Polypeptide Materials: Methods and Protocols [1st ed.] 9781071609279, 9781071609286

Table of contents :
Front Matter ....Pages i-xiv
A Shortcut to the Synthesis of Peptide Thioesters (Richard Raz, John Offer)....Pages 1-12
SnoopLigase-Mediated Peptide–Peptide Conjugation and Purification (Can M. Buldun, Irsyad N. A. Khairil Anuar, Mark Howarth)....Pages 13-31
Peptide Nanoparticles for Gene Packaging and Intracellular Delivery (Paula Vila-Gómez, James E. Noble, Maxim G. Ryadnov)....Pages 33-48
Synthesis and Application of Peptide–siRNA Nanoparticles from Disulfide-Constrained Cyclic Amphipathic Peptides for the Functional Delivery of Therapeutic Oligonucleotides to the Lung (Jade J. Welch, David A. Dean, Bradley L. Nilsson)....Pages 49-67
FRET-Mediated Observation of Protein-Triggered Conformational Changes in DNA Nanostructures (Simon Chi-Chin Shiu, Yusuke Sakai, Julian A. Tanner, Jonathan G. Heddle)....Pages 69-80
Molecular Simulations Guidelines for Biological Nanomaterials: From Peptides to Membranes (Irene Marzuoli, Franca Fraternali)....Pages 81-100
Functional Peptide Nanocapsules Self-Assembled from β-Annulus Peptides (Hiroshi Inaba, Kazunori Matsuura)....Pages 101-121
Electrostatic Self-Assembly of Protein Cage Arrays (Soumyananda Chakraborti, Antti Korpi, Jonathan G. Heddle, Mauri A. Kostiainen)....Pages 123-133
Design and Generation of Self-Assembling Peptide Virus-like Particles with Intrinsic GPCR Inhibitory Activity (Sergey G. Tarasov, Marzena Dyba, Joshua Yu, Nadya Tarasova)....Pages 135-148
Imaging and 3D Reconstruction of De Novo Peptide Capsids (Emiliana De Santis, Maxim G. Ryadnov)....Pages 149-165
Monitoring the Assembly and Aggregation of Polypeptide Materials by Time-Resolved Emission Spectra (Abeer Alghamdi, Li Hung C. Chung, Olaf J. Rolinski)....Pages 167-177
Forming Low-Molecular-Weight Hydrogels by Electrochemical Methods (Emily R. Cross, Kate McAulay, Dave J. Adams)....Pages 179-188
In Situ Measurements of Polypeptide Samples by Dynamic Light Scattering: Membrane Proteins, a Case Study (Tristan O. C. Kwan, Rosana Reis, Isabel Moraes)....Pages 189-202
Measurement of Peptide Coating Thickness and Chemical Composition Using XPS (David J. H. Cant, Alexander G. Shard, Caterina Minelli)....Pages 203-224
Imaging the Effects of Peptide Materials on Phospholipid Membranes by Atomic Force Microscopy (Katharine Hammond, Georgina Benn, Isabel Bennett, Edward S. Parsons, Maxim G. Ryadnov, Bart W. Hoogenboom et al.)....Pages 225-235
Microfluidic Single-Cell Phenotyping of the Activity of Peptide-Based Antimicrobials (Jehangir Cama, Stefano Pagliara)....Pages 237-253
Ultramicrotomy Analysis of Peptide-Treated Cells (Stephanie Rey, Nilofar Faruqui, Maxim G. Ryadnov)....Pages 255-264
Back Matter ....Pages 265-268

Citation preview

Methods in Molecular Biology 2208

Maxim G. Ryadnov Editor

Polypeptide Materials Methods and Protocols

METHODS

IN

MOLECULAR BIOLOGY

Series Editor John M. Walker School of Life and Medical Sciences University of Hertfordshire Hatfield, Hertfordshire, UK

For further volumes: http://www.springer.com/series/7651

For over 35 years, biological scientists have come to rely on the research protocols and methodologies in the critically acclaimed Methods in Molecular Biology series. The series was the first to introduce the step-by-step protocols approach that has become the standard in all biomedical protocol publishing. Each protocol is provided in readily-reproducible step-bystep fashion, opening with an introductory overview, a list of the materials and reagents needed to complete the experiment, and followed by a detailed procedure that is supported with a helpful notes section offering tips and tricks of the trade as well as troubleshooting advice. These hallmark features were introduced by series editor Dr. John Walker and constitute the key ingredient in each and every volume of the Methods in Molecular Biology series. Tested and trusted, comprehensive and reliable, all protocols from the series are indexed in PubMed.

Polypeptide Materials Methods and Protocols

Edited by

Maxim G. Ryadnov National Physical Laboratory, Teddington, Middlesex, UK

Editor Maxim G. Ryadnov National Physical Laboratory Teddington, Middlesex, UK

ISSN 1064-3745 ISSN 1940-6029 (electronic) Methods in Molecular Biology ISBN 978-1-0716-0927-9 ISBN 978-1-0716-0928-6 (eBook) https://doi.org/10.1007/978-1-0716-0928-6 © Springer Science+Business Media, LLC, part of Springer Nature 2021 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Cover Illustration Caption: A self-assembled peptidic virus-like capsid. Courtesy of Irene Marzuoli and Franca Fraternali. This Humana imprint is published by the registered company Springer Science+Business Media, LLC, part of Springer Nature. The registered company address is: 1 New York Plaza, New York, NY 10004, U.S.A.

Preface This volume continues the misson of the book series to provide the research community with a systematic overview of contemporary research methods in the area of molecular biology. The volume details the synthesis and assembly of polypeptide materials (proteins and peptides), their precursors, conjugates, and derivatives. A particular emphasis is made on the measurement tools and procedures for material characterization, both physicochemical and functional. Often overlooked in the literature, such a focus can address the growing demand for improving reproducibility in biomedical research. It can also help to better reflect the interdisciplinary nature of molecular biology and the importance of developing innovative measurement methods to advance the area. Written as step-by-step instructions, protocols in this volume are not limited to the given examples and are applicable to other materials of the same and similar types. This is also meant to help the reader to keep abreast with achievements beyond the scope of each topic avoiding both generalization, which would otherwise omit important nuances in measurement results, e.g., sources of variation, and instantiation, which would narrow them down to specialist cases. The volume opens with a protocol for the chemical synthesis of peptide thioesters. These are routinely used as building blocks for complex structures ranging from proteins to advanced functional materials (Raz and Offer). Traditional approaches toward peptide thioesters use highly hazardous chemistries, which are increasingly problematic to implement in the modern, risk-averse laboratory practice, whereas alternative routes continue to suffer from poor reproducibility, giving products of low purity and yields. All of this is addressed in the first chapter that describes a shortcut methodology to peptide thioesters, which can enable post-synthesis applications such as native chemical ligation. Unlike chemical methods, which require achieving optimum stoichiometry, native peptide–peptide ligations may proceed to completion at nearly equimolar ratios of ligating fragments. These reactions are self-catalyzed, while their efficiency can be improved using an auxiliary means such as conformationally driven ligations that bring ligating ends in close proximity allowing the two components to fuse even at low concentrations. An example of these reactions is given in the following chapter, which offers a straightforward method for peptide–peptide ligation (Buldun et al.). The method termed SnoopLigase conjugation facilitates the sitespecific fusion of proteins via the formation of an isopeptide bond with >95% efficiency. The described strategy relies on the use of peptide tags and has already been demonstrated to enhance enzyme resilience and the efficacy of vaccines through antigen oligomerization. The protocol details a range of conditions for SnoopLigase reactions together with optimized purification steps for the ligated products and the quantification of ligation efficiency. The key advantages of this method are highligted to include the maximum conversion of reacting components, low concentrations used, and precise tag location which prove to be effective even inside cells. Both the chapters advocate for covalent conjugates as preferable forms of peptide materials. This holds true for many. For other technologies of biological relevance, covalent modifications may not be necessary or even prohibitive, while both covalent and non-covalent macromolecular systems are equally common. Gene therapy is one of such application areas. Its main purpose is to deliver functional genetic material into the cell. Genes, be these RNA or DNA, cannot traverse cellular membranes by themselves and require a carrier vehicle to mediate the transfer. A variety of peptides with similar properties

v

vi

Preface

have been shown to provide suitable packaging systems for various nucleic acids and oligonucleotides alike. The next two chapters describe protocols for the preparation and characterization of non-covalent gene packaging systems. Gomez et al. make stress on nanoparticle formulations resulting from cell-penetrating peptides that can either condense or encapsulate small interference RNA (siRNA) and DNA. The protocol is exemplified by a number of commercial and experimental reagents and provides an exhaustive workflow for their preparation and analysis including peptide synthesis, purification, efficiency and stability of particle formation, and intracellular uptake. Welch et al. follows upon this chapter with nanoparticulate formulations of siRNA enabled by condensation with bespoke disulfide-constrained cyclic peptides. The peptides act as a switch system that opens up in the reducing environments of the cell and is degraded by proteases releasing siRNA into the cytoplasm. This protocol details a step-by-step preparation of the formulation, from synthesis to siRNA complexation, and ramps up the biological validation of peptide packaging systems of this type to animal model experiments, demonstrating effective gene knowdown, which confirms the relevance and applicability of non-covalent peptide systems in vivo. The nature of forces that drive nanoparticle formation is not specifically addressed in these two chapters. Nevertheless, for all the systems described, conformational changes are involved in complexation, intracellular uptake, and release. In some cases, the changes are a direct result of binding to nucleic acids; in others, it is the folding of peptides that mediates the complexation or predetermines encapsulation. Complementary to this, Chapter 5 reports quantitative examples of conformationally driven functions mediated by proteins and peptides (Shiu et al.). This protocol describes the protein-triggered opening of a nanoscale box assembled using the principles of DNA origami. The construction of the DNA box is detailed in line with the functional assessment of the opening process that is monitored by Fo¨rster resonance energy transfer (FRET). Lactate dehydrogenase from Plasmodium falciparum specifically binds to a DNA aptamer serving as a “lock” in the box. When the target protein binds to the aptamer, the box opens, which leads to a change in FRET signal. All the processes supporting the methodology are mediated by highly specific molecular recognition events between biological materials of different chemistries. Exploiting such events for developing and studying functional materials is the subject of the following chapter which broadens the application of conformational changes to the behavior prediction of self-assembling systems (Marzuoli and Fraternali). This chapter presents detailed guidelines for molecular dynamic simulations of biological nanoscale systems with a particular emphasis on their hierarchical nature. The latter is neatly matched with force fields used in simulations, which are categorized according to the accuracy and speed they offer. Each force field is assigned to a specific type of simulation, thus probing the pros and cons of their prediction power in comparison, and to a specific hierarchical type of biological assembly as complex as protein cages. Experimental protocols in emulating, predicting, and exploiting the protein shell architecture are given in the next four chapters. Chapter 7 (Inaba and Matsuura) describes nanocapsulate assembly of protein shells from β-annulus peptides. These peptides derive from naturally occuring viruses such as tomato bushy stunt virus and represent a structurally conserved topology allowing for symmetric assembly into spherical hollow nanoparticles. The protocol details all aspects of their production and analysis starting with peptide synthesis protocols. Conjugations and non-covalent decorations of the assembled particles with DNA, metal nanoparticles, and peptides are also discussed. Both exterior (surface) and interior (encapsulation) functionalizations are presented, rendering this design one of the

Preface

vii

most characterized discrete systems reported to date. The next chapter (Chakraborti et al.) takes a different approach with a protocol describing the use of protein cages in constructing regular arrays of metal nanoparticles. The methodology uses the same principles of electrostatic interactions, but instead of encapsulating metal particles in cages or decorating cages with a metal corona, this approach employs the interactions to build three-dimensional (3D) geometric objects, in which protein cages and metal particles alternate forming superlattices. Protein cages with biological activities may present an obvious choice for design. This is covered in the following protocol (Tarasov et al.), which describes synthetic analogs of a portion of a transmembrane domain of G-protein-coupled receptors (GPCR). The analogs are designed to assemble into virus-like particles retaining an intrisic GPCR inhibitory activity. The methodology takes a structure-based design route to implement both the functions, biological and self-assembly, in the constituent monomer of the particles. Comprehensive instructions are provided for the structure, design, synthesis, assembly, and physicochemical characterization of the particles. The final chapter in this quartet (De Santis and Ryadnov) continues this direction with a protocol for imaging and 3D reconstruction of virus-like particles using high-resolution and cryogenic electron microscopy. Unlike examples used in previous chapters, these particles are de novo capsids constructed without prior structural information. Designed from scratch, the capsids need ultrastructural elucidation to support and improve the design principles. Therefore, this protocol provides major steps in the structural analysis of the particles as a continuum of measurement procedures, from assembly to 3D reconstruction. Virus-like particles represent one of the major nanomaterial morphologies that find use in various applications and show consistent progress in new designs and the development of bespoke methodologies to advance the precision and accuracy of analysis. Since most, if not all, protein and peptide nanoscale materials rely on self-assembly processes, methodologies that are able to monitor the processes in relation to resulting morphologies are of particular interest. In this regard, the next protocol desribes the application of time-resolved emission spectroscopy to track the protein assembly exemplified by fibrillogenesis (Alghamdi et al.). Protein filaments constitute another major nanomaterial morphology, which can also be of natural and synthetic origin. The protocol is built around case studies for both types and extends measurements to post-assembly events such as aggregation, with relevance to pathological processes including amyloidosis. The method follows changes in fluorescence as a function of time, which gives a continuous measure of growth kinetics and, when applicable, aggregation (e.g., conversion of amyloid fibrils to plaques). Representative timeresolved spectra with data interpretation are also discussed to demonstrate the advantages of the methodology. There exists a subtle borderline between the different outcomes of peptide assembly and aggregation. The former assumes a higher level of control to enable the fabrication of reproducible materials that deliver the desired functions. The rate and type of assembly is defined at the primary structure level but can also be directed by external stimuli and fabrication methods. An exemplar method is described in the next chapter (Cross et al.). This protocol focuses on the fabrication and characterization of nanostructured hydrogel materials. Typically designed to support 3D cell culture, both in vitro and in vivo, peptide hydrogels result from the formation of 3D filamentous networks. Filamentous assembly does not necessarily lead to gelation. Therefore, successful hydrogel designs are often complemented with effective fabrication methods. The protocol provides an additional value of producing hydrogels in different forms—thin films prepared in seconds and thicker gels prepared in minutes. Protein assembly is initiated within 1 h, but can mature to an

viii

Preface

equilibrium phase in a few hours. Monitoring nanomaterial formation under assembly conditions from the outset is important and is of particular interest if it allows to probe size changes. Such changes are a universal indicator not only for the growth kinetics but also for the quality of the produced material as they can inform preventive measures against undesired progessions, e.g., aggregation. Dynamic light scattering is a common tool for size measurements, with yet limited use due to the requirements of using large sample volumes and low throughput. These challenges are addressed in the protocol by Kwan et al., which describes dynamic light scattering measurements executed in a fully automated fashion for low-volume samples and in solution, i.e., in situ. The protocol offers a generic pre-screening method for downstream structural studies of polypeptide and other biomolecular materials using higher-resolution approaches including X-ray crystallography, electron microscopy, small-angle X-ray scattering, and NMR. Since in most cases nanoscale peptide materials are produced under aqueous conditions, solution measurements remain a prerequisite for their characterization. Finally, assembled products can be used under different conditions, which necessitate the development of other specialist methods and protocols. Peptide coatings are increasingly relevant for applications where accurate control over thickness, homogeneity, and density is imperative. X-ray photoelectron spectroscopy offers one of the most informative methods to probe these properties. This technique is highly sensitive to changes on surfaces providing quantitative information on the chemical composition of materials within 10-nm depths from their surfaces. The use of this method for flat and 3D substrates (e.g., nanoparticulate materials) is detailed in the protocol by Cant et al. Surface analysis approaches are explored for more complex measurements aiming to characterize different events and under different conditions. Atomic force microscopy (AFM) has revolutionalized the area of protein assembly at interfaces and is rapidly progressing toward live cell and high-speed measurements of various biologically relevant meachanims mediated by peptide materials (Hammond et al.). This development is characteristic of a timely and complementary response to advances in material designs and biotechnologies that demand more precise mechanistic investigations in native and near-native environments. Broader developments concern microscopy capabilities as a whole. One drive here is to enable correlative imaging of the same biological processes or systems using microscopy modalities that differ by imaging speed, resolution, and sample size. Correlative imaging is viewed as an ultimate solution to the characterization of nanoscale materials in cellular media. AFM is one of such modalities. It can cover the nano-to-micrometer length scale but has limitations in depth resolution and screening throughput. Other approaches lack the resolution of AFM but can support highthroughput measurements, ensuring statistically significant analysis of cell populations, while being able to support single-cell measurements. A highly effective strategy is to empower optical fluorescence microscopy with microfluidic devices such as one described in the penultimate protocol of the volume (Cama and Pagliara). Here, a single chip of thousands of micro-channels can accommodate thousands of individual cells whose response to peptide treatment can be imaged over a set period of time. This allows kinetic measurements of cell responses (growth and lysis) as well as concentration-dependent peptide–cell interactions. Exemplified by antimicrobial effects of peptide materials on bacterial phenotypes, the strategy is adaptable to other cell-based measurements including processes that seek new insights at the intracellular level. However, revealing or confirming the details at this level demands much higher resolution, whereas the number of available modalities is limited. Encouragingly enough, ultramicrotomy analysis of peptide-treated

Preface

ix

cells using electron microscopy proves to provide access to critical information at highest resolution possible today. Described in the last chapter by Rey et al., this methodology provides a step-by-step instruction for sample preparation and processing, imaging and analysis of bacterial and mammalian cells following peptide treatment. Furthermore, examples of intracellular infections and their targeting by bespoke peptide materials is also discussed, thus culminating this volume with arguably the most challenging and complex application measurement of peptide materials at the nanoscale. Teddington, Middlesex, UK

Maxim G. Ryadnov

Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v xiii

1 A Shortcut to the Synthesis of Peptide Thioesters . . . . . . . . . . . . . . . . . . . . . . . . . . . Richard Raz and John Offer 2 SnoopLigase-Mediated Peptide–Peptide Conjugation and Purification . . . . . . . . Can M. Buldun, Irsyad N. A. Khairil Anuar, and Mark Howarth 3 Peptide Nanoparticles for Gene Packaging and Intracellular Delivery . . . . . . . . . . Paula Vila-Go mez, James E. Noble, and Maxim G. Ryadnov 4 Synthesis and Application of Peptide–siRNA Nanoparticles from Disulfide-Constrained Cyclic Amphipathic Peptides for the Functional Delivery of Therapeutic Oligonucleotides to the Lung . . . . . . . . . . . . Jade J. Welch, David A. Dean, and Bradley L. Nilsson 5 FRET-Mediated Observation of Protein-Triggered Conformational Changes in DNA Nanostructures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simon Chi-Chin Shiu, Yusuke Sakai, Julian A. Tanner, and Jonathan G. Heddle 6 Molecular Simulations Guidelines for Biological Nanomaterials: From Peptides to Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Irene Marzuoli and Franca Fraternali 7 Functional Peptide Nanocapsules Self-Assembled from β-Annulus Peptides. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hiroshi Inaba and Kazunori Matsuura 8 Electrostatic Self-Assembly of Protein Cage Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . Soumyananda Chakraborti, Antti Korpi, Jonathan G. Heddle, and Mauri A. Kostiainen 9 Design and Generation of Self-Assembling Peptide Virus-like Particles with Intrinsic GPCR Inhibitory Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . Sergey G. Tarasov, Marzena Dyba, Joshua Yu, and Nadya Tarasova 10 Imaging and 3D Reconstruction of De Novo Peptide Capsids . . . . . . . . . . . . . . . Emiliana De Santis and Maxim G. Ryadnov 11 Monitoring the Assembly and Aggregation of Polypeptide Materials by Time-Resolved Emission Spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Abeer Alghamdi, Li Hung C. Chung, and Olaf J. Rolinski 12 Forming Low-Molecular-Weight Hydrogels by Electrochemical Methods . . . . . Emily R. Cross, Kate McAulay, and Dave J. Adams 13 In Situ Measurements of Polypeptide Samples by Dynamic Light Scattering: Membrane Proteins, a Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tristan O. C. Kwan, Rosana Reis, and Isabel Moraes

1

xi

13 33

49

69

81

101 123

135 149

167 179

189

xii

14

15

16

17

Contents

Measurement of Peptide Coating Thickness and Chemical Composition Using XPS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . David J. H. Cant, Alexander G. Shard, and Caterina Minelli Imaging the Effects of Peptide Materials on Phospholipid Membranes by Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Katharine Hammond, Georgina Benn, Isabel Bennett, Edward S. Parsons, Maxim G. Ryadnov, Bart W. Hoogenboom, and Alice L. B. Pyne Microfluidic Single-Cell Phenotyping of the Activity of Peptide-Based Antimicrobials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jehangir Cama and Stefano Pagliara Ultramicrotomy Analysis of Peptide-Treated Cells . . . . . . . . . . . . . . . . . . . . . . . . . . Stephanie Rey, Nilofar Faruqui, and Maxim G. Ryadnov

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

203

225

237 255 265

Contributors DAVE J. ADAMS • School of Chemistry, University of Glasgow, Glasgow, UK ABEER ALGHAMDI • Department of Physics, Scottish Universities Physics Alliance, University of Strathclyde, Glasgow, UK ISABEL BENNETT • London Centre for Nanotechnology, University College London, London, UK GEORGINA BENN • London Centre for Nanotechnology, University College London, London, UK; National Physical Laboratory, Teddington, UK; Institute of Structural and Molecular Biology, University College London, London, UK CAN M. BULDUN • Department of Biochemistry, University of Oxford, Oxford, UK; Roche Diagnostics GmbH, Penzberg, Germany JEHANGIR CAMA • Living Systems Institute, University of Exeter, Exeter, UK; College of Engineering, Mathematics and Physical Sciences, University of Exeter, Exeter, UK DAVID J. H. CANT • National Physical Laboratory, Teddington, UK SOUMYANANDA CHAKRABORTI • Malopolska Centre of Biotechnology, Jagiellonian University, Krakow, Poland LI HUNG C. CHUNG • Department of Physics, Scottish Universities Physics Alliance, University of Strathclyde, Glasgow, UK EMILY R. CROSS • School of Chemistry, University of Glasgow, Glasgow, UK DAVID A. DEAN • Department of Pediatrics and Neonatology, Biomedical Engineering, and Pharmacology & Physiology, School of Medicine and Dentistry, University of Rochester, Rochester, NY, USA EMILIANA DE SANTIS • National Physical Laboratory, Teddington, Middlesex, UK MARZENA DYBA • Basic Science Program, Frederick National Laboratory for Cancer Research, Frederick, MD, USA NILOFAR FARUQUI • National Physical Laboratory, Teddington, Middlesex, UK FRANCA FRATERNALI • Randall Centre for Cell & Molecular Biophysics, King’s College London, London, UK KATHARINE HAMMOND • London Centre for Nanotechnology, University College London, London, UK; National Physical Laboratory, Teddington, UK; Department of Physics and Astronomy, University College London, London, UK JONATHAN G. HEDDLE • Malopolska Centre of Biotechnology, Jagiellonian University, Krakow, Poland BART W. HOOGENBOOM • London Centre for Nanotechnology, University College London, London, UK; Institute of Structural and Molecular Biology, University College London, London, UK; Department of Physics and Astronomy, University College London, London, UK MARK HOWARTH • Department of Biochemistry, University of Oxford, Oxford, UK HIROSHI INABA • Department of Chemistry and Biotechnology, Graduate School of Engineering, Tottori University, Tottori, Japan; Centre for Research on Green Sustainable Chemistry, Tottori University, Tottori, Japan IRSYAD N. A. KHAIRIL ANUAR • Department of Biochemistry, University of Oxford, Oxford, UK

xiii

xiv

Contributors

ANTTI KORPI • Biohybrid Materials, Department of Bioproducts and Biosystems, Aalto University, Aalto, Finland MAURI A. KOSTIAINEN • Biohybrid Materials, Department of Bioproducts and Biosystems, Aalto University, Aalto, Finland TRISTAN O. C. KWAN • National Physical Laboratory, Teddington, UK; Research Complex at Harwell Rutherford, Appleton Laboratory, Oxford, UK IRENE MARZUOLI • Randall Centre for Cell & Molecular Biophysics, King’s College London, London, UK KAZUNORI MATSUURA • Department of Chemistry and Biotechnology, Graduate School of Engineering, Tottori University, Tottori, Japan; Centre for Research on Green Sustainable Chemistry, Tottori University, Tottori, Japan KATE MCAULAY • School of Chemistry, University of Glasgow, Glasgow, UK CATERINA MINELLI • National Physical Laboratory, Teddington, UK ISABEL MORAES • National Physical Laboratory, Teddington, UK; Research Complex at Harwell Rutherford, Appleton Laboratory, Oxford, UK BRADLEY L. NILSSON • Department of Chemistry, University of Rochester, Rochester, NY, USA JAMES E. NOBLE • National Physical Laboratory, Teddington, Middlesex, UK JOHN OFFER • The Francis Crick Institute, London, UK STEFANO PAGLIARA • Living Systems Institute, University of Exeter, Exeter, UK; College of Life and Environmental Sciences, University of Exeter, Exeter, UK EDWARD S. PARSONS • London Centre for Nanotechnology, University College London, London, UK ALICE L. B. PYNE • London Centre for Nanotechnology, University College London, London, UK; Department of Materials Science and Engineering, University of Sheffield, Sheffield, UK RICHARD RAZ • Department of Paediatrics, University of Oxford, Oxford, UK ROSANA REIS • National Physical Laboratory, Teddington, UK; Research Complex at Harwell Rutherford, Appleton Laboratory, Oxford, UK STEPHANIE REY • National Physical Laboratory, Teddington, Middlesex, UK OLAF J. ROLINSKI • Department of Physics, Scottish Universities Physics Alliance, University of Strathclyde, Glasgow, UK MAXIM G. RYADNOV • National Physical Laboratory, Teddington, Middlesex, UK YUSUKE SAKAI • Malopolska Centre of Biotechnology, Jagiellonian University, Krakow, Poland ALEXANDER G. SHARD • National Physical Laboratory, Teddington, UK SIMON CHI-CHIN SHIU • School of Biomedical Sciences, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Pokfulam, Hong Kong SAR, China JULIAN A. TANNER • School of Biomedical Sciences, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Pokfulam, Hong Kong SAR, China NADYA TARASOVA • Laboratory of Cancer Immunometabolism, National Cancer Institute, Frederick, MD, USA SERGEY G. TARASOV • Structural Biophysics Laboratory, National Cancer Institute, Frederick, MD, USA PAULA VILA-GO´MEZ • National Physical Laboratory, Teddington, Middlesex, UK JADE J. WELCH • Department of Chemistry, University of Rochester, Rochester, NY, USA JOSHUA YU • Laboratory of Cancer Immunometabolism, National Cancer Institute, Frederick, MD, USA

Chapter 1 A Shortcut to the Synthesis of Peptide Thioesters Richard Raz and John Offer Abstract Peptide thioesters serve as fundamental building blocks for the synthesis of proteins and cyclic peptides. Classically, methods to synthesize thioesters have been based on acid-labile amino-protecting groups for which final side-chain deprotection required the use of hazardous hydrogen fluoride (HF). Alternative protection schemes based on base-labile amino-protecting groups have become preferred methods but are not suitable due to the lability of thioester bonds toward bases. In this method, we employ a trifluoracetic acid/trimethylsilyl bromide (TFA/TMSBr) protocol using a hydroxymethyl resin obviating the need for HF. TFA/TMSBr is volatile enough to be easily removed yet less hazardous than HF, making it more practical for general peptide chemists. We describe optimized cleavage procedures and appropriate protecting group schemes and discuss in situ neutralization protocols. The method is relatively simple, straightforward, and easily scalable, allowing the facile preparation of alkyl and aryl thioesters. Key words Peptide thioesters, Protein synthesis, Polypeptide synthesis, Cyclic peptides, Native chemical ligation, Solid-phase peptide synthesis

1

Introduction Peptide thioesters are indispensable tools for chemical biologists. Through native chemical ligation (NCL), they serve as essential building block precursors for large polypeptide and protein synthesis [1]. They also serve as excellent activated esters for the facile preparation of cyclic peptides [2]. Their utility is countered by their difficulty to synthesize and prepare. Direct synthesis of peptide thioesters using standard fluorenylmethyloxycarbonyl solid-phase peptide synthesis (Fmoc SPPS) methods is limited [3] due to the lability of thioesters in the presence of a base that is required for Fmoc removal [4], and typically Fmoc-based methods rely on indirect thioester formation [5]. Initially, peptide thioesters were prepared directly by tertbutyloxycarbonyl (Boc) SPPS facilitated by the stability of the thioester bond to trifluoroacetic acid (TFA), used for Boc deprotection cycles [6]. The major limitation of Boc SPPS in the past has been its requirement for anhydrous hydrogen fluoride (HF) for the

Maxim G. Ryadnov (ed.), Polypeptide Materials: Methods and Protocols, Methods in Molecular Biology, vol. 2208, https://doi.org/10.1007/978-1-0716-0928-6_1, © Springer Science+Business Media, LLC, part of Springer Nature 2021

1

2

Richard Raz and John Offer

deprotection and cleavage of the peptide from the resin. HF is extremely hazardous and requires specialized apparatus and training [7]. This also limits the ability of synthesis in parallel format, and it is difficult and impractical to perform microcleavages to check a synthesis before completion. In addition, HF is not compatible with the incorporation of many post-translational modifications such as phosphorylation. In general, this has led to Fmoc SPPS becoming the method of choice for the routine synthesis of peptides [8]. Despite the prevalence of Fmoc-based methods, Boc SPPS possesses a number of advantages when compared with Fmoc SPPS. Boc amino acids are normally more soluble than their Fmoc counterparts, which can allow a higher concentration resulting in faster coupling [9]. Boc SPPS has a much lower rate of failure sequences and is particularly useful for peptides that are high in ß-sheet structure that are prone to aggregation with Fmoc SPPS. This is owing to the powerful solvation properties of TFA at each Boc deprotection cycle [10]. Direct comparison of Boc SPPS and Fmoc SPPS reveals that a cleaner product is obtained by Boc SPPS, deprotection of the Boc group using TFA occurs more rapidly than the removal of Fmoc with piperidine and always goes to completion, and the resulting TFA salt protects against aspartimide formation during chain assembly [11]. Practically, Boc SPPS preparation of peptide thioesters is scalable; they can be reliably synthesized in large amounts, which is crucial for protein synthesis. There have been attempts to substitute HF in Boc SPPS with trifluoromethanesulfonic acid (TFMSA) [12]. Though moderately successful, these studies have underscored the need for a more volatile strong acid that can be easily removed. Residual amounts of unremoved TFMSA can degrade the synthesized peptide after concentration. A two-step approach to get around this combines TFMSA with thiolysis to yield the final thioester [13]. Clearly, a simple straightforward method that removes HF, while still producing clean thioester products, is desirable. We have adapted many protocols from existing literature on manual Boc SPPS, taking where appropriate, cues from insight and lessons learned. We have implemented TFA/trimethylsilyl bromide (TMSBr) not only for the deprotection of side-chain protecting groups but also for the removal of the peptide from the hydroxymethyl resin [14]. The following TFA/TMSBr protocols represent a universal approach, which any lab can easily implement to prepare these extremely useful building blocks [15].

