin-based polymers. Lignin is a highly abundant natural polymer, receiving much attention in the fields of chemistry and polymer science due to its promise of sustainability. Also attractive is its unique chemistry that can provide variation from the standard petroleum-based chemicals on which we rely (8). When depolymerized, lignin can be processed into a variety of aromatic compounds for use as high-value chemicals and products—a better use for this very abundant source (9). To date, we have successfully utilized lignin-based molecules in the development of methacrylates, epoxy resins, polycarbonates, polyesters, and polymers assembled via thiol-ene chemistries. The following sections provide brief snapshots of these successes 266

with their associated references. All in all, our passion is to create newer, better, and safer materials for society, while keeping green chemistry and engineering metrics in mind, as well as to provide impetus for future scientists and engineers to advance biomass feedstock processing and utilization out of its infancy.

Lignin-Based Methacrylates Methacrylate functional monomers have a wide variety of uses in both thermoplastic and thermosetting applications (10–15). We investigated the use of lignin-based single phenolics (including phenol, guaiacol, and 4-propylguaiacol [4PG]) as platform chemicals to potentially replace styrene (St) as the reactive diluent (RD) in vinyl ester resins (10). St is classified as both a volatile organic compound (VOC) and hazardous air pollutant (HAP) and is suspected to be carcinogenic (16, 17). Due to underlying environmental and human health concerns related to St, less hazardous replacements are being pursued. Phenyl methacrylate (PM), methacrylated guaiacol (MG), methacrylated 4-propylguaiacol (M4PG), and St (shown in Figure 1) were blended with a commercial vinyl ester resin (bismethacryl glycidyl ether of bisphenol A [BPA] epoxy, VE828) in varying weight ratios, and the effect of RD structure on polymer performance was evaluated.

Figure 1. Structures of RDs studied.

The lignin-derived RDs possessed higher viscosities relative to St (0.7 cP at 25 °C) where, in general, the RD viscosity increased with an increase in monomer molecular weight. When blended with a commercial vinyl ester resin and cured, all RDs displayed comparable thermal properties (Table 1). The substituents present on the aromatic rings of the lignin-based diluents have small effects on thermomechanical performance of the cured vinyl ester system, suggesting that PM, MG, and M4PG are all suitable replacements for St as RDs in vinyl ester resins. 267

Table 1. Thermal Properties of Cured Vinyl Ester Resin Systema Sample

Tmax (°C)b

E′ (25 °C) (GPa)c

Tg (°C)d

1:1 VE828:St

423

2.9

135

1:1 VE828:PM

413

3.2

129

1:1 VE828:MG

412

3.5

127

1:1 VE828:M4PG

413

2.7

124

a

Tmax = temperature at maximum decomposition rate, E′ = storage modulus, Tg = glass transition temperature. Data adapted from Bassett et al. (10) b From TGA in N2 at 10 °C min-1. c From DMA at 2 °C min-1. d From DMA at 2 °C min-1 measured at peak maximum of the loss modulus (E″) curve.

In addition to utilizing lignin-based single phenolics in the development of St replacements for vinyl ester resins [with the Epps Research Group at the University of Delaware (UD)], we have investigated the use of a series of methacrylates, including lignin-inspired methacrylates as monomers, in the synthesis of renewable homopolymers and block polymers via controlled reversible addition-fragmentation chain transfer (RAFT) polymerizations (14, 15). Specifically, methacrylated vanillin was incorporated as the rigid block in RAFT polymerized block polymers that also incorporated lauryl methacrylate (a triglyceride-based monomer) as the soft block. The resulting annealed block polymers self-assembled into body-centered, cubic arrays of vanillin-based nanospheres in a poly(lauryl methacrylate) matrix, demonstrating the potential of high bio-based content block polymers for thermoplastic elastomeric applications. Moreover, the Epps Research Group has demonstrated the ability to RAFT polymerize a series of guaiacyl lignin-based methacrylate polymers that possessed tunable properties based on naturally inherent functionalities present on the lignin-based building blocks (namely, the methoxy moieties) (15, 18). These low-pispersity homopolymers exhibited controllable glass transition temperatures (Tgs) that are comparable to, and higher than, polystyrene and poly(methyl methacrylate), as well as desirable thermal stabilities (>100 °C above Tg) (18). More recently, and again in collaboration with UD, a series of lignin-inspired poly(dimethoxyphenol methacrylate) homopolymers and copolymers were synthesized via RAFT polymerization (14). A strategic set of dimethoxyphenyl methacrylate isomers (3,5-dimethoxyphenol-, 2,3-dimethoxyphenol-, 2,4-dimethoxyphenol-, and 2,6-dimethoxyphenol-based methacrylates) were synthesized and utilized as monomers to eludicate the effect of substituent placements on the phenyl ring on the resulting polymer properties and their ultimate manipulability for usages in a variety of applications. These materials possessed well-controlled molecular weights and dispersities, and exhibited thermal oxidative stabilities that were at least 100 °C greater than their respective Tgs. Based on substituent placement, the Tgs of the resulting polymers, including copolymers of the dimethoxyphenol methacrylates, spanned 268

over a 100 °C range [Tg of poly(3,5-dimethoxyphenol methacrylate) = 82 °C and Tg of poly(2,6-dimethoxyphenol methacrylate) = 203 °C] (14). Moreover, the poly(dimethoxyphenol methacrylate)s demonstrated higher-than-anticipated solvent resistances to common organic solvents, including tetrahydrofuran and CHCls, with the ability to tune copolymer solvation based on monomer selections and concentrations.

Bioinspired Epoxy Resins Epoxy resins make up approximately 70% of the thermosetting polymer market with BPA (4,4’-isopropylidenediphenol) serving as the platform chemical for over 85% of epoxy resins produced (19). BPA is a synthetic organic chemical compound that is typically synthesized via an acid catalyzed electrophilic aromatic condensation of phenol and acetone (Figure 2) (20, 21). The stoichiometric ratio for this reaction is 2:1 (phenol:acetone); however, the process for synthesizing BPA uses large excesses of phenol to prevent the formation of higher molecular weight compounds and oligomers. The excess phenol is typically removed via distillation processes and recycled (21, 22). BPA is further purified via recrystallization and other downstream processes (20–22).

Figure 2. Conventional synthesis of BPA.

In 2003, over 1.9 billion pounds of BPA were produced, of which 406 million pounds were utilized to produce epoxy resins (23). BPA is known to be a human endocrine disruptor or xenoestrogen and has been shown to mimic estradiol, the primary female sex hormone, due to its structural similarities (24). BPA mimics estradiol by binding to two estrogen receptors, alpha and beta (ER-α and ER-β), as an agonist (24, 25). The ability of BPA to act as an endocrine disruptor has raised numerous concerns and has initiated several debates due to its use in consumer products (26). Canada became the first country to ban the use of BPA in toys and baby bottles in 2010 and was quickly followed by all European countries in 2011 (24). Bisphenol F (BPF) is occasionally used as a BPA alternative in epoxies and other applications and is synthesized in a similar pathway to BPA; however, BPF synthesis uses formaldehyde instead of acetone. Unfortunately, BPF has been shown to be an endocrine disruptor like BPA and, therefore, is not considered a safer alternative to BPA (27). 269

The properties exhibited by BPA-based polymeric materials are due to the rigidity provided by their aromaticity and bisphenolic structure; therefore, the utilization of biomass alternatives with aromatic content similar to BPA have become sources of interest. In addition, while upholding aromaticity, BPA alternatives are desired to have substituents on the aromatic rings, thus ultimately reducing the toxicity of the resulting compounds (28). We have investigated using lignin (an abundant natural resource) with a complex, three-dimensional, substituted polyphenol structure for the development of BPA alternatives. Lignin, a waste product of the paper and pulp industry, accounts for up to 18–35% by weight of wood with an annual production of up to 50 million tons (29, 30). Ongoing research has shown that the strategic depolymerization of lignin can result in a variety of aromatics and methoxy phenols that can be more easily processed (31). Industrially, vanillin is the most widely produced lignin derivative, and we have utilized its reduced form, vanillyl alcohol (VA), as a platform chemical for the production of bisphenolic analogues (32, 33). The hydroxymethyl group on VA provides the necessary reactivity for phenolic coupling and avoids the use of volatile and carcinogenic molecules (e.g., formaldehyde and acetone). The electrophilic aromatic condensation of VA with guaiacol was performed to produce bisguaiacol (BG) as a mixture of isomers, as shown in Figure 3.

Figure 3. Electrophilic aromatic condensation of VA and guaiacol to produce BG.

The major structural isomer of BG was para-para, based on 1H-NMR analysis (32). BG has structural similarities to that of BPA and BPF; however, BG contains methoxy moieties ortho to both hydroxyls. The methoxy moieties present on BG are anticipated to reduce the estrogenic activity of the compound relative to BPA and BPF. We prepared diglycidyl ether (DGE) of BG (DGE BG) to elucidate the structure–property relations of resulting cured resins relative to that of commercial BPA (Epon 828) and BPF (Epon 862) cured epoxy resins. Additionally, we prepared DGEs of VA (DGE VA) and gastrodigenin (DGE Gd) to study the influence of the methoxy substituent on polymer properties. All epoxies were cured with stoichiometric equivalents of a commercial cycloaliphatic diamine [Amicure PACM, (4,4′-methylenebiscyclohexanamine)]. Thermal analyses of the cured epoxy-amine thermosets showed that DGE BG containing thermoset had a lower Tg and thermal resistance relative to those containing Epon 828 and 862; however, it possessed a significantly higher storage modulus (E′) at 25 °C (Table 2) (32). 270

Table 2. TGA and DMA Results of Epoxy Resins Cured with Amicure PACM. Data adapted from Hernandez et al. (32) Tmax (°C)a

E′ (25 °C) (GPa)b

Tg (°C)c

DGE VA

352

3.02

100

DGE GD

354

2.29

123

DGE BG

376

3.35

104

Epon 862

388

2.57

131

Epon 828

392

Sample

a

min-1.

2.37 b

From TGA in N2 at 10 °C From DMA at 2 °C min-1 measured at peak maximum of the E″ curve.

149 min-1.

c

From DMA at 2 °C

The methoxy substituents present on BG were shown to lower the Tg and increase the E′ in epoxy-amine systems, and this was further confirmed via the comparison of cured resin systems containing DGE VA and DGE Gd. Nonetheless, DGE BG displayed promising properties, inspiring the use of BG in other materials applications.