2

Materials In general, it is advisable to use high-quality amino acids, reagents, and solvents to obtain consistent results. Standard Boc amino acids can be used for the synthesis; however, care must be taken in

A Shortcut to the Synthesis of Peptide Thioesters

3

Table 1 Protecting group strategy Amino acid

Side-chain protection HF cleavagea

Side-chain protection TFA/TMSBr

Deprotection time using TFA/TMSBrb

Asp

cHx

Bzlc

1h

Arg

Tos

Mts

1–3 h

Asn

Xan

Xan

30 A˚ apart) may require a longer linker than that suggested in Subheading 3.1 to minimize intermolecular conjugation. If initial reaction conditions show a high level of intermolecular conjugation by SDS-PAGE, choose lower substrate concentrations. 2. Stoichiometry: The reaction works at 1:1:1 stoichiometry. However, to maximize the conjugation efficiency of one substrate, double the concentration of SnoopLigase and the other substrate. 3. pH: The reaction works best at pH 6.5–8.5 (Fig. 3a, b). 4. Temperature: The reaction works well at 4–25
C (Fig. 3c). 5. Glycerol: Glycerol is not essential for the reaction but slightly enhances the reaction rate (Fig. 3d). 6. NaCl: Avoid NaCl if possible. NaCl concentrations above 10 mM reduce the SnoopLigase reaction rate (Fig. 3e, f). Preferably, dialyze your SnoopTagJr-fusion and DogTagfusion into a buffer without sodium or chloride ions before reaction [17].

SPeptide Gene Delivery Packagers

23

Fig. 3 SnoopLigase reaction in a range of conditions. (a) pH-dependence. SnoopTagJr-AffiHER2 and SUMODogTag were ligated using SnoopLigase (10 μM each) for 2 h in Tris phosphate citrate (25 mM phosphoric acid and 25 mM citric acid, pH-adjusted with Tris) with indicated pH. (b) pH-dependence. SnoopTagJr-AffiHER2 and SUMO-DogTag were ligated using SnoopLigase (10 μM each) for 1.5 h at 4
C in Tris borate with 15% (v/v) glycerol with the indicated pH. (c) Temperature-dependence. As in (b) at pH 7.25 at 4–37
C. (d) Glyceroldependence. As in (b) at pH 7.25 with 0–40% glycerol. (e) NaCl-dependence. As in (b) at pH 7.25 with 0–512 mM NaCl. (f) Effect of NaCl on time-course. As in (b) at pH 7.25 with or without 137 mM NaCl. Data are mean of triplicate 1 SD. Some error bars are too small to be visible. (Data were based on [8, 17])

7. Other additives: Reducing agents do not affect SnoopLigase reaction, so you can use any optimal conditions for your POI. Tween 20 and Triton X-100 are tolerated up to at least 2% (v/v).

24

Can M. Buldun et al.

8. Reaction time: Choose reaction time according to substrate concentration. At 10 μM concentration of components, we usually use 24–48 h. Cyclization reactions typically occur faster: 0.5–24 h [8]. 3.5 Determining the Efficiency of Ligation

1. Samples to be analyzed by SDS-PAGE: (a) SnoopLigase reaction mix and (b) substrate mix, i.e., the same as (a) but no SnoopLigase added. Add 10 μL of sample to a PCR tube. 2. Add 2 μL of 6 SDS loading buffer. 3. Heat samples for 5 min at 95
C in a PCR block with heated lid. 4. Let samples cool down to 25
C and briefly centrifuge. 5. Run samples on SDS-PAGE. 6. Stain the gel using InstantBlue or Coomassie. 7. Acquire an image of the gel, e.g., using a ChemiDoc XRS+ Imager (Bio-Rad). 8. Quantify the intensity of protein bands (densitometry) using suitable software, e.g., Image Lab Software (Bio-Rad). 9. Calculate reaction efficiency using band intensities (see Note 4): l If equimolar substrate concentrations are used, calculate the reaction efficiency based only on the bands in the reaction sample: %Product formed ¼

½Product  100% ½Product þ ½Substrates

If one substrate is used in excess, calculate the reaction efficiency by loss of the less concentrated substrate:   ½Substrate after reaction %Substrate reacted ¼ 1   100% ½Substrate before reaction l

3.6 Purification of Conjugated Product

Since the SnoopLigase system is derived from a tripartite split of protein RrgA from Streptococcus pneumoniae, the binding and conjugation of DogTag:SnoopTagJr by SnoopLigase reconstitutes a folded complex similar to the parent protein. As a consequence, we find that SnoopLigase still binds strongly to the reaction product. We capitalized on this product inhibition to remove excess unreacted substrates by solid-phase immobilization of SnoopLigase bound to the reaction product. We can then elute the purified reaction product by disrupting the SnoopLigase–product interaction (Fig. 4a) [8]. In fact, it is not always necessary to remove SnoopLigase from the reaction product, as we showed with our enzyme and vaccine applications [8, 11].

SPeptide Gene Delivery Packagers

25

Fig. 4 Purification of SnoopLigase reaction product. (a) Cartoon of solid-phase SnoopLigase purification. SnoopTagJr- and DogTag-linked proteins are covalently conjugated using biotin-SnoopLigase. Then streptavidin agarose (SA-agarose) binds biotin-SnoopLigase, unreacted proteins are washed away, and ligated proteins are subsequently eluted. Red represents the isopeptide bond. (b) Cartoon of peptide competitor production. SUMO-DogTag:SnoopTagJr covalent conjugate is produced by SnoopLigase and imidazole elution of the conjugate. SUMO is then cleaved from this conjugate by SUMO-protease Ulp1 (gray). Incubation with Ni-NTA resin depletes the His-tagged SUMO and Ulp1, yielding purified DogTag:SnoopTagJr peptide. (c) Analysis of product from SnoopLigase purification using three different elution methods. SnoopTagJr-AffiHER2 and SUMO-DogTag were ligated using biotin-SnoopLigase (10 μM each) for 16 h at 4
C. Biotin-SnoopLigase was captured with streptavidin agarose, followed by elution with glycine (pH 2.0) or imidazole or peptide. Analysis by SDS-PAGE with Coomassie staining. (Data adapted from [8]) 3.6.1 SnoopLigase Capture by Solid-Phase Immobilization

1. To 0.5 mL of SnoopLigase reaction mix containing 10 μM biotin-SnoopLigase or HaloTag7-SnoopLigase, add 0.5 μL of 10% (v/v) Tween 20. Tween 20 will decrease the resin sticking to plastic surfaces.

26

Can M. Buldun et al.

2. Add 25 μL (packed volume) of equilibrated HiCap Streptavidin agarose to reactions containing biotin-SnoopLigase, or 50 μL (packed volume) of equilibrated HaloLink resin to reactions containing HaloTag7-SnoopLigase. If SnoopLigase concentrations other than 10 μM are used, adjust the resin volumes accordingly. 3. Incubate for 30 min at 25
C with end-over-end rotation. Incubation time may need to be prolonged, depending on the conjugated proteins. 4. Equilibrate a 1-mL polyprep column with 0.5 mL of 50 mM Tris borate (pH 8.0) with 0.01% (v/v) Tween 20. 5. Transfer the sample containing the resin into the column. Place the column into a 1.5-mL tube and centrifuge for 1 min at 300  g. 6. Test the capture efficiency by SDS-PAGE and Coomassie staining, followed by densitometry analysis. Compare the amount of SnoopLigase and product in the reaction mix before the addition of resin, and the amount in the flow-through after SnoopLigase capture. 3.6.2 Product Elution

We have established three different methods for product elution. The elution method should be chosen based on the tolerance of the substrate molecules. Imidazole and low pH elution are the more convenient methods, while elution by peptide competition is more laborious and expensive but gentler. The procedures described below are for monomeric conjugates, i.e., one SnoopTagJr and one DogTag per product molecule. For multimeric conjugates, the procedure may need to be adjusted (see Note 3).

Product Elution by Imidazole

Perform all steps at room temperature (~25
C). 1. To the polyprep column containing the resin from Subheading 3.6.1, add five resin volumes of Tris phosphate (pH 7.0) containing 300 mM NaCl, 500 mM imidazole (pH 7.0), and 0.01% (v/v) Tween 20. Centrifuge for 1 min at 300  g. Repeat the wash once more. 2. Add five resin volumes of Tris phosphate (pH 7.0) containing 500 mM imidazole (pH 7.0). Centrifuge for 1 min at 300  g. Repeat the wash twice more. 3. Centrifuge for 1 min at 300  g to remove the remaining buffer. 4. Plug the bottom of the column. To elute the conjugated product, add one resin volume of Tris phosphate (pH 7.0) containing 2 M imidazole (pH 7.0). Incubate for 5 min on a Thermomixer at 500 rpm. Unplug the column and centrifuge

SPeptide Gene Delivery Packagers

27

for 1 min at 300  g. The flow-through contains the eluted product. 5. Repeat the elution step twice more. 6. Dialyze the eluted protein three times against 1000 volumes of a suitable buffer, e.g., PBS, at 4
C using a 3500-Da MWCO membrane. Product Elution by Low pH

Perform all steps at 4
C or on ice. 1. To the polyprep column containing the resin from Subheading 3.6.1, add five resin volumes of ice-cold 50 mM glycine (pH 3.0) containing 300 mM NaCl and 0.01% (v/v) Tween 20. Centrifuge for 1 min at 300  g, 4
C. Repeat the wash once more. 2. Add five resin volumes of ice-cold 50 mM glycine (pH 3.0). Centrifuge for 1 min at 300  g, 4
C. Repeat the wash twice more. 3. Centrifuge for 1 min at 300  g to remove remaining buffer. 4. Plug the bottom of the column. To elute the conjugated product, add one resin volume of ice-cold 50 mM glycine (pH 2.0) and incubate for 1 min. Unplug the column and centrifuge for 1 min at 300  g, 4
C into a 1.5-mL tube containing 0.3 resin volumes 1 M Tris–HCl, pH 9.5. 5. Repeat the elution step twice with centrifugation into the same 1.5-mL tube and without the addition of fresh Tris–HCl. 6. Dialyze the eluted protein three times against 1000-fold excess volume of a suitable buffer, e.g., PBS, at 4
C using a 3500-Da MWCO membrane.

Product Elution by Peptide Competition

Peptide Competitor Generation

The DogTag:SnoopTagJr competitor peptide is generated from His-tagged SUMO-DogTag and SnoopTagJr peptide (Fig. 4b). Here, we describe the liquid phase production of competitor peptide. Alternatively, a solid-phase approach could be used, as described previously [8]. 1. Express and purify SUMO-DogTag in E. coli from pET28aSUMO-DogTag, as described above for AviTag-SnoopLigase and HaloTag7-SnoopLigase. 2. Perform a conjugation reaction using 75 μM biotinSnoopLigase or HaloTag7-SnoopLigase, 75 μM SUMODogTag, and 150 μM SnoopTagJr peptide in 50 mM Tris borate (pH 7.25) with 15% (v/v) glycerol at 4
C for 4 h. 3. Capture the ligase and elute the reaction product using the imidazole elution method described above. Dialyze the eluted product three times against 1000 volumes of 50 mM Tris borate (pH 7.5) at 4
C using a 3500-Da MWCO membrane.

28

Can M. Buldun et al.

4. Concentrate the SUMO-DogTag:SnoopTagJr conjugate to 100 μM using a 10-kDa MWCO spin filter. Measure the protein concentration using a Nanodrop at OD280. 5. To the concentrated SUMO-DogTag:SnoopTagJr conjugate, add His-tagged SUMO-protease Ulp1 at 1:50 molar ratio. 6. Incubate for 45 min at 25
C. 7. Add Tween 20 to a final concentration of 0.01% (v/v). 8. To deplete the His-tagged proteins (SUMO and Ulp1), add Ni-NTA agarose at 1:4 reaction volume. Incubate for 1 h at 25
C, rotating end-over-end. 9. Centrifuge the sample for 1 min at 16,900  g. Collect the supernatant containing the DogTag:SnoopTagJr competitor peptide. The competitor peptide contains 3 tyrosine residues and therefore allows concentration determination by OD280. Measure the OD280 and calculate the molar concentration using the molar extinction coefficient of 4470 M1 cm1. Product Elution by Peptide Competition

1. To the polyprep column containing the resin from Subheading 3.6.1, add five resin volumes of Tris phosphate (pH 7.0) containing 0.01% (v/v) Tween 20. Centrifuge for 1 min at 300 g. Repeat the wash four more times. 2. Plug the bottom of the column. To elute the conjugated product, add two resin volumes of 100 μM DogTag:SnoopTagJr competitor solution generated above. 3. Incubate for 4 h at 37
C on a Thermomixer at 500 rpm. 4. Unplug the column and centrifuge for 1 min at 300  g to collect the supernatant. 5. Dialyze the eluted protein three times against 1000 volumes of a suitable buffer, e.g., PBS, at 4
C using a 3500-Da MWCO membrane.

3.6.3 Resin Regeneration and Reuse

Having eluted the reaction product from immobilized biotinSnoopLigase or HaloTag7-SnoopLigase, the remaining SnoopLigase resin can be regenerated, stored for at least 2 weeks at 4
C, and used for new conjugations at least eight times [8]. Perform all steps at 4
C or on ice to minimize protein hydrolysis. 1. To the polyprep column containing SnoopLigase-linked resin (from Subheading 3.6.2), add five resin volumes of ice-cold 50 mM glycine, pH 2.0. Centrifuge for 1 min at 300  g, 4
C. Repeat the wash once more. 2. Add five resin volumes of 50 mM Tris borate, pH 8.0. Centrifuge for 1 min at 300  g, 4
C. Repeat the wash once more. 3. Optional: To store the resin, plug the bottom of the column and add two resin volumes of 50 mM Tris borate (pH 8.0)

SPeptide Gene Delivery Packagers

29

containing 0.05% (w/v) sodium azide (Caution: sodium azide is toxic). Store at 4
C. No obvious loss in SnoopLigase activity was detected for storage up to 2 weeks. Longer storage times have not been tested. 4. To reuse the SnoopLigase resin, unplug the column and centrifuge the column for 1 min at 300  g, 4
C. Add five resin volumes of 50 mM Tris borate (pH 8.0) with 0.01% (v/v) Tween 20. Centrifuge for 1 min at 300  g, 4
C. Repeat the wash twice more. 5. To start a new ligation reaction, add 10–20 resin volumes of 10 μM of each SnoopTagJr-substrate and DogTag-substrate in 50 mM Tris borate (pH 7.25) with 15% (v/v) glycerol. Incubate for 24 h at 4
C on a Thermomixer at 800 rpm. Reaction conditions can be adjusted as described in Subheading 3.4. 6. Elute the product as described in Subheading 3.6.2 (Fig. 4c). 7. Repeat the regeneration process if required.

4

Notes 1. Biotin-SnoopLigase requires enzymatic biotinylation of AviTag-SnoopLigase, to enable subsequent solid-phase capture. HaloTag7-SnoopLigase does not require further processing, but the size of HaloTag7 might cause steric hindrance with some substrates, causing inefficient conjugation. HaloTag7 contains an intramolecular disulfide bond. During expression and purification, a fraction of the proteins might form undesirable intermolecular disulfide bonds. If homodimerization of HaloTag7 is problematic for the anticipated application, e.g., when purifying multimeric SnoopTagJr:DogTag conjugates, we suggest isolating the monomeric version by size exclusion chromatography or using the C61S C262S mutant HaloTag7SS instead [8, 22]. 2. AviTag-SnoopLigase and HaloTag7-SnoopLigase are soluble in a variety of buffers. However, SnoopLigase reaction works most efficiently in the absence of sodium or chloride ions, so we recommend dialyzing the proteins into a buffer without these ions [17]. 3. When purifying multimeric conjugates (e.g., ligating an antigen onto a heptameric scaffold) [11], it is important to use 100% biotinylated AviTag-SnoopLigase (streptavidin shift assay, Subheading 3.3) or well-purified monomeric HaloTag7-SnoopLigase, to prevent SnoopLigase contamination in the eluted product. Furthermore, the SnoopLigase capture and elution protocols need to be adjusted for multimeric conjugates. Use double the amount of resin to capture

30

Can M. Buldun et al.

SnoopLigase and increase the incubation period on the resin to 12 h. For imidazole elution, use 1 M instead of 0.5 M imidazole in the wash buffer, and 3.5 M instead of 2 M imidazole in the elution buffer. 4. SDS-PAGE densitometry, as an analytical method, has a limited dynamic range. If substrate concentrations higher than 10 μM have been used in the conjugation reaction, dilute the samples to be analyzed on SDS-PAGE to 10 μM substrate concentration before loading on the gel.

Acknowledgments Can M. Buldun and Irsyad N.A. Khairil Anuar contributed equally to this work. Funding for C.M.B. was provided by the Engineering and Physical Sciences Research Council (EPSRC) and Corpus Christi College Oxford. Funding for I.N.A.K.A. was provided by Yayasan Khazanah, Oxford Centre for Islamic Studies, and St. John’s College Oxford. M.H. was funded by the Biotechnology and Biological Sciences Research Council (BBSRC, BB/S007369/1). References 1. Banerjee A, Howarth M (2018) Nanoteamwork: covalent protein assembly beyond duets towards protein ensembles and orchestras. Curr Opin Biotechnol 51:16–23 2. Mootz HD (2009) Split inteins as versatile tools for protein semisynthesis. Chembiochem 10:2579–2589 3. Pishesha N, Ingram JR, Ploegh HL (2018) Sortase A: a model for transpeptidation and its biological applications. Annu Rev Cell Dev Biol 34:163–188 4. Nguyen GK, Wang S, Qiu Y et al (2014) Butelase 1 is an Asx-specific ligase enabling peptide macrocyclization and synthesis. Nat Chem Biol 10:732–738 5. Reddington SC, Howarth M (2015) Secrets of a covalent interaction for biomaterials and biotechnology: SpyTag and SpyCatcher. Curr Opin Chem Biol 29:94–99 6. Fierer JO, Veggiani G, Howarth M (2014) SpyLigase peptide–peptide ligation polymerizes affibodies to enhance magnetic cancer cell capture. Proc Natl Acad Sci 111: E1176–E1181 7. Wu XL, Liu Y, Liu D et al (2018) An intrinsically disordered peptide-peptide stapler for highly efficient protein ligation both in vivo

and in vitro. J Am Chem Soc 140:17474–17483 8. Buldun CM, Jean JX, Bedford MR et al (2018) SnoopLigase catalyzes peptide-peptide locking and enables solid-phase conjugate isolation. J Am Chem Soc 140:3008–3018 9. Schoene C, Bennett SP, Howarth M (2016) SpyRings declassified. Methods Enzymol 580:149–167 10. Brune KD, Howarth M (2018) New routes and opportunities for modular construction of particulate vaccines: stick, click, and glue. Front Immunol 9:1432 11. Andersson A-MC, Buldun CM, Pattinson DJ et al (2019) SnoopLigase peptide-peptide conjugation enables modular vaccine assembly. Sci Rep 9:4625 12. Fairhead M, Howarth M (2015) Site-specific biotinylation of purified proteins using BirA. Methods Mol Biol 1266:171–184 13. Veggiani G, Nakamura T, Brenner MD et al (2016) Programmable polyproteams built using twin peptide superglues. Proc Natl Acad Sci 113:1202–1207 14. Wieduwild R, Howarth M (2018) Assembling and decorating hyaluronan hydrogels with twin protein superglues to mimic cell-cell interactions. Biomaterials 180:253–264

SPeptide Gene Delivery Packagers 15. Wriggers W, Chakravarty S, Jennings PA (2005) Control of protein functional dynamics by peptide linkers. Biopolymers 80:736–746 16. van Rosmalen M, Krom M, Merkx M (2017) Tuning the flexibility of glycine-serine linkers to allow rational design of multidomain proteins. Biochemistry 56:6565–6574 17. Buldun CM (2017) Synthetic biology engineering to catalyse unbreakable linkage between peptide building blocks. DPhil Thesis, University of Oxford 18. Oesterle S, Roberts TM, Widmer LA et al (2017) Sequence-based prediction of permissive stretches for internal protein tagging and knockdown. BMC Biol 15:100

31

19. Salis HM (2011) The ribosome binding site calculator. Methods Enzymol 498:19–42 20. O’ Callaghan CA, Byford MF, Wyer JR et al (1999) BirA enzyme: production and application in the study of membrane receptor–ligand interactions by site-specific biotinylation. Anal Biochem 266:9–15 21. Zhang WB, Sun F, Tirrell DA et al (2013) Controlling macromolecular topology with genetically encoded SpyTag-SpyCatcher chemistry. J Am Chem Soc 135:13988–13997 22. Ke N, Landgraf D, Paulsson J et al (2016) Visualization of periplasmic and cytoplasmic proteins with a self-labeling protein tag. J Bacteriol 198:1035–1043

Chapter 3 Peptide Nanoparticles for Gene Packaging and Intracellular Delivery Paula Vila-Go´mez, James E. Noble, and Maxim G. Ryadnov Abstract Efficient gene transfer is necessary for advanced biotechnologies ranging from gene therapy to synthetic biology. Peptide nanoparticles provide suitable packaging systems promoting targeted gene expression or silencing. Though these systems have yet to match the transfection efficacy of viruses, they are typically devoid of drawbacks characteristic of virus-based vectors, including insertional mutagenesis, low packaging capacities, and strong immune responses. Given the promise nanoparticle formulations hold for gene delivery, methods of their preparation and accurate analysis of their physicochemical and biological properties become indispensable for progress toward systems that seek to outperform viral vectors. Herein, we report a comprehensive protocol for the preparation and characterization of archetypal peptide nanoparticles resulting from nonspecific and noncovalent complexation with RNA and DNA. Key words Nanoparticles, Gene therapy, Synthetic biology, Intracellular gene delivery, Peptide selfassembly

1

Introduction Nucleic acids (NAs) and oligonucleotides hold promise as drugs with a potential to uproot virtually any disease [1]. Quantitative and functional transfer of these molecules into cells and tissues presents an ultimate test of this potential. As a minimum, this requires an ability to traverse cellular and subcellular membranes, which nucleic acids lack. Various delivery systems have been proposed to endow them with cell-penetrating properties [2]. Therapeutic genes vary in size, chemistry, and folding topologies and are typically based on different RNA types and DNAs, from plasmids to chromosome-sized constructs [3]. Gene knockdown, gene editing, and protein synthesis are common processes underpinning therapeutic responses. Irrespective of their origin or size, these are anionic polymers that can complex with cationic systems. Therefore, and unsurprisingly, polycations are utilized as NA complexation agents [4]. These are chemically inert and water-soluble

Maxim G. Ryadnov (ed.), Polypeptide Materials: Methods and Protocols, Methods in Molecular Biology, vol. 2208, https://doi.org/10.1007/978-1-0716-0928-6_3, © Springer Science+Business Media, LLC, part of Springer Nature 2021

33

34

Paula Vila-Go´mez et al.

compounds, which make them compatible with cell and tissue cultures. Polycations based on polymers, however, tend to form polydisperse complexes that readily agglomerate and exhibit poor surface chemistry failing to support biological functionalization that is often required for specific targeting. Cell-penetrating polypeptides, unlike polymers, are devoid of most of these shortcomings and have an additional advantage of supporting foldingdependent complexation, which can mask nucleic acids via secondary and tertiary contacts, protecting them from enzymatic degradation. By tuning such contacts, it is possible to minimize aggregation effects and limit sizes to narrow distributions, though size ranges of 50–300 nm are usually viewed as optimal [5, 6]. The ability to penetrate cell membranes comes from cellpenetrating motifs identified in several proteins [7]. These are typically short helices derived from protein transduction domains such as the Antennapedia homeodomain [8], VP22 [9], and HIV-1 Tat protein [10], or bipartite structures such as transportan [11] and others [12]. These motifs are not typically used to complex NAs but rather are conjugated covalently, which imposes an additional requirement of breaking down the conjugates in the cytoplasm. However, their structure–activity relationships are reasonably well understood allowing for the design of purely artificial peptide sequences that exhibit similar cell-penetrating properties and acquire the capacity to package NAs [4]. Broadly, two main strategies of peptide-mediated gene packaging are used—condensation and encapsulation—both resulting in the formation of particulate nanostructures and both relying on noncovalent interactions to enable packaging. In complexation methods, peptides condense or wrap around nucleic acids, collapsing them into spheroid-like, solid aggregates, whereas encapsulation serves to encase NAs inside a formed particle in a manner similar to that of viruses. A detailed study on the comparative efficiency of the strategies in protecting NAs from degradation has yet to be reported. Transfection efficacies are routinely measured for every new variant of the two and to a large extent show comparable effects. However, other effects such as cytotoxicity and dependence of gene expression on the disassembly of the complexes, inclusive of nonpeptide origin, in the cytoplasm are often overlooked in performance assessments. This protocol considers both types and provides a workflow of synthesis and measurement methodologies that are typically used to produce and characterize peptide nanoparticle formulations. An emphasis is also made on the biological analysis of their properties including the determination of internalization mechanisms, highlighting potential adverse effects of endosomal entrapment [13]. Table 1 and Fig. 1 show NA-encapsulating and NA-condensing sequences exemplified in the protocol.

Peptide Gene Delivery Packagers

35

Table 1 Examples of condensing and encapsulating cell-penetrating peptides (CPPs) Cargo

Sequence

Encapsulating CPPs TecVir siRNA, pDNA [14]

CGGEIARLEQEIARLEQEIARLEYEIARLE-am

Capzip siRNA [15]

RRWTWE-βAla-K(RRWTWE)K(RRWTWE)-am

Condensing CPPs GeT

ssDNA, pDNA [4]

KIAKLKAKIQKLKQKIAKLK-am

CADY siRNA [6, 16]

ac-GLWRALWRLLRSLWRLLWKA-cysteamide [6] ac-GLWRALWRLLRSLWRLLWRA-cysteamide [16]

MPG

ac-GALFLGFLGAAGSTMGAWSQPKKKRKVcysteamide

siRNA, pDNA [17]

Pepfect SCOs [5, 18], pDNA [5], siRNA [5]

stearyl-AGYLLGKLLOOLAAAALOOLL-am

HALo

siRNA [19]

ac-CKLLKWLAKWLKWLAKWLKL-am

RALA

siRNA [20], pDNA [20, 21]

WEARLARALARALARHLARALARALRACEA

pDNA and ssDNA represent plasmid and single-stranded DNA, respectively, and SCO stands for splice correcting oligonucleotides

Fig. 1 A schematic of examples for peptide nanoparticle reagents. (a) and (b) show the condensation and selfassembly into particles of pDNA and mRNA with the peptide TC and polymer polyoxaine (Reprinted by permission from Springer-Nature: Nature Nanotechnology, Self-assembled peptide–poloxamine nanoparticles enable in vitro and in vivo genome restoration for cystic fibrosis, Guan, S. Munder, A. Hedtfeld, S. et al), (COPYRIGHT2019) [22]. (c) shows self-assembled peptides of TecVir forming a capsule and encapsulating siRNA (Reprinted with permission from (A De Novo Virus-Like Topology for Synthetic Virons). Copyright (2016) American Chemical Society [14])

36

2 2.1

Paula Vila-Go´mez et al.

Materials Equipment

1. An automated peptide synthesizer: Liberty microwave peptide synthesizer (CEM). 2. High-performance liquid chromatographic (HPLC) system (Jasco-2080 with UV 2072 detector). 3. MALDI-TOF (Bruker Daltonics) Autoflex III smartbeam, with flex control as proprietary software. 4. Malvern Zetasizer Nano Series Nano ZS with proprietary software. 5. Philips BioTwin transmission electron microscopy with 80 kV accelerating voltage or FEI Tecnai 20 with 200 kV accelerating voltage. 6. Nanodrop FluostarOmega BMG Labtech equipped with an LVis BMG Labtech plate and the Omega Mars software. 7. Cytoflex S (Beckman Coulter) flow cytometer equipped with the CytExpert software. 8. Microscope: FEI Tecnai 20 with 200 kV accelerating voltage and fitted with an Eagle 4 k camera (FEI). 9. Power bank (BioRad Power Pac 200) and gel imager (Gel Doc™ EZ Gel). 10. Synchrotron SAXS beamline B21 (Diamond Light Source, UK): B21 operated with a fixed camera length (4 m) and fixed energy (12.4 keV) allowing data collection for ˚ –1 (q ¼ 4πsinΘ/λ, with Θ: scattering angle q ¼ 0.004–0.4 A ˚ and λ ¼ 1 A); Pilatus 2 M detector for image acquisition. Sasfit software to model SAXS data. 11. Image analysis: ImageJ software. 12. Jasco J-810 spectropolarimeter or Chirascan Plus (Applied Photophysics, UK) spectropolarimeters fitted with a Peltier temperature controller. 13. Acquisition software: Chirascan in the case of the Chirascan Plus and Spectra Manager with the Jasco J-810.

2.2 Solid-Phase Peptide Synthesis

1. Solvents: Dimethylformamide (DMF), N,N-diisopropylethylamine (DIPEA), dichloromethane (DCM), trifluoroacetic acid (TFA), and triisopropyl silane (TIS). 2. Standard fluorenylmethoxycarbonyl (Fmoc) amino acids and Fmoc-Lys(Mtt)-OH for orthogonal synthesis for Capzip (Table 1). 3. O-benzotriazole-N,N,N0 ,N0 -tetramethyluronium hexafluorophosphate (HBTU), or O-(1H-6-chlorobenzotriazole-1-yl)1,1,3,3-tetramethyluronium hexafluorophosphate (HCTU).

Peptide Gene Delivery Packagers

37

4. 100–200 mesh (0.36–0.67 mmol/g) Fmoc-Gln(Trt) Wang resin. 5. 100–200 mesh (0.36–0.67 mmol/g) 4-methylbenzhydrylamine (MBHA) resin.

Rink

Amide

6. 20% piperidine in DMF. 7. Allyl deprotection: Tetrakis(triphenylphosphine)palladium (0) (4 eq) in CHCl3/AcOH/N,N-diisopropylethylamine (DIPEA)—3.6/0.2/0.2 mL. 8. diethyldithiocarbamate. 9. Phenylsilane. 10. Cleavage cocktail: Trifluoroacetic acid (TFA)/triisopropylsilane (TIS)/water (95:2.5:2.5). 2.3 Peptide Material Purification and Characterization

1. Reversed-phase high-performance liquid chromatography (RP-HPLC) mobile phases: 5% acetonitrile (AcN) + 95% water (buffer A), 5% water +95% AcN (buffer B) both containing 0.1% TFA. All solvents are HPLC grade. 2. Reversed-phase chromatography stationary phase: Vydac C8 and C18 reverse-phase columns. Analytical (5 μm, 4.6 mm i.d. 250 mm) and semi-preparative (5 μm, 10 mm i.d. 250 mm). 3. 3x3 mm light path, 9.65-mm Zentrum Hellma quartz cuvette and folded capillary cell for zeta potential measurements (Malvern Scientific). 4. 2,5-Dihydroxybenzoic acid and sinapinic acids as matrices for mass-spectrometry analysis. 5. 0.05-cm cuvette for circular dichroism spectroscopy

2.4 Transfection Reagents

1. Filtered water (0.22 μm) with a resistivity of 18.2 MΩ cm. 2. RNase-free water. 3. Synthetic peptides and DNA or RNA: peptides in Table 1 are dissolved in water; DNA is prepared in RNase-free water. 4. Lipofectamine LTX (Thermo Fisher Scientific). 5. phMGFP plasmid (Promega). 6. Plasmid labeling kit (Myrus). 7. siRNA (Eurogentec). Control siRNA duplex negative control. Control siRNA duplex GFP (jellyfish). Alexa-labeled siRNA: 50 -30 ! AF647-GCAAGCUGACCCUGAAGUUCTT (sense strand).

38

Paula Vila-Go´mez et al.

8. Lipofectamine RNAiMAX® (Thermo Fisher scientific). 9. N-TER™ (Sigma). 10. RNaseZap RNase decontamination solution (Thermo Fisher Scientific). 2.5 Cell Culture Reagents and Buffers

1. Dulbecco’s modified Eagle medium supplemented with 10% fetal bovine serum (FBS) and antibiotics (gentamicin, amphotericin B). 2. Serum-reduced medium (OptiMEM I) with 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 2.4 g/L sodium bicarbonate, L-glutamine, Hanks’ balanced salt solution (HBSS), CaCl2, MgCl2, TrypLE 1x (Gibco)32. Thermo Scientific™ BioLite cell culture treated flasks. 3. Incubator Heal Force HF90. 4. Cell counter Scepter™ (Millipore). 5. 96-well tissue culture coated plates (Corning incorporated Costar 5398). 6. Fluorescence plate reader. FluostarOmega Nanodrop (BMG Labtech) with the Omega Mars software. 7. 10 mM 3-(N-morpholino)propane sulfonic acid (MOPS), pH 7.4, in filtered water.

2.6 Flow Cytometry and Microscopy Reagents

1. 48 well tissue culture coated polystyrene plates. 2. U-shaped-bottom 96-well plates. 3. Gibco™ trypsin-EDTA (0.05%) and Gibco™ trypsin neutralizer solution. 4. CytoFLEX Daily QC fluorospheres and 510/20 + OD1 and 660 filters. 5. Olympus IX81 confocal microscope with the Olympus FluoView V4.2b software. 6. Oil for image acquisition at 60 magnification. Olympus Type F immersion oil in the case of the glass Ibidi chambers and Immersol™ 518 F oil in the case of collagen IV coated Ibidi chambers. 7. 15-μm slide 8-well chambers (Ibidi)—Collagen IV-coated for HEK293 cells and glass-bottom for HeLa cells. 8. Lysotracker DND 99 (Molecular Probes). 9. EM support grid: Formvar/carbon-coated copper/palladium support grids or copper grids coated with Pioloform and carbon film. All subjected to plasma glow discharge before use. 10. Staining reagent: 2% (w/v) phosphotungstic acid or 1% (w/v) uranyl acetate in filtered water.

Peptide Gene Delivery Packagers

39

11. Cryo-EM support grid: Lacey carbon grids (EM Resolutions Ltd.) subjected to plasma glow discharge before use. 12. Sample freezing: Plunge freezing into liquid ethane cooled with liquid N2 and performed using a VITROBOT mark IV (FEI Company). 2.7 Gel Retardation Assays

1. Anhydrous sodium acetate (Fisher). 2. Ethylenediaminetetraacetic acid (EDTA) (Sigma), glycerol (Fisher), SYBR Gold (Molecular Probes). 3. Agarose (Fisher). 4. 1-Kb Plus DNA ladder and a dsRNA ladder (NEB).

3

Methods

3.1 Peptide Preparation: Synthesis and Characterization

1. Assemble peptide sequences (Table 1) at a 100-μM scale in the automated peptide synthesizer on a Rink Amide MBHA resin. Use Fmoc/tBu solid-phase synthesis protocols and HBTU/ DIPEA as coupling reagents (see Note 1). 2. Wash resin with DCM to remove residual DMF. 3. Keep the resin under cleavage cocktail (10 mL/g of resin) for 3–4 h. 4. Recover the liquid phase by filtration (main filtrate) and rinse the resin with another 5 mL of the cocktail (second filtrate). 5. Add a threefold excess of diethyl ether to each filtrate. 6. Centrifuge the filtrates at 3500 rpm for 45 min at 4  C. 7. Discharge the supernatant to give peptide precipitates. 8. Dissolve the obtained precipitates in distilled water and freeze. 9. Lyophilize the solidified sample under vacuum. 10. Solubilize the lyophilized crude peptide in distilled water. 11. Inject 5 mL of the obtained peptide solution into a semipreparative RP-HPLC column (Vydac C18, 5 μm). 12. Run a 10–70% gradient of 95% aq. AcN containing 0.1% TFA over 50 min, with a flow rate of 4.7 mL/min. 13. Monitor the gradient at different wavelengths: 214 and 230 nm (peptide bonds), 280 nm (aromatic side chains), and 490 nm (fluorescein). 14. Collect eluted fractions every 30 s. 15. Mix 1 μL of DHB matrix (10 mg/mL in AcN/water 50:50 containing 0.1% TFA) with peptide (1 μL) and spot 1 μL of the obtained mixture on a MALDI-TOF plate to confirm peptide identity.

40

Paula Vila-Go´mez et al.

16. Combine pure fractions, freeze, and lyophilize. 17. Solubilize an aliquot of the purified peptide to confirm and quantify its purity by analytical RP-HPLC (Vydac C18, 5 μm) at a flow rate of 1 mL/min. 3.2 Nanoparticle Formation

1. Dissolve peptide powder (>90% purity) at ~3 mM in filtered (0.22 μm filter) MilliQ water.

3.2.1 Encapsulation

2. Measure peptide concentration via absorbance at 216, or 280 nm (if aromatic amino acids are present). 3. Dilute the peptide to 100 μM in 10 mM MOPS, pH 7.4. 4. Immediately after dilution, add either 30 pmol of siRNA, or 500 ng of plasmid DNA to form a complex with a charge ratio of +2 (see Note 2) and incubate it to allow capsule formation. For Capzip and TecVir capsule formation, incubate over 30 min or overnight, respectively.

3.2.2 Complexation

1. Dissolve peptide powder (>90% purity) at ~3 mM in filtered (0.22 μm filter) MilliQ water. 2. Measure the peptide concentration via absorbance at 216, or 280 nm (if aromatic amino acids are present). 3. Dilute the peptide to concentrations from 0.1 to 1.5 mM in MilliQ water, buffer, or OptiMEM. Specific solvents for peptides include PBS [19], 1 mM TFA (PepFect), or 5% glucose water solutions [6]. 4. Check plasmid purity by gel electrophoresis of the linearized plasmid as well as with the A260/280 absorbance ratio, kept at ~1.8. 5. Label plasmids using 5 times less Label IT fluorescent reagent than described in the manufacturer’s protocol to allow reporter gene expression. 6. Resuspend siRNA in RNase-free water at 100 μM. 7. Dilute the specific amount of oligonucleotide (typically 50–1000 ng) in RNase-free water targeting a final concentration of 100 μM for RNA or 50 ng/μL for pDNA, respectively (water is replaced with OptiMEM for Pepfect). 8. Add the peptide to the oligonucleotide dropwise and then mix at the desired charge ratio (see Note 2). 9. Incubate the mixture at room temperature or 37  C [6] for 30 min, 15 min (RALA protocol), or 1 h [18]. Vortex for 15 s for RALA. 10. The amount of complex required for transfection depends on cell numbers and cell type (see Note 3).

Peptide Gene Delivery Packagers

3.3 Cell Culture and Transfection

41

1. Culture HEK293 and HeLa cells in DMEM supplemented with 10% FBS and antibiotics at 37  C, 5% CO2. If using a different cell line, medium and medium supplements should be changed and optimized accordingly. 2. Seed the cells in complete DMEM in order to reach 60–90% confluency on the day of transfection. 3. For mRNA analysis, seed into 24-well plates; for flow cytometry experiments, seed into 48-well plates; for cell viability assays, seed into 96-well plates; and for confocal imaging analysis, seed into 8-well chambers. 4. On the day of the transfection experiment, remove the complete medium and wash the cells once with HBSS. 5. Add the required amount of OptiMEM to the preformed complexes at 250 μL final volume per cm2 of the well surface. 6. Remove HBSS used for washing the cells and add OptiMEM containing the complexes dropwise to the well. 7. Incubate the transfected cells under the optimal culture conditions for 3 h. Optimize incubation times depending on the system under study, ranging from 2 h to 6 h (RALA protocol) [18], to correlate the efficacy of the uptake [23]. 8. Remove the medium containing complexes and replace them with 250 μL of DMEM per cm2 of the well surface (see Note 4). 9. Incubate cells for an additional 24 h or 48 h if long incubations are required (e.g., gene expression, knockdown, or splice correction) or for shorter periods (1–5 h) (e.g., uptake, intracellular localization).

3.4

Flow Cytometry

Flow cytometry supports the assessment of transfection efficiencies when using a fluorescently labeled oligonucleotide or peptide, a plasmid encoding a fluorescent protein or siRNAs targeting mRNAs encoding for fluorescent proteins e.g., GFP. 1. Calibrate the flow cytometer using the manufacturer’s fluorescent beads. 2. Wash the cells once with HBSS and add 150 μL of trypsin to each well (for 1-cm2 wells). 3. Incubate the plate at 37  C in 5% CO2 for 10 min until the cells detach from the well after agitation, i.e., tapping the plate, as observed using microscopy. 4. Immediately add 150 μL of trypsin neutralizer (for 1-cm2 wells) to each well and again gently dislodge the cells, transfer to a flow cytometer compatible tube, or plate (see Note 5). 5. Set fluorescence acquisition parameters to FITC gain 50 and APC gain 326 (see Note 6).

42

Paula Vila-Go´mez et al.

6. Record 104 singlet cell events per sample. 7. Using the negative controls, cells treated with plasmid only, set a gate within the singlet population to assess the number of fluorescence-positive cells. From this singlet gated population, obtain the mean and median fluorescent intensity. 3.5 Confocal Microscopy

Use confocal microscopy to monitor the uptake of the fluorescently labeled transfection complex, to monitor endosomal entrapment using co-staining with CellLight™ or Lysotracker™ (Fig. 2), or to count the number of cells that display GFP expression. 1. After uptake or transfection, remove the medium and wash the cells in a staining compatible buffer. 2. Select dyes that are spectrally unique compared to those used to label the peptide or oligonucleotide, so no bleed-through between channels is observed when suitable lasers and filters are selected on the confocal microscope. 3. Prepare the staining according to the manufacturer’s instructions and incubate for the recommended amount of time. 4. If required, remove the stain, wash the cells with HBSS, and replace it with a fluorescence compatible medium. For longer or multi-time-point acquisition, supplement the medium with an antifade reagent. 5. Acquire the images with the Olympus IX81 confocal microscope under optimal cell culture conditions (37  C and 5% CO2). 6. Process the images using the ImageJ software to identify the number of transfection-positive cells, or co-location of the

Fig. 2 Endocytic trafficking and complexation of HAL -siRNA. (A) Confocal fluorescence micrographs of FLp-In TREx-293 cells with stained endosomal vesicles incubated for 2 h with HAL /siRNA at a 3/2 P/N charge ratio. Key: AF647-siRNA (red), nuclei stained with H33342 (blue), and endosomes and lysosomes stained with GFP (green). White dashed boxes are enlarged at the bottom left corner of each image. Original image reproduced with permission from Guyader, CP. Lamarre, B. De Santis, E. et al. [19]

Peptide Gene Delivery Packagers

43

membrane, nucleus, or lysozyme dyes with the transfection complex. 7. Optional: Use additional methods to assess transfection; e.g., luciferase reporters for both gene expression and recombination and RNA knockdown. Measure the luciferase activity either after cell lysis or sampled directly from the medium. Use qPCR to measure DNA uptake [23], or mRNA amount from cell lysates, quantify protein extraction from knockdown assays using western blotting, or MALDI-MS [24, 25]. 3.6

Cell Viability

Use cell viability assays to assess the potential toxicity associated with the use of nonviral transfection reagents. 1. Seed cells in 96-well plates in order to target 60–90% confluence on the day of transfection. Avoid using the wells placed at the plate edges. Establish a calibration curve within the sample plate starting at double the optimal cell seeding density and performing 1:2 dilutions from it. 2. Perform the transfection as in Subheading 3.3, adjusting the reagent amounts. Use OptiMEM for the untreated controls. 3. After 24 h or 48 h of incubation, check cell viability using the Presto Blue reagent (see Note 7). Prepare the cell viability reagent by diluting it 1:10 into complete DMEM. 4. Remove the transfection complex, wash the well twice in HBSS and replace it with the diluted cell viability reagent, and incubate following the manufacturer’s protocol. 5. Add the diluted cell viability reagent to 3 wells without cells. 6. Measure the plate absorbance or fluorescence, ensure that the calibration well with the highest number of cells returns a signal within the dynamic range of the instrument. 7. Subtract the background signal obtained for the reagent alone from all samples. Plot the calibration curve and ensure that all assay sample data lie within the linear portion of this plot. 8. Express the results as % of signal taking the untreated control samples as the reference.

3.7 Gel Retardation Assays

1. Form complexes as in Subheading 3.2 with a constant amount of DNA or RNA at 1 or 2 μg, respectively, across different samples; i.e., add different amounts of the peptide to vary the charge ratios of the complexes. 2. Mix 5 μL of the complex with 1 μl of glycerol and load onto the gel. 3. To spare lanes, add 1-Kb Plus ladder to estimate the DNA band sizes. Include in wells control samples of DNA only and peptide only to estimate changes in DNA retardation due to

Paula Vila-Go´mez et al.

DNA ladder

44

kDa

P/N charge ratio

HAL°

0

1/2

1

3/2

2

10 3 1 0.5

0.1

Fig. 3 Agarose gel electrophoresis of HAL0/siRNA complexation. Peptide/siRNA complexes are run on 2% agarose gel, with retardation of the DNA observed at peptide/nucleic acid charge ratios >1. HAL0 and 0 represent control incubations of peptide and DNA only (no retardation), respectively. (Original image reproduced with permission from Guyader, CP. Lamarre, B. De Santis, E. et al. [19])

peptide complex formation. For siRNA samples, use the dsRNA ladder. 4. Run the gel at 75 V for 45 min. 5. Acquire images of the gel on the BioRad Gel Doc EZ Imager. Changes in DNA or siRNA migration on the gel suggest complex formation with the peptide (Fig. 3). 3.8 Nanoparticle Characterization

1. Add OptiMEM or 10 mM MOPS (pH 7.4) to the preformed nanoparticle complexes. Use folded capillary cells to load the samples to acquire two measurements in a Zetasizer Nano. 2. Use triplicates of ten recordings of three independent preparations to obtain Z potential measurements. For size distribution, acquire five readings of twenty recordings, each with a 2-min separation for three independent preparation samples. 3. Use TEM to analyze particle sizes. Deposit droplets (8.0 μl) of preformed complexes on carbon-coated copper grids (1 min) and dry by blotting with filter paper. Stain the grids with filtered 0.75% (w/v) uranyl acetate (8 μL, 10 s). Blot excess staining with filter paper. Load samples onto a Philips BioTwin transmission electron microscope (see Note 8).

Peptide Gene Delivery Packagers

45

4. For cryo-EM: Prepare samples in a climate chamber with 100% relative humidity. Load a droplet (5 μL) on a glow-discharged lacey carbon grid and let the content of the droplet settle for 2–10 s before blotting. Plunge into liquid ethane cooled with liquid N2 for freezing. Peptide–oligonucleotide nanoparticles become embedded in vitreous ice suspended inside the holes of the carbon grid. Transfer the sample grid (without warming) into a Gatan cryo-holder filled with liquid N2 and image. 5. Use small-angle X-ray scattering (SAXS) to measure the capsule shell dimensions including wall thickness. Prepare capsules, oligonucleotides, and filled capsules in 10 mM MOPS (pH 7.4) to a final peptide concentration of 200–300 μM (oligonucleotide cargo concentration adjusted to obtain the desired charge ratio). Inject samples via an automated sample exchanger at a slow and reproducible flux into a quartz capillary (1.8 mm internal diameter), which is then placed in front of the X-ray beam. The quartz capillary is enclosed in a vacuum chamber in order to avoid parasitic scattering. After the sample was injected in the capillary and reached the X-ray beam, stop the flow during the SAXS data acquisition. 6. Capture images using a Pilatus 2 M detector. Model SAXS intensity using Sasfit software, considering the contribution of spherical shells with two different size distributions coexisting with cylinders in solution. 7. Use circular dichroism (CD) spectroscopy to probe folding properties of the formed complexes. Set acquisition conditions to 1 nm step, 1 nm spectral bandwidth, and 1 s collection time per step, and the Peltier temperature controller is set to 20  C. Collect four scans per sample. 8. Take CD measurements in millidegrees. Apply for data processing baseline correction and convert obtained values to mean residue ellipticities by normalizing for the concentration of peptide bonds and cuvette path length. Estimate the percent helix using the following equation [26]:
100 ½θ222 þ 3000 %helix ¼ 33, 000 9. Estimate the affinity of peptide–DNA and peptide–RNA interactions by preparing a series of complexes with fixed DNA (RNA) concentrations and by varying peptide concentration. Fit data points to an appropriate binding equation to derive the affinity constant for the interaction measured.

46

4

Paula Vila-Go´mez et al.

Notes 1. For the synthesis of Capzip (Table 1), Fmoc-Lys(Mtt)-OH is used to enable orthogonal conjugation via a tri-functional dendritic hub – βAla-KK-am. After the synthesis, add tetrakis (triphenylphosphine) palladium (0) (400 μM) in chloroform/ DIPEA/acetic acid (36:0.2:0.2, v) and phenylsilane (4 eq.) as a scavenger. Incubate for 2–16 h with gentle agitation in the dark. Wash with 0.5% sodium diethyldithiocarbamate (w/v in DMF), 0.5% DIPEA (v/v in DMF), and DMF. Repeat washing. 2. Charge ratio is the molar charge ratio and is calculated by determining the overall charge of the peptide and nucleic acid (P:N, or P/N). The optimum ratio is determined for each peptide system based on a biological effect, i.e., knockdown and potential toxicity. A complex will need an overall positive charge to be >1 to bind to negatively charged carbohydrates on the cell surface. An optimal charge ratio is determined for a particular cell line, peptide, and the length of oligonucleotide used and usually falls within the 1:1 to 10:1 range. Another approach to calculate the ratios between peptide and nucleic acids is to use molar concentrations [6]. For RNA, the literature usually references molar ratios in the range of 20–40 (WRAPS, CADY) or 3–20 (PepFect). 3. The volume of complex to be produced is calculated based on the amount of oligonucleotide to be added per surface area seeded with cells. This is typically 125 ng of DNA or 15 pmol of RNA per 1 cm2 of the well surface. Optimization is required for peptide to cell ratios for morphologically different cell lines [23]. 4. A variant of this step can also be to top up the well containing the transfection complex with serum-supplemented medium to the final concentration of 10% FBS per well [6, 14, 18]. 5. A better cell recovery is achieved by sampling from a U-bottom 96-well plate, compared to a flat-bottom plate. 6. Settings for flow cytometry are defined using positive (e.g., Lipofectamine®) and negative (untreated cells) control cell samples to allow effective gating and ensure that the positive control can be measured within the dynamic range of the instrument. For GFP knockdown assays, a neutral density filter may be required to ensure that fluorescence is within the dynamic range of the instrument. 7. Other methods to assess cell viability can also be used: cytotoxicity detection kit (LDH, Roche Diagnostics [6], AlamarBlue

Peptide Gene Delivery Packagers

47

reagent (ThermoFisher Scientific) [14, 19], Roche Wst-1 proliferation assay [18]). 8. Atomic force microscopy (AFM) imaging of nanoparticles dried onto a mica surface and in a liquid cell (in liquid AFM) can also be used to measure particle dimensions. For smaller capsules (200 nm), capzip, cryo-SEM is used to fracture the capsules and expose the lumen for characterization [15].

Acknowledgments We acknowledge funding from the UK’s Department for Business, Energy and Industrial Strategy. References 1. Sridharan K, Gogtay NJ (2016) Therapeutic nucleic acids: current clinical status. Br J Clin Pharmacol 82:659–672 2. Walsh G (2018) Biopharmaceutical benchmarks. Nat Biotechnol 36:1136–1145 3. Pahle J, Walther W (2016) Vectors and strategies for nonviral cancer gene therapy. Expert Opin Biol Ther 16:443–461 4. Lamarre B, Ravi J, Ryadnov MG (2011) GeT peptides: a single domain approach to gene delivery. Chem Commun 32:9045–9047 € Pooga M 5. Margus H, Arukuusk P, Langel U, (2016) Characteristics of cell-penetrating peptide/nucleic acid nanoparticles. Mol Pharm 13:172–179 6. Konate K, Lindberg MF, Vaissiere A, Jourdan C, Aldrian G, Margeat E, Deshayes S, Boisguerin P (2016) Optimisation of vectorisation property: a comparative study for a secondary amphipathic peptide. Int J Pharm 509:71–84 7. Deshayes S, Morris MC, Divita G, Heitz F (2005) Cell-penetrating peptides: tools for intracellular delivery of therapeutics. Cell Mol Life Sci 62:1839–1849 8. Derossi D, Joliot AH, Chassaing G, Prochiantz A (1994) The third helix of the Antennapedia homeodomain translocates through biological membranes. J Biol Chem 269:10444–10450 9. Normand N, Leeuwen H, Drew J, Phelan A, Brewis N, O’Hare P (1999) VP22-mediated

delivery of oligonucleotides into cells. Nat Biotechnol 17:40 10. Vive`s E, Brodin P, Lebleu B (1997) A truncated HIV-1 Tat protein basic domain rapidly translocates through the plasma membrane and accumulates in the cell nucleus. J Biol Chem 272:16010–16017 11. Pooga M, H€allbrink M, Zorko M, Langel U (1998) Cell penetration by transportan. FASEB J 12:67–77 12. Wender P, Mitchell DJ, Pattabiraman K, Pelkey ET, Steinman L, Rothbard JB (2000) The design, synthesis, and evaluation of molecules that enable or enhance cellular uptake: peptoid molecular transporters. Proc Natl Acad Sci 97:13003–13008 13. Li Z, Zhang Y, Zhu D, Li S, Yu X, Zhao Y, Ouyang X, Xie Z, Li L (2017) Transporting carriers for intracellular targeting delivery via nonendocytic uptake pathways. Drug Deliv 24:45–55 14. Noble JE, De Santis E, Ravi J, Lamarre B, Castelletto V, Mantell J, Ray S, Ryadnov MG (2016) A de novo virus-like topology for synthetic virions. J Am Chem Soc 138:12202–12210 15. Castelletto V, De Santis E, Alkassem H, Lamarre B, Noble JE, Ray S, Bella A, Burns JR, Hoogenboom BW, Ryadnov MG (2016) Structurally plastic peptide capsules for synthetic antimicrobial viruses. Chem Sci 7:1707–1711

48

Paula Vila-Go´mez et al.

16. Crombez L, Aldrian-Herrada G, Konate K, Nguyen QN, McMaster GK, Brasseur R, Heitz F, Divita G (2009) A new potent secondary amphipathic cell-penetrating peptide for siRNA delivery into mammalian cells. Mol Ther J Am Soc Gene Ther 17:95–103 17. Simeoni F, Morris MC, Heitz F, Divita G (2003) Insight into the mechanism of the peptide-based gene delivery system MPG: implications for delivery of siRNA into mammalian cells. Nucleic Acids Res 31:2717–2724 18. Ezzat K, Andaloussi SE, Zaghloul EM, Lehto T, Lindberg S, Moreno PM, Viola JR, Magdy T, Abdo R, Guterstam P, Sillard R, Hammond SM, Wood MJ, Arzumanov AA, € Gait MJ, Smith CI, H€allbrink M, Langel U (2011) PepFect 14, a novel cell-penetrating peptide for oligonucleotide delivery in solution and as solid formulation. Nucleic Acids Res 39:5284–5298 19. Guyader CP, Lamarre B, De Santis E, Noble JE, Slater NK, Ryadnov MG (2016) Autonomously folded α-helical lockers promote RNAi. Sci Rep 6:35012 20. Bennett R, Yakkundi A, McKeen HD, McClements L, McKeogh TJ, McCrudden CM, Arthur K, Robson T, McCarthy HO (2015) RALA-mediated delivery of FKBPL nucleic acid therapeutics. Nanomed 10:2989–3001

21. McCarthy HO, McCaffrey J, McCrudden CM, Zholobenko A, Ali AA, McBride JW, Massey AS, Pentlavalli S, Chen KH, Cole G, Loughran SP, Dunne NJ, Donnelly RF, Kett VL, Robson T (2014) Development and characterization of self-assembling nanoparticles using a bio-inspired amphipathic peptide for gene delivery. J Control Release 189:141–149 22. Guan S, Munder A, Hedtfeld S, Braubach P, Glage S, Zhang L, Lienenklaus S, Schultze A, Hasenpusch G, Garrels W, Stanke F, Miskey C, Johler SM, Kumar Y, Tu¨mmler B, Rudolph C, Ivics Z, Rosenecker J (2019) Self-assembled peptide-poloxamine nanoparticles enable in vitro and in vivo genome restoration for cystic fibrosis. Nat Nanotechnol 14:287–297 23. H€allbrink M, Oehlke J, Papsdorf G, Bienert M (2004) Uptake of cell-penetrating peptides is dependent on peptide-to-cell ratio rather than on peptide concentration. Biochim Biophys Acta 1667:222–228 24. Burns JR, Lamarre B, Pyne ALB, Noble JE, Ryadnov MG (2018) DNA origami inside-out viruses. ACS Synth Biol 7:767–773 25. Rakowska PD, Lamarre B, Ryadnov MG (2014) Probing label-free intracellular quantification of free peptide by MALDI-TOF mass spectrometry. Methods 68:331–337 26. Morrisett JD, Jackson RL, Gotto AM (1977) Lipid-protein interactions in the plasma lipoproteins. Biochim Biophys Acta 472:93–133

Chapter 4 Synthesis and Application of Peptide–siRNA Nanoparticles from Disulfide-Constrained Cyclic Amphipathic Peptides for the Functional Delivery of Therapeutic Oligonucleotides to the Lung Jade J. Welch, David A. Dean, and Bradley L. Nilsson Abstract The potential of RNAi therapies has been largely impeded by the inherent challenges in the functional delivery of siRNA to cells. Herein, we describe protocols for the synthesis and characterization of novel peptide–siRNA nanoparticles prepared from disulfide-constrained amphipathic peptides complexed with siRNA as promising siRNA delivery vectors. We also describe protocols for the application of these nanoparticles to the in vitro and in vivo delivery of siRNA to lung cells for the functional knockdown of lung proteins. Key words Amphipathic peptide, Cyclic peptide, Nanoparticle, Oligonucleotide delivery, RNAi, siRNA

1

Introduction RNA interference (RNAi) is an endogenous cellular process in which short interfering RNA (siRNA), double-stranded RNA segments approximately 21 nucleotides in length, is incorporated into the RNA induced silencing complex (RISC) to inhibit expression of complementary gene targets [1–4]. After the discovery of the RNAi mechanism for gene regulation in 1998, the potential for this strategy in the development of gene-based therapies prompted significant investment in the subsequent decade by both federal agencies and the pharmaceutical industry [5–7]. However, challenges with the delivery of siRNA into cells subsequently slowed the development of siRNA-based pharmaceuticals and curtailed investment in the technology. Recently, a revival in siRNA development has resulted in the first FDA-approved siRNA therapeutic [8]. This is exciting progress toward more widespread siRNA therapies, but significant barriers to delivery remain.