Lignin-Derived Thermoplastics In the last two decades, much research has focussed on the synthesis of bio-based monomers for thermoplastics (2, 4). As aromatic groups generally impart heat resistance, rigidity, and hydrophobicity to polymers, phenolic compounds derived from lignin represent an undoubtedly valuable alternative to fossil-based chemicals for the preparation of monomers capable of undergoing condensation polymerization (34, 35). Examples of commercially available and widely used thermoplastics containing aromatic units are polycarbonates and (semi)aromatic polyesters. Recent investigations conducted in the Sustainable Materials Research Laboratory at Rowan Univeristy revealed that the hydroxyl groups present in BG can fruitfully be exploited for the synthesis of polycarbonates and polyesters through step-growth polymerizations (36, 37). Polycarbonates were produced by reacting BG with p-nitrophenyl chlorophormate (pNC) in acetonitrile in the presence of triethylamine (TEA) and using 4-dimethylaminopyridine (DMAP) as a catalyst at 70 °C (36). For comparison purposes, polycarbonate analogues containing BPA and BPF were synthesized. Thermal analyses showed that polycarbonates containing BPA were typically characterized by a higher Tg and degradation temperatures, most probably linked to the absence of isomer distribution and methoxy substituents. Nonetheless, the BG-containing polycarbonate showed an adequate thermal resistance and Tg (~ 100 °C) suitable for moderate to high temperature applications (Table 3). 271

Table 3. TGA and DSC Results of BG, BPA, and BPF Polycarbonates. Data adapted from Mauck et al. (36) Tmax (°C)a

Char Content (%)a

Tg (°C)b

BG–PC

450

24

100

BPA–PC

545

23

150

BPF–PC

501

Sample

a

From TGA in N2 at 10 °C

min-1.

40 b

From DSC in N2 at 10 °C

119 min-1.

Furthermore, aromatic and semiaromatic polyesters were produced using BG together with different activated diacids (Figure 4) (37). The polymerization reactions were conducted in a biphasic system dichloromethane/NaOHaq in the presence of a phase transfer catalyst.

Figure 4. BG containing semi and fully aromatic polyesters. (a) Poly(BG adipate), (b) poly(BG succinate), (c) poly(BG 2,5-furandicarboxylate), (d) poly(BG terephthalate). Adapted with permission from reference (37). Copyright 2018 Wiley.

Polyesters containing aliphatic and aromatic comonomers were successfully obtained in good yield, and thermal analyses showed remarkable thermal resistance and tunable Tgs ranging from 40 °C up to ~160 °C. As expected, aliphatic comonomers afforded more flexible polymers, with Tg values of 42 °C and 75 °C for poly(BG adipate) and poly(BG succinate), respectively. On the other hand, poly(BG furandicarboxylate) and poly(BG terephthalate) showed much higher Tgs, with values approaching those typical of engineering thermoplastics (Figure 5). 272

Figure 5. DSC traces of poly(BG furandicarboxylate) and poly(BG terephthalate). Adapted with permission from reference (37). Copyright 2018 Wiley. Similarly, TGA analyses showed excellent thermal stabilities with maximum degradation temperature of 318 °C for poly(BG adipate) and around 400 °C for poly(BG succinate), poly(BG furandicarboxylate), and poly(BG terephthalate). Additionally, solubility studies showed that these four polyesters were easily solublized in a range of organic solvents commonly employed for polymer processing and modification, which undoubtedly imparts an added value compared to the incumbent, petroleum-derived, analogous (semi)aromatic polyesters. Finally, biodegradation studies conducted in collaboration with the Austrian Centre of Industrial Biotechnology (ACIB) and the University of Natural Resources and Life Sciences (BOKU) showed that enzymes from the compost-derived bacteria Thermobifida cellulosilytica led to the release of building blocks and, hence, effectively hydrolyzed the polymer chains. These investigations confirmed that BG can be used to prepare useful thermoplastics with interesting thermal attributes, and that BG-based polyesters are susceptible to enzymatic hydrolysis while retaining high thermal resistance and rigidity.

Xylochemicals and Thiol-ene Polymerizations While some members of the scientific community have focused on using bio-based sources to achieve unique and sustainable materials, others have turned to exploring underutilized chemistries as a way to introduce novelty and environmentalism to advanced applications. The resurgence in research centered around thiol chemistries is one such area of study. Specifically, thiol-ene polymerizations have seen much use because of their unique advantages, including reduced inhibition due to oxygen, delayed gelation point, rapid rate, and high yield (38–40). Additionally, thiol-ene polymerizations can be photo-intiated, leading to fast, solvent-free, environmentally-friendly manufacturing. 273

Combining the use of bio-based sources with thiol-ene polymer synthesis is a logical step already taken by some researchers; and it is in this arena that the following section aims to focus (41, 42). The resulting thermomechanical properties of our polymers are described, as well as one specific application we have explored. However, by reporting our data and observations, we hope others are able to apply bio-based thiol-ene polymers to their research; thereby, increasing the versatility both of bio-based sustainable resources and of thiol-related chemistries. Lignin-Derived Single Phenolics As seen in the previous sections, vanillin, a commercially available phenolic produced from lignin, can easily be transformed into other advantageous single phenolic compounds (43). When used in tandem with thiol-ene chemistry, these single phenolics can impart structural and thermal stability to typically flexible thiol-ene polymer networks. To test this, we functionalized VA, which can be derived from vanillin (43), with allyl groups to make di-allylated VA (DAVA), which can then be used in thiol-ene polymerizations (44). To see if the methoxy group on the ring of VA had any effect on thermal or mechanical properties, we did the same with gastrodigenin (DAGd), which can also be derived from lignin (see Figure 6) (44).

Figure 6. Conversion of aromatic lignin-derived compounds into allylated form. Adapted with permission from reference (44). Copyright 2018 Elsevier. Thiols are commercially available in a wide variety of sizes and shapes. Three different thiols were explored in this work, as can be seen in Figure 7. Trifunctional thiols ETTMP 700 and ETTMP 1300, with molecular weights of 700 g/mol and 1300 g/mol, respectively, were chosen in addition to tetra-functional PETMP, which has a lower molecular weight of 488.66 g/mol. The DAVA and 274

DAGd monomers were polymerized with each of the three thiol compounds individually. Dynamic mechanical analysis results of the gel-polymers are shown in Table 4.

Figure 7. Trifunctional thiols ETTMP 700 and ETTMP 1300 (top) and tetrafunctional thiol PETMP (bottom). Adapted with permission from reference (44). Copyright 2018 Elsevier.

Table 4. Thermomechanical and Other Properties of the Six Thiol-ene Gel-Polymers. Data adapted from Baroncini et al. (44) Tg (°C)b

E′ (MPa)c

ρ (g/cm3)d

ν (mmol/cm3)e

DAVA + ETTMP 1300

-37.1 ± 0.4

1.90 ± 0.2

1.15 ± 0.01

0.26 ± 0.03

DAVA + ETTMP 700

-22.7 ± 0.6

4.36 ± 0.68

1.19 ± 0.01

0.60 ± 0.10

DAVA + PETMP

-5.0 ± 2.2

4.93 ± 0.75

1.25 ± 0.02

0.66 ± 0.09

DAGd + ETTMP 1300

-42.1 ± 0.8

1.56 ± 0.16

1.15 ± 0.00

0.21 ± 0.02

DAGd + ETTMP 700

-29.2 ± 1.1

5.08 ± 0.91

1.17 ± 0.00

0.68 ± 0.13

0.3 ± 1.5

10.08 ± 0.64

1.24 ± 0.00

Gel-Polymera

DAGd + PETMP a

1.36 ± 0.99 b

Cured with 2 wt % of total ene + thiol weight of photo-initiator. Tg measured as temperature at peak of loss modulus. c E’ measured at 25 ˚C. d ρ measured according to Archimedes’ principle. e ν measured at point where storage modulus began to increase from ν = E’/3RT.

As can be seen by the Tg, E′, and crosslink density (ν) values, these are very flexible gel-polymers. The gel-polymers containing the highest molecular weight thiol, ETTMP 1300, had the lowest Tgs and were the most flexible. Gel-polymers 275

are crosslinked polymer networks with applications in a wide variety of industries, including use as electrolytes in batteries. In a battery, these gel-polymer electrolytes (GPEs) act both as a separator between the anode and the cathode but also as the electrolyte that allows for ionic flow between the electrodes. The gel-polymers made from xylochemical derivatives proved interesting for this application because of the structural stability of the polymers imparted by the aromaticity of the lignin-derived monomers. Previous gel-polymers have often showed lack of mechanical integrity (45). Additionally, the thiols chosen for the polymerization had potentially conductivity-enhancing features (e.g., mobile chains with repeating ethoxy units) (46). The gel-polymers were imparted with conducting ability by swelling each in a liquid electrolyte solution (see Table 5) (44). The highly flexible ETTMP 1300 containing polymers did swell the most, but also exhibited the highest conductivity values even when not swollen to maximum. Although successful, the results revealed the need for slightly higher conductivity values for industrial application. We believe better performance can be achieved through incorporation of more aromaticity, selection of novel thiols, and alternate methods of incorporating the conducting lithium salt. However, the study sheds light on the effect of aromatic substituents on the resulting GPEs, as well as the influence of thiol size and functionality on conducting abilties, giving our group a path forward for future battery applications of xylochemical-derived polymers.

Table 5. Thermomechanical and Other Properties of the Six Thiol-ene Gel-Polymers. Data adapted from Baroncini et al. (44) Normalized Swelling Ratio at Equilibrium (gsolvent/ggel) in EC-DECb

Conductivity (× 107 S/cm)c

DAVA + ETTMP 1300

4.14 ± 0.21

92.01 ± 18.95

DAVA + ETTMP 700

1.37 ± 0.01

9.28 ± 3.16

DAVA + PETMP

0.75 ± 0.03

7.66 ± 6.59

DAGd + ETTMP 1300

3.56 ± 0.22

102.73 ± 33.28

DAGd + ETTMP 700

1.09 ± 0.05

18.72 ± 6.27

DAGd + PETMP

0.47 ± 0.07

7.04 ± 3.06

Gel-Polymera

a

Cured with 2 wt % of total ene + thiol weight of photo-initiator. b Swelling measured in 1:1 (w:w) of ethylene carbonate (EC) diethyl carbonate (DEC) solution; c Conductivity measured of polymers swollen to 80% swelling capacity in 1 M LiPF6 in 1:1 (w:w) EC-DEC via 4 point probe method.