Maxim G. Ryadnov (ed.), Polypeptide Materials: Methods and Protocols, Methods in Molecular Biology, vol. 2208, https://doi.org/10.1007/978-1-0716-0928-6_4, © Springer Science+Business Media, LLC, part of Springer Nature 2021

49

50

Jade J. Welch et al.

Fig. 1 Disulfide-constrained amphipathic peptides mixed with siRNA condense into peptide–siRNA nanoparticles capable of transporting siRNA into the cell. In the reducing environment of the cell, the disulfide bond is reduced, the peptide is degraded by proteases, and the siRNA is released to the RNAi machinery

The translocation of siRNA and other therapeutic oligonucleotides across cellular membranes in vivo is met by a number of impediments [2, 9]. These include the inherent size and charge of oligonucleotides as well as the lack of stability of naked oligonucleotides to nucleolytic degradation [10–12]. In order to meet these challenges, translocation vectors that include lipid micelles and cell-penetrating peptides have been developed to aid transport of oligonucleotides across cell membranes and to facilitate the endosomal escape of the functional RNA [13–15]. We recently reported a new strategy for peptide-based siRNA delivery [16]. Disulfide-constrained cyclic amphipathic peptides can be noncovalently complexed with siRNA to form discreet nanoparticles. These nanoparticles protect siRNA from nucleolytic degradation and efficiently translocate siRNA into cells, facilitating the functional knockdown of target protein expression. The disulfideconstrained cyclic peptides are stable to protease degradation in the cyclic form. However, in the reducing environment of the cell, the disulfide-constraint is cleaved and the peptide is proteolytically degraded, releasing the siRNA to the RNAi machinery (Fig. 1) [16]. The precise mechanism of siRNA translocation has yet to be elucidated. The cyclic peptides used to form the peptide–siRNA nanoparticles are amphipathic peptides composed of alternating hydrophobic and hydrophilic/charged residues [16, 17]. The amphipathic segments are eight amino acids in length flanked by cysteine (Cys) residues that form the cyclic peptide by formation of an intramolecular disulfide bond. While several peptides have been tested for siRNA delivery in our lab, in this chapter we discuss the most effective of these, Ac-C(WR)4CG-NH2. This sequence was most

Peptide-siRNA Nanoparticles for Functional Delivery of siRNA

51

efficient for binding siRNA and for translocating siRNA across cell membranes [16, 18]. Herein, we provide detailed protocols for the solid-phase peptide synthesis of this peptide (referred to as the linear-amphipathic peptide or linear-AP) and oxidative disulfide bond formation to give the cyclic-amphipathic peptide (cyclic-AP as shown in Scheme 1). We also discuss methods for the formation of cyclic-AP–siRNA nanoparticles and the use of these nanoparticles for in vitro and in vivo delivery of siRNA to lung cells for functional knockdown of lung protein targets. The in vivo lung delivery studies are described for mouse models. These protocols will enable readers to apply these peptide-derived biomaterials for functional knockdown of gene targets in lung and potentially in other target tissues and organs.

Scheme 1 Disulfide-constrained amphipathic peptides are flanked at the N- and C-termini with cysteine residues, which facilitates peptide cyclization by the formation of an intramolecular disulfide bond under oxidizing conditions. The peptide contains an alternating sequence of hydrophobic and cationic residues, providing a hydrophobic face and a hydrophilic cationic face in the cyclic form

52

2

Jade J. Welch et al.

Materials

2.1 Peptide Synthesis and Cyclization

1. Low-loading Rink amide resin (0.1 mmol based on commercial resin loading) (see Note 1). 2. Fritted syringe (15 mL). 3. Rotisserie or tube rotator. 4. N,N-Dimethylformamide (DMF). 5. Dichloromethane (DCM). 6. Deprotection solution: 20% piperidine in DMF (v,v). 7. Cysteine: N-(9-Fluorenylmethoxycarbonyl)-S-trityl-L-cysteine (Fmoc-Cys(Trt)-OH). 8. Arginine: Nα-(9-Fluorenylmethoxycarbonyl)-Nω-(2,2,4,6,7pentamethyldihydrobenzofuran-5-sulfonyl)-L-arginine (Fmoc-Arg(Pbf)-OH). 9. Tryptophan: Nα-(9-Fluorenylmethoxycarbonyl)-N(in)-(tertbutoxycarbonyl)-L-tryptophan (Fmoc-Trp(Boc)-OH). 10. Glycine: N-(9-Fluorenylmethoxycarbonyl)-L-glycine (FmocGly-OH). 11. O-(Benzotriazol-1-yl)-N,N,N0 ,N0 -tetramethyluronium hexafluorophosphate (HBTU). 12. 1-[Bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b] pyridinium 3-oxid hexafluorophosphate (HATU). 13. Diisopropylcarbodiimide (DIC). 14. 1-Hydroxybenzotriazole (HOBt). 15. 1-Hydroxy-7-azabenzotriazole (HOAt). 16. N,N-Diisopropylethylamine (DIPEA). 17. Capping solution: 20% acetic anhydride in DMF (v,v). 18. Methanol. 19. Cleavage solution: 95% trifluoroacetic acid (TFA), 2.5% triisopropylsilane, 2.5% water (v,v,v) (see Note 2). 20. 5,50 -Dithiobis(2-nitrobenzoic acid) (Ellman’s reagent). 21. Phosphate-buffered saline solution (PBS): 0.1 M phosphate, 1 M NaCl, pH 8. 22. Diethyl ether. 23. Acetonitrile. 24. Water purified by filtration to a resistance of 18.2 MΩ. 25. 4,40 -Dipyridyl disulfide (PDS). 26. Aspirator attached to sidearm flask with an adaptor fitted to the size of a fritted syringe. 27. Benchtop centrifuge.

Peptide-siRNA Nanoparticles for Functional Delivery of siRNA

53

28. 50-mL conical tubes. 29. Lyophilizer. 30. Rotary evaporator. 31. MALDI Matrix solution: Saturated solution of α-cyano-4hydroxycinnamic acid in 60% acetonitrile in water. 32. MALDI-TOF mass spectrometer. 33. Semi-preparatory high-performance liquid chromatography instrument, Shimadzu LC-6 AD, with RP-C18 column (Phenomenex, Gemini 5u C18 110A, 250  4.6 mm). 34. Analytical high-performance liquid chromatography instrument, Shimadzu LC 2010-A, with analytical RP-C18 column (Phenomenex, Gemini 10 μm C18 110A, 250  21.2 mm). 35. Organic HPLC solution: HPLC-grade acetonitrile containing 0.1% TFA (semi-prep scale) or 0.05% TFA (analytical scale). 36. Aqueous HPLC solution: Water purified to a resistance of 18.2 MΩ containing 0.1% TFA (semi-prep scale) or 0.05% TFA (analytical scale). 2.2 siRNA–Peptide Nanoparticle Preparation

1. RNase-free water. 2. siRNA for the desired gene target, 3. 200 Mesh carbon-coated copper TEM grids 4. Uranyl acetate solution: 2% uranyl acetate in water, freshly filtered. 5. Transmission electron microscope, Hitachi 7650 TEM. 6. Dynamic light scattering (DLS) instrument, DynaPro Plate Reader II. 7. DLS plate, such as SensoPlate 1536-well glass-bottom plate. 8. 1.2% agarose gel impregnated with ethidium bromide 9. TBE buffer: 89 mM Tris-borate, 2 mM EDTA, pH 8.3. 10. UV illuminator.

2.3 In Vitro and In Vivo siRNA Delivery

1. Phosphate-buffered saline (PBS): 137 mM NaCl, 10 mM phosphate, 2.7 mM KCl, pH 7.4. 2. Serum-free Dulbecco’s modified Eagle’s medium (DMEM). 3. Roswell Park Memorial Institute medium (RPMI). 4. DMEM or RPMI-1640 with 10% fetal bovine serum (FBS). 5. Passive lysis buffer (Promega, Madison WI). 6. Cell lines—A549 and H441. 7. 12-well tissue culture plates 8. Humidified CO2 Waltham, MA).

Incubator

(ThermoFisher

Scientific,

54

Jade J. Welch et al.

9. Blunt-ended forceps. 10. Scissors for dissection. 11. Pipette. 12. Mouse anesthesia: 2–3% isoflurane with oxygen. 13. Mouse euthanasia: Sodium pentobarbital (100 mg/kg body weight). 14. 25-mL syringe fitted with a 25G needle 15. Liquid nitrogen bath. 16. Bead-o-lyzer or similar device. 17. Microcentrifuge.

3

Methods

3.1 Synthesis of Cyclic-Amphipathic Peptides 3.1.1 Solid-Phase Peptide Synthesis of Linear-Amphipathic Peptides (See Note 3)

1. Dispense the requisite amount of Rink amide resin (weigh out based on the indicated loading density of the resin) and place the resin into a fritted syringe. In this protocol, we illustrate the process for the synthesis of a peptide on a 0.1-mmol scale. 2. Swell resin in 10 mL of DCM for 5 min with gentle agitation on a rotisserie or tube rotator. Filter resin and rinse with 10 mL of DMF; repeat rinsing 3 times. Filtration and rinsing are facilitated by removal of liquids from the fritted syringe by vacuum filtration using a sidearm flask attached to a vacuum aspirator. 3. Add 10 mL of deprotection solution to resin and mix for 20 min with gentle agitation on a rotisserie or tube rotator. After 20 min, filter the resin and rinse 6 times with 10 mL of DMF. After the addition of each volume of rinse solution, vigorously agitate the resin for at least 30 s. 4. Activate Fmoc-Gly-OH (see Note 4): For the synthesis of a peptide on a 0.1-mmol scale, dissolve 0.4 mmol Fmoc-Gly-OH (4 equivalents), 0.4 mmol HOBt (4 equivalents), and 0.38 mmol HBTU in 10 mL of DMF. To this solution, add 0.4 mL of DIPEA. The solution should turn a dark orange color upon the addition of DIPEA. Allow this solution to stand for at least 10 min before proceeding to the next step. 5. Add the activated Fmoc-Gly-OH solution from step 4 to resin and agitate gently for 1.5 h using a rotisserie or tube rotator. After 1.5 h, filter the resin and rinse 3 times with 10 mL of DMF, mixing the resin for at least 30 s in each rinse step. 6. Repeat steps 4 and 5. 7. Repeat step 3 to deprotect the Fmoc-Gly residue on the resin.

Peptide-siRNA Nanoparticles for Functional Delivery of siRNA

55

8. Activate Fmoc-Cys(Trt)-OH: For the synthesis of a peptide on a 0.1-mmol scale, dissolve 0.5 mmol Fmoc-Cys(Trt)-OH and 0.5 mmol HOBt in 10 mL of DMF. Add 0.5 mmol of DIC. Allow mixture to stand for at least 10 min before proceeding to step 9. 9. Add activated Fmoc-Cys(Trt)-OH solution from step 8 to resin from step 7 and mix with gentle agitation for 2 h. Filter the resin and rinse 3 times with 10 mL of DMF, agitating the resin/DMF suspension for at least 30 s in each rinse step. 10. Repeat steps 8 and 9 twice. 11. Add 10 mL of capping solution to resin and mix with gentle agitation for 20 min. Filter the resin and rinse resin 3 times with 10 mL of DMF (see Note 5). 12. Repeat step 3. 13. Activate Fmoc-Arg(Pbf)-OH: Dissolve 0.4 mmol (4 equivalents) Fmoc-Arg(Pbf)-OH, 0.4 mmol HOAt, and 0.38 mmol HATU in 10 mL of DMF. Add 0.4 mL of DIPEA to this solution; the solution should turn dark orange. Allow this mixture to stand for 10 min before proceeding to step 14. 14. Add activated Fmoc-Arg(Pbf)-OH solution from step 13 to resin from step 12 and gently agitate for 1 h. After 1 h, filter the resin and rinse the resin 3 times with 10 mL of DMF, agitating the resin/DMF suspension for at least 30 s in each rinse step. 15. Deprotect as described in step 3. 16. Activate Fmoc-Trp(Boc)-OH: Dissolve 0.4 mmol (4 equivalents) Fmoc-Trp(Boc)-OH, 0.4 mmol HOAt, and 0.38 mmol HATU in 10 mL of DMF. Add 0.4 mL of DIPEA to this solution; the solution should turn dark orange. Allow this mixture to stand for 10 min before proceeding to step 17. 17. Add activated Fmoc-Trp(Boc)-OH solution from step 16 to resin from step 15 and gently agitate for 1 h. After 1 h, filter the resin and rinse the resin 3 times with 10 mL of DMF, agitating the resin/DMF suspension for at least 30 s in each rinse step. 18. Deprotect as described in step 3. 19. Repeat steps 13–18 3 times until the peptide sequence WRWRWRWRCG has been completed on the resin. 20. Add the final Cys residue by repeating steps 8–12. 21. After deprotection of the final amino acid, terminate the peptide with an acetyl cap by adding 10 mL of capping solution to the resin followed by gentle agitation for 20 min. 22. Filter the resin and rinse 3 times with 10 mL of DMF. Rinse the resin with 10 mL of DCM and then twice with 10 mL of

56

Jade J. Welch et al.

methanol, agitating the resin for at least 30 s for each rinse step. After the final rinse step, allow the resin to dry completely under vacuum for at least 1 h. It is essential that the resin is completely dried before the addition of cleavage solution (next step). 23. Add 8 mL of cleavage solution to resin and gently agitate the resin for 1 h. 24. Filter the resin and collect the cleavage solution in a 50-mL conical tube. 25. Repeat steps 23 and 24 and combine the cleavage solutions from the two resin treatments in the same 50-mL conical tube. 26. Reduce the volume of the cleavage solution by approximately 75% using a rotary evaporator. 27. Precipitate the product peptide by adding 15 mL of diethyl ether that has been cooled in an ice bath to the remaining volume of cleavage solution from step 26. 28. Collect the peptide product by centrifugation of the suspension from step 27 and carefully decant the supernatant solution from the peptide pellet. 29. Resuspend the crude peptide pellet in 15 mL of cold diethyl ether. Repeat step 28. 30. Dissolve the peptide pellet in 50 mL of 60% acetonitrile in water. 3.1.2 Disulfide Bond Formation for Cyclic-Amphipathic Peptides

1. Determine the concentration of peptide in the solution from step 30 in Subheading 3.1.1 by quantifying the sulfhydryl groups using Ellman’s solution (see Note 6) [19]. 2. Prepare Ellman’s solution: Dissolve 0.01 g of Ellman’s reagent in 10 mL of phosphate-buffered saline (PBS). 3. Prepare blank solution: Mix 1.25 mL of PBS, 0.025 mL of Ellman’s solution, and 0.025 mL of 60% acetonitrile in water. 4. Prepare peptide solution: Mix 1.25 mL of PBS, 0.025 mL of Ellman’s solution, and 0.025 mL of peptide solution from step 30 in Subheading 3.1.1. 5. Allow solutions from steps 3 and 4 to stand for 15 min at room temperature. 6. Measure the UV absorbance of solutions from steps 3 and 4 at 412 nm, subtracting the value of the blank solution (3) from that of the peptide solution (4). 7. Determine the concentration of sulfhydryls using Beer’s law (A421 nm ¼ εcl) with a molar extinction coefficient of ε ¼ 14,150 cm1 M1 [19]. The concentration of the peptide in the solution from step 30 in Subheading 3.1.1 is half the concentration of sulfhydryl groups.

Peptide-siRNA Nanoparticles for Functional Delivery of siRNA

57

8. Dilute the peptide from step 30 in Subheading 3.1.1 with 60% acetonitrile in water to a concentration of less than 90 μM to minimize intermolecular disulfide formation during cyclization (see Note 7). 9. Remove a 1-mL aliquot from the solution in step 8 for later testing. 10. Make a solution of PDS in 20 mL of methanol using the formula below to determine the amount of PDS to use (see Notes 8 and 9). Amount of PDS ðmolÞ ¼ 4  mol peptide from step 1 11. Slowly add 5 mL of this PDS solution to the peptide solution from step 8 every 12 h. Addition of each 5-mL aliquot of PDS should occur over 2 min. 12. Monitor the cyclization reaction by analytical HPLC, comparing to the aliquot removed in step 9. The cyclized peptide will have an earlier retention time, as shown in Fig. 2. We conducted analytical HPLC using a Shimadzu LC 2010-A instrument equipped with a Phenomenex, Gemini 10 μm C18 110A, 250  21.2 mm reverse-phase column. A flow rate of 1 mL min1 with a gradient of 5% organic HPLC solution isocratic for 5 min followed by a gradient of 5–95% organic

Cyclic-AP Linear-AP

10.5

11

11.5

12

12.5

13

13.5

14

14.5

Time (min.)

Fig. 2 Representative analytical HPLC trace depicting the shift in retention time between the linear- and cyclic-amphipathic peptides. These HPLC data were acquired using a Shimadzu 2010-A HPLC instrument (Phenomenex, Gemini 5u C18 110A, 250  4.6 mm) equipped with a reverse-phase column at a flow rate of 1 mL min1 with a gradient of isocratic 5% organic HPLC solution followed by a gradient of 5% to 95% organic HPLC solution over 10 min

58

Jade J. Welch et al.

HPLC solution over 10 min was used. The injection volume of the reaction solution was 100 μL. 13. Once analytical HPLC indicates that cyclization is complete, freeze and lyophilize the cyclized crude cyclic-amphipathic peptide directly from the cyclization solution (see Note 10). 3.1.3 Purification and Analysis of Cyclic-AP

1. Dissolve lyophilized peptide from cyclization protocol in 10 mL of 60% acetonitrile in water with 1% TFA (see Note 11) and purify by HPLC. 2. For a semi-preparative-scale purification, we used a Shimadzu LC-6 AD HPLC instrument with a Gemini C18 column. A flow rate of 10 mL min1 was used with a gradient of 21–24% organic HPLC solution (see item 32 and 33 in Subheading 2.1) over 15 min. A typical injection volume is 0.3 mL. 3. Monitor eluent with a UV detector at absorbances of 215 nm and 254 nm and collect fractions in test tubes or conical tubes. 4. Determine the identity of collected peptides by MALDI-TOF mass spectrometric analysis (see Note 12). 5. Deposit 0.5 μL of MALDI matrix solution onto a MALDI plate using a pipettor. 6. To each deposited spot in step 5, add 0.5 μL of HPLC eluent for each collected peak from step 1. 7. Allow these co-spotted samples to air-dry. 8. Perform MALDI-TOF analysis. We performed this analysis using a Bruker MALDI-TOF instrument. 9. Pool fractions of cyclic-AP. Other byproduct peptides can be discarded (see Note 12 for a discussion of some common byproducts). 10. Confirm the purity of cyclic-AP by analytical HPLC. 11. For an analytical-scale analysis, we used a Shimadzu LC-2010A HPLC instrument with a Phenomenex, Gemini 5u C18 110A, 250  4.6 mm column. A flow rate of 1 mL min1 was used with a gradient of 5% organic HPLC solution (see items 32 and 33 in Subheading 2.1) to 95% organic HPLC solution over 10 min. 12. Monitor eluent by using a UV detector at absorbances of 215 nm and 254 nm. 13. Combine, freeze, and lyophilize all fractions containing cyclicAP peptide (see Note 13). 14. Prepare a concentration curve for determination of cyclic-AP concentration for siRNA–peptide nanoparticle preparation (see Note 14) [16, 20].

Peptide-siRNA Nanoparticles for Functional Delivery of siRNA

59

15. Prepare a stock solution of the purified peptide from step 13 in 60% acetonitrile in water (v/v). 16. Remove 100 μL from this stock solution and make 10 serial 1:2 dilutions. 17. Analyze each dilution in triplicate by analytical HPLC. Use the same or similar conditions as described in step 11. Take note of the injection volume so that the peak area can be correlated to the concentration of the dilution (see subsequent steps). 18. Determine the concentration of the peptide stock solution by amino acid analysis. We use a commercial vendor that conducts amino acid analyses as a service. Using the amino acid concentration, extrapolate the concentration of each dilution from step 16. 19. Create the concentration curve by plotting the average peak areas for each dilution against the peptide concentration (or amount of peptide) for each dilution (see Note 14). The peptide concentration for each point is extrapolated from the amino acid analysis data and the volume of the injection. 3.2 Formation of Cyclic-AP–siRNA Nanoparticles 3.2.1 Nanoparticle Formation

1. Make siRNA stock by dissolving in RNase-free water according to the manufacturer’s instructions. 2. Add appropriate volume of siRNA to cyclic-AP dissolved in RNAse-free water so that the resulting mixture has the desired ratio of cyclic-AP:siRNA. 3. For nanoparticle characterization: Use a 1000:1 ratio of cyclicAP:siRNA (1250 μM of cyclic-AP, 1.25 μM of siRNA). 4. For in vitro and in vivo delivery studies: Use a cyclic-AP concentration of 100 μM and an siRNA concentration of 600 nM. 5. Allow the cyclic-AP:siRNA to stand at room temperature for 30 min.

3.2.2 Nanoparticle Characterization: Agarose Gel Electrophoresis

1. A gel shift analysis is used to assess cyclic-AP binding to siRNA. 2. Prepare the agarose gel setup with a 1.2% agarose gel (w/v) impregnated with ethidium bromide and TBE buffer as the running buffer. 3. Load 10 μL of the cyclic-AP:siRNA mixture onto the agarose gel. Also, load naked siRNA onto the gel for comparison. 4. Run the gel for 30 min at 70 V. 5. Visualize the gel with a UV illuminator to show the comparative positions of the naked siRNA and the cyclic-AP:siRNA particles. Successful nanoparticle formation causes a dramatic shift of the siRNA band, as shown in Fig. 3a.

60

Jade J. Welch et al.

a

b

c

35% 30% 25%

Intensity

100 nm

20% 15% 10% 5% 0.01

0.1

1

10 100 Radius (nm)

1000

10000

. siR 1 eq

0.5

eq.

NA siRN A siRN C(F A + 100 KF x siRN E)2 CG A+ C(F 1 KFE 000x ) siRN 2 CG A C(W + 100 R) C x 4 G

200 nm

Fig. 3 (a) Gel shift assay showing the shifting of the siRNA band when nanoparticles are formed. (b) TEM images of peptide–siRNA nanoparticles. (c) Representative DLS data showing hydrodynamic radii and distribution of nanoparticles. (Adapted with permission from Welch JJ, Swanekamp RJ, King C, et al. (2016) Functional Delivery of siRNA by Disulfide-Constrained Cyclic Amphipathic Peptides. ACS Med Chem Lett 7:584–589 Copyright 2016 American Chemical Society) 3.2.3 Nanoparticle Characterization: Transmission Electron Microscopy

1. Pipette 5 μL of the cyclic-AP:siRNA mixture onto a 200-mesh carbon-coated copper grid and allow it to stand for 60 s. Carefully remove the liquid by capillary action with filter paper. 2. Pipette 5 μL of uranyl acetate solution onto the grid and allow this to stand for 60 s. Carefully remove the liquid by capillary action with filter paper. 3. Allow the grid to air-dry for 10 min. 4. Obtain images using a transmission electron microscope; we performed these analyses with a Hitachi 7650 transmission electron microscope at an accelerating voltage of 80 kV. Representative nanoparticle images are shown in Fig. 3b. 5. Use ImageJ software to measure nanoparticle dimensions.

3.2.4 Nanoparticle Characterization: Dynamic Light Scattering

1. Pipette a sample of the cyclic-AP:siRNA mixture into a dynamic light-scattering plate fitted to the instrument you will use (see Note 15). The volume of the sample depends on the size of the plate used. We used a 1536-well plate and used 10 μL of the sample per well. Pipette carefully to avoid air bubbles, which will interfere with measurements. 2. Characterize the hydrodynamic radius of the cyclic-AP:siRNA nanoparticles using the plate reader. We used a DynaPro Plate Reader II, and the Dynamics software was used for data analysis.

Peptide-siRNA Nanoparticles for Functional Delivery of siRNA

61

3. Set the Mw-R model to “Globular Proteins,” acquisition time to 5 s, and the number of acquisitions to 10. 4. The hydrodynamic radius should be averaged from several readings in different wells. Representative data are shown in Fig. 3c. Note that these sizes are larger than the particles observed by TEM images in Fig. 3b. This is because the particles aggregate and form larger clusters in solution. 3.3 siRNA Delivery Using siRNA–Peptide Nanoparticles 3.3.1 In Vitro Delivery

1. Grow cells to confluency in 12-well plates and wash 2 times with 3 mL of PBS per well. 2. Allow nanoparticles to form for 30 min as described in step 4 in Subheading 3.2.1. 3. Suspend cyclic-AP:siRNA nanoparticles in 50 μL of PBS. 4. Immediately before addition to cells, dilute nanoparticles into 1 mL of serum-free DMEM. 5. Add 300 μL of diluted nanoparticles to each well (3 wells per sample) and incubate for 4 h. 6. Wash cells: Remove media by aspiration and wash twice with 3 mL of PBS per well. Aspirate last PBS wash. 7. Add 3 mL per well of DMEM or RPMI-1640 (for A549 and H441 cells, respectively) containing 10% FBS to cells. 8. Wash cells as in step 6 after 48 h. 9. Lyse using 0.4 mL of lysis buffer. 10. Protein concentration can be determined by SDS-PAGE for western blots using antibodies against appropriate protein target and β-actin or GAPDH (control) as shown in Fig. 4a, b.

3.3.2 In Vivo Delivery to Mouse Lung

1. Follow instructions given in step 4 in Subheading 3.2.1 to prepare the cyclic-AP:siRNA nanoparticles. 2. Anesthetize mice in an induction chamber with isofluorine (2–3% with oxygen) and place in the supine position. 3. Gently pull out tongue using blunt-ended forceps. 4. Using a pipette, deposit the cyclic-AP:siRNA nanoparticles suspended in 50 μL of saline from step 1 onto the back of the tongue while continuing to hold the animal’s tongue. 5. Using a second set of forceps, gently pinch the mouse’s nose until the animal has aspirated the cyclic-AP:siRNA nanoparticle fluid. 6. Upon aspiration, release the tongue and nose. 7. Allow the animal to recover for 48 h. 8. Euthanize mice by sodium pentobarbital overdose (100 mg/ kg BW).

62

Jade J. Welch et al.

Fig. 4 (a, b) TTF-1 siRNA delivered to H441 cells. (a) Raw western blot showing TTF-1 compared to the GAPDH control. (b) Relative TTF-1 protein concentration to show functional knockdown based on the siRNA delivery method. (c) In vivo delivery to mouse lung shows effective TTF-1 knockdown using the cyclic-amphipathicsiRNA nanoparticles. (Adapted with permission from Welch JJ, Swanekamp RJ, King C, et al. (2016) Functional Delivery of siRNA by Disulfide-Constrained Cyclic Amphipathic Peptides. ACS Med Chem Lett 7:584–589; Copyright 2016 American Chemical Society)

9. Following euthanasia, open the chest cavity and cut away ribs with scissors to reveal the heart and lungs. 10. Cut the renal artery to facilitate the removal of blood by perfusion. 11. Perfuse the lungs by inserting a 25G needle into the right ventricle and deliver 20 mL of PBS in a slow steady stream (blood and perfusate will pool from the kidneys). 12. Deliver PBS until the lungs are white and the fluid coming out of the renal artery is clear—this usually takes 10–15 mL. 13. Remove the lungs from the chest and snap-freeze the individual lobes in a liquid nitrogen bath. 14. Transfer frozen lobes to microcentrifuge tubes containing 0.4 mL of frozen lysis buffer and freeze at 80 until further use.