276

Conclusions Lignin is an essential biopolymer and, to date, is highly underutilized. As demonstrated in this chapter and that which has been reported in the literature, as well as company communications, government reports, and the media, the valorization of lignin as a viable renewable resource in the development of sustainable polymers, and with continued progress, composites is alive and thriving—and rightfully so. The strategic depolymerization of lignin in next-generation lignocellulosic biorefineries will play a significant role both in the success of the biomass to advance the materials industry and in the sustainable development of society. One may argue that our successes, described above, are micro- or even nano-steps toward progressing advanced biomass feedstock processing and utilization out of its infancy. Nonetheless, these are significant steps forward and we hope that they will encourage others to take similar and even bigger strides.

Acknowledgments The authors gratefully acknowledge the U.S Army Research Laboratory for financial support through Cooperative Agreements W911NF-06-2-001, W911NF-14-2-0086, and W911NF-16-2-0225 and the Strategic Environmental Research and Development Program (SERDP) WP-1758. In addition, J.F.S. and A.W.B. wish to acknowledge the NSF CHE-1507010 for funding support for the collaborative efforts with the Epps Research Group at the University of Delaware. Moreover, E.A.B. would like to acknowledge the Department of Defense Science, Mathematics And Research for Transformation (SMART) Scholarship Program. Lastly, J.F.S. would like to personally thank Richard P. Wool, Ph.D., Professor, FRSC, Dr. John J. La Scala, Rowan University, and the Sustainable Materials Research Laboratory and all of its collaborators for all their open-mindedness, dedication, inspiration, and hard work. Without these folks, this work would not have been possible. For this, J.F.S. is truly grateful!

References 1. 2. 3. 4. 5. 6. 7.

Pellis, A.; Herrero Acero, E.; Gardossi, L.; Ferrario, V.; Guebitz, G. M. Polym. Int. 2016, 65, 861–871. Pion, F.; Ducrot, P.-H.; Allais, F. Macromol. Chem. Phys. 2014, 215, 431–439. Frech, C. B. J. Chem. Educ. 2002, 79, 1072. Belgacem, M. N.; Gandini, A. Monomers, Polymers and Composites from Renewable Resources; Elsevier Science: Oxford, 2011. Smith, P. B.; Gross, R. B. Biobased Monomers, Polymers, and Materials; American Chemical Society: Washington, DC, 2013. de Jong, E.; Higson, A.; Walsh, P.; Wellish, M. Bio-based Chemicals: Value Added Products from Biorefineries; 2012. BIO-TIC. Overcoming hurdles for innovation in industrial biotechnology in Europe. Non-technological Roadmap: Draft 3; European Union: 2015. 277

8.

9. 10. 11. 12. 13. 14. 15. 16. 17.

18. 19. 20. 21. 22. 23.

24. 25.

26. 27.

28. 29.

Graichen, F. H. M.; Grigsby, W. J.; Hill, S. J.; Raymond, L. G.; Sanglard, M.; Smith, D. A.; Thorlby, G. J.; Torr, K. M.; Warnes, J. M. Ind. Crops Prod. 2017, 106, 74–85. Custodis, V. B. F.; Bahrle, C.; Vogel, F.; van Bokhoven, J. A. J. Anal. Appl. Pyrolysis 2015, 115, 214–223. Bassett, A. W.; Rogers, D. P.; Sadler, J. M.; La Scala, J. J.; Wool, R. P.; Stanzione, J. F. J. Appl. Polym. Sci. 2016, 133. Bassett, A. W.; La Scala, J. J.; Stanzione, J. F. J. Appl. Polym. Sci. 2016, 133. Stanzione, J. F.; Sadler, J. M.; La Scala, J. J.; Reno, K. H.; Wool, R. P. Green Chem. 2012, 14, 2346–2352. Stanzione, J. F.; Sadler, J. M.; La Scala, J. J.; Wool, R. P. ChemSusChem 2012, 5, 1291–1297. Wang, S.; Bassett, A. W.; Wieber, G. V.; Stanzione, J. F.; Epps, T. H. ACS Macro Lett. 2017, 6, 802–807. Holmberg, A. L.; Stanzione, J. F.; Wool, R. P.; Epps, T. H. ACS Sus. Chem. Eng. 2014, 2, 569–573. Sumner, S. J.; Fennell, T. R. Crit. Rev. Toxicol. 1994, 24, S11–S33. National Emissions Standards for Hazardous Air Pollutants: Reinforced Plastic Composites Production. Environmental Protection Agency. Federal Register: 2003; pp 19375–19443. Holmberg, A. L.; Nguyen, N. A.; Karavolias, M. G.; Reno, K. H.; Wool, R. P.; Epps, T. H. Macromolecules 2016, 49, 1286–1295. Auvergne, R.; Caillol, S.; David, G.; Boutevin, B.; Pascault, J. P. Chem. Rev. 2014, 114, 1082–1115. Dugan, G. F.; Widiger, J. A. H. Process for purifying 4, 4′-isopropylidenediphenol. US3326986, 1967. Neagu, L. Synthesis of Bisphenol A with Heterogeneous Catalysts. Masters Thesis, Queens University, Canada, 1998. Zhang, W.; Li, Y.; He, M.; Chen, Q. Huagong Jinzhan 2007, 26, 1032–1035. NTP-CERHR Expert Panel Report on the Reproductive and Developmental Toxcity of Bisphenol A; National Toxicology Program: Research Triangle Park, NC, 2007. Rogers, J. A.; Metz, L.; Yong, V. W. Mol. Immunol. 2013, 53, 421–430. European Union Summary Risk Assessment Report - 4,4′isopropylidenediphenol (Bisphenol A); European Chemicals Bureau: United Kingdom, 2003. Vandenberg, L. N.; Hauser, R.; Marcus, M.; Olea, N.; Welshons, W. V. Reprod. Toxicol. 2007, 24, 139–177. Eladak, S.; Grisin, T.; Moison, D.; Guerquin, M.-J.; N’Tumba-Byn, T.; Pozzi-Gaudin, S.; Benachi, A.; Livera, G.; Rouiller-Fabre, V.; Habert, R. Fertil. Steril. 2015, 103, 11–21. Hong, H.; Harvey, B.; Palmese, G.; Stanzione, J.; Ng, H.; Sakkiah, S.; Tong, W.; Sadler, J. Int. J. Environ. Res. Public Health 2016, 13, 705–721. Saito, T.; Brown, R. H.; Hunt, M. A.; Pickel, D. L.; Pickel, J. M.; Messman, J. M.; Baker, F. S.; Keller, M.; Naskar, A. K. Green Chem. 2012, 14, 3295–3303. 278

30. Thielemans, W. Lignin and carbon nanotube utilization in bio-based composites. Ph.D. Dissertation, University of Delaware, 2004. 31. Shen, D. K.; Gu, S.; Luo, K. H.; Wang, S. R.; Fang, M. X. Bioresour. Technol. 2010, 101, 6136–6146. 32. Hernandez, E. D.; Bassett, A. W.; Sadler, J. M.; La Scala, J. J.; Stanzione, J. F. ACS Sus. Chem. Eng. 2016, 4, 4328–4339. 33. Araujo, J. D. P.; Grande, C. A.; Rodrigues, A. E. Chem. Eng. Res. Des. 2010, 88, 1024–1032. 34. Llevot, A.; Grau, E.; Carlotti, S.; Grelier, S.; Cramail, H. Polym. Chem. 2015, 6, 6058–6066. 35. Llevot, A.; Grau, E.; Carlotti, S.; Grelier, S.; Cramail, H. Macromol. Rapid Commun. 2016, 37, 9–28. 36. Mauck, J. R.; Bassett, A. W.; Sadler, J.; La Scala, J. J.; Napadensky, E.; Reno, K. H.; Stanzione, J. F. J. Biobased Mater. Bioenergy 2018, 12, 482–492. 37. Curia, S.; Biundo, A.; Fischer, I.; Braunschmid, V.; Guebitz, G. M.; Stanzione, J. F. ChemSusChem 2018, 11, 2529–2539. 38. Hoyle, C. E.; Lee, T. Y.; Roper, T. J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 5301–5338. 39. Hoyle, C. E.; Lowe, A. B.; Bowman, C. N. Chem. Soc. Rev. 2010, 39, 1355–1387. 40. Lowe, A. B. Polym. Chem. 2014, 5, 4820–4870. 41. Uemura, Y.; Shimasaki, T.; Teramoto, N.; Shibata, M. J. Polym. Res. 2016, 23, 215–225. 42. Trita, A. S.; Over, L. C.; Pollini, J.; Baader, S.; Riegsinger, S.; Meier, M. A. R.; Goossen, L. J. Green Chem. 2017, 19, 3051–3060. 43. Fache, M.; Auvergne, R.; Boutevin, B.; Caillol, S. Eur. Polym. J. 2015, 67, 527–538. 44. Baroncini, E. A.; Stanzione, J. F. Int. J. Biol. Macromol. 2018, 113, 1041–1051. 45. Cheng, X. L.; Pan, J.; Zhao, Y.; Liao, M.; Peng, H. S. Adv. Energy Mater. 2018, 8, 1702184. 46. Willgert, M.; Kjell, M. H.; Lindbergh, G.; Johansson, M. Solid State Ionics 2013, 236, 22–29.

279

Chapter 18

Enhancing the Sustainability of High-Performance Fiber Composites Christopher N. Kuncho,1 Wenhao Liu,1 Johannes Möller,1 Julia Kammleiter,3 Julia Stehle,3 Akshay Kokil,1 Emmanuelle Reynaud,*,2 and Daniel F. Schmidt1,4 1Department

of Plastics Engineering and University of Massachusetts Lowell, One University Avenue, Lowell, Massachusetts 01854, United States 2Department of Mechanical Engineering, University of Massachusetts Lowell, One University Avenue, Lowell, Massachusetts 01854, United States 3Department of Mechanical and Plastics Engineering, Darmstadt University of Applied Sciences, Haardtring 100, 64295 Darmstadt, Germany 4Current address: Department of Materials Research and Technology, Luxembourg Institute of Science and Technology, 5, rue Bommel, Z.A.E. Robert Steichen, 4940 Hautcharage, Luxembourg *E-mail: [email protected].