Peptide-siRNA Nanoparticles for Functional Delivery of siRNA

63

15. Lyse tissue by mechanical disruption (e.g., bead-o-lyzer or silar device) and remove debris by centrifugation for 2 min at 16,000  g in a microcentrifuge. 16. Transfer lysate to a fresh tube. 17. Store at 80  C until use. 18. Protein expression can be quantified by western blot analysis, as shown in Fig. 4c.

4

Notes 1. Determine appropriate resin mass based on the loading reported on the certificate of analysis. Amount of resin ðgÞ ¼

Desired mmol peptide
 Reported resin loading mmol g

We experienced significantly higher yields of the synthetic peptide by reducing the loading density of the resin to 0.1 mmol/g. Typical commercial low-loading rink amide resin is approximately 0.3 mmol/g. To lower the resin loading density, acetic anhydride is used to add an acetyl cap to a portion of the active amine sites on the resin, effectively reducing peptide density on the resin, which improves yields of these peptides. The process for lowering the resin loading density is as follows: (a) Add an amount of resin to a fritted syringe that matches your required amount of the peptide: Amount of resin ðgÞ ¼

Desired mmol peptide 0:1 mmol
¼ mmol 0:1 mmol Desired resin loading g g

(b) Follow Steps 1–3 in Subheading 3.1.1. (c) On the resin certificate of analysis, find the commercial resin loading and use this to determine the volume of acetic anhydride to be added to the Rink amide resin: Volume of acetic anhydride ðmLÞ ¼     mmol Amount of resin ðgÞ  Reported resin loading  Desired mmol peptide g ! g 102:09 mol  g 1000 mmol mol  1:08 mL (d) Add calculated volume acetic anhydride to 10 mL of DMF.

64

Jade J. Welch et al.

(e) Add acetic anhydride solution to resin for 1 h with gentle agitation. (f) Rinse with 10 mL of DMF; repeat rinsing process 3 times. Continue to step 4 in Subheading 3.1.1. 2. A cleavage solution containing 2.5% ethanedithiol (EDT) (instead of triisopropylsilane) by volume is recommended for cleavage of peptides containing cysteine residues. In our experience, the use of EDT has not changed the yield or purity of these peptides [21]. 3. Instructions are given for manual solid-phase peptide synthesis. This process can also be performed using an automated peptide synthesizer. We found that peptides made on an automated synthesizer had a lower purity and isolated yield than those that were prepared manually as described. If an automated synthesizer is used, we recommend coupling the cysteine residues manually instead of on the synthesizer, especially if the synthesizer utilizes heating in any form. 4. The peptide sequences are flanked at each end by cysteine. Cysteine has a high propensity for racemization during solidphase peptide synthesis, and this can be exacerbated when the cysteine is being attached to resin [22, 23]. Thus, we do not attach cysteine residues directly onto the resin. Instead, we attach a glycine residue to the Rink amide resin and then add cysteine, which reduces problems with cysteine racemization. 5. We have found that cysteine couplings can be problematic. For this reason, we recommend including an acetyl capping step at this point of the protocol to terminate any nascent peptide strands in which cysteine has not been correctly incorporated. This simplifies the purification of the peptide. 6. We have found that the crude peptide is sometimes poorly soluble in PBS, which complicates determination of sulfhydryl group concentration as described. In this case, we estimate the concentration of the peptide by assuming a synthetic yield of approximately 75% based on resin loading. For example, in a 0.1-mmol-scale synthesis, an assumed 0.075 mmol of the peptide should be diluted to 800 mL to reach an approximate concentration of 90 μM. 7. The typical pH of these solutions is 3, although the addition of up to 0.5% TFA by volume to improve peptide dissolution still allows cyclization to proceed without hindrance at pH 1 [24]. 8. If the crude peptide mixture is insoluble in PBS, estimate the amount of PDS needed by taking the estimated amount of the peptide (see Note 6) and multiplying by 4. For example, for a 0.1-mmol-scale peptide synthesis, 0.4 mmol of PDS should be added to the cyclization reaction.

Peptide-siRNA Nanoparticles for Functional Delivery of siRNA

65

Fig. 5 Structures and masses of byproducts indicative of poor-quality 4,40 -PDS. Linear-AP, Ac-C(WR)4CG-NH2, has a mass of 1690 g mol1. These observed byproducts have masses of 1799.8 g mol1 (1690.8 + 109) and 1908.8 g mol1 (1690.8 + 218)

9. Dilution of PDS into 20 mL of methanol (to approximately 10 mM) and slow addition reduces the occurrence of peptide dimerization or higher-order oligomerization [24]. 10. In our experience, reducing the volume of the cyclization reaction by rotary evaporation has led to lower yield and lower purity of the cyclic peptide. Freezing the cyclization solution followed by lyophilization results in higher quality and yield of the cyclic-amphipathic peptide. 11. Addition of TFA aids in dissolution of peptides. If you encounter problems with the complete dissolution of the peptide, the crude cyclized peptide can also be dissolved in DMSO prior to HPLC purification. 12. We have observed certain byproducts that are indicative of poor-quality PDS reagent (see Fig. 5). These byproducts have masses that are +109 and + 218 g mol1 relative to the mass of the desired product. These masses correlate to uncyclized linear-AP that have formed disulfide bonds with pyridine-4-thiol from PDS. These peptides can be converted to linear-AP by treatment with an excess of reducing agent, such as tris(2-carboxyethyl)phosphine (TCEP); the recovered linear-AP can

66

Jade J. Welch et al.

then be resubjected to the cyclization conditions to form cyclicAP. When these byproducts are observed, PDS should be purified or repurchased. These byproducts are only observed with aged PDS, which has presumably been contaminated with oxidized material. 13. Long-term storage of peptides should be under argon at 20  C to avoid oxidation. 14. At low concentration dilutions, the curve will be linear; at higher concentrations, the curve will begin to diverge from linearity. Include only values that appear in the linear portion of the curve to ensure accurate concentration determination. Use of concentration curves to quantify peptide is only reliable for the same instrument, column, and conditions that were used to construct the curve. We recommend reconstructing concentration curves every 6 months to account for column changes upon aging. 15. It may be necessary to remove air bubbles by centrifuging the dynamic light-scattering plate since any air bubbles will invalidate the measurements. Each sample should be divided into 4 wells and centrifuged using a microplate centrifuge or plate rotor at a spin rate of 600  g. Any wells containing bubbles or particulates should be excluded from analysis.

Acknowledgments This work was supported by the National Science Foundation (DMR-1148836) and the National Institutes of Health (R01HL138538, R01HL120521, and EB9903). We thank Karen Bentley and Gayle Schneider of the University of Rochester Medical Center Electron Microscopy Core for assistance with transmission electron microscopy and Jermaine Jenkins of the University of Rochester Medical Center for assistance with dynamic light scattering. References 1. Schwarz DS, Hutvagner G, Du T et al (2003) Asymmetry in the assembly of the RNAi enzyme complex. Cell 115:199–208 2. Tatiparti K, Sau S, Kashaw SK, Iyer AK (2017) siRNA delivery strategies: a comprehensive review of recent developments. Nanomaterials 7(4):77 3. Fire A, Xu S, Montgomery MK et al (1998) Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature 391:806–811

4. Whitehead KA, Langer R, Anderson DG (2009) Knocking down barriers: advances in siRNA delivery. Nat Rev Drug Discov 8:129 5. Beierlein JM, Mcnamee LM, Ledley FD (2017) As Technologies for nucleotide therapeutics mature, products emerge. Mol Ther Nucleic Acid 9:379–386 6. Garber K (2017) Worth the RISC? Nat Biotechnol 35:198 7. Chakraborty C, Sharma AR, Sharma G et al (2017) Therapeutic miRNA and siRNA:

Peptide-siRNA Nanoparticles for Functional Delivery of siRNA moving from bench to clinic as next generation medicine. Mol Ther Nucleic Acids 8:132–143 8. Garber K (2018) Alnylam launches era of RNAi drugs. Nat Biotechnol 36:777 9. Ho W, Zhang X-Q, Xu X (2016) Biomaterials in siRNA delivery: a comprehensive review. Adv Healthc Mater 5:2715–2731 10. Osborn MF, Khvorova A (2018) Improving siRNA delivery in vivo through lipid conjugation. Nucleic Acid Ther 28:128–136 11. Juliano RL, Ming X, Nakagawa O (2012) Cellular uptake and intracellular trafficking of antisense and siRNA oligonucleotides. Bioconjug Chem 23:147–157 12. Wittrup A, Lieberman J (2015) Knocking down disease: a progress report on siRNA therapeutics. Nat Rev Genet 16:543–552 13. Endoh T, Ohtsuki T (2009) Cellular siRNA delivery using cell-penetrating peptides modified for endosomal escape. Adv Drug Deliv Rev 61:704–709 14. Tai W, Gao X (2017) Functional peptides for siRNA delivery. Adv Drug Deliv Rev 110–111:157–168 15. Ahmadzada T, Reid G, McKenzie DR (2018) Fundamentals of siRNA and miRNA therapeutics and a review of targeted nanoparticle delivery systems in breast cancer. Biophys Rev 10:69–86 16. Welch JJ, Swanekamp RJ, King C et al (2016) Functional delivery of siRNA by disulfideconstrained cyclic amphipathic peptides. ACS Med Chem Lett 7:584–589

67

17. Bowerman CJ, Nilsson BL (2010) A reductive trigger for peptide self-assembly and hydrogelation. J Am Chem Soc 132:9526–9527 18. Mandal D, Nasrolahi Shirazi A, Parang K (2011) Cell-penetrating homochiral cyclic peptides as nuclear-targeting molecular transporters. Angew Chem Int Ed Engl 50:9633–9637 19. Aitken A, Learmonth M (2009) In: Walker JM (ed) Estimation of disulfide bonds using Ellman’s reagent BT - the protein protocols handbook. Humana Press, Totowa, NJ, pp 1053–1055 20. O’Nuallain B, Thakur AK, Williams AD et al (2006) Kinetics and thermodynamics of amyloid assembly using a high-performance liquid chromatography-based sedimentation assay. Methods Enzymol 413:34–74 21. Chan W, White P (1999) Fmoc solid phase peptide synthesis: a practical approach. OUP, Oxford 22. Han Y, Albericio F, Barany G (1997) Occurrence and minimization of cysteine racemization during stepwise solid-phase peptide Synthesis1,2. J Org Chem 62:4307–4312 23. Lukszo J, Patterson D, Albericio F, Kates SA (1996) 3-(1-Piperidinyl)alanine formation during the preparation ofC-terminal cysteine peptides with the Fmoc/t-Bu strategy. Lett Pept Sci 3:157–166 24. Cline DJ, Thorpe C, P Schneider J (2004) General method for facile intramolecular disulfide formation in synthetic peptides. Anal Biochem 335:168–170

Chapter 5 FRET-Mediated Observation of Protein-Triggered Conformational Changes in DNA Nanostructures Simon Chi-Chin Shiu, Yusuke Sakai, Julian A. Tanner, and Jonathan G. Heddle Abstract DNA origami is a powerful technique, which allows virtually limitless 2D or 3D nanostructure designs to be constructed from DNA strands. Such nanostructures can even include programmable nanorobots, which are able to respond to the environment in predetermined ways. DNA aptamers hold particular promise as interfaces, which can enable proteins, peptides, and other non-nucleic acid biomolecules to trigger conformational changes in DNA nanostructures for diagnostic, biosensing, or therapeutic applications. Here, we provide the methodology for FRET-mediated observation of aptamer-triggered conformational change in a DNA origami box nanostructure. The method described can, in principle, be adapted to a wide variety of experimental circumstances where the DNA nanostructure conformational change is mediated by molecular or environmental cues. Key words DNA origami, DNA–protein interaction, aptamers, molecular logic gate

1

Introduction DNA origami was invented in 2006 by Paul Rothemund. In its first iteration, it allowed the design and facile production of arbitrary 2D shapes [1]. This was achieved by taking a long single strand of DNA (M13 phage genomic DNA) and folding it using several hundred short “staple” strands complementary to the template. Thus, the lowest energy base-paired state leads to “stapling” of the oligonucleotide into the predetermined shape. The technique quickly evolved to enable the construction of 3D objects and containers [2] and programmable nanorobots, including a DNA barrel [3] and DNA boxes [4]. The opening, closing, and other structural changes of such nanostructures can be actuated by DNA aptamers

Simon Chi-Chin Shiu and Yusuke Sakai contributed equally to this work. Julian A. Tanner and Jonathan G. Heddle contributed equally to this work. Maxim G. Ryadnov (ed.), Polypeptide Materials: Methods and Protocols, Methods in Molecular Biology, vol. 2208, https://doi.org/10.1007/978-1-0716-0928-6_5, © Springer Science+Business Media, LLC, part of Springer Nature 2021

69

70

Simon Chi-Chin Shiu et al.

[5]. Aptamers are produced using an evolutionary process called Systematic Evolution of Ligands by EXponential enrichment (SELEX) [6, 7]. Target molecules are incubated with a library of random DNA sequences, and sequences with higher binding affinity are separated and amplified for the next round of selection. The specificity of aptamers can be tailored through counter-selection approaches, which can prevent binding of an aptamer to similar structures through evolutionary selection and optimization. There is a significant potential utility in hollow DNA containers that can be opened and closed in response to specific molecular signals. Such containers could be used to deliver a therapeutic (the contents of the box) or to give a detectable signal in sensing or diagnostic applications. Biophysical sources of detectable signals may include fluorescence resonance energy transfer (FRET) such that fluorescence changes with the distance between lid and box body during opening and closing. We have previously developed a DNA aptamer that specifically binds to lactate dehydrogenase from Plasmodium falciparum (PfLDH) [8], which is a biomarker for malaria diagnosis. We showed that it is able to specifically bind to the target protein even when part of a DNA origami structure [9]. This was further developed into an aptamer “lock” for a DNA origami box consisting of an aptamer attached to the box lid and another partially complementary strand on the box body. Base pairing between the two strands keeps the lid closed in the absence of the target protein. When the target protein binds the aptamer, base-pairing interactions are disrupted, leading to lid opening, resulting in a change in the FRET signal [10] (Fig. 1a). Here, we provide a detailed methodology, based on our published work [10], for the interaction of an aptamer-modified, FRET-labeled DNA origami box with PfLDH and detection of the resulting FRET signal change. The FRET assay was designed using Cy3 and Cy5 FRET label pairs. One label is placed on the box lid, and the other on the box body such that when the lid is closed, the fluorescence is quenched by FRET. Conversely, fluorescence increases when the lid is opened (Fig. 2a).

2

Materials Prepare all solutions using pyrogen-free, DNase-free, and RNasefree ultrapure water and analytical-grade reagents. Prepare and store all reagents at room temperature (unless indicated otherwise). Diligently follow all waste disposal regulations when disposing of waste materials.

2.1 Buffers and Reagents

1. 0.5 M ethylenediaminetetraacetic acid (EDTA), pH 8.0: Weigh 73.06 g EDTA and transfer to a glass beaker. Add water to a

Protein-Actuated DNA Nanostructures

71

Fig. 1 PfLDH-mediated dissociation of an aptamer-lock module. (a) Scheme showing the principle of aptamerlock displacement. The 50 12 nt of the PfLDH aptamer (red) are base-paired to a complementary ssDNA sequence (blue). In the presence of PfLDH (orange), the aptamer–target interaction induces a conformational change of the aptamer, releasing the complementary strand. (b) Typical gel image of electrophoretic mobility shift assay for aptamer-lock module optimization (see Note 6). Without (w/o) PfLDH, no shifted band corresponding to the aptamer–protein complex is observed. With (w/) PfLDH, for the optimal complementary strand (12 bp1), the aptamer duplex still retains affinity to PfLDH and releases the strand in its presence. (Reprinted from Nanomedicine: Nanotechnology, Biology, and Medicine, 14, 1161–1168, Copyright (2018), with permission from Elsevier)

volume of 400 mL. Mix and adjust pH with NaOH (see Note 1). Make up to 500 mL with water. Store at room temperature. 2. 0.5 M MgCl2: Weigh 203.3 g of MgCl2·6H2O and transfer to a glass beaker. Add water to a volume of ca. 800 mL and dissolve. Transfer to a graduated measuring cylinder and make up to 1 L with water. Store at room temperature. 3. 10
phosphate-buffered saline (PBS): Purchase PBS tablets and follow the protocol of reconstitution to make up 1 L of 10
concentration. Store at room temperature. 4. 10
TAEM buffer: 400 mM Tris, 200 mM acetic acid, 10 mM EDTA (pH 8.0), 125 mM MgCl2. Weigh 48.44 g of Tris and transfer to a glass beaker filled with ca. 600 mL water (see Note 2). Add and mix 11.45 mL of acetic acid, 20 mL of 0.5 M EDTA, and 250 mL of 0.5 M MgCl2. Transfer to a graduated

72

Simon Chi-Chin Shiu et al.

FRET

Locked

Opened

Fig. 2 (a) Scheme of FRET-based detection of DNA box opening operated by protein–aptamer-lock interaction. Two aptamer keys are incorporated into the box lid to close the DNA box, and a pair of dye-labeled staple DNAs are integrated on the edge of the DNA box. FRET occurs only when the lid of the DNA box is closed. Unlocking occurs upon binding of PfLDH to the aptamer strands, leading to lid opening and loss of FRET signal. (b) Typical time course of DNA box opening observed as a decline of FRET signal (Blue: DNA box with PfLDH; Red: DNA box without the addition of PfLDH; Black: background noise). (Reprinted from Nanomedicine: Nanotechnology, Biology, and Medicine, 14, 1161–1168, Copyright (2018), with permission from Elsevier)

Protein-Actuated DNA Nanostructures

73

cylinder and make up to 1 L with water. Store at room temperature. Concentrated TAE buffer (e.g., 50
TAE buffer) is also available commercially as an alternative to mixing individual solutions. 5. 1% agarose gel: Add 1 g of agarose powder to 100 mL of TAEM buffer and microwave until all powder is dissolved. Pour the gel solution into a gel casting tray and wait until the gel is set. 6. 10% ammonium persulfate (APS): Add 1 g APS into 10 mL water. Mix thoroughly until APS is completely dissolved. Store at 20  C. Thaw completely before use (see Note 3). 7. Native polyacrylamide gel: To prepare a 12% resolving gel, add 4 mL of 30% acrylamide/bis-acrylamide Solution 29:1 (Merck), 1 mL of 10
TBE (Invitrogen), 4.89 mL of H2O, 0.1 mL of 10% APS, and 10 μL of tetramethylethylenediamine (TEMED) into a 50-mL falcon tube. Gently swirl the tube to prevent bubble formation and mix completely. Add the mixture to a gel caster and wait until it is set. To prepare a 6% stacking gel on top of the resolving gel, add 2 mL of 30% Acrylamide/ Bis-acrylamide solution 29:1, 1 mL of 10
TBE, 6.89 mL of H2O, 0.1 mL of 10% APS, and 10 μL of TEMED into 50-mL falcon tube (see Note 4). Gently swirl the tube to prevent bubble formation and mix completely. Add the mixture to a gel caster and place a 10-well comb in it. Wait until it is set. 8. Staining buffer: Add 4 μL of 10,000
SYBR Gold nucleic acid gel stain (Thermo Fisher Scientific) into 40 mL water (see Note 5). Swirl the mixture completely for mixing. 2.2 DNA Strands and Protein

1. 100 μM single-stranded DNA: Add the correct amount of water to the freeze-dried DNA with mixing in order to make a 100 μM stock solution. Single-stranded oligo DNA (desalted grade) is commercially available (e.g., from Integrated DNA Technologies). See supplementary information of published work [10] for the sequence of staple strands for a DNA origami box. The PfLDH aptamer sequence (CTGGGCGGTAGAACCATAGTGACCCAGCCGTCTAC) and its partially complementary sequence (TCTACCGCCCAG) are directly connected to 30 end or 50 end of staple strands, respectively, as aptamer-lock modules. Fluorescence dye-labeled DNAs (“Cy3-PolyA” with 50 end Cy3 and “Cy5-PolyT” with 30 end Cy5 modification for standard curve, 30 end Cy5 labeled “D4” staple strand and 30 end Cy3 labeled “B3” staple strand for DNA origami box, HPLC purified grade) can also be produced commercially (e.g., from Biomers GmbH). Store at 4  C in the dark for short term or at 20  C in the dark for long periods.

74

Simon Chi-Chin Shiu et al.

2. Scaffold DNA: M13mp18 single-stranded DNA (circular, 7249 nucleotides in length) is used and is commercially available (e.g., from New England Biolabs). 3. Purified recombinant PfLDH: Express and purify protein as reported (full methods text available online) [8].

3

Methods Carry out all procedures at room temperature unless otherwise specified.

3.1 Aptamer-Lock Module Verification (See Note 6)

1. Mix 50 μL of 100 nM PfLDH aptamer with 50 μL of 400 nM complementary DNA (12 bp1) in TAEM buffer (see Note 7). 2. In a thermal cycler, heat the mixture to 95  C and then cool to 20  C at a steady rate of 1 K per min. Keep the product at 20  C (see Note 8). 3. Evaluate assembly by running native PAGE. Load 10 μL of the product on a native polyacrylamide gel and run it at 100 V for 1 h at 4  C to prevent overheating of samples (see Note 9). Stain the gel in staining buffer and image using a gel documentation system (see Fig. 1, Notes 10 and 11). 4. Dilute assembled aptamer-lock module with TAEM buffer to 50 nM. 5. Mix 50 nM aptamer-lock with 2 μM PfLDH in 1:1 ratio (see Note 7). 6. Incubate the mixture for 1 h at 25  C. 7. Evaluate displacement by native PAGE as step 3.

3.2 DNA Box Assembly

1. Mix 50 nM staple strands and 10 nM scaffold DNA in 1
TAEM buffer (see Notes 7 and 12). 2. In a thermal cycler, anneal the mixture by heating to 95  C and then cooling to 25  C at a steady rate of 1 K per min (see Note 8). 3. Remove excess staple strands by using an Illustra MicroSpin S-400 HR column (GE Healthcare). Resuspend the resin in the column by vortexing. Spin at 750
g for 2 min to elute all buffer from the column. Carefully add 500 μL TAEM buffer to the column. Spin at 750
g for 2 min. Discard the eluent. Repeat this step THREE times. For the last centrifugation, set a longer spinning time (e.g., 2.5 min) to remove residual TAEM. Add 50–100 μL of DNA origami sample carefully to the center of the column. Avoid disturbing the beads inside. Spin at 750
g for 2 min. Collect flow-through as a purified sample.

Protein-Actuated DNA Nanostructures

75

4. Optional: Quantify the concentration of DNA origami sample by measuring absorption at 260 nm by, e.g., Nanodrop™ (ThermoFisher Scientific). We regard the DNA origami structure as a 7.2 kbp length of double-stranded DNA (ca. 4.8 MDa). 5. Optional: If necessary, concentrate the sample using an Amicon Ultra-0.5 mL 50 K filter (Merck) (see Note 13). To reduce unspecific binding to the filter membrane, load 500 μL of TAEM buffer on the top filter unit and centrifuge at 14,000
g for 4 min. Discard the flow-through. 6. Load the assembled DNA box (up to 500 μL) and centrifuge at 14,000
g for 1–4 min. The DNA box will be concentrated in the top unit. 7. Invert the top unit over a new collection tube and centrifuge at 1000
g for 1 min. The concentrated sample will be eluted into the collection tube. 8. Optionally, elute again using 20 μL of TAEM buffer as follows: put 10 μL of TAEM buffer on both filter membrane surfaces (total 20 μL), incubate at room temperature for 3 min, and collect the elution by centrifugation of the inverted top unit at 1000
g as above. 9. To evaluate the assembly by gel imaging, run ca. 30 fmol of samples on a 1% agarose gel in TAEM buffer. First, load 5 μL of DNA sample with DNA gel loading dye (Thermo Scientific) into the well of the gel and operate at 80 V for 1 h at 4  C to prevent overheating of the gel (see Note 14). Stain the gel in staining buffer and image using a gel documentation system (see Note 11). 10. Optional: To evaluate the assembly by in-liquid AFM for planar DNA origami structures (see Note 15), apply 1.5 μL of ca. 3 nM sample on the center of freshly cleaved mica (e.g., Electron Microscopy Sciences 71856-02) immobilized on a glass slide and incubate the plate at room temperature for 1 min; gently add 20 μL of TAEM buffer and optionally 1.5 μL of 100 mM NiCl2 (see Note 16) and incubate at room temperature for 3 min; put the sample glass plate on the stage of an AFM (we use Dimension Icon (Bruker) and our instructions are optimized for that machine) and carefully add TAEM buffer to make a droplet that covers the whole of the cantilever holder; and scan the image using a SCANASYST-FLUID+ probe (Bruker) with PeakForce QNM in Fluid mode. 11. Optional: To evaluate the assembly of 3D DNA origami structures by TEM (see Note 15), glow-discharge formvar/carbon film on copper 400 mesh (EM Resolutions) using an HDT-400 hydrophilic treatment device (JEOL) for 160 s (Grid mode) and apply 4 μL of ca. 5 nM of sample on the

76

Simon Chi-Chin Shiu et al.

grid chip carefully to cover whole grid region by the liquid and remove the excess of liquid using blotting paper. Incubate at room temperature for 1 min (should be optimized for individual samples, longer incubation times give more adsorption of samples) and apply 4 μL of 3% uranyl acetate and remove with blotting paper as described above. Incubate for 1 min (should be optimized for individual samples, longer incubation (up to 2 min) can result in enhanced contrast) and wash the grid using water as described above for staining. Repeat once to remove the excess of uranyl acetate (should be optimized for individual samples, the number of washes can be reduced, or the step may be eliminated in some cases). Visualize the DNA origami sample using a transmission electron microscope (we used a JEM-2100LaB6 (JEOL)) at 30,000
magnification and 200 kV. 3.3 Fluorescence Resonance Energy Transfer (FRET) Standard Curve for Quantification of DNA Nanobox

1. Mix 50 μL of 40 nM Cy3-PolyA with 50 μL of 40 nM Cy5-PolyT in PBS. 2. In a thermal cycler, heat the mixture to 95  C and reduce to 20  C at a steady rate of 1  C per min. 3. Dilute the product with PBS into different concentrations such as 2.5 nM, 5 nM, 10 nM, 15 nM, and 20 nM. 4. Gently transfer 20 μL of the sample into a 384-well plate (see Note 17). 5. Insert the 384-well plate into the plate reader. Measure the fluorescence signal with excitation at 558 nm and emission at 680 nm (see Note 18). 6. Plot the results on a graph as concentration (x-axis) against emission (y-axis) using appropriate software (e.g., Origin). Fit the data with linear regression (y ¼ mx + c). 7. Compare the signal response from 20 μL of 10 nM DNA nanobox to the regression line to find the experimental concentration (Fig. 3).

3.4 Protein-Mediated Opening of DNA Origami Box Assessed by FRET Assay

1. Mix 20 nM DNA nanobox with 1 μM PfLDH in a 1:1 ratio in PBS (see Note 7). 2. Prepare the plate reader by setting the temperature to 25  C and wait until the temperature is stable. 3. Gently transfer 20 μL of the sample into a 384-well plate (see Note 17). 4. Insert the 384-well plate into the plate reader. Incubate for 3 h with shaking at 60 rpm. 5. Measure the fluorescence signal with excitation at 558 nm and emission at 680 nm. Make the measurement at 30-s intervals (see Note 18, Fig. 2b).

Protein-Actuated DNA Nanostructures

77

Fig. 3 Quantification of DNA origami box with FRET dye incorporation. DNA origami box concentration was experimentally measured based on FRET signal intensity. (Reprinted with minor adaptation from Nanomedicine: Nanotechnology, Biology, and Medicine, 14, 1161–1168, Copyright (2018), with permission from Elsevier)

4

Notes 1. Concentrated NaOH (10 N) can be used initially to equilibrate the pH close to 8. Then, it is suggested to use a lower concentration of NaOH to avoid a sudden increase in pH above the required pH. 2. Use a magnetic stirrer for mixing. 3. Aliquot the 10% APS solution to avoid repetitive freeze–thaw cycles. 4. The addition of reagents has to follow the sequence as mentioned. The addition of TEMED has to be performed in a fume hood with good ventilation because of its toxicity. 5. Containers of SYBR gold stain should be covered in foil to protect from light and kept at 20  C for long-term storage and to avoid bleaching. 6. Subheading 3.1 is for readers trying new aptamer-lock modules. Before integrating a new aptamer-lock module into a DNA origami structure, evaluate its specificity and kinetics based on EMSA as described here. We recently produced a

78

Simon Chi-Chin Shiu et al.

general guide for designing aptamer modules to integrate into DNA nanostructures [11] as well as DNA origami [5]. 7. Pipette up and down (without bubbles) to ensure complete mixing of all components in the solution. 8. Set the ramp rate of the thermal cycler to 0.1  C per 6 s. For precise temperature control, keep sample volume at no more than 50 μL and avoiding wells around the edge of heat block. 9. Make sure the running buffer in the gel tank is enough to cover the entire gel and pre-run the gel at 4  C for 30 min before loading samples. 10. To optimize the competitive interaction between PfLDH and complementary ssDNA, we first examined the length of the complementary sequence from 8 bp to 24 bp in 4-bp steps and determined 12 bp as the optimal length. A length of 8 bp was too short to hybridize to aptamer while strands longer than 16 bp hybridized too tightly and were not able to be displaced by PfLDH interaction. We next scanned for the optimal position on the aptamer for the complementary sequence in 6-bp steps and concluded that 12 bp1 (TCTACCGCCCAG, covering the 50 -end region of aptamer) was the most optimal sequence for the aptamer key module. 11. The gel staining process needs to be carried out in the dark to avoid unnecessary bleaching of SYBR Gold stain. 12. Scaffold DNA is assembled to form a box shape [4] by use of a molar excess of staple DNA strands (fivefold excess in our typical method). Other groups employ from fourfold [12] to 100-fold excess [1]). 13. Since staple strands can pass through a 50 K filter, this concentration step can replace the purification step described in Subheading 3.2, step 3. To remove free staple strands thoroughly, at least three cycles of buffer exchange should be carried out between steps 2 and 3. The final optional elution step will increase the yield while it will dilute the final concentration of the sample. Be aware that there is a loss of DNA origami sample due to nonspecific adsorption on the filter. If the yield is poor, PEG precipitation is an alternative approach for purification and concentration of DNA origami structures [12]. During the whole procedure, avoid touching the membrane with the edge of the pipette tips so as not to damage the filter. 14. Electrophoresis with TAEM buffer generates heat due to high electric current and will generate chlorine gas at the anode. This is potentially dangerous. Consult your local safety protocols before attempting this and bear in mind temperature control and air circulation. Assembled DNA nanostructures, in general, migrate faster than unassembled structures on the

Protein-Actuated DNA Nanostructures

79

gel, but this can vary depending on the specific shape of the assembled structure [2]. We used a reported DNA origami structure and employed an assembling method from the literature. However, correct DNA origami assembly conditions will vary from structure to structure. A typical optimization procedure would be to test a range of Mg2+ (and optionally Na+) ion concentrations, temperature ramps for annealing, and varying stoichiometry of staple and scaffold [12]. 15. This book chapter focusses on FRET assay-based DNA nanostructure analysis. However, we describe typical sample preparation methods for AFM and TEM here for readers trying new DNA origami structures. In general, AFM is suitable for a planar DNA nanostructure, which is easy to immobilize on mica surface, while TEM works well for visualization of solid (3D) DNA nanostructures having higher electron density. The methods should be optimized to get high-quality images. Follow the manufacturer’s instructions and caution warnings during the handling of microscopy devices. 16. NiCl2 addition improves immobilization of multiple-layered DNA origami nanostructure on mica [13, 14]. The method is adopted from a previous report [15] and adapted. 17. Avoid the introduction of air bubbles, which may affect the fluorescence readout. 18. Measure the fluorescence signal with 500 ms measurement time and 12 nm bandwidth. FRET signal intensity will gradually decrease, reflecting PfLDH-mediated opening of the lid of the DNA box structure.

Acknowledgments JGH was supported by the TEAM programme of the Foundation for Polish Science co-financed by the European Union under the European Regional Development Fund (TEAM/2016-3/23). YS was supported by POLONEZ3 (2016/23/P/NZ1/04097) of Polish National Science Centre (NCN) co-funded by the Marie Skłodowska-Curie action of the European Union (665778). S.C.C.S. and J.A.T. were funded by grants from the General Research Fund (17163416 and 17127515) of the Hong Kong University Grants Council. Simon Chi-Chin Shiu and Yusuke Sakai contributed equally to this work. References 1. Rothemund PW (2006) Folding DNA to create nanoscale shapes and patterns. Nature 440 (7082):297–302. https://doi.org/10.1038/ nature04586

2. Douglas SM, Dietz H, Liedl T, Hogberg B, Graf F, Shih WM (2009) Self-assembly of DNA into nanoscale three-dimensional shapes.

80

Simon Chi-Chin Shiu et al.

Nature 459(7245):414–418. https://doi.org/ 10.1038/nature08016 3. Douglas SM, Bachelet I, Church GM (2012) A logic-gated nanorobot for targeted transport of molecular payloads. Science 335 (6070):831–834. https://doi.org/10.1126/ science.1214081 4. Andersen ES, Dong M, Nielsen MM, Jahn K, Subramani R, Mamdouh W, Golas MM, Sander B, Stark H, Oliveira CL, Pedersen JS, Birkedal V, Besenbacher F, Gothelf KV, Kjems J (2009) Self-assembly of a nanoscale DNA box with a controllable lid. Nature 459 (7243):73–76. https://doi.org/10.1038/ nature07971 5. Sakai Y, Islam MS, Adamiak M, Shiu SC-C, Tanner JA, Heddle JG (2018) DNA aptamers for the functionalisation of DNA origami nanostructures. Genes 9(12):571. https:// doi.org/10.3390/genes9120571 6. Ellington AD, Szostak JW (1990) In vitro selection of RNA molecules that bind specific ligands. Nature 346:818. https://doi.org/10. 1038/346818a0 7. Tuerk C, Gold L (1990) Systematic evolution of ligands by exponential enrichment: RNA ligands to bacteriophage T4 DNA polymerase. Science 249(4968):505–510. https://doi. org/10.1126/science.2200121 8. Cheung YW, Kwok J, Law AW, Watt RM, Kotaka M, Tanner JA (2013) Structural basis for discriminatory recognition of Plasmodium lactate dehydrogenase by a DNA aptamer. Proc Natl Acad Sci U S A 110(40):15967–15972. https://doi.org/10.1073/pnas.1309538110 9. Godonoga M, Lin TY, Oshima A, Sumitomo K, Tang MSL, Cheung YW, Kinghorn AB, Dirkzwager RM, Zhou CS, Kuzuya A, Tanner JA, Heddle JG (2016) A DNA aptamer recognising a malaria protein biomarker can function as part of a DNA

origami assembly. Sci Rep 6:21266. https:// doi.org/10.1038/srep21266 10. Tang MSL, Shiu SC-C, Godonoga M, Cheung Y-W, Liang S, Dirkzwager RM, Kinghorn AB, Fraser LA, Heddle JG, Tanner JA (2018) An aptamer-enabled DNA nanobox for protein sensing. Nanomedicine 14(4):1161–1168. https://doi.org/10.1016/j.nano.2018.01. 018 11. Shiu SC-C, Kinghorn AB, Sakai Y, Cheung Y-W, Heddle JG, Tanner JA (2018) The three S’s for aptamer-mediated control of DNA nanostructure dynamics: shape, selfcomplementarity, and spatial flexibility. Chembiochem 19(18):1900–1906. https://doi. org/10.1002/cbic.201800308 12. Wagenbauer KF, Engelhardt FAS, Stahl E, Hechtl VK, Stommer P, Seebacher F, Meregalli L, Ketterer P, Gerling T, Dietz H (2017) How we make DNA origami. Chembiochem 18(19):1873–1885. https://doi. org/10.1002/cbic.201700377 13. Billingsley DJ, Lee AJ, Johansson NA, Walton A, Stanger L, Crampton N, Bonass WA, Thomson NH (2014) Patchiness of ion-exchanged mica revealed by DNA binding dynamics at short length scales. Nanotechnology 25(2):025704. https://doi.org/10. 1088/0957-4484/25/2/025704 14. Tang T-C, Amadei CA, Thomson NH, Chiesa M (2014) Ion exchange and DNA molecular dip sticks: studying the nanoscale surface wetting of muscovite mica. J Phys Chem C 118 (9):4695–4701. https://doi.org/10.1021/ jp411125n 15. Veneziano R, Ratanalert S, Zhang K, Zhang F, Yan H, Chiu W, Bathe M (2016) Designer nanoscale DNA assemblies programmed from the top down. Science 352(6293):1534. https://doi.org/10.1126/science.aaf4388

Chapter 6 Molecular Simulations Guidelines for Biological Nanomaterials: From Peptides to Membranes Irene Marzuoli and Franca Fraternali Abstract In studying biological processes and focusing on the molecular mechanisms at the basis of these, molecular dynamics (MD) simulations have demonstrated to be a very useful tool for the past 50 years. This suite of computational methods calculates the time-dependent evolution of a molecular system using physics-based first principles. In this chapter, we give a brief introduction to the theory and practical use of molecular dynamics simulations, highlighting the different models and algorithms that have been developed to tackle specific problems, with a special focus on classical force fields. Some examples of how simulations have been used in the past will help the reader in discerning their power, limitations, and significance. Key words Molecular dynamics, Simulations, Force fields, Multiscale modeling, Coarse-grained parametrization, Proteins, Lipids

1

Introduction Molecular dynamics (MD) simulations have been rightly defined as the “computational microscope” [1, 2] as they offer otherwise inaccessible insights into the molecular details underlying conformational changes of biomolecules. Computational methods and tools based on MD are routinely applied in structural biology to quantitatively characterize the dynamics and thermodynamics of proteins and their complexes. The increasing computational power available, and the ease of implementation of simulation algorithms, have made it possible to access molecular and dynamical properties inaccessible to experiments. MD simulations and the associated force fields are commonly used in the process of structure determination from NMR data or of theoretical structure prediction from homology models [3, 4]. In particular, simulations of the structure suggested from the method of choice help in relieving the artifacts deriving from the experiments, or in determining the correct conformation in cases when the experimental measure or the homology modeling has a large uncertainty.

Maxim G. Ryadnov (ed.), Polypeptide Materials: Methods and Protocols, Methods in Molecular Biology, vol. 2208, https://doi.org/10.1007/978-1-0716-0928-6_6, © Springer Science+Business Media, LLC, part of Springer Nature 2021

81

82

Irene Marzuoli and Franca Fraternali

Modeling and simulating a biological system consists in describing its components and its mutual interactions, by implementing the laws of physics to reproduce the dynamics of phenomena observed in nature. A quantum mechanic description would be most accurate but expensive to achieve for large systems. To facilitate the task, several simplified models have been devised, each most suitable to investigate particular cases. In particular, a popular approach is a classical mechanics description of the dynamics: for increasing sizes of the systems and longer simulation time, the classical approximations will become more accurate, and also the only possible model which is computationally affordable. Understanding the methodology of classical molecular dynamics (MD) provides an interpretative key with which simulations must be designed, run, and interpreted in each specific case. We offer here some general background and directions on the most commonly used algorithms and their applicability in certain selected scenarios that we consider representative and challenging.

2

Materials

2.1 Extracting the Dynamics of a Molecular System: The Time Evolution Algorithm

In a classical MD framework, Newton’s second law of motion rules the dynamics, stating that the acceleration a that a particle is subject to at time t depends on the total force F acting on the particle itself and on its mass m (bold denotes vectorial quantities): Fðt Þ ¼ m aðt Þ:

ð1Þ

As the acceleration a(t) is the second derivative of the position r (t) with respect to time, given the initial position and velocity of the particle (r(t0), v(t0)), their temporal evolution can be computed by integrating a(t) ¼ F(t)/m as follows: Z t Fðt 0 Þ 0 dt ; ð2Þ vðt Þ ¼ vðt 0 Þ þ t0 m Z t Z t Z t0 Fðt 00 Þ 00 0 0 dt dt : ð3Þ rðtÞ ¼ rðt 0 Þ þ vðt 0 Þ dt þ t0 m t0 t0 In the case of complex biomolecular systems with many particles and multiple interactions acting between them, it is impossible to integrate analytically Eq. 2, while a different and feasible approach consists in discretizing Eq. 1. The idea is to consider very short time steps of length Δt so that in such intervals the forces are (almost) constant, and the integration of Eq. 2 becomes trivial. A careful choice of the values to integrate allows reducing the approximations derived from such an approach. For example, choosing the velocity value at time t0 + Δt/2 (and not at t0) reduces the error to orders of (Δt)4 (rather than (Δt)2). This framework is at

Molecular Simulations Guidelines for Biological Nanomaterials: From. . .

83

the basis of the leap-frog algorithm, which is used in the vast majority of MD simulations engines:     Fðt Þ Δt Δt ¼ v t0  þ Δt; ð4Þ v t0 þ 2 2 m   Δt Δt: ð5Þ rðt 0 þ Δt Þ ¼ rðt 0 Þ þ v t 0 þ 2 This algorithm can thus “solve” every possible Newton equation at the expense of precision. 2.2 Thermostats and Barostats: Rescuing the Approximation Limit

As the integration procedure is not exact, specific algorithms have been developed to compensate for the necessary approximations and in order to realistically reproduce the simulation’s conditions of choice, for example, temperature and pressure. To set up a temperature, at the beginning of a simulation, all particles are given random initial velocities according to the Maxwell–Boltzmann distribution, which describes noble gas atom velocities at temperature T. Their velocity will be influenced by the specific interactions occurring in the system but, in a constant temperature environment, the total average kinetic energy hEki (proportional to T) must remain constant. Even in the absence of any dissipative term in the dynamics, the approximations performed by MD algorithms lead to energy/temperature drift from its initial value. Therefore, to ensure that the temperature is maintained throughout the simulation, thermostat algorithms have been devised. The principle behind a thermostat consists in rescaling the velocity of all or a few selected particles to restore the correct average kinetic energy. The rescaling must not adjust the kinetic energy at the target value for each time step, as the goal of a thermostat is to maintain the average temperature, and fluctuations are allowed in natural systems. Moreover, it is strongly recommended to couple solute and solvent to separate heat baths, to ensure that both maintain the correct temperature. Indeed, it is possible that the energy exchange between solute and water (or other components) is not perfect due to different conditions adopted for their simulations, like, for example, restraints. The most used thermostat algorithms are the Berendsen [5], Nose´– Hoover [6, 7], Andersen [8], and velocity rescale [9] ones, which differ in the way they perform their velocity adjustments. For example, the Berendsen algorithm [5] rescales the velocities of all the particles in the simulation at each time step. The rescaling factor λ is computed, imposing that the corrected kinetic energy Ek∗ is equal to:   1 X Δt Δt 0 2 ∗ ðλvi Þ ¼ 1  E ð6Þ Ek þ Ek ¼ 2 i τT τT K

84

Irene Marzuoli and Franca Fraternali

where Ek0 is the target kinetic energy and τT is a time interval multiple of Δt, which regulates the strength of the coupling. Due to its efficiency, this thermostat is one of the most used in the initial phases of a MD simulation. However, if used with a value of τT equal to Δt, it suppresses all thermal fluctuations, and even with larger τT, it cannot maintain the correct distribution of velocities during the evolution of the dynamics. For these reasons, after the equilibration stages, which bring the temperature around the correct value, other algorithms are preferred. For example, the velocity rescale one [9] takes the Berendsen framework, adding a stochastic term to the rescaling factor shown in Eq. 6: sffiffiffiffiffiffiffiffiffiffiffiffiffi E k E 0k 2 dW ð7Þ τT N f with Nf being the number of degrees of freedom in the system and dW a Wiener noise [10], which ensure that fluctuations are sampled correctly. Another condition one wishes to maintain is either the volume or the pressure of the system. While maintaining a constant volume is straightforward (and, combined with constant temperature, gives the NVT ensemble), pressure regulation (i.e., maintaining an NPT ensemble) requires a barostat. Pressure is directly proportional to the average quantity of motion exchanged between the particles and the walls of the box they are confined to, which depends on the frequency of collision and thus on the extent of the box. Barostats rescale the box size (and thus the positions of all the particles inside it) to regulate the pressure, ideally allowing fluctuations around the target value. Usually, all the box dimensions are rescaled by the same amount. In the case of anisotropic systems like lipids, the directions parallel to the membrane plane can be rescaled separately with respect to the one perpendicular to it. It has to be noticed that most MD simulations are run under periodic boundary conditions, i.e., a particle which exits from the simulation box during a move is brought back on the opposite side of the box, leaving the box density constant. This mimics the presence of an infinite number of equivalent boxes one next to the other and alleviates the finite-size effects that arise when simulating small systems. In this scenario, particles are not bouncing on the box walls; rather, a virtual pressure is computed from the velocities of the ones trespassing the box boundaries during a move. Also, for pressure coupling, several algorithms can be used: the Berendsen [5], Parrinello–Rahman [11], and Martyna–Tuckerman–Tobias–Klein (MTTK) [12] are popular ones. The Berendsen barostat is analogous to its thermostat counterpart, as it defines a scaling factor for the velocities (and thus the coordinates) based on the target and effective pressure P0 and P, and a coupling time τP:

Molecular Simulations Guidelines for Biological Nanomaterials: From. . .

1=3 β Δt ðP 0  P Þ v ¼v 1 τP ∗

85



ð8Þ

where β is the isothermal compressibility. As also the Berendsen barostat suffers from poor sampling pressure fluctuations, the Parrinello–Rahman [11] is very often used for production runs after the initial equilibration. This algorithm introduces the constraint of constant pressure in the dynamical equations of the systems (in a Lagrangian approach), which allows reproducing the correct distribution of the pressure fluctuations. 2.3

Force Fields

2.3.1 Force Field Definition

Force fields for classical MD simulations provide the expression of the potential energy of a system. Thus, they determine the forces employed in Newton’s law ruling the dynamics. They usually rely on the breakdown of interactions into several independent, additive, and derivable terms, identified on an empirical physical basis. We report here the functional form of the GROMOS force field [13, 14] as implemented in the GROMACS MD engine [15–17], as an example of a classical force field. Covalent (Bonded) Interactions Covalent interactions are modeled with potential energy terms representing bond stretching, angle bending, and improper and proper dihedral angle torsion. The functional forms of potential energy functions aim a simplified, classical description of the atomic motion of molecules. Often, it is modeled as a harmonic-like vibration around the equilibrium position, regulated by a constant. In the GROMOS force field, this translates in the equations displayed in Table 1, where for proper dihedrals, the convention states that φijkl is the angle between the (i, j, k) and ( j, k, l) planes; with i, j, k, and l being four subsequent atoms, for example, along a protein backbone. A value of zero for φijkl corresponds to a cis configuration and π to a trans. The integer n denotes the number  of equally spaced energy minima available in a 360 turn. The same conventions hold for improper dihedrals ξijkl, which are used to ensure ring planarity and control the chirality of tetrahedric centers. It must be noticed that these types of potentials cannot model bond breaking: for this, more sophisticated descriptions are needed. Nonbonded Interactions Nonbonded interactions include the short-range Pauli repulsion, the “mid”-range van der Waals attraction, and the long-range electrostatic term. The first two terms can be modeled together by a Lennard– Jones potential. Its functional form, describing the interaction between two neutral atoms at distance r, models the long-range dispersion with an r6 behavior typical of the dipole–dipole interactions found in noble gases (London dispersion forces), while the Pauli term is represented by an r12 behavior to ease the computation in relation with the previous one:

86

Irene Marzuoli and Franca Fraternali

Table 1 Equations used for covalent bonds simulated in GROMOS force field Type

Eq. pos.

Const.

[Const.]

Functional form

Bond

bij

kijb

kJ mol m2

  2

2 V b rij ¼ 14 kbij rij   b ij 2

Angle

θijk0

kijkθ

kJ mol

 2

V a θijk ¼ 12 kθijk cos θijk  cos θ0ijk

Dihedral

ϕijkl0

kijklϕ

kJ mol rad2

Improper

ξijkl0

kijklξ

kJ mol

     V d ϕijkl ¼ kϕijkl 1 þ cos nϕijkl  ϕ0ijkl    2 V i ξijkl ¼ 12 kξijkl ξijkl  ξ0ijkl

   6  σ 12 σ : V LJ ðr Þ ¼ 4ϵ  r r

ð9Þ

Two parameters, ε and σ, tune the interaction strength and the equilibrium distance, respectively. They are parametrized by fit to experimental data and are specific of each pair of atom species. The Coulomb energy between two charges q1 and q2 at distance r is represented by the Coulomb law: 1 qi q j V C ðr Þ ¼ : ð10Þ 4πε0 εr r ij with ε0 being the dielectric constant of vacuum and εr the relative dielectric constant, introduced to properly take into account the screening provided by the material surrounding the object. Indeed, as electrons are not present in classical force fields, the screening effect due to the atom polarization must be modeled with this mean-field approach. The treatment of nonbonded interactions requires particular care because of their long-range nature: in every point of the simulation box, many forces from distant atoms are acting at the same time, making the prediction of the outcome difficult. The van der Waals forces decay fast; therefore, the tail of their functional can be cut after a threshold distance with little impact on the outcome, while Coulomb interactions, with their slower decay, must be taken into account throughout the whole simulation box. Many algorithms have been devised to efficiently compute them like the Particle Mesh Ewald [18] or the Reaction Field [19] approaches. Finally, force fields designed to describe biomolecules are parametrized to describe systems at room temperature. Therefore, care should be taken when interpreting simulations performed at substantially different temperatures.

Molecular Simulations Guidelines for Biological Nanomaterials: From. . . 2.3.2 Force Fields: Classifications

87

Many force fields for classical MD simulations adopt a functional form equal or similar to the one described above. Their difference lies in the number of degrees of freedom modeled, in a hierarchy of descriptions proceeding from detailed to coarse (Fig. 1). Three possible classes of descriptions are as follows: l

All-atom force fields, where all the atoms are presented in the description and represented as spheres of variable size according to their van der Waals radius (e.g., proportional to σ in a Lennard–Jones model). Examples of all-atom force fields are AMBER [24–26] and CHARMM [27–29]. In addition, OPLS has a united-atom version [30].

l

United-atom force fields, similar to the previous ones but where nonpolar hydrogens are incorporated in the heavy atom they are bonded to. The “united atom” is given a new σ parameter and increased mass according to how many hydrogens it includes. The GROMOS force field [13, 14] follows this philosophy, and OPLS has also a united-atom version [31].

l

Coarse-grained force fields, which group together a few atoms in a unique bead, to reduce the number of variables to compute. The clustered atoms are such that their mutual distances are expected to vary little with respect to the ample movements of components of the system far away from each other (which will be grouped in different beads). The MARTINI [32–34] and SIRAH [35, 36] force fields belong to this category.

We now give a more detailed insight into the characteristics and parametrization strategies of an atomistic and a coarse-grained force field among the ones mentioned. 2.3.3 The GROMOS Force Field

All-atom and united-atom force fields are parametrized against experimental values. While for the all-atom force fields AMBER and CHARMM the parametrization is based on ab initio quantum mechanics calculations refined against experimental data [24–28], the united-atom GROMOS force field relies on the reproduction of free enthalpies of solvation and heat of vaporization of small molecules at physiological temperatures and pressures [14, 37, 38]. This procedure sets not only the constant of the bonded interactions but also the partial charges of the atoms inside a molecule: as no electrons are included for the sake of efficiency, their redistribution across atoms that are bonded is modeled through fractional charges assigned to each atom (while the total charge of a molecule must sum to an integer). Moreover, it is assumed that the parametrization performed for small moieties can be transferred to a larger compound including these moieties. This limits the number of chemical groups to be described in order to simulate biomolecules. In every MD simulation, the description of water is crucial. Out of the many water models proposed, the GROMOS

88

Irene Marzuoli and Franca Fraternali

Fig. 1 List of most popular simulation force field for biomolecules, ordered from detailed to coarse (reference to the relative papers in Subheading 2.3.2). On the left, snapshot of notable systems simulated with the force fields CHARMM (Adapted with permission from [20]. Copyright (2017) American Chemical Society); GROMOS (Adapted with permission from [21]); SIRAH (Adapted with permission from [22]. Copyright (2017) American Chemical Society) and MARTINI (Adapted with permission from [23]). Copyright (2017) Elsevier)

parametrization has been performed with a flexible simple point charge (SPC [39]) model. This description represents water as a three-atom molecule, with a negative charge on the oxygen and a

Molecular Simulations Guidelines for Biological Nanomaterials: From. . .

89

positive complementary charge on the two hydrogen atoms, and allows flexible hydrogen–oxygen bonds. This model reproduces correctly the density and dielectric permittivity of water [40]. The improvement of computational techniques and reparametrization strategies prompts the periodical release of newer versions of force fields. Accordingly, the latest version of the GROMOS force field, version 54a8 [41], was released in 2012, and further refinements have been proposed recently [42, 43]. 2.3.4 The MARTINI Force Field

The MARTINI force field is a popular coarse-grained description of biological molecules [32–34]: developed originally with a focus on lipids, it has been then extended to include proteins, small ligands, and DNA/RNA molecules. MARTINI opts for a four-to-one approach; i.e., four heavy atoms are grouped in one bead. The number of bead types has been kept to the minimum necessary to represent biological molecules. They are organized systematically in polar, nonpolar, apolar, or charged, and each type has a number of subtypes with increasing polarity to differentiate the chemical nature of the underlying atomistic structures. This systematic approach can easily be transferred to new compounds, without the need for introducing new bead types. Ring molecules represent the only exception, where a two-to-one approach is needed to maintain the circular topology. Two different approaches are taken to develop a coarse-grained force field: top-down and bottom-up. In the first, parameters are fitted directly to global quantities derived from experiments, as performed in the atomistic GROMOS parametrization. In the second, coarse-grained simulations results are fitted to outcomes from atomistic ones. The MARTINI force field chooses a top-down approach to parametrize nonbonded interactions, tuning them against experimental partitioning free energies between polar and apolar phases, while bonded interactions are derived from reference all-atom simulations in a bottom-up approach. The four-to-one mapping implies that the amino acid backbone is represented by one bead only, preventing the description of directional bonds, which are key to reproduce the secondary structure. The bonded parameters partially account for this, favoring for each residue type the backbone conformation in which it is most likely found (as computed from the Protein Data Bank—PDB [44]). When this is not sufficient, the protein can be constrained around a given structure through an elastic network model approach (ElNeDyn [45]). However, both the backbone parametrization and the use of ElNeDyn imply that local energy minima of the natural structure are not well sampled in MARTINI simulations, biasing the understanding of the structure dynamics. The MARTINI force field provides two water models. The standard one groups four water molecules in one bead only, losing

90

Irene Marzuoli and Franca Fraternali

the polarizability typical of water, the effect of which is partially restored with the use of a high dielectric constant. The polarizable water model [46] maps instead four water molecules to a single “inflated” water, i.e., a three-bead molecule with the same shape of a single molecule, but expanded, and a charge splitting which can account for the water dipole. Backmapping Techniques Coarse-grained descriptions are very effective in reproducing long timescales; however, to retrieve finer details after such extensive exploration, backmapping techniques have been designed to obtain atomistic configurations from the coarse-grained ones [47]. The easy transfer between the two resolutions gave rise to many multiscale studies applied to biomolecular systems [48]. 2.4 Beyond a Classical Atomistic Framework

Without entering into the details, we want to bring to the reader’s attention two possible refinements of the aforementioned models and two computational strategies, which, on the contrary, speed up the calculations at the expense of the loss of some details. Regarding the accuracy of simulations, it must be noticed that none of the force field mentioned above takes into account polarizability, i.e., the displacement of electrons with respect to the nucleus, as a consequence of the surrounding electrostatic environment, because electrons and nucleus of an atom are modeled as a single object. Specific force fields have been modeled to include this effect, on top of atomistic descriptions, as in the AMOEBA [49, 50], Drude polarisable CHARMM [51], or AMBERff02 [52], or in combinations with a coarse-grained description, as in the ELBA force field [53]. Polarizability does improve the accuracy of simulations, but it can significantly slow down simulations. Going beyond the classical approximation, for biological processes governed by quantum mechanics—such as photosynthesis, DNA mutation processes, or enzymatic activities—many semiclassical hybrid techniques have been developed [54]. They combine computational quantum mechanical modeling methods, such as density functional theory (DFT) or Hartree–Fock computations (HF) [55], with classical molecular dynamics to gain the accuracy of a quantum description in the region of interest and the speedup of a classical one in the surrounding areas. Tackling instead efficiency issues, an implicit solvent model can be used to speed up simulations. The solvent is represented as a continuous medium, as opposed to explicit models, which include all its particles [56]. Models of implicit solvent can be based on different assumptions: for example, the solute–solvent interactions can be taken as proportional to the solvent-accessible surface area (SASA) of every particle of solute [57–60], or instead can be derived from a solution of the Poisson–Boltzmann equation governing the charge density in a material, for example, in the form of the generalised Born equation [61] which is valid under particularly

Molecular Simulations Guidelines for Biological Nanomaterials: From. . .

91

simple conditions. Hybrid particle-field algorithms constitute another speedup technique. The idea is to treat nonbonded interactions through a mean-field approach, where atoms/beads move in the field generated by the others. The field does not need to be updated at every time step, as it is a collective and thus slowly evolving variable; moreover, for each particle, only the interaction with the field, and not with all the neighbor particles, needs to be computed, reducing the computation effort further. This approach has been employed with a coarse-grained description of polymers and biological molecules in the OCCAM software [62]. Finally, further strategies are possible to enhance the sampling performed by a simulation in the case that the one obtained by the natural evolution of the system would exceed the computational time available. As a noncomprehensive list of these techniques, we mention replica-exchange algorithms [63], which combine together multiple simulations held at different conditions, local potential-energy elevation (or metadynamics) [64, 65], which avoids the re-sampling of already visited conformations adding an energy penalty to them, umbrella sampling [66], which reconstructs free-energy barriers from simulations performed at specific values of the coordinate along which the barrier exists, or finally the simple use of higher temperature to overcome energy barriers [67].

3

Methods Several established procedures have been devised for MD simulations of biomolecules. Nevertheless, each case requires a careful investigation to find the simulation conditions that suit best. We list below the key components and routines necessary to run a simulation, suggesting the user to tune the protocols referring to validated literature examples, as, for example, the ones in Subheading 3.2.

3.1

System Setup

To set up a simulation, the desired resolution, i.e., force field, must be chosen according to the system to simulate. A few examples of this will be given in Subheading 3.2; however, for any choice of parameters, some common “ingredients” and steps are necessary to prepare the system: l

Structural file: A pdb file (or any structural file format suitable with the MD engine of choice), which contains only the elements to be simulated, with a correct format of names and positions. The choice of the initial conformation is particularly important, especially for atomistic resolution, which can access limited timescales and thus is likely to sample conformations in the vicinity of the initial position.