This chapter provides an overview of our recent efforts to provide new, bio-based options for the sourcing of structural epoxies and to introduce reworkability into these systems, both in the neat form and as glass fiber composites. In the first section, the work focuses on understanding how to formulate anhydride-cured epoxidized linseed oil (ELO) for maximum mechanical performance. The second section describes the adaptation of the bio-based epoxy resin system to the preparation of glass fiber composites via vacuum-assisted resin transfer molding (VARTM). The third section summarizes efforts to realize reworkability in both conventional and bio-based epoxy resins through the introduction of transesterification catalysts. Together, these efforts enable the formation of high-performance bio-based epoxy/glass fiber composites. These composites may be mechanically recycled by grinding and hot-pressing and chemically recycled by fiber reclamation and further infusion.

© 2018 American Chemical Society

Introduction Thanks to an attractive combination of performance characteristics, epoxies can be used in a range of demanding applications including coatings, adhesives, binders, and encapsulants. They are the preferred matrix materials for high-performance long-fiber composites, which are increasingly integrated into complex applications (1–3). With the growing use of such epoxy composites, the question becomes sustainability of these materials. The main issues are twofold: (1) conventional high-performance epoxy resins are generally petroleum-derived (4), and (2) the end-of-life options for epoxy/long-fiber composites are limited, especially from the standpoint of resource recovery and reuse. This is particularly important in the context of recent predictions concerning wind turbine blade waste, for example (5). With enhancement of sustainability in mind, the goals of our recent efforts in the domain of high-performance long-fiber composites have been to: (1) diversify the available sources of high-performance epoxies by identifying optimal compositions derived from a renewable base; (2) develop approaches to more effectively reuse/recycle the components of epoxy/long-fiber composites; and (3) combine the aforementioned approaches to more effectively address the overall life cycle of these materials. This chapter is an overview of these parallel and convergent efforts. The first section describes the approach followed to optimize the thermomechanical properties of a bio-based epoxy matrix, benchmarked in terms of hardness, modulus, and main relaxation temperature vs. a commercial epoxy with excellent properties. The second section reviews findings from the preparation of long-fiber composites based on the aforementioned bio-based and commercial epoxies, and compares their properties. The third section focuses on recyclability from the standpoint of vitrimer formation: Metal transesterification catalysts are introduced into the commercial resin and their effect on high-temperature reprocessing is evaluated. The concept is then translated to composites, based on both the commercial and bio-based resins. Vitrimer formation enables recycling both mechanically, thanks to the reworkable nature of the materials, and chemically, via wholesale reversion of the cross-links, liquefaction of the cured resin, and recovery of the intact reinforcing fabric. An overarching theme identified through these parallel efforts is the importance of matching the resin formulation with the desired performance characteristics (thermomechanical properties, processability, recyclability, etc.).

Epoxidized Linseed Oil as Renewable Resin Numerous research groups have worked to expand the selection of available epoxy resins with sustainability in mind (4, 6). As in the field of bio-based polymers more generally, however, industrial adoption of such technologies is often limited by concerns over cost, performance and consistency. With this in mind, the following criteria were applied in the context of base resin selection: 282

• •

The resin should be commercially available in large quantities and are of good quality and consistency. The feedstock for the resin should be affordable, and the number of chemical steps involved should be minimal.

In this context, epoxidized vegetable oils (EVOs) appear quite attractive, as its production and subsequent conversion to fully cured epoxies involve only two steps—epoxidation and curing. Among readily available EVOs, epoxidized linseed oil (ELO) was identified as the best choice for this work (7, 8). Linseed oil is produced from a nonfood crop and epoxidized via a clean, efficient process, minimizing its initial cost. It is readily available in industrial quantities from multiple suppliers. It has a lower viscosity than conventional epoxy resins (less than 1000 mPa·s, favoring effective infiltration during composite processing). Its functionally f is roughly 6, favoring the formation of a rigid, highly cross-linkable network appropriate for structural applications, and its epoxy equivalent weight (EEW) is between 170 and 180, very similar to conventional epoxy resins. It is minimally toxic and is approved by the U.S. Food and Drug Administration for food-contact applications. The biggest disadvantage of this and other EVOs is their inherently low reactivity, given that they contain only secondary epoxy groups. The results presented herein are based on Epoxol 9-5 from American Chemical Services with an oxirane oxygen content of 9.57%. Control specimens were based on the Momentive EPIKOTE MGS 145 system (consisting of RIMR 145 resin, RIMH 145 hardener, and RIMC 145 catalyst), an anhydride-cured conventional epoxy resin provided by Momentive, German Lloyd approved and aimed at wind energy applications. While the initial focus was on amine curatives (9), liquid anhydrides were preferred, given their high performance, lack of side reactions, and ease of incorporation (10, 11). After an initial screening of all readily available liquid anhydrides typically used in epoxy formulation, MTHPA (Huntsman Aradur 917 US), was identified as the most attractive hardener from the standpoint of stiffness and glass transition temperature. A schematic of the reaction between these two components is shown in Figure 1.

Figure 1. Schematic representation of the reaction between representative isomers of epoxidized linseed oil (ELO) and methyltetrahydrophthalic anhydride (MTHPA), with the structure of the resultant network simplified for clarity.

283

After screening a large number of cure catalysts, the focus was reduced to two: 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU), a liquid at room temperature, and 2-ethyl-4-methylimidazole (2E4MI), which requires preheating prior to introduction. Formulation and process conditions were then optimized to maximize hardness and homogeneity (as tested by Shore D hardness tests). In this section, results are presented from the characterization of two anhydride-cured ELO-based bioepoxy systems: DBU will refer to a stoichiometric ratio of ELO to MTHPA with 4.5 phr of DBU; 2E4MI will refer to a stoichiometric ratio of ELO to MTHPA with 6 phr of 2E4MI. The control refers to the RIM 145 resin, cured according to the manufacturer’s recommendations (12). Dynamic mechanical analysis allowed for a comparison of the temperature and width of the main relaxation (measured based on the location and breadth of the largest loss modulus peak), as presented in Table 1. While the bioepoxies exhibit a main relaxation temperature ~13–20 °C lower than that of the conventional resin, the values obtained are very high for a thermoset based solely on a bio-based epoxy resin and are sufficient for a broad range of structural applications. However, it is noteworthy that the width of the main relaxation is two to three times larger in the bioepoxies than in the conventional system. This highlights a common feature of a number of bio-based polymeric materials, namely, that they are based on complex mixtures of components. Here, this results in a broadening of the main relaxation, something that must be accounted for when applying such systems, given that it is typically the onset of the main relaxation that controls heat distortion behavior.

Table 1. Comparison of the Temperature and Width of the Main Relaxation Between Bio-Based and Conventional Epoxiesa

a

System

Main Relaxation T (°C)

Main Relaxation Width (°C)

DBU

64 ± 3

40

2E4MI

70 ± 3

61

Control

83 ± 4

22

Calculated based on DMA loss modulus data.

In addition to DMA, room temperature tensile testing was performed on all systems; the results are presented in Figure 2. As shown, the bio-based resins both provide a modulus of ~2 GPa and a tensile strength of ~40 MPa. While these values are only ~60% of the conventional resin, it should be emphasized that the latter is an extremely high-performance material that is a product of decades of research and optimization. Additionally, because bio-based resins fail in a nonfragile (i.e., noncatastrophic) manner, their break strains are higher than that of conventional resins. 284

Figure 2. Comparison of the room temperature tensile properties of the bio-based and conventional epoxies.

Figure 3. Integrated fluidity vs. temperature for the bio-based epoxies are shown with a conventional control; integrated fluidity represents the integral of isothermal fluidity (inverse viscosity) vs. time data and is proportional to the volume of material that may be infused under a given set of conditions. The lines represent three parameter exponential fits to guide the eye. 285

Turning now from pure resin performance to composite processing, parallel plate rheometry was used to assess the flow behavior of the resins prior to curing. In order to derive a single numerical metric that would account for the overall viscosity evolution and gelation behavior of the resins under any given set of circumstances, families of isothermal cure curves (viscosity vs. time) were collected, then converted to fluidity (inverse viscosity) vs. time. From Darcy’s Law, the volumetric infusion rate should be proportional to the fluidity. Therefore, integration of data on fluidity vs. time will yield a value directly proportional to the total volume of material that may be infused under a certain set of conditions. Figure 3 presents integrated fluidity values of the bio-based and conventional epoxies as a function of temperature. The curves clearly demonstrate that bioepoxies infuse at least as well as conventional resins under most circumstances—substantially more in some instances. This was demonstrated by the bio-based resins infusing as much as five times faster than the conventional resins, depending on the formulation and process conditions.

Bioepoxy-Based Long-Fiber Composites A vacuum-assisted resin transfer molding (VARTM) process was used to manufacture composite specimens. Unidirectional (UD) stitched E-glass (Saertex 955) was chosen as the reinforcement, with plies of glass fabric stacked on top of one another in the confines of the mold. A vacuum was applied to the mold, causing resin to flow from its pot through the mold, infusing the glass fabric at room temperature. Once the mold was filled, the system was sealed and the mold was transferred into an oven for resin curing. While the DBU-based bioepoxy gives very good performance on its own, excessive voiding was observed in parts based on this formulation during the VARTM process. This was ascribed to a decarboxylation reaction specific to the combination of MTHPA and DBU, as this was not observed in the case of a DBU-catalyzed system where MTHPA had been replaced with nadic methyl anhydride (NMA), nor in systems based on MTHPA in combination with other catalysts. Regardless, this result led to the elimination of the DBU-based bioepoxy formulation from consideration in the context of composite manufacturing. Instead, a stoichiometric ELO/MTHPA system containing 6 phr of 2E4MI as catalyst was favored. A constituent component analysis on cured composites showed that the fiber volume fraction was between 52 and 57% in these samples, the resin fraction was 42 to 46%, and the void fraction was between 0.7 and 1.4%, enabling meaningful properties comparisons. The mechanical behavior of these composites was assessed in flexure, both parallel to the fiber (i.e., axial) direction and in the orthogonal (i.e., transverse) direction. Typical stress–strain curves in the axial direction are presented in Figure 4, whereas Figure 5 presents the flexural modulus and flexural strength. The work of fracture before (solid domain) and after (patterned domain) the first fracture event are presented in Figure 6.

286

Figure 4. Typical flexural stress–strain curves for composites based on the optimized bioepoxy formulation (left) and the control formulation (right), measured in the axial direction and with the work of deformation before (solid) and after (pattern) the first fracture event.

Figure 5. Flexural modulus and flexural strength for the bioepoxy and control composites, in both the axial and transverse directions.

While the transverse modulus of the bio-based composite is somewhat lower than the conventional control, the modulus values are effectively identical in the axial direction. This is logical given the initial property differences between the matrix materials coupled with the fact that the matrix modulus affects the transverse composite modulus much more strongly than the axial composite modulus. In contrast, while the axial strength of the bio-based composite is limited to approximately two-thirds of the conventional composite, the values for transverse strength are effectively identical between the two composites. This implies a lack of efficient stress transfer between the matrix and the fibers in the bio-based composite, suggesting that modifications to fiber sizing could be used to address this issue. 287

Figure 6. Work of deformation before (solid) and after (pattern) the first fracture event for the bioepoxy and control composites, in both the axial and transverse directions.

In sum, the characteristics of the optimized bio-based composite are quite attractive, including substantially faster infusion coupled with a level of mechanical performance, easily sufficient for structural applications, and subject to further enhancements via improvements in fiber sizing. Given an acceptable main relaxation temperature and superior damage tolerance vs. the conventional system resulting in an equivalent work of deformation between the two, these materials show significant promise as the basis for more sustainable composites.

Introducing Reworkability Covalent adaptive networks (13, 14), in which external stimuli can trigger molecular-level reorganization, appear as a platform for addressing end-of-life issues in thermosetting systems. With this in mind, the second goal was to take prior work on epoxy vitrimers (15, 16), validate its applicability in high-performance resins based on industrially relevant components and stoichiometries, and eventually apply it to optimized bio-based systems as well. In service of this goal, an initial screening of metal salt-based transesterification catalysts was performed using the control epoxy (17) by performing constant load creep experiments at elevated temperatures for a fixed amount of time. 288

Based on the observed compressive creep strain values, four different catalysts were selected, and the associated samples were subjected to stress relaxation experiments in a parallel plate geometry at a range of different temperatures. From the resultant isothermal stress relaxation data, activation energies and characteristic relaxation times (τ*) were determined and are presented in Table 2.

Table 2. Activation Energies and Relaxation Times for Selected Transesterification Catalysts in a RIM 145-Based Epoxy Vitrimera

a

Catalyst

EA (kJ/mol)

τ* (120 °C)

τ* (270 °C)

Dibutyltin dilaurate

121 ± 15

120 days

499 s

Dibutyltin diacetate

170 ± 21

25 years

330 s

Dibutyltin bis(2,4-pentanedionate)

114 ± 16

42 days

287 s

n-Butyltris(2-ethylhexanoate)tin

95 ± 8

4 days

78 s

Catalyst loading was 5 mol% vs. moles epoxy groups in all cases.

While all selected catalysts enable significant reworkability, n-butyltris(2ethylhexanoate)tin (Sn_th) was selected as the most efficient. Dibutyltin diacetate (Sn_a) was also investigated for comparison purposes, given its low molecular weight and reduced tendency to plasticize the network. At this stage, two routes of recycling were explored: mechanical recycling, whereby the material was ground into a powder then reformed via hot-pressing, and chemical recycling, where the matrix was transesterified with a low volatility alcohol at elevated temperatures to convert the network to a mixture of liquid esters of known composition. This process enabled reclamation of the intact reinforcing fabric. Results from both approaches are presented below. Plates of unreinforced RIM 145 containing Sn_th as a transesterification catalyst were ground mechanically, and the resulting powder was then passed through sieves of different aperture diameters (5, 2, and 0.2 mm) to collect fractions with different particle sizes. Plates of RIM 145 / glass fiber composites containing Sn_th were ground in an analogous fashion, then passed through a sieve with a 5-mm aperture size. The powders were then consolidated in a dedicated mold at 260 °C using a maximum pressure of 22.5 MPa. Figure 7 presents the appearance of the consolidated samples.

289

Figure 7. Mechanically recycled samples of pristine RIM 145 and its glass fiber composite, using n-butyltris(2-ethylhexanoate)tin (Sn_th) as a transesterification catalyst to enable reworkability in these materials.

While specimens based on coarse powders present visually evident “grain boundaries,” the specimen shown based on the finest powder appears quite homogeneous. Figure 8 and Figure 9 present the results of DMA and room temperature flexural testing of these mechanically recycled samples in comparison with the properties of the pristine cast RIM 145 resin.

Figure 8. DMA-derived room temperature storage moduli and main relaxation temperatures for the pristine cast RIM 145 resin and mechanically recycled specimens of the pristine cast resin and a glass fiber reinforced composite, using n-butyltris(2-ethylhexanoate)tin (Sn_th) as a transesterification catalyst to enable reworkability in these materials. 290

Figure 9. Flexural stress and strain at failure for the pristine cast RIM 145 resin and mechanically recycled specimens of the pristine cast resin and a glass fiber reinforced composite, using n-butyltris(2-ethylhexanoate)tin (Sn_th) as a transesterification catalyst to enable reworkability in these materials.

These analyses show the storage modulus and main relaxation temperature of the resin remain effectively unchanged following mechanical recycling. The grinding procedures do not result in irreversible damage to the network, and the visual inhomogeneities observed following mechanical recycling of samples based on all but the finest powder are inconsequential in this context. In contrast, ultimate properties are more sensitive to the particle size of the material used to produce the mechanically recycled samples. The sample based on the most finely ground powder gave significantly higher flexural stress and strain at failure values than the other mechanically recycled samples; its values were closer to those of the pristine resin. The second process applied was chemical recycling, which consisted of placing a RIM 145-based glass fiber reinforced composite plate containing the Sn_th catalyst into a heated solution of Sn_th in 1-dodecanol for 12 h. This was followed by a rinse in solvent, then in water, to yield the original reinforcing fabric minus the resin, as illustrated in Figure 10. Reinforcing fabric reclaimed in this fashion was then reinfused with conventional epoxy to form a new composite plate. The flexural properties of such reinfused plates are presented in Figure 11, with the equivalent composites based on pristine glass fabric shown for comparison. 291

Figure 10. Visual of a composite plate after chemical recycling.

Figure 11. Comparison of the (axial) flexural modulus and strength of RIM 145/glass fiber composites based on pristine and chemically recycled unidirectional E-glass fabrics, using n-butyltris(2-ethylhexanoate)tin (Sn_th) as the transesterification catalyst for recycling. 292

What this work shows is that within experimental uncertainty, the stiffness of the composite was unaffected by the chemical recycling step. However, the strength of the composite based on recycled glass fabric is only ~60% of the composite based on pristine glass fabric. This suggests that the fiber sizing was modified by the chemical recycling process, resulting in a weaker interface with the recycled glass fabric and the observed reduction in strength. The observation of a significant darkening in color of the recycled glass fabric upon composite formation further hints at chemical-level changes during this process. At the same time, these results also highlight the potential of this approach as a way to simultaneously recover a reinforcing fabric and modify its surface chemistry. The chemical recycling of glass fiber composites based on the bio-based resin (i.e., a stoichiometric ratio of ELO and MTHPA with 6 phr 2E4MI) used anisyl alcohol as the solvent. Both the Sn_th and Sn_a catalysts provided interesting results. Figure 12 presents a comparison of the flexural properties of bio-based composites based on pristine and recovered glass fabrics.

Figure 12. Comparison of the (axial) flexural modulus and strength of bioepoxy/glass fiber composites based on pristine and chemically recycled unidirectional E-glass fabrics, using either n-butyltris(2-ethylhexanoate)tin (Sn_th) or dibutyltin diacetate (Sn_a) as the transesterification catalyst for recycling. In marked contrast with the results obtained using the conventional epoxy resin, here the composites combining the bioepoxy resin with reclaimed glass fabric possess superior flexural properties vs. those based on pristine glass fabric. This may arise for the same reasons as the degraded properties in the conventional resin system, namely via modification in fiber surface chemistry, further highlighting the potential to modify fiber surface during the chemical recycling process. 293

Conclusions Within the framework of sustainable polymeric materials, this chapter highlights the possibility to produce high-performance structural thermosets based solely on ELO as epoxy resin component. In particular, we find that curing of ELO with an appropriate anhydride and catalyst gives final parts that display a high level of performance, both in terms of main relaxation temperature and tensile properties. While proper catalyst selection is even more important for composite processing in an industrially relevant setup, the rheological behavior of the bio-based system studied here is ideal for rapid infusion. In this context, we propose the use of a new metric, termed “integrated fluidity” as a measure of resin infusibility, and show that the value of this metric can be substantially higher for the bioepoxy than the control, a result confirmed by actual infusion times (up to 5 times shorter in practice). The resulting bioepoxy composites provide a level of mechanical performance (as measured y flexural stiffness and strength) easily sufficient for structural applications and competitive with analogous composites based on a highly optimized conventional epoxy resin system. In particular, the bioepoxy composites exhibit a noncatastrophic failure mechanism in flexure, providing a similar work of deformation to the conventional composite but with greater damage tolerance. As optimization of fiber sizing was outside the scope of our efforts, a weak interface may account for the interfacial debonding and lower axial strength observed in the bio-based composite specimens, and represents an area where additional work could further enhance the performance of these materials. Finally, we demonstrate that metal transesterification catalysts may be used to enable material recovery, thus helping to address the end-of-life issues associated with the use of epoxy thermosets and their composites. In the case of mechanical recycling (i.e., grinding and hot-pressing of the resultant powders), stiffness and main relaxation temperature were unaffected, while flexural stress and strain at failure were retained most effectively when the material was finely ground. Chemical recycling enabled the successful reclamation of intact glass fabric which could then be reinfused to form new composites without a loss in composite stiffness in the recycled system. Additionally, the bio-based epoxy matrix experienced significantly more rapid chemical recycling, and bio-based composites based on reclaimed glass fabric were shown to give enhanced properties vs. those based on pristine fabric, further emphasizing the promise of this approach as well as the importance of changes in resin formulation. In sum, these results are highly encouraging as far as the sustainability of structural epoxies and epoxy composites are concerned. Selected areas of particular promise moving forward include the following: the use of EVOs with higher functionality (which would be expected to close the gap in resin properties vs. the conventional control); optimization of fiber sizing (enabling improvements in bio-based composite strength in particular); acceleration of the chemical recycling process (thus enhancing the rate of fiber reclamation); and recovery and reuse of the catalyst and resin components in the context of chemical recycling.