92

Irene Marzuoli and Franca Fraternali l

Topology file: Given a structural file, every MD engine has dedicated commands to retrieve the list of parameters for bonded atoms. The presence of exotic residues or a nonstandard network of bonds might need case-to-case manual parametrization.

l

Simulation box: For simulations of a protein, it is good practice to place it within a box large enough to avoid interference of the protein with its periodic images, for example, allowing a minimum distance between protein and box of at least the cutoff chosen for nonbonded interactions. However, it must be considered that the protein might adopt extended conformations during the simulation; therefore, one should find a compromise between a sufficient box size and the cost of increasing the effective number of particles to calculate. In the case of lipids instead, to reproduce an infinite membrane and account for lateral tension, the box can be chosen exactly as big as the membrane patch (in the directions parallel to its plane) so that periodic images merge together.

l

Solvation: The prepared box is filled with molecules of the water model of choice, except in the spaces already occupied by the solute.

l

Ions: Ions are added in replacement of a suitable number of water molecules to neutralize the charge of the system. Additional ones can be added to reach the experimental ion concentration (though there will be a slight imbalance between the positive and negative species due to the counter ions already added).

l

Energy minimization: The system prepared according to the steps above might have several atom clashes due to artifacts in the initial structure or imperfect packing of the solvent around the solute. To alleviate this, an energy minimization is performed initially. Usually, a protocol restraining the solute and letting relax solvent positions in the first instance, followed by a second run that releases the solute, is used in these cases.

l

Equilibration: Even after energy minimization, the prepared system is far away from the ideal equilibrium configuration. For this reason, several rounds of simulations with specific conditions are run before the final production. The specific protocol depends on the system and the force field; however, for proteins in solution, it is quite common to start with NVT ensemble runs with position restraints at increasing temperatures, followed by NPT runs at the same temperatures. For a lipid system, the simulated ensemble is preferably NPT, as an NVT ensemble with a fixed box size would allow unphysical penetration of water in the bilayer if it shrinks. The equilibration then consists in a temperature increase protocol only. In some cases, as for the MARTINI force field with standard water, it is not possible to

Molecular Simulations Guidelines for Biological Nanomaterials: From. . .

93

simulate low temperatures; thus, the equilibration consists in a short production employing different thermostats and barostats with respect to the production. Some of these protocols are more suitable for approaching efficiently the correct temperature or pressure value from a configuration far from equilibrium, while others for maintaining the correct ensemble properties. Generally, every force field is provided with validated equilibration procedures. 3.2

Examples

One of the challenges of MD simulations consists in choosing the most suitable granularity of the description, together with the choice of the system to simulate. We list here a few examples, which might convey better the idea of suitable setup for systems of current interest. Simulations of Protein and Their Interactions Simulations of protein have been crucial to understand small molecular details of their structure and functioning, which influence their macroscopic behavior. A challenging subject in the field of enzymatic regulation is constituted by allosteric regulation, as the pathways involved are of difficult exploration. Molecular dynamics simulations can shed light on the residues responsible for transferring the information [68]. For example, atomistic simulations of pre-assembled pyruvate kinase M2 (PKM2) tetramers elucidated the mechanism of allosteric activation by fructose 1,6-bisphosphate (FBP) [21]. In this work, the correlated motions extracted from simulations of the apo and FBP bound tetramer were processed within an information-theoretical framework. Such a framework indicates a set of specific residue hubs. These were experimentally demonstrated to be responsible for the communication between orthosteric and allosteric pockets. Another particularly interesting concept explored by MD is the one of ensemble: a protein in solution adopts a variety of conformations, which can differ significantly from the experimentally determined X-ray structures available. Such flexibility could be proven using techniques such as small-angle X-ray scattering (SAXS) or nuclear magnetic resonance (NMR) [69–71]. As such, simulations can provide “snapshots” of all these conformations and their mutual interchange. Some of these changes are crucial for the protein function, as in the case of prions, proteins which become pathogenic under conversion from an α-helix-rich form to a β-sheet isoform, which is prone to aggregation. Atomistic MD simulations of the prion protein showed the (reversible) unfolding of the protein between the two structural states [72], starting from the helical crystal structure available. Moreover, they identified in the C-terminus region the trigger of the conformational change, corresponding to a region that is enriched in pathogenic mutations.

94

Irene Marzuoli and Franca Fraternali

The results were then used to investigate the aggregation process after the conversion, again identifying critical residues promoting it [73]. Interestingly, the water distribution density around the prion molecule was demonstrated to have a role in identifying possible spots prone to aggregation [74, 75]. Such examples show that one should consider the solvent, or the forces induced by the protein surrounding media, as an important determinant in the conformational equilibria and recognition processes. Assembly simulations applied to amyloid-forming peptides have provided recently a map of all the relevant states of the assembly process [76], suggesting that the constant improvement of computational power will make this possible also for more complex systems like the prion in the near future. Simulations of Model Membranes Membrane simulations had a key role in elucidating the mechanical and dynamical characteristics of these important organelles, clarifying, for example, how the lipids composing the membrane influence its fluidity [77, 78], or elucidating the interactions between lipids and surface proteins, transmembrane ones or membrane-active peptides, such as antimicrobial ones [79–81]. The last point, in particular, raised the interest in simulating bacterial membranes. In simulating the action of antimicrobial peptides, it is important to consider carefully the characteristics of bacterial membranes. However, as these objects are very complex, simplified models can be used to approach the problem. Very often, these models use only one or two lipid species to represent membranes; e.g., bacterial ones contain neutral phospholipids with a percentage of negatively charged ones [20, 82–84] as key characteristic distinguishing them from mammal membranes, which possess only zwitterionic phospholipids, with the occasional inclusion of cholesterol, deemed important in achieving the flexibility typical of mammal membranes [20, 77, 82–84]. Because of their simplicity, these systems are extensively used in experiments [85–87], making a direct comparison with simulations possible. Nevertheless, attempts in accurately modeling cell membranes have been pursued [88]. For membrane systems, the atomistic level is suitable to reproduce single components of the membrane, e.g., the isolated inner membrane of Gram-negative bacteria [89], even if atomistic computations started in recent years to be affordable also for larger systems. For example, atomistic simulation of the outer membrane of Gram-negative bacteria combined with the peptidoglycan layer (which is positioned between the two membranes) elucidated how their distance is variable, thanks to the presence of Braun’s lipoproteins, which act as a bridge between the two, and can bring them closer by bending and tilting [23]. Moreover, the permeability of membranes to ions and small compounds needs to be assessed at the atomistic level to get sufficient accuracy, and because of the

Molecular Simulations Guidelines for Biological Nanomaterials: From. . .

95

intensive task, often enhanced MD techniques such as umbrella sampling are employed [90]. Coarse-grained descriptions are instead the most suitable to represent the full bacterial envelope, especially in the case of Gram-negative bacteria, as the inclusion of all its elements results in large systems. Accordingly, MARTINI simulations have been able to reproduce the behavior of several transmembrane proteins, which spanned both the inner and outer membranes [91], and to model all the different components of the Gram-negative cell envelope [88]. Similarly, the ability of coarse-grained simulations to investigate very large systems makes them suitable to assess elastic properties of membranes, as they can access low-frequency undulations with little influence from finite-size effects [92].

4

Notes Validation of MD simulations is performed by comparison with experiments: the same properties obtained experimentally are computed from the MD trajectory as well, and the two compared. If these are correctly reproduced, it is usually assumed that the simulation is sampling the correct ensemble of states and then one can identify in the simulation the determinants responsible for the experimental outcome of interest, which are not accessible by the experiment itself. The comparison however is not always easy: often, the experimentally measured quantities are temporal or spatial averages (for example, circular dichroism spectra or SAXS profiles of a peptide in solution) and many different combinations of computationally derived structure ensembles can produce compatible results. It is still challenging to compare experimentally derived ensembles and the ones derived by molecular simulations. Indeed, the extracted average properties may be different, so it is important to understand which are the relevant ones playing a role in the measured ensemble before attempting comparisons. Thus, in the validation of MD outcomes, it is important to have a critical attitude and to interpret the result within the validity of the approximation [93]. Agreement may indeed arise from a simulation that reflects correctly the experimental system, but it can also be achieved when the property examined is insensitive to the details of the simulated trajectory. Again, it can be a result of compensation of errors, which is more likely to occur for systems with a high number of degrees of freedom (as biomolecular ones). Similarly, disagreement may hint at an error in the simulation (either in the model, the implementation, or simply the estimated simulation’s convergence) or an error in the interpretation and/or conditions of experimental setup (either in the result itself or its interpretation) so that both

96

Irene Marzuoli and Franca Fraternali

must be carefully checked to improve a convergence in the agreement. Some apparently negative results may suggest or stimulate new experimental settings to validate the hypotheses one was set to test [94, 95]. Additionally, simulations still suffer from the limited computational time available: most of the times, the real experimental system is simply too large to be reproduced and the timescale of the process too long. Simulations are thus confined to explore a restricted space, implying that the initial conditions must be chosen carefully to optimize the search and avoid any bias that might persist for the whole length of the simulation. The use of enhanced MD techniques can increase the chances of sampling relevant states; however, it introduces a bias that must be removed or properly accounted for in the interpretation of the results [96–100]. Finally, one should keep in mind that the force fields used are far from optimal, partly because they rely on approximate functional forms, and partly because it is difficult to find experimental observables measured with the desired resolution able to discriminate between sets of parameters. Nevertheless, it remains important to note the contribution that molecular dynamics simulations have played in elucidating important details behind biological processes and in unraveling molecular details not accessible to experiments. References 1. Lee EH et al (2009) Discovery through the computational microscope. Structure 17 (10):1295–1306 2. Dror RO et al (2012) Biomolecular simulation: a computational microscope for molecular biology. Annu Rev Biophys 41(1):429 3. Vogel A, Huster D (2017) Combining NMR spectroscopy and molecular dynamics simulation to investigate the structure and dynamics of membrane-associated proteins. In: Chattopadhyay A (ed) Membrane organization and dynamics. Springer International Publishing, Cham, pp 311–350 4. Heo L, Feig M (2018) Experimental accuracy in protein structure refinement via molecular dynamics simulations. Proc Natl Acad Sci 115 (52):13276–13281 5. Berendsen HJC et al (1984) Molecular dynamics with coupling to an external bath. J Chem Phys 81(8):3684–3690 6. Nose´ S, Klein M (1983) Constant pressure molecular dynamics for molecular systems. Mol Phys 50(5):1055–1076 7. Hoover WG (1985) Canonical dynamics: equilibrium phase-space distributions. Phys Rev A 31(3):1695–1697

8. Andersen HC (1980) Molecular dynamics simulations at constant pressure and/or temperature. J Chem Phys 72(4):2384–2393 9. Bussi G, Donadio D, Parrinello M (2007) Canonical sampling through velocity rescaling. J Chem Phys 126(1):014101 10. Durrett R (2010) Probability: theory and examples. Cambridge University Press, Cambridge 11. Parrinello M, Rahman A (1981) Polymorphic transitions in single crystals: a new molecular dynamics method. J Appl Phys 52 (12):7182–7190 12. Martyna GJ et al (1996) Explicit reversible integrators for extended systems dynamics. Mol Phys 87(5):1117–1157 13. Oostenbrink C et al (2004) A biomolecular force field based on the free enthalpy of hydration and solvation: the GROMOS force-field parameter sets 53A5 and 53A6. J Comput Chem 25(13):1656–1676 14. Schmid N et al (2011) Definition and testing of the GROMOS force-field versions 54A7 and 54B7. Eur Biophys J 40(7):843–856

Molecular Simulations Guidelines for Biological Nanomaterials: From. . . 15. Berendsen H, van der Spoel D, van Drunen R (1995) GROMACS: a message-passing parallel molecular dynamics implementation. Comput Phys Commun 91(1–3):43–56 16. Abraham MJ et al (2015) GROMACS: high performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX 1-2:19–25 17. Abraham MJ et al (2018) GROMACS user manual version 2016. www.gromacs.org 18. Essmann U et al (1995) A smooth particle mesh Ewald method. J Chem Phys 103 (19):8577–8593 19. Tironi IG et al (1995) A generalized reaction field method for molecular dynamics simulations. J Chem Phys 102(13):5451–5459 20. Lipkin R, Pino-Angeles A, Lazaridis T (2017) Transmembrane pore structures of β-hairpin antimicrobial peptides by all-atom simulations. J Phys Chem B 121(39):9126–9140 21. Macpherson JA et al (2019) Functional crosstalk between allosteric effects of activating and inhibiting ligands underlies PKM2 regulation. elife 8:e45068 22. Machado MR, Gonza´lez HC, Pantano S (2017) MD simulations of virus-like particles with supra CG solvation affordable to desktop computers. J Chem Theory Comput 13 (10):5106–5116 23. Samsudin F et al (2017) Braun’s lipoprotein facilitates OmpA interaction with the Escherichia coli cell wall. Biophys J 113 (7):1496–1504 24. Maier JA et al (2015) ff14SB: improving the accuracy of protein side chain and backbone parameters from ff99SB. J Chem Theory Comput 11(8):3696–3713 25. Dickson CJ et al (2014) Lipid14: the amber lipid force field. J Chem Theory Comput 10 (2):865–879 26. Wang J et al (2004) Development and testing of a general amber force field. J Comput Chem 25(9):1157–1174 27. MacKerell AD et al (1998) All-atom empirical potential for molecular mod eling and dynamics studies of proteins. J Phys Chem B 102 (18):3586–3616 28. Klauda JB et al (2010) Update of the CHARMM all-atom additive force field for lipids: validation on six lipid types. J Phys Chem B 114(23):7830–7843 29. Huang J, MacKerell AD (2013) CHARMM36 all-atom additive protein force field: validation based on comparison to NMR data. J Comput Chem 34(25):2135–2145 30. Jorgensen WL, Tirado-Rives J (1988) The OPLS [optimized potentials for liquid simulations] potential functions for proteins, energy

97

minimizations for crystals of cyclic peptides and crambin. J Am Chem Soc 110 (6):1657–1666 31. Jorgensen WL, Maxwell DS, Tirado-Rives J (1996) Development and testing of the OPLS all-atom force field on conformational energetics and properties of organic liquids. J Am Chem Soc 118(45):11225–11236 32. Marrink SJ et al (2007) The MARTINI force field: coarse-grained model for biomolecular simulations. J Phys Chem B 111 (27):7812–7824 33. Monticelli L et al (2008) The MARTINI coarse-grained force field: Exten Sion to proteins. J Chem Theory Comput 4(5):819–834 34. De Jong DH et al (2013) Improved parameters for the Martini coarse-grained protein force field. J Chem Theory Comput 9 (1):687–697 35. Machado MR et al (2019) The SIRAH forcefield 2.0: Altius, Fortius, Citius. J Chem Theory Comput 15(4):2719–2733 36. Barrera EE, Machado MR, Pantano S (2019) Fat SIRAH: coarse-grained phospholipids to explore membrane-protein dynamics. J Chem Theory Comput 15(10):5674–5688. https://doi.org/10.1021/acs.jctc.9b00435 37. Oostenbrink C et al (2005) Validation of the 53A6 GROMOS force field. Eur Biophys J 34 (4):273–284 38. Reif MM, Winger M, Oostenbrink C (2013) Testing of the GROMOS force-field parameter set 54A8: structural properties of electrolyte solutions, lipid bilayers, and proteins. J Chem Theory Comput 9(2):1247–1264 39. Berendsen H et al (1981) Interaction models for water in relation to protein hydration. In: Pullman B (ed) Intermolecular forces. D. Reidel Publishing Company, Dordrecht, pp 331–342 40. Mark P, Nilsson L (2001) Structure and dynamics of the tip3p, spc, and spc/e water models at 298 k. Chem A Eur J 105 (43):9954–9960 41. Reif MM, Hu¨nenberger PH, Oostenbrink C (2012) New interaction parameters for charged amino acid side chains in the GROMOS force field. J Chem Theory Comput 8 (10):3705–3723 42. Margreitter C, Reif MM, Oostenbrink C (2017) Update on phosphate and charged post-translationally modified amino acid parameters in the GROMOS force field. J Comput Chem 38(10):714–720 43. Marzuoli I, Margreitter C, Fraternali F (2019) Lipid head group parameterization for GROMOS 54A8: a consistent approach

98

Irene Marzuoli and Franca Fraternali

with protein force field description. J Chem Theory Comput 15(10):5175–5193 44. Berman HM et al (2000) The protein data bank. Nucleic Acids Res 28:235–242. http://www.rcsb.org/ 45. Periole X et al (2009) Combining an elastic network with a coarse-grained molecular force field: structure, dynamics, and intermolecular recognition. J Chem Theory Comput 5(9):2531–2543 46. Yesylevskyy SO et al (2010) Polarizable water model for the coarse-grained MARTINI force field. PLoS Comput Biol 6(6):e1000810 47. Wassenaar TA et al (2015) Computational Lipidomics with insane: a versatile tool for generating custom membranes for molecular simulations. J Chem Theory Comput 11 (5):2144–2155 48. Lee OS, Cho V, Schatz GC (2012) Modeling the self-assembly of peptide Amphiphiles into fibers using coarse-grained molecular dynamics. Nano Lett 12(9):4907–4913 49. Ren P, Ponder JW (2003) Polarizable atomic multipole water model for molecular mechanics simulation. J Phys Chem B 24 (107):5933–5947 50. Ponder JW et al (2010) Current status of the AMOEBA polarizable force field. J Phys Chem B 114(8):2549–2564 51. Anisimov VM et al (2005) Determination of electrostatic parameters for a polarizable force field based on the classical Drude oscillator. J Chem Theory Comput 1(1):153–168 52. Cieplak P, Caldwell J, Kollman P (2001) Molecular mechanical models for organic and biological systems going beyond the atom centered two body additive approximation: aqueous solution free energies of methanol and N-methyl acetamide, nucleic acid base, and amide hydrogen bonding and chloroform/water partition coefficients of the nucleic acid bases. J Comput Chem 22 (10):1048–1057 53. Orsi M, Essex JW (2011) The ELBA force field for coarse-grain modeling of lipid membranes. PLoS One 6(12):e28637 54. Ahmadi S et al (2018) Multiscale modeling of enzymes: QM-cluster, QM/MM, and QM/MM/MD: a tutorial review. Int J Quantum Chem 118(9):e25558 55. Shao Y et al (2015) Advances in molecular quantum chemistry contained in the Q-Chem 4 program package. Mol Phys 113 (2):184–215 56. Kleinjung J, Fraternali F (2014) Design and application of implicit solvent models in

biomolecular simulations. Curr Opin Struct Biol 25:126–134 57. Fraternali F, van Gunsteren W (1996) An efficient mean solvation force model for use in molecular dynamics simulations of proteins in aqueous solution. J Mol Biol 256 (5):939–948 58. Kleinjung J et al (2003) Thermal unfolding simulations of apo-calmodulin using leapdynamics. Proteins 50(4):648–656 59. Kleinjung J et al (2012) Implicit solvation parameters derived from explicit water forces in large-scale molecular dynamics simulations. J Chem Theory Comput 8(7):2391–2403 60. Fornili A et al (2012) Protein-water interactions in MD simulations: POPS/POPSCOMP solvent accessibility analysis, solvation forces and hydration sites. In: Baron R (ed) Computational drug discovery and design. Springer New York, New York, pp 375–392 61. Zhu J, Alexov E, Honig B (2005) Comparative study of generalized born models: born radii and peptide folding. J Phys Chem B 109 (7):3008–3022 62. Milano G, Kawakatsu T (2009) Hybrid particle-field molecular dynamics simulations for dense polymer systems. J Chem Phys 130 (21):214106 63. Okamoto Y (2004) Generalized-ensemble algorithms: enhanced sampling techniques for Monte Carlo and molecular dynamics simulations. J Mol Graph Model 22 (5):425–439 64. Huber T, Torda AE, van Gunsteren WF (1994) Local elevation: a method for improving the searching properties of molecular dynamics simulation. Chem A Eur J 101 (33):5926–5930 65. Laio A, Parrinello M (2002) Escaping freeenergy minima. Proc Natl Acad Sci 99 (20):12562–12566 66. Torrie G, Valleau J (1977) Nonphysical sampling distributions in Monte Carlo freeenergy estimation: umbrella sampling. J Comput Phys 23(2):187–199 67. Kirkpatrick S, Gelatt CD, Vecchi MP (1983) Optimization by simulated annealing. Science 220(4598):671–680 68. Pandini A et al (2012) Detection of allosteric signal transmission by information-theoretic analysis of protein dynamics. FASEB J 26 (2):868–881 69. Bonomi M et al (2017) Principles of protein structural ensemble determination. Curr Opin Struct Biol 42:106–116

Molecular Simulations Guidelines for Biological Nanomaterials: From. . . 70. Kikhney AG, Svergun DI (2015) A practical guide to small angle X-ray scattering (SAXS) of flexible and intrinsically disordered proteins. FEBS Lett 589(19PartA):2570–2577 71. Kleckner IR, Foster MP (2011) An introduction to NMR-based approaches for measuring protein dynamics. Biochim Biophys Acta 1814(8):942–968 72. Chakroun N et al (2013) Decrypting prion protein conversion into a β- rich conformer by molecular dynamics. J Chem Theory Comput 9(5):2455–2465 73. Collu F et al (2018) Probing the early stages of prion protein (PrP)aggregation with atomistic molecular dynamics simulations. Chem Commun 54(57):8007–8010 74. De Simone A et al (2005) Prion and water: tight and dynamical hydration sites have a key role in structural stability. Proc Natl Acad Sci 102(21):7535–7540 75. De Simone A et al (2006) Water molecules as structural determinants among prions of low sequence identity. FEBS Lett 580 (10):2488–2494 76. Sengupta U, Carballo-Pacheco M, Strodel B (2019) Automated Markov state models for molecular dynamics simulations of aggregation and self- assembly. J Chem Phys 150 (11):115101 77. Risselada HJ, Marrink SJ (2008) The molecular face of lipid rafts in model membranes. Proc Natl Acad Sci 105(45):17367–17372 78. Song C, de Groot BL, Sansom MS (2019) Lipid bilayer composition influences the activity of the antimicrobial peptide dermcidin channel. Biophys J 116(9):1658–1666 79. Leontiadou H, Mark AE, Marrink SJ (2006) Antimicrobial peptides in action. J Am Chem Soc 128(37):12156–12161 80. Ulmschneider JP (2017) Charged antimicrobial peptides can translocate across membranes without forming channel-like pores. Biophys J 113(1):73–81 81. Sun D, Forsman J, Woodward CE (2015) Atomistic molecular simulations suggest a kinetic model for membrane translocation by arginine-rich peptides. J Phys Chem B 119 (45):14413–14420 82. Wang Y et al (2012) Comparative molecular dynamics simulations of the antimicrobial peptide CM15 in model lipid bilayers. Biochim Biophys Acta Biomembr 1818 (5):1402–1409, NIHMS150003 83. Zhao L et al (2018) Molecular dynamics simulations of human antimicrobial peptide LL-37 in model POPC and POPG lipid bilayers. Int J Mol Sci 19(4):1186

99

84. Chen C et al (2019) Simulation-guided rational de novo design of a small pore- forming antimicrobial peptide. J Am Chem Soc 141 (12):4839–4848 85. Castelletto V et al (2016) Structurally plastic peptide capsules for synthetic antimicrobial viruses. Chem Sci 7(3):1707–1711 86. Tang M, Hong M (2009) Structure and mechanism of β-hairpin antimicrobial peptides in lipid bilayers from solid-state NMR spectroscopy. Mol BioSyst 5(4):317 87. Glukhov E et al (2005) Basis for selectivity of cationic antimicrobial pep tides for bacterial versus mammalian membranes. J Biol Chem 280(40):33960–33967 88. Khalid S, Piggot TJ, Samsudin F (2019) Atomistic and coarse grain simulations of the cell envelope of gram-negative bacteria: what have we learned? Acc Chem Res 52 (1):180–188 89. Piggot TJ, Holdbrook DA, Khalid S (2011) Electroporation of the E. coli and S. Aureus membranes: molecular dynamics simulations of complex bacterial membranes. J Phys Chem B 115(45):13381–13388 90. Carpenter TS, Parkin J, Khalid S (2016) The free energy of small solute permeation through the Escherichia coli outer membrane has a distinctly asymmetric profile. J Phys Chem Lett 7(17):3446–3451 91. Hsu PC et al (2017) It is complicated: curvature, diffusion, and lipid sorting within the two membranes of Escherichia coli. J Phys Chem Lett 8(22):5513–5518 92. Fowler PW et al (2016) Membrane stiffness is modified by integral membrane proteins. Soft Matter 12:7792 93. Van Gunsteren WF, Dolenc J, Mark AE (2008) Molecular simulation as an aid to experimentalists. Curr Opin Struct Biol 18 (2):149–153 94. Gonc¸alves MB et al (2013) Structural prediction of a rhodamine-based biosensor and comparison with biophysical data. Phys Chem Chem Phys 15(6):2177–2183 95. Meißner RH et al (2014) Computational prediction of circular Dichroism spectra and quantification of helicity loss upon peptide adsorption on silica. Langmuir 30 (12):3487–3494 96. Bernardi RC, Melo MCR, Schulten K (2015) Enhanced sampling techniques in molecular dynamics simulations of biological systems. Biochim Biophys Acta Gen Subj 1850 (5):872–877 97. Best RB, Clarke J, Karplus M (2005) What contributions to protein side-chain dynamics

100

Irene Marzuoli and Franca Fraternali

are probed by NMR experiments? A molecular dynamics simulation analysis. J Mol Biol 349(1):185–203 98. Barducci A, Bonomi M, Parrinello M (2010) Linking well-tempered metadynamics simulations with experiments. Biophys J 98(9): L44–L46

99. Barducci A, Bonomi M, Parrinello M (2011) Metadynamics. Wiley Interdiscip Rev Comput Mol Sci 1(5):826–843 100. Mills M, Andricioaei I (2008) An experimentally guided umbrella sampling protocol for biomolecules. J Chem Phys 129(11):114101

Chapter 7 Functional Peptide Nanocapsules Self-Assembled from β-Annulus Peptides Hiroshi Inaba and Kazunori Matsuura Abstract Spherical viruses are unique nanocapsules formed by self-assembly of coat proteins (capsids). By mimicking natural spherical capsids, various artificial viral capsids are developed by using self-assembled proteins and peptides as building blocks. We developed an artificial viral capsid consisting of a β-annulus peptide designed from natural viruses. The “β-annulus capsid” can be functionalized by encapsulating guest molecules to the inside and decoration of exogenous molecules on the outside. Here, we describe the encapsulation and decoration on the β-annulus capsids by connecting additional sequences to the β-annulus peptide, conjugation with objective molecules, and subsequent self-assembly in aqueous solutions. Key words β-annulus peptide, Nanocapsules, Artificial viral capsid, Self-assembly, Encapsulation, Decoration, Nanomaterials, Nanostructures

1

Introduction Viruses are natural nanoarchitectures consisting of coat proteins called capsids and the encapsulated nucleic acids. The viral capsids are formed by self-assembly even in the absence of the genomes, inspiring scientists to construct artificial viral capsids (AVCs) that are self-assembled viruslike nanostructures [1–8]. Among various AVCs, spherical nanocapsules have been widely developed by using proteins [9–13] and peptides [14–19] as building blocks. We developed peptide-based AVCs by designing a trimeric annular β-structure called “β-annulus motif” of spherical viruses [20, 21]. Tomato bushy stunt virus (TBSV) consists of 180 quasiequivalent protein subunits containing 388 amino acids each to form a spherical capsid (ca. 33 nm in diameter) [22–24]. We developed a 24-mer β-annulus peptide (INHVGGTGGAIMAPVAVTRQLVGS) from a β-annulus motif of TBSV (Ile69-Ser92), which self-assembles in water to form a hollow nanocapsule (β-annulus capsid) with 30–50 nm diameter (Figs. 1, 2, and 3) [20]. The β-annulus capsid is a useful scaffold for

Maxim G. Ryadnov (ed.), Polypeptide Materials: Methods and Protocols, Methods in Molecular Biology, vol. 2208, https://doi.org/10.1007/978-1-0716-0928-6_7, © Springer Science+Business Media, LLC, part of Springer Nature 2021

101

102

Hiroshi Inaba and Kazunori Matsuura

Fig. 1 Self-assembly of β-annulus peptide from tomato bushy stunt virus (TBSV) to form β-annulus capsid. (Reproduced with permission from Matsuura et al. (2010) Self-assembled synthetic viral capsids from a 24-mer viral peptide fragment. Angew Chem Int Ed 49:9662–9665 [20])

Fig. 2 TEM image of β-annulus capsid stained with uranyl acetate. (Reproduced with permission from Matsuura et al. (2010) Self-assembled synthetic viral capsids from a 24-mer viral peptide fragment. Angew Chem Int Ed 49:9662–9665 [20])

nanotechnological applications because the inside and outside of the capsid can be modified with exogenous molecules. Here, we describe three topics: (1) synthesis of the β-annulus and derived peptides, (2) encapsulation of guest molecules inside the β-annulus capsid [25–28], and (3) decoration of exogenous molecules on the outer surface of the β-annulus capsid [29–31]. An important point for the design of functionalization on the capsid is that the N- and C-termini of the β-annulus peptide are directed to the inside and outside of the self-assembled capsid, respectively [25]. Thus, the Nand C-termini of the β-annulus peptide are connected with additional peptides designed for the encapsulation and decoration of objective molecules, respectively (Tables 1 and 2). The derivatives

Nanocapsules from β-Annulus Peptides

103

Fig. 3 (a) Size distribution of β-annulus capsid self-assembled from 0.1 mM β-annulus peptide and (b) concentration dependence of β-annulus peptide on scattering intensity determined by DLS. (Reproduced with permission from Matsuura et al. (2010) Self-assembled synthetic viral capsids from a 24-mer viral peptide fragment. Angew Chem Int Ed 49:9662–9665 [20]) Table 1 β-annulus peptides for encapsulation of guest molecules Entry Guest

Sequence

Conjugation

Reference

1

M13 phage DNA, CdTe NP

INHVGGTGGAIMAPVAVTRQLVGS (original β-annulus peptide)

–

[25, 26]

2

ZnO NP

HCVAHRGGGINHVGGTGGAIMAPVA VTRQLVGS

–

[27]

3

His-tag GFP

CINHVGGTGGAIMAPVAVTRQLVGS

NTA[28] maleimide to Cys

The inserted sequences are underlined

Table 2 β-annulus peptides for decoration of exogenous molecules Decorated Entry molecule

Sequence

Conjugation

Reference

INHVGGTGGAIMAPVAVTRQLVG SGGGCG

AuNP to Cys

[29]

1

AuNP

2

DNA (dA20 INHVGGTGGAIMAPVAVTRQLVCS or dT20)

dA20-maleimide or dT20-maleimide to Cys

[30]

3

Coiled-coil INHVGGTGGAIMAPVAVTRQL B peptide VGGCGGGKIAALKKKNAALK QKIAALKQ

–

[31]

The inserted sequences are underlined

104

Hiroshi Inaba and Kazunori Matsuura

of the β-annulus peptide are conjugated with the objective molecules, and subsequently self-assemble to form the functionalized β-annulus capsids in aqueous solutions. The β-annulus capsids are characterized by several techniques such as transmission electron microscopy (TEM), dynamic light scattering (DLS), ζ-potentials, and fluorescence correlation spectroscopy (FCS).