294

Acknowledgments This material is based upon work supported by the National Science Foundation under Grant No. 1230884. The authors thank Prof. Meg Sobkowicz-Kline for her assistance and advice, as well as Hexion/Momentive Performance Materials, Huntsman Advanced Materials, Reaxis, Inc., and PMC Organometallix for supplying the epoxy resin system, the MTHPA, and several catalysts. The authors extend the greatest of thanks to Mr. Stephen Driscoll, Mr. David Rondeau, Mr. Xun Chen, and Mr. Patrick Casey.

References 1. 2. 3.

4. 5. 6. 7. 8. 9. 10. 11. 12.

13. 14. 15. 16. 17.

Mallick, P. K. Fiber-Reinforced Composites: Materials, Manufacturing, and Design, 3rd ed.; CRC Press: Boca Raton, FL 2008. Campbell, F. C. Structural Composite Materials; ASM International, 2010. Hollaway, L. C.; Teng, J. G. Strengthening and Rehabilitation of Civil Infrastructures Using Fibre-Reinforced Polymer (FRP) Composites; Elsevier, 2008. Raquez, J. M.; Deléglise, M.; Lacrampe, M. F.; Krawczak, P. Prog. Polym. Sci. 2010, 35, 487–509. Pu Liu, C.; Barlow, Y. Waste Manage. 2017, 62, 229–240. Lligadas, G.; Ronda, J. C.; Galià, M.; Cádiz, V. Mater. Today 2013, 16, 337–343. Boquillon, N.; Fringant, C. Polymer 2000, 41, 8603–8613. Pin, J. M.; Sbirrazzuoli, N.; Mija, A. ChemSusChem 2015, 8, 1232–1243. Sivasubramanian, S.; Jafferji, K.; Schmidt, D. F.; Reynaud, E. Green Mater. 2014, 2, 2–10. Kuncho, C. N.; Schmidt, D. F.; Reynaud, E. Ind. Eng. Chem. Res. 2017, 56, 2658–2666. Möller, J.; Kuncho, C. N.; Schmidt, D. F.; Reynaud, E. Ind. Eng. Chem. Res. 2017, 56, 2673–2679. Wind Energy Resin Product Bulletin, http://abbymillager.com/wp-content/ uploads/2013/08/WindEnergy_Resin_productbulletin.pdf, 2012 (accessed Aug 25, 2018). Bowman, C. N.; Kloxin, C. J. Ang. Chem., Int. Ed. 2012, 51, 4272–4274. Kloxin, C. J.; Scott, T. F.; Adzima, B. J.; Bowman, C. N. Macromolecules 2010, 43, 2643–2653. Capelot, M.; Montarnal, D.; Tournilhac, F.; Leibler, L. J. Am. Chem. Soc. 2012, 134, 7664–7667. Montarnal, D.; Capelot, M.; Tournilhac, F.; Leibler, L. Science 2011, 334, 965–968. Liu, W.; Schmidt, D. F.; Reynaud, E. Ind. Eng. Chem. Res. 2017, 56, 2667–2672.

295

Bio-Based Monomers and Resulting Products

Chapter 19

Divinylglycol, a Glycerol-Based Monomer: Valorization, Properties, and Applications Léa Bonnot,1,2 Christophe Len,3,4 Etienne Grau,1,2 and Henri Cramail1,2,* 1Laboratoire

de Chimie des Polymères Organiques (LCPO), UMR 5629, Bordeaux INP/ENSCBP, University of Bordeaux, F-33607 Pessac Cedex, France 2Laboratoire de Chimie des Polymères Organiques (LCPO), UMR 5629, Centre National de la Recherche Scientifique (CNRS), F-33607 Pessac Cedex, France 3Centre de Recherches de Royallieu, CS 60319, Université de Technologie de Compiègne, Sorbonne Universités, F-60203 Compiègne Cedex, France 4Institut de Recherche de Chimie Paris, Centre National de la Recherche Scientifique (CNRS), Chimie ParisTech, PSL Research University, F-75231 Paris Cedex 05, France *E-mail: [email protected].

In the context of the development of bio-refineries, glycerol and its derivatives are co-products of the oleochemistry for which new valorization routes must be found. In this chapter, the polymerizability of divinylglycol (DVG), a symmetrical C-6 glycerol derivative that bears a vicinal diol and two vinyl functions was investigated. The reactivity of the hydroxyl and vinyl functions of DVG through polycondensation and polyaddition reactions was evaluated. First, the synthesis of polyesters was carried out by reaction of DVG with various biosourced diesters. Second, DVG was polymerized through its vinyl functions by acyclic diene metathesis polymerization and thiol-ene addition. Finally, three-dimensional epoxy–amine networks were prepared from a series of diamines and bis-epoxidized DVG, the latter being prepared by oxidation of the DVG double bonds. These different polymerization © 2018 American Chemical Society

reactions showed that the DVG double bonds were more reactive than the alcohol bonds and that a panel of original polymers could be obtained from this bio-sourced synthon.

Introduction Polymers are ubiquitous in our daily lives; the automotive, packaging, health, and textile sectors, to name just a few, are meaningful sources of polymers. With the depletion of fossil resources and increasing environmental concerns, the search for more sustainable solutions is becoming a necessity. Since the beginning of the 21st century, bio-sourced chemistry has been expanding widely with the valorization of biomass, which is an abundant source of carbon structures (1). In this regard, the concept of a biorefinery was created with the objective of producing feed, food, fuels, and molecule platforms for industry (2). Vegetable oils are the second most used renewable resource after ligno-cellulosic biomass (3). Vegetable oils are inexpensive and nontoxic, and they have great potential as precursors of bio-based polymers. These oils are composed of triglycerides, which after reaction with an alcohol, water, or a base, will produce three molecules of fatty acids (or fatty esters) and a molecule of glycerol. Fatty acids have been widely studied as precursors of thermoplastic or thermosetting polymers (4, 5). Glycerol is a simple, nontoxic, versatile molecule that provides access to a wide range of high-value bio-based molecules (6). Divinylglycol (DVG), or 1,5-hexadiene-3,4-diol, is a symmetrical C-6 monomer that bears a vicinal diol and two vinyl functions. It can be synthesized from mannitol (7–10) or tartaric acid (11), but this requires protection/deprotection steps and has a rather low yield of 20 to 52%. Another route to DVG synthesis is by a pinacol coupling of acrolein and this has a yield of 90% (Scheme 1) (12, 13). Acrolein can be synthesized from glycerol, but acrolein is toxic; however, Len and colleagues have developed an eco-responsible direct synthesis of DVG from glycerol (7). Divinylglycol is used in the total synthesis of several analogs of natural molecules (14–16). As a polymer precursor, DVG is actually used as cross-linker for thermosets (e.g., Noveon® Polycarbophil from Lubrizol) (17–19). However, data concerning the reactivity of DVG and the physicochemical properties of polymeric materials are scarce in the literature (20–26). In this chapter, we investigate the reactivity of DVG (mixture of d/l and meso) with respect to chain-growth and step-wise growth polymerizations through the synthesis of linear or cross-linked polymers from the two types of available functions of DVG: the hydroxyl and unsaturation functions.

300

Scheme 1. Synthetic routes to DVG.

Divinylglycol Polymerization via Hydroxyl Reactivity: Example of Polyester We first investigated the synthesis of DVG-based polyesters by polytransesterification. DVG has two secondary alcohol functions, and both are known to be of lower reactivity than that of primary alcohols, particularly in the esterification or transesterification reactions (27, 28). Examples of polyester syntheses from short secondary diols can be found in the literature, in particular, 2,3-butanediol, which can be obtained by the fermentation of glycerol. In the 1950s, Watson and colleagues (29) described the synthesis of oligoesters from 2,3-butanediol that exhibited molar masses between 450 and 2600 g/mol. Bio-sourced polyesters based on 2,5-furandicarboxylic acid and 2,3-butanediol have also been synthesized (30). Although the formation of rings was demonstrated, polyesters of relatively high molar masses of up to 7000 g/mol were obtained, depending on the nature of the catalyst used. Avérous and colleagues (31) studied the esterification kinetics between 2,3-butanediol and 1,4-butanediol with adipic acid and described the synthesis of (co)polyesters from this mixture of diols (32). In the present study, the polyesters were prepared in bulk by copolymerization of DVG with dimethylsuccinate (DMSu) or dimethylsebacate (DMSe) as bio-sourced diesters (Scheme 2). Several catalysts were tested, including 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD), sodium methanolate (MeONa), titanium butoxide [Ti(OBu)4], and titanium isopropoxide [Ti(OiPr)4], all of which are commonly used in polyester synthesis (33, 34). 301

Scheme 2. Polytransesterification reaction of DVG with DMSu (x = 1) or DMSe (x = 4).

These experiments show that whichever diester and catalyst are used only oligomers are formed (Table 1). The apparent molar masses measured by SEC do not exceed 1840 g/mol for polyesters obtained with DMSe and 1200 g/mol with DMSu in the presence of TBD, which was the most effective catalyst. Titanium catalysts lead to low conversions of hydroxyl functions of DVG (23 to 50%). These results can be explained by a possible formation of a complex between the diol and the titanium, which prevents the reaction from taking place, as previously reported by Gau and colleagues (35). Nevertheless, various NMR analyses have confirmed the presence of pendant double bonds along the polyester chains. A chemical shift from 5.81 to 5.74 ppm of the peak corresponding to the a′ protons of the double bond of DVG is noticed and c′ protons in alpha of the alcohols are shifted from 4 to 5.35 ppm, confirming the reaction of the alcohol functions (Figure 1). Divinylglycol was further used as a co-monomer (co-diol) with 1,3propanediol (1,3-PD) or 1,12-dodecanediol (1,12-DD) in the course of polyester synthesis; DMSu or DMSe were kept as diesters and TBD as a catalyst (Figure 2). Data are shown in Table A1 in the Appendix. 302

Table 1. Characteristics of Polyesters Synthesized with DVG and DMSe or DMSu as the Diester in the Presence of Different Catalysts Entry

Diester

Mna (g mol–1)

Diester

DVG

Ða

Td5% (°C)

Tg (°C)

Tf (°C)

TBD

100

96

950

2

220

–39

—

MeONa

100

100

1840

2

260

–54

—

Ti(OBu)Ti(OBu)4

41

23

500

1.2

97

—

–40, 0

4

Ti(OiPr)4

50

50

700

1.6

120

—

–37, –16

5

TBD

100

90

1200

1.2

199

–17

—

MeONa

37

84

800

1.8

147

–47

—

Ti(OBu)Ti(OBu)4

b

b

b

b

88

—

—

Ti(OiPr)4

b

b

b

b

88

—

—

1 2 3

6

303

7 8 a

Conversion (%)

Catalyst (10 mol %)

DMSe

DMSu

Measured by SEC in tetrahydrofuran (THF), PS calibration.

b

Insoluble polymer.

Figure 1. 1H NMR in CDCl3 of DVG, DMSe, and polyesters synthesized (Entries 1–4, Table 1): catalyzed by (a) TBD, (b) MeONa, (c) Ti(OBu)4, and (d) Ti(OiPr)4.

Figure 2. Co-monomers used for polytransesterification of DVG. 304

Similar results were obtained with the two diesters; incorporation of DVG is difficult in both cases and does not exceed 67% with DMSu or 58% with DMSe. Moreover, the incorporation of DVG leads to a decrease of the polymers’ molecular masses. The copolyesters synthesized with 1,12-DD are semi-crystalline with a Tg of around room temperature and several melting temperatures, indicating a heterogeneous composition of the chains. With the short co-diol, 1,3-PD, the copolyesters formed were all amorphous with a Tg of between –52 and –40 °C. In conclusion, the synthesis of high-molar-mass polyesters from DVG has proven to be impossible. Nevertheless, the advantage of using this monomer is the provision of pendant double bonds in the copolymers, allowing a potential post-functionalization that was not addressed in this chapter.

Divinylglycol Polymerization via Double Bond Reactivity Acyclic Diene Metathesis Polymerization of DVG Acyclic diene metathesis (ADMET) polymerization follows a step-wise growth mechanism to produce linear polyalcenamers from α,ω-dienes (36). Acyclic diene metathesis proceeds through a transalkylidenation reaction with the release of ethylene, which can be removed by applying a vacuum or a constant flow of an inert gas in order to obtain high conversions and high-molecular-weight polymers (37).

Scheme 3. ADMET polymerization of DVG.

Acyclic diene metathesis polymerization enables the synthesis of polyalcenamers with well-defined architectures and various functions (e.g., alcohol, halogen, amine) that give them unique properties (38, 39). We investigated the reactivity of DVG and its derivatives through ADMET polymerization. Knowing that DVG is a short diene with alcohol functions, we analyzed the possibility of the negative neighboring group effect (NNGE) with catalysts. Wagener and colleagues (40) has shown that close proximity of functional groups strongly reduces the monomer reactivity, presumably due to complexation of the heteroatom’s nonbonded electrons with the metal center. In addition, at higher temperatures, the DVG may be evaporated (Tb = 198 °C under 1 atm). 305

Table 2. Characteristics of poly(DVG) Synthesized by ADMET Polymerization with Different Catalysts Condition

Bulk 24 h, 35 °C Under vacuum

306

a

Catalyst

Quantity of catalyst (mol %)

pa (%)

Mnb (g/mol)

Ðb

Tg (°C)

Td5% (°C)

Xn

Schrock

1

—

—

—

—

—

—

G1

1

—

—

—

—

—

—

G2

1

26

780

1.2

–49

74

1.35

HG1

1

—

—

—

—

—

—

HG2

1

65

300

1.5

–21

126

2.9

HG2

5

68

350

1.1

–10

77

3.1

HG2

0.5

75

890

1.2

36

122

4

Conversion of vinyl group calculated by 1H NMR.

b

Measured by SEC in DMF, LiBr, PS calibration.

The polymerization of DVG by ADMET (Scheme 3) was tested with five different catalysts [i.e., the Schrock catalyst, first- and second-generation Grubbs catalysts (G1 and G2), and Hoveyda-Grubbs catalysts (HG1 and HG2)]. The reaction was carried out at 35 °C, without solvent and under dynamic vacuum to remove the ethylene formed and shift the equilibrium of the reaction (Table 2). As expected, no reaction occurred with the Schrock catalyst, which was presumably deactivated by the alcohol functions of DVG. Polymerization also did not occur with G1 and HG1, although these catalysts are more robust and polar resistant than Schrock’s catalyst. With the second-generation catalysts, we observed an oligomerization reaction. In the case of the G2 catalyst, however, the double bond conversion of DVG remained very low (26%), probably due to the low polymerization temperature; indeed, G2 is truly active above 45 °C, but its activity is one-tenth that the Hoveyda-Grubbs catalyst (41, 42). On the other hand, the reaction with HG2 was more favorable, as evidenced by the release of ethylene that was observable after the addition of the catalyst. Even if this value is still too low, a 65% conversion to poly(DVG) was reached when HG2 was set to 1 mol %. Other tests were carried out at HG2 concentrations of 0.5 and 5 mol %; however, in all cases, the conversion remained of the same order of magnitude and was therefore still insufficient to achieve the “correct” molar masses. Other reaction conditions that have been tested include the addition of a polar solvent [e.g., tetrahydrofuran (THF) or water] or an additive [e.g., tetrachloro-1,4-benzoquinone or titanium butoxide) to avoid an isomerization reaction or to prevent NNGE by hydroxyl screening, respectively (Table A2). Better results were obtained when ADMET polymerization was performed in THF with tetrachloro-1,4-benzoquinone (5 mol %). A maximum of 80% of double bond conversion was obtained, leading to the formation of oligomers with a degree of polymerization of Xn = 5. Because these last polymerization attempts were carried out in diluted medium, the possibility of forming cycles was favored. MALDI-TOF analysis of the poly(DVG) synthesized in THF in the presence of HG2 as a catalyst demonstrated that no ring was formed and that the repeating pattern of poly(DVG), 86 g/mol, was clearly visible (Figure 3). This first experiment showed that HG2 was the best catalyst for ADMET polymerization of DVG. Even though DVG exhibits an immediate reactivity with HG2, only oligomers were obtained, with a maximum conversion of vinyl functions of 80%. Various reactions were tested by adding additives to the medium to suppress the isomerization reaction induced by the catalyst or the NNGE effect of the alcohol functions of DVG. Although the isomerization reaction was suppressed, similar conversions were calculated. The same results were observed when carrying out the reactions in diluted medium (i.e., in THF). Again, we found DVG quite difficult to polymerize by ADMET. DVG was copolymerized with hydrophobic undecyl undecenoate (UndU) (Figure 4) in order to increase the polymer molar mass and to prepare copolymers exhibiting amphiphilic properties. Indeed, polymers obtained by ADMET polymerization of DVG are water soluble. In the first stage, UndU was polymerized by ADMET at 80 °C for 24 h under dynamic vacuum using HG2 as a catalyst. Calculation of the conversion of the vinyl functions of UndU (88%) by NMR analysis (Figure A1) allowed us to determine that Xn = 8.3 (Mn NMR = 2788 307

g/mol), while a value of 1430 g/mol was measured by SEC in THF (Table 3). The difference between the molar mass values is due to the double bond isomerization reaction caused by the ruthenium catalyst (HG2) as confirmed by MALDI-TOF analysis of poly(UndU) on which we detected a distribution of the characteristic peaks of the isomerization at an interval of 14 g/mol (Figure A2).

Figure 3. MALDI-TOF of poly(DVG) synthesized by ADMET polymerization with HG2 as the catalyst.

Figure 4. Structure of undecyl undecenoate (UndU).

The poly(UndU) formed was then used for the synthesis of multi-sequenced copolymers by ADMET copolymerization with DVG employed at different ratios (using 70, 50, and 30 mol %). The 1H NMR analysis of the copolymers that were formed was carried out in a CDCl3/DMSO-d6 mixture (50/50 v/v) in order to identify the signal of the two moities (Figure A3). Low conversions of vinylic functions were obtained in all cases: In the case of the DVG/UndU mixture at a ratio 30/70 mol %, the conversion in the terminal double bond was high (98%), but the signals corresponding to DVG were not detectable by NMR. For the other two copolymers, DVG signals could be observed and conversions of 24.5 and 36% were calculated for a DVG/UndU ratio of 70/30 and 50/50 mol %, respectively. These conversions lead only to a degree of polymerization (Xn) of less than 2 (Table 3). 308

Table 3. Characteristics of Multi-Sequenced Polymers Synthesized by ADMET Polymerization from poly(UndU) and DVG Product

Ratio DVG/UndU (mol %)

UndU

Total

Mnb (g/mol)

Ðb

DVG

ΔH (J/g)

Tf (°C)

Td5% (°C)

Xn

Poly(UndU)

0/100

—

—

88

1430

1.6

96

34

227

8.3

70/30

32

17

24.5

1750

1.3

101

33

143

1.3

50/50

45

26

36

1870

1.4

60

34

151

1.5

0

98

98

1470

1.5

56

33

204

50

Poly(DVG-b-UndU)

pa (%)

30/70 a

Conversion of vinyl group calculated by RMN

1H.

b

Measured by SEC in THF, PS calibration.

309

The self-assembly properties of the synthesized copolymers composed of hydrophobic (UndU) and hydrophilic (DVG) sequences were evaluated in water by nanoprecipitation and analyzed by dynamic light scattering (DLS). The copolymers were dissolved in 0.5 mL of THF at a concentration of 5 mg/mL, and the solution was added dropwise (every second) into 4.5 mL of water with constant stirring at 250 rpm. THF was then evaporated at room temperature and a cloudy solution was obtained, visually indicating the presence of dispersed particles (Figure 5). The solution was then filtered through a 0.8-μm filter. We performed DLS analysis three times for each sample to test for reproducibility.

Figure 5. DLS analysis of multi-block polymers synthesized from poly(UndU) and DVG.

These analyses confirm that objects of the order of 200 to 500 nm are formed with polydispersity (PDI) on the order of 0.1 (Figure 5). These dispersions are stable for several months, except for the 50/50 copolymer that contains the largest objects. One should note that the use of poly(UndU) in this process does not produce stable particles, indicating that the 30/70 copolymer mixture contains a small fraction of DVG units. Therefore, multi-sequenced copolymers can be synthesized from DVG and UndU, with DVG again showing a low reactivity by ADMET. However, these copolymers exhibit amphiphilic behaviors as confirmed by NMR and DLS analyses. 310

Polymerization of DVG by Thiol-ene Reaction We tested a series of aliphatic dithiols in the course of thiol-ene addition to DVG following two procedures. The monomers were used at a 1:1 stoichiometry and the reaction was conducted either at 80 °C for 24 h or under UV (65 mW/cm2 at 365 nm) for 15 min; azobisisobutyronitrile (AIBN) and Irgacure 2659 were used as the thermo- and the photoinitiator, respectively (Figure 6).

Figure 6. Thiol-ene reaction of DVG and structure of dithiols and initiators used.

Photochemical initiation (procedure b) was more efficient than the thermal initiation (procedure a) (Table 4). Indeed, by photopolymerization, polymers of high molar masses (6000 g/mol) were obtained in a short time (i.e., within 15 min). In the case of polymerizations carried out under UV, no kinetic monitoring was carried out, but the polymers were analyzed by 1H NMR after 15 min of irradiation and conversions of double bonds from 93 to 96% were calculated. The various reactions tested indicate that the chosen aliphatic dithiols have different reactivities with DVG. The best system was with HDT under photochemical activation (Table 4, entry 4); a total conversion was obtained within a few hours. NMR analysis also showed that the alcohol functions of DVG (characteristic signal at 4.4 ppm) were still present (Figure 7). Moreover, whatever the initiation procedure, the NMR spectra of the polymers formed revealed the presence of multiplets at 0.9 ppm that are characteristic of methyl groups (-CH3). This phenomenon indicates the possible addition of the radical on the most substituted carbon of the DVG double bond, whereas anti-Markovnikov addition was generally expected. In summary, DVG exhibits a good reactivity toward thiol-ene reactions through thermal or photochemical initiation. The light-curing process enables the synthesis of polymers in a simple and rapid manner with conversions of the DVG double bonds higher than 93% within 15 min. Although DVG is hydrophilic and provides hydroxyl functions, the polymers synthesized by thiol-ene addition from DVG and linear dithiols were insoluble in water. 311

Table 4. Characteristics of Polymers Synthesized by Thiol-ene Reaction with DVG Entry 1

Dithiol BDT

2 3

HDT

4 5 6 a

NDT

Procedure

p (%)

Xn[Mn RMN]

Mna (g mol–1)

Ðc

Td5% (°C)

Tg (°C)

Tf (°C)

b

93

14.3[3375]

2900

1.7

262

–26

—

c

93

14.3[3375]

7400

1.5

277

–22

—

b

93

14.3[3775]

4000

2

232

–14

73

c

96

33[8712]

6600

1.6

279

–25

83

b

83

5.9[1800]

3700

1.7

267

35

89

c

94

16.7[5100]

6000

1.5

275

35

90

Measured by SEC DMF, LiBr, PS calibration.

b

AIBN (5 mol %), 24 h at 80 °C.

c

Irgacure (0.05 mol %), 15 min under UV (365 nm).

312

Figure 7. 1H NMR in DMSO of polymers synthesized with DVG and (a) BTD, (b) HDT, or (c) NDT by photopolymerization.

The thiol-ene reaction provides access to DVG-based polymers with relatively high molar masses, indicating that the vinylic functions of DVG display a good reactivity. The synthesis of cross-linked materials by the thiol-ene reaction was thus carried out in the presence of a symmetric tetrathiol [i.e., pentaerythritol tetrakis(3-mercaptopropionate) (PTM)] (which has the advantage of not being odorous) and Irgacure 2659 (as a photo-initiator) (Figure 8). The thiol-ene reaction was conducted by mixing DVG and the thiol monomer at different ratios. The initiator was added (0.5 mol %) and the mixture was poured into a Teflon mold measuring 7.2 × 5.1 cm and then cross-linked under UV (365 nm) at 65 mW/cm2 for 15 min to form films that were 2-mm thick.

Figure 8. Structure of the tetrathiol and the initiator used for the synthesis of DVG-based thermosets. 313

Cross-linked materials were thus synthesized by the thiol-ene reaction from DVG and tetrafunctional PTM. The thiol/double bond ratio was varied from 50 to 100% in order to measure the impact of the amount of PTM used on the crosslinking reaction and the properties of the networks obtained (Table 5). Swelling tests were carried out in THF for 24 h. The swelling ratios varied from 189 to 290%, in agreement with the cross-linking density, ρ. Similarly, the soluble fraction decreased with the percentage of PTM used, from 39 to 8%. These experiments confirm that a high content of tetrathiol (PTM) is required to obtain the formation of a 3D-network. Because DVG brings some hydrophilicity through its OH groups, swelling tests in water were also carried out. The swelling ratios, around 10%, were lower in water than THF, but they followed the same trend according to the PTM content. These results were confirmed by FTIR analysis of the films (Figure A4): A characteristic signal of the double bond of DVG was observed at 3080 and 1640 cm–1 and it decreased with higher PTM concentrations, demonstrating the almost complete cross-linking of the polymers. In summary, 3D networks based on DVG and PTM were prepared. The latter have different mechanical properties whose properties vary with the degree of cross-linking. It was also possible to vary the hydrophilicity of these materials by varying the thiol used.

DVG-Derivate Polymerization: Example of an Epoxy–Amine Network While DVG direct polymerization remains of low efficiency when using either double bond or hydroxyl reactivity, we investigated the derivatization of DVG to a more reactive function. At first, DVG was derivatized by the oxidation of the double bonds in order to obtain a bis-epoxidized precursor (DAG). DVG was oxidized in dichloromethane during 24 h in the presence of an excess of metachloroperoxybenzoic acid (mCPBA) (Scheme 4). A maximum yield of 50% was obtained with incomplete conversion of DVG. DAG was then reacted with a series of diamines to prepare the epoxy networks.

Scheme 4. Epoxidation reaction of DVG with mCPBA. rt, room temperature. 314

Syntheses with short alkyl diamines having primary or secondary amino groups (e.g., ethylene diamine, propylene diamine, isophorone diamine, diethylene triamine) lead to “burned” materials. Indeed, the enthalpies of the reaction values measured by differential scanning calorimetry (DSC) are quite high (422 to 647 J g–1), and the reaction starts at room temperature (Table A3). These preliminary investigations clearly demonstrated the high reactivity of DAG. To avoid this issue, diamines with long alkyl chains were selected, that is, Priamine® 1075, which is a bio-sourced derivative from fatty acid, and Jeffamine ED600 and ED900, which are oligo-ethers with molar masses equal to 678.8 and 972 g/mol, respectively. The curing reaction was followed by DSC. The enthalpy of reaction (ΔH), reaction start temperature (Ton set), and Tg were determined by DSC (Table 6). These analyses revealed relatively low Tg values, in agreement with the Tg of Jeffamine precursors, –39 and –45 °C, respectively. A Tg of 26 °C was measured for the networks formed with Priamine® 1075, which adds more rigidity. The formulation of networks was carried out by using the epoxy equivalent weight (EEW, g/equiv) value of DAG and by using the amine hydrogen equivalent weight (AHEW, g/equiv) value of the diamine. The curing of the epoxy–amine formulation is theoretically performed with a ratio of 1 mole of epoxy group to 1 mole of active hydrogen from amine. The network swelling was higher in THF (180%) and revealed a fairly large soluble fraction of 34%, demonstrating that the cross-linking was not complete. Due to its chemical structure, DAG provides hydroxyl functions in water and the opening of the epoxide ring generates new rings, which explains a certain affinity of this material with water. Nevertheless, the proportion of the soluble fraction remained very low at 2%. Mechanical measurements by dynamic mechanical analysis (DMA) allowed us to estimate a cross-link density of 1.6 mol/dm3 and to measure a phase transition, Tα, of around 20 °C. With the Jeffamines, the networks were cured for 1 h at 120 °C and for 2 h at 140 °C, and the complete polymerization was confirmed by DSC (Table A4). The network prepared from Jeffamine ED600 was similar to a viscous fluid, whereas an aspect close to an elastomeric gum was obtained from Jeffamine ED900. Sticky materials were obtained, so the tack properties of these networks were measured by DMA (Figure A5). Tack is measured as the force required to separate an adhesive from a probe at the interface shortly after they have been brought into contact under a defined duration at a specific temperature (here, 10 s at room temperature). Diglycidyl ether of bisphenol-A (DGEBA) and isophorone diamine (IPDA) are classical components of epoxy–amine networks. One of the major issues with this type of material is the impossibility of degrading them in “soft” conditions. To test this, DAG was incorporated into the DGEBA/IPDA network at different percentages (1, 5, and 10 wt %) in order to provide 1,2-diol units within the network and then to evaluate whether the latter could be cleaved after treatment with a strong acid, such as periodic acid (Table 8). Networks obtained with these diamines display different properties. With Priamine® 1075, the curing was performed at room temperature over a 24-h period. Swelling tests were conducted in THF and water (Table 7).

315

Table 5. Properties of Networks Synthesized by Thiol-ene with DVG and PTM

a

PTM (%)

Swelling

50

290 [10]

70

Soluble part

Td5% (°C)

Tg (°C)

E′Tδ+30°C (MPa)

ρ (mol/dm3)

Strain (MPa)a

Stress (%)a

Modulus (MPa)a

39 [15]

120

–33

0.15

0057

0.1

99 ± 7.8

0.13

191 [7]

14 [6]

170

–22

1.8

0.67

0.6 ± 0.1

35.8 ± 1.4

1.96 ± 0.3

90

163 [6]

14 [3]

285

–17

2.5

0.93

0.6

28 ± 3.8

2.2 ± 0.3

100

189 [10]

8 [3]

290

–16

3.5

1.3

1.2 ± 0.2

33.6 ± 4.2

4.1 ± 0.6

Average values.

THF[H2O] (%)

316

Table 6. Characteristics of the Reactivity of Diamines with DAG Diamine

f

AHEW (g equiv–1)

ΔH (J g–1)

Ton set (°C)

Tg (°C)

Priamine® 1075

4

136.17

433

36

26

Jeffamine ED600

4

169.7

353

62

–39

Jeffamine ED900

4

243

263

71

–45

317

Table 7. Characteristics of DAG/Priamine® 1075 Network Diamine

Swelling (%)

Soluble part (%)

Td5% (%)

Ta (Tg) (°C)

E′Ta+30°C (MPa)

ρ (mol/dm3)

287

22 (19)

6.2

1.6

THF [H2O] Priamine® 1075

180 [20]

34 [2]

318

Table 8. Properties of DGEBA/IPDA Materials with Addition of 1, 5, and 10 wt % of DAG DAG (wt %)

a

Td5% (%)

Ta (Tg) (°C)

ρ (mol/dm3)

Strain (MPa)a

Stress (%)a

Young modulus (MPa)a

Swelling

Soluble part

THF [H2O] (%)

0

352

148 (137)

10.6

74 ± 11

6.3 ± 1.1

1337 ± 102

3 [1]