2

Materials 1. Fmoc-resins: Fmoc-Ser(tBu)-Alko-PEG resin; Fmoc-GlyAlko-PEG resin; Fmoc-NH-SAL resin; Dawson Dbz AM resin. 2. Fmoc-amino acids: Fmoc-Ile-OH; Fmoc-Asn(Trt)-OH; Fmoc-His(Trt)-OH; Fmoc-Val-OH; Fmoc-Gly-OH; Fmoc-Thr(tBu)-OH; Fmoc-Ala-OH; Fmoc-Met-OH; FmocPro-OH; Fmoc-Arg(Pbf)-OH; Fmoc-Gln(Trt)-OH; FmocLeu-OH; Fmoc-Cys(Trt)-OH; Fmoc-Lys(Boc)-OH. 3. Activating agents: 2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HBTU); 1-[bis (dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b] pyridinium 3-oxide hexafluorophosphate (HATU); or (1-cyano-2-ethoxy-2-oxoethylidenaminooxy) dimethylaminomorpholinocarbenium hexafluorophosphate (COMU). When HBTU is used, use HOBt (1-hydroxybenzotriazole) together to accelerate condensation reaction and inhibit racemization. 4. Solvents: dimethylformamide (DMF); N-methylpyrrolidone (NMP). 5. Neutralizer: N,N0 -diisopropylethylamine (DIPEA). 6. Deprotecting agent for Fmoc group: piperidine. 7. Agents for qualitative test of amino group: TNBS test kit or chloranil test kit. 8. Agents for deprotection and deresination: trifluoroacetic acid (TFA); 1,2-ethanedithiol (EDT); triisopropylsilane (TIPS); thioanisole; distilled water. 9. Reverse-phase HPLC: Inertsil WP300 C18 or ODS-3 column (4.6 mm  250 mm for analysis and 20 mm  250 mm for purification). 10. Solvents for HPLC: acetonitrile (HPLC-grade); trifluoroacetic acid (TFA); distilled water. 11. Matrix for MALDI-TOF-MS: α-cyano-4-hydroxycinnamic acid (α-CHCA) for peptide; 3-hydroxypicolinic acid (3-HPA) with diammonium hydrogen citrate for DNA. 12. Dialysis kit: Mini Dialysis Kit 1 kDa or 50 kDa cutoff.

Nanocapsules from β-Annulus Peptides

105

13. Dyes: sodium 8-anilino-1-naphthalenesulfonic acid (ANS); uranine (fluorescein); Congo red; methyl orange; pyrogallol red; rhodamine 6G; crystal violet; methylene blue; thioflavin T. 14. M13 phage DNA: M13 mp18 RF DNA (7249 bp). 15. His-tag GFP: recombinant enhanced green fluorescent protein with N-terminal His  6 tags (His-tag GFP, MW ¼ 27 kDa) expressed in an E. coli expression system, purified using Ni-NTA resin. 16. Gold nanoparticles (AuNPs): a stabilized suspension of AuNPs (5 nm diameter) in citrate buffer. 17. TEM grids: C-SMART hydrophilic TEM grids. 18. Stain solution for TEM: 2% uranyl acetate, 0.1 mM cisplatin (cis-diammine dichloro platinum (II)), 2% sodium phosphotungstate, 0.5% ruthenium tetroxide.

3

Methods

3.1 Synthesis of β-Annulus Peptides

Derivatives of β-annulus peptides (Table 1) are prepared by Fmoc solid-phase synthesis.

3.1.1 Fmoc Solid-Phase Synthesis of the β-Annulus Peptides

1. Incubate a resin (typically 0.1 mmol) containing a C-terminal Fmoc-amino acid of respective peptides with NMP (2 mL) in a plastic column with stirring for at least 1 h. 2. Deprotect Fmoc groups from the resin by stirring with 20% piperidine in DMF (2 mL) for 15 min. Repeat this step two times and wash the resin with NMP 5 times. 3. Confirm the deprotection of Fmoc groups from the small amount of the resin using a TNBS test kit (red color appears by reaction with a free amine). When the deprotection is not enough, repeat step 2. Use a chloranil test kit to confirm the deprotection of a Fmoc group from Fmoc-proline introduced on the resin. 4. Add NMP solution (2 mL) of 4 equivalents of respective Fmocamino acid, 4 equivalents of activating agents (COMU, HATU, or HBTU/HOBt), and 8 equivalents of DIPEA to the resin. Perform the condensation reaction by stirring at room temperature for 2 h or by a microwave reaction with 35 W microwave power at 75  C for 5 min. Wash the resin with NMP 5 times. 5. Confirm the completion of the condensation reaction by using a small amount of the resin and TNBS test kit (no color change is observed due to introduction of the Fmoc-amino acid). When the condensation reaction is not enough, repeat step 4.

106

Hiroshi Inaba and Kazunori Matsuura

6. Repeat steps 2–5 to reach the objective peptide sequences. Deprotect a Fmoc group of the N-terminal amino acid by performing step 2, wash the resulting resin with NMP, and then dry it under vacuum. 3.1.2 Deprotection and Cleavage from Resin

1. Prepare a cleavage cocktail containing TFA/thioanisole/ water/EDT/TIPS (2.04 mL/0.125 mL/0.125 mL/ 0.0625 mL/0.025 mL). 2. Add the resin prepared as in Subheading 3.1.1 to the cleavage cocktail and incubate it with stirring at room temperature for 3 h. 3. Filtrate the resulting mixture. 4. Add the ice-cooled tert-butylmethylether (15 mL) to the filtrate to precipitate the objective peptide. 5. Centrifuge the peptide at 2000 rpm for 10 min and discard the supernatant. 6. Repeat steps 4 and 5 three times. 7. Freeze dry the crude peptide as a white powder. Confirm the objective peptide by MALDI-TOF-MS (matrix: α-CHCA) (see Note 1).

3.1.3 Purification of the Peptides

1. Dissolve the crude peptide prepared as in Subheading 3.1.2 in water. If the peptide is less soluble in water, sonicate the mixture and/or add organic solvents such as DMF and acetonitrile. 2. Purify the crude peptide by reverse-phase HPLC using a column such as Inertsil WP300 C18 and Inertsil ODS-3 (20  250 mm) with water/acetonitrile (both containing 0.1% TFA, linear gradient from 75:25 to 65:35, v/v, 10 mL/ min, 100 min, detected at 220 nm) (see Note 2). 3. Confirm the objective peptide by MALDI-TOF-MS (matrix: α-CHCA). 4. Freeze dry the peptide as a white powder and store it in a freezer.

3.2 Molecular Encapsulation into β-Annulus Capsid 3.2.1 Encapsulation of Anionic Guests (Dyes, DNAs, and CdTe Nanoparticles)

Since the interior of the β-annulus capsid is cationic at neutral pH [25], anionic guests such as dyes [25], DNAs (Fig. 4) [25], and negatively charged CdTe nanoparticles (NPs) (Fig. 5) [26] can be encapsulated in the capsid by electrostatic interaction. 1. Synthesize the β-annulus peptide (Table 1, Entry 1) according to Subheading 3.1. 2. Dissolve the β-annulus peptide (typically 100 μM) in water and freeze dry it. 3. Add an aqueous solution/dispersion of the guests in water or 10 mM Tris–HCl buffer, pH 7.4, to the powdered β-annulus

Nanocapsules from β-Annulus Peptides

107

Fig. 4 (a) Encapsulation of M13 phage DNA in β-annulus capsid. (b) TEM image of the DNA-encapsulated capsid stained with cisplatin followed by uranyl acetate. (c) Size distribution of the DNA-encapsulated capsid obtained from DLS

peptide (typical final concentrations: [β-annulus peptide] ¼ 100 μM, [guest] ¼ 100 μM) (see Notes 3 and 4). 4. Incubate the mixture at 25  C for 1 h to encapsulate the guest molecules (Figs. 4a and 5a). 5. Characterize the guest-encapsulated β-annulus capsid by TEM (Figs. 4b and 5b), DLS (Fig. 4c), and FCS (Fig. 5c). 3.2.2 Encapsulation of Ligand-Binding Guests

Since the N-terminus of the β-annulus peptide is located on the interior surface of the self-assembled capsid [25], binding motifs are fused to the N-terminus for the encapsulation of ligand-binding guests for ZnO NP, to give rise to ZnO NP-encapsulated capsids [27]. 1. Synthesize and purify the β-annulus peptide bearing a ZnO-binding peptide (HCVAHR) at the N-terminus (ZnO-binding β-annulus peptide, Table 1, Entry 2, see Subheading 3.1). 2. Prepare ZnO NPs from Zn(OH)2 according to a reported procedure [32]. Confirm the formation of ZnO NPs of a

108

Hiroshi Inaba and Kazunori Matsuura

Fig. 5 (a) Encapsulation of CdTe nanoparticle (NP) in β-annulus capsid. (b) TEM image of the CdTe NP-encapsulated capsid stained with ruthenium tetroxide. Yellow arrows show the positions of CdTe NPs. (c) FCS of the CdTe NP-encapsulated capsid consisting of 0.1 μM CdTe NP and 0–500 μM β-annulus peptide. (Reproduced with permission from Fujita S, Matsuura K (2016) Encapsulation of CdTe quantum dots into synthetic viral capsids. Chem Lett 45:922–924 [26])

quantum size by observation of an absorption peak at 340 nm in the UV–Vis spectrum. 3. Add an aqueous dispersion of ZnO NPs in 10 mM Tris–HCl buffer, pH 7.4, to the powdered peptide prepared in step 1 (final concentrations: [ZnO] ¼ [peptide] ¼ 0.1 mM) (see Note 4). 4. Incubate the mixture at room temperature for 10 min to encapsulate ZnO NPs. 5. Characterize the ZnO NP-encapsulated β-annulus capsid by TEM and DLS. 3.2.3 Encapsulation of Ligand-Binding Guests

Binding motifs are fused to the N-terminus for the encapsulation of ligand-binding guests for His-tag GFPs ZnO NP, to give rise to His-tag GFP-encapsulated capsids (Fig. 6) [28].

Fig. 6 (a) Structure of Ni-NTA-β-annulus peptide. (b) Encapsulation of His-tag GFP in the Ni-NTA-β-annulus capsid. (c) Size distributions obtained from DLS and (d) size-exclusion chromatography of Ni-NTA-β-annulus capsid (top), His-tag GFP (middle), and an equimolar mixture of His-tagg GFP and Ni-NTA-β-annulus peptide (bottom)

110

Hiroshi Inaba and Kazunori Matsuura

1. Synthesize and purify the β-annulus peptide bearing a cysteine at the N-terminus (Cys-β-annulus, Table 1, Entry 3) according to Subheading 3.1. 2. Mix aqueous solution of 0.5 mM Cys-β-annulus (0.5 mL) with aqueous solution of 1.0 mM maleimido-C3-NTA (0.5 mL). 3. Incubate the mixture with stirring at room temperature for 12 h to conjugate NTA moiety to Cys-β-annulus (NTA-β-annulus). 4. Purify NTA-β-annulus by using reverse-phase HPLC (Inertsil ODS-3), eluted with a linear gradient of acetonitrile/water (25/75 to 27/73 over 100 min) containing 0.1% TFA. 5. Confirm NTA-β-annulus by MALDI-TOF-MS (matrix: α-CHCA). 6. Dissolve NTA-β-annulus (3.2 mg) in 10 mM Tris–HCl buffer, pH 7.3, to prepare 5.6 mM solutions. 7. Add 50 mM NiCl2 solution in 10 mM Tris–HCl buffer, pH 7.3, to NTA-β-annulus solution (final concentrations: [Ni2+] ¼ [peptide] ¼ 5.0 mM) to obtain the Ni-NTA-modified β-annulus peptide (Ni-NTA-β-annulus, Fig. 6a). 8. Confirm the Ni-NTA-β-annulus by MALDI-TOF-MS (matrix: α-CHCA). 9. Add aqueous solution of His-tag GFP in 10 mM Tris–HCl buffer, pH 7.3, to the powdered Ni-NTA-β-annulus (final concentrations: [His-tag GFP] ¼ [peptide] ¼ 0.1 mM) to encapsulate GFP via the interaction between His-tag and Ni-NTA (Fig. 6b) (see Note 4). 10. Characterize the GFP-encapsulated β-annulus capsid by DLS (Fig. 6c) and size-exclusion chromatography (Fig. 6d). 3.3 Decoration of Exogenous Molecules on the Outer Surface of β-Annulus Capsid

Since the C-terminus of the β-annulus peptide is exposed on the outer surface of the capsid [25], exogenous molecules such as gold nanoparticles (AuNPs) [29], DNAs [30], and coiled-coil peptides [31] can be decorated on the outer surface by covalent binding to the C-terminus of the β-annulus peptide. Decoration of AuNPs (Fig. 7) [29]. 1. Synthesize and purify the β-annulus peptide bearing GGGCG at the C-terminus (Table 2, Entry 1) according to Subheading 3.1. 2. Dilute an aqueous solution of the β-annulus-GGGCG peptide with water to 2 μM, which is lower than the critical aggregation concentration (CAC) of the peptide (see Note 5). 3. Mix an aliquot of the aqueous solution of the peptide (2 μM, 2 mL) with a diluted dispersion of AuNPs ([AuNP] ¼ 1 μM, 2 mL).

Nanocapsules from β-Annulus Peptides

111

Fig. 7 (a) Decoration of AuNPs on the outer surface of the β-annulus capsid. (b) TEM image of the AuNPdecorated capsid without the use of a stain. (c) Size distribution of the AuNP-decorated capsid obtained from DLS

4. Incubate the mixture at 25 AuNP–β-annulus conjugate.



C for 60 min to construct

5. Add an aliquot (0.5 mL) of 20 mM thioctic acid solution in ethanol/water (4/1) to the mixture. 6. Incubate the mixture at 25  C for 10 min to protect the surface of AuNPs by thioctic acid to prevent aggregation of AuNPs. 7. Evaporate water in the mixture. 8. Add water (80 μL) to disperse the residue to form the AuNPdecorated β-annulus capsid by increasing the concentration of the peptide above CAC (Fig. 7a) (final concentrations: [peptide] ¼ 50 μM, [AuNP] ¼ 25 μM). 9. Dialyze the solution against water (cutoff Mw ¼ 50 kDa) to remove the unassembled AuNPs, peptides, and unassembled AuNP–β-annulus conjugates. 10. Characterize the AuNP-decorated β-annulus capsid by TEM (Fig. 7b), DLS (Fig. 7c), and ζ-potential measurement. 3.3.1 Decoration of DNAs (Fig. 8) [30]

Synthesis of dA20-maleimide conjugate 1. Mix an aqueous solution of dA20-(CH2)6-NH2 (1.0 mM, 0.1 mL) with 9.1 mg of N-(4-maleimidobutyryloxy)-

112

Hiroshi Inaba and Kazunori Matsuura

sulfosuccinimide sodium salt (the activated ester of a heterofunctional linker, Sulfo-GMBS, 2.4 μmol (240 equiv.)) and 0.1 M sodium bicarbonate aqueous solution (0.5 mL). 2. Incubate the mixture at 25  C for 2 h to conjugate dA20 with maleimide moiety (Fig. 8a, dA20-maleimide). Monitor the formation of dA20-maleimide by reverse-phase HPLC (Inertsil ODS-3), eluting with a linear gradient of acetonitrile/0.1 M ammonium formate aqueous solution (0/100 to 100/0 over 105 min, detected at 260 nm). Add water into the elution fraction, freeze the mixture using liquid N2, and concentrate it using a centrifugal evaporator to remove ammonium formate and acetonitrile. Repeat the step five times. Confirm dA20maleimide by MALDI-TOF-MS (matrix: 3-HPA with diammonium hydrogen citrate). 3. Dialyze the solution against water (cutoff Mw ¼ 1 kDa) to remove the excess Sulfo-GMBS. 4. Lyophilize dA20-maleimide for the next reaction without further purification. 5. The dT20-maleimide is obtained by the same method using dT20-(CH2)6-NH2 instead of dA20-(CH2)6-NH2. Construction of DNA-decorated capsid

1. Synthesize and purify the β-annulus peptide bearing CS at the C-terminus (β-annulus-Cys, Table 2, Entry 2) according to Subheading 3.1. 2. Add β-annulus-Cys (1 mg, 0.47 μmol) in water/acetonitrile (2/3, v/v, 1 mL) to dA20-maleimide dissolved in 0.2 mL of 0.1 M phosphate buffer, pH 6.6. 3. Incubate the mixture at 40  C for 48 h to conjugate the β-annulus peptide with dA20 moiety (Fig. 8a, dA20-β-annulus). 4. Purify dA20-β-annulus by reverse-phase HPLC (Inertsil ODS-3), eluting with a linear gradient of acetonitrile/0.1 M ammonium formate aqueous solution (0/100 to 100/0 over 105 min). 5. Freeze the elution fraction by liquid N2 and concentrate it by a centrifugal evaporator. 6. Dialyze the solution against water (cutoff Mw ¼ 1 kDa) to remove ammonium formate and acetonitrile. 7. Lyophilize the solution to give a flocculent solid. 8. Determine the concentration from the absorbance at 260 nm and confirm dA20-β-annulus by MALDI-TOF-MS (matrix: 3-HPA with diammonium hydrogen citrate) upon dissolving in water.

Nanocapsules from β-Annulus Peptides

113

Fig. 8 (a) Synthesis of dA20-β-annulus peptide by the reaction of β-annulus-Cys with dA20-maleimide. (b) Decoration of dA20 on the outer surface of the β-annulus capsid. (c) TEM images stained with sodium phosphotungstate and size distributions obtained from DLS of the dA20-decorated capsids (top), the dA20decorated capsids with polydT (middle), and the dA20-decorated capsids with polydA (bottom). (Reproduced with permission from Nakamura Y et al. (2017) DNA-modified artificial viral capsids self-assembled from DNA-conjugated β-annulus peptide. J Pept Sci 23:636–643 [30])

114

Hiroshi Inaba and Kazunori Matsuura

The dT20-modified β-annulus peptide (dT20-β-annulus) is obtained according to the same method using dT20-maleimide instead of dA20-maleimide. 9. Dissolve dA20-β-annulus (25 μM) in 10 mM phosphate buffer, pH 7.1, and sonicate it for 5 min to form dA20-decorated β-annulus capsid (Fig. 8b). 10. Characterize the dA20-decorated β-annulus capsid by TEM (Fig. 8c, middle), DLS (Fig. 8c, right), and ζ-potential measurement. 11. For the analysis of the binding of the dA20-β-annulus capsid to the polydeoxyadenylic acid sodium salt (polydA) and polythymidylic acid sodium salt (polydT), mix 25 μM dA20-β-annulus peptide and polydA or polydT ([nucleotide] ¼ 0.5 mM) in 10 mM phosphate buffer, pH 7.1, sonicate the mixture for 5 min, and then characterize it by TEM (Fig. 8c, middle) and DLS (Fig. 8c, right). 3.3.2 Decoration of Coiled-Coil Peptides (Fig. 9) [31]

Since the coiled-coil B-modified β-annulus peptide (coiled-coil B-β-annulus) is relatively long (Table 2, Entry 3, 49-mer), it is synthesized by a native chemical ligation (NCL) of a β-annulusSBn peptide with a Cys-containing coiled-coil B peptide. 1. Synthesize and purify the coiled-coil B peptide bearing CGGG at the N-terminus (CGGGKIAALKKKNAALKQKIAALKQ) according to Subheading 3.1. 2. Synthesize and purify the β-annulus peptide bearing N-acylbenzimidazolinone at the C-terminus (β-annulus-Nbz, INHVGGTGGAIMAPVAVTRQLVGG-Nbz) using Dawson Dbz AM resin. Protect the amino group of the resin by allyloxycarbonate (Alloc) by incubating with a mixture of 7 equivalents of allylchloroformate and 1 equivalent of DIPEA at room temperature for 24 h. Load the Fmoc-amino acids and deprotect the Fmoc group according to Subheading 3.1. Deprotect the Alloc group by adding a mixture of 20 equivalents of PhSiH3 and 0.35 equivalent of Pd (PPh3)4 in CH2Cl2. Add 5 equivalents of p-nitrophenylchloroformate in CH2Cl2 followed by 0.5 M DIPEA in DMF to convert to the Nbz group. Deprotect and cleave the peptide from the resin by treatment with a cleavage cocktail for 3 h as described in Subheading 3.1. Confirm β-annulus-Nbz by MALDI-TOFMS (matrix: α-CHCA). 3. Since the NCL reaction between β-annulus-Nbz and coiledcoil B peptide minimally proceed, β-annulus-Nbz should be converted into β-annulus-SBn (Fig. 9a) (see Note 6). Add crude β-annulus-Nbz peptide to a mixture of 10% benzyl mercaptan and 20 mM triethylamine in DMSO to convert to the benzylthioesterified β-annulus peptide at the C-terminus. After

Nanocapsules from β-Annulus Peptides

115

Fig. 9 (a) Synthesis of coiled-coil B-β-annulus peptide by NCL reaction of β-annulus-SBn with coiled-coil B peptide. (b) Decoration of coiled-coil B on the outer surface of the β-annulus capsid and subsequent conjugation with coiled-coil A. (c) TEM images stained with sodium phosphotungstate and size distributions obtained from DLS of the coiled-coil B-decorated capsid (top), and 4:1 (middle) and 1:1 mixtures (bottom) of coiled-coil B-β-annulus and coiled-coil-A

116

Hiroshi Inaba and Kazunori Matsuura

incubation at 37  C for 15 min, precipitate the peptide by adding ethyl acetate and decant the supernatant. Repeat the washing with ethyl acetate three times. Dry the peptide under vacuum and use it for NCL without further purification. 4. Prepare NCL buffer, pH 7.0, by dissolving 6 M guanidinium hydrochloride (GdmCl), 20 mM tris(2-carboxyethyl) phosphine (TCEP), 200 mM 4-mercaptophenylacetic acid (MPAA), and 200 mM sodium phosphate in water and degassing the solution with N2. 5. Mix the coiled-coil B peptide bearing cysteine at the N-terminus (0.75 μmol) and β-annulus-SBn (0.5 μmol) dissolved in NCL buffer. 6. Incubate the mixture at 37  C for 1 h to perform NCL reaction (Fig. 9a). 7. Purify the objective coiled-coil B-β-annulus by reverse-phase HPLC (Inertsil ODS-3), eluted with a linear gradient of acetonitrile/water (20/80 to 40/60 over 100 min) containing 0.1% TFA. 8. Confirm coiled-coil B-β-annulus by MALDI-TOF-MS (matrix: α-CHCA). 9. Dissolve coiled-coil B-β-annulus (50 μM) in 10 mM Tris–HCl buffer, pH 7.4, and incubate it at 25  C for 1 h to form coiledcoil B-decorated β-annulus capsid (Fig. 9b). 10. Characterize the coiled-coil B-decorated β-annulus capsid by TEM (Fig. 9c, middle) and DLS (Fig. 9c, right). 11. For the analysis of the binding of the coiled-coil B-β-annulus capsid to the coiled-coil A peptide, mix 50 μM coiled-coil B-β-annulus and 12.5 μM or 50 μM coiled-coil A peptide in 10 mM Tris–HCl buffer, pH 7.4, and characterize them by TEM (Fig. 9c, middle) and DLS (Fig. 9c, right). 3.4 Characterization of β-Annulus Capsids 3.4.1 Transmission Electron Microscopy (TEM)

The morphology of the β-annulus capsids is directly investigated using TEM. Hydrophilized carbon-coated Cu-grids are used because the capsids are hard to retain on non-hydrophilized grids. 1. Apply an aliquot (5 μL) of the solution of the β-annulus capsids to the grids. 2. Incubate for 1 min for absorption and wick away excess solution using a filter paper. 3. Apply a drop of stain solution for TEM on each of the grids and incubate for 1 min (see Note 7). 4. Dry the sample-loaded grids in a vacuum desiccator before TEM measurement using an accelerating voltage of 80 kV.

Nanocapsules from β-Annulus Peptides

117

3.4.2 Dynamic Light Scattering (DLS)

DLS is a powerful tool to estimate the size of the β-annulus capsids in aqueous solution. The intensity autocorrelation data obtained by DLS enable the determination of a diffusion coefficient, which can be used to calculate the effective hydrodynamic radius by the Stokes–Einstein equation. Typically, the diameter of the β-annulus capsids detected by DLS is between 30 and 50 nm in the concentration range of 25 μM–6 mM (Fig. 3a). The obtained count rates (scattering intensity) are used to estimate critical aggregation concentration (CAC). A plot of count rates as a function of concentration of the β-annulus peptide shows a sharp increase and CAC is determined as 25 μM from the inflection point (Fig. 3b) (see Note 8). The encapsulation and decoration of exogenous molecules possibly affect the size of the β-annulus capsids, which can be detected by DLS.

3.4.3 ζ-Potentials

Surface potentials (ζ-potentials) of the β-annulus capsids provide useful information to estimate whether the exogenous molecules are located inside or outside of the capsids. For instance, the original β-annulus capsid has neutral surface potentials [25] whereas the decoration of negatively charged AuNPs [29] and DNAs [30] on the outer surface of the capsids induces the negative surface potentials. The ζ-potentials are determined by measuring the electrophoretic mobility in disposable Zeta cells.

3.4.4 Fluorescence Correlation Spectroscopy (FCS)

FCS is a sensitive and versatile technique to provide information about fluorescent NPs, such as the diameter and number of molecules, by measuring spontaneous fluorescence intensity fluctuations in a microscopic detection volume of about 1015 L. Small molecules rapidly diffuse and produce rapidly fluctuating intensity patterns, whereas large molecules slowly diffuse and produce slowly fluctuating intensity patterns. By using FCS, encapsulation of fluorescent CdTe NPs inside the β-annulus capsid can be investigated [26]. Figure 5c shows the autocorrelation function curve of 0.1 μM CdTe NPs (see Note 9) with various concentrations of the β-annulus peptide. At 0–25 μM β-annulus peptide, the autocorrelation function shows a single-compartment curve with the diffusion time of 0.08 ms, indicating no encapsulation of CdTe NPs. At 50–200 μM, the autocorrelation function shows a two-step decay curve, which can be fitted to a dual component model, indicating the coexistence of fast (free CdTe NPs) and slow (encapsulated CdTe NPs) components. At 500 μM, the autocorrelation function curve can be fitted to a single component model with a diffusion time of 1.31 ms. These results indicate that CdTe NPs are encapsulated in the β-annulus capsid at peptide concentrations above 50 μM, which corresponds to the CAC of the β-annulus peptide (25 μM). The encapsulation efficiency is determined by analyzing the ratios of the fast and slow components. Thus, FCS is a useful technique to analyze the encapsulation of fluorescent guests in the β-annulus capsid in situ.

118

4

Hiroshi Inaba and Kazunori Matsuura

Notes 1. Weigh the peptide at this stage to determine the crude yield. When the crude yield is low, the efficiency of the introduction of Fmoc-amino acids to the resin may be low. The yield is possibly improved by using other resins with long linkers and/or adding more Fmoc-amino acids, activating agents, and DIPEA. When peptide with remaining protecting groups is observed by MALDI-TOF-MS, repeat the deprotection step in Subheading 3.1.2 using a cleavage cocktail. 2. The gradient of water/acetonitrile and analyzing time should be optimized depending on the types of the peptides and the purity of the crude peptides. 3. The concentration of the guests should be optimized depending on the types of guests and analyzing methods. For instance, the same concentrations of anionic dyes (100 μM) and the β-annulus peptide (100 μM) are used in order to set the same concentrations of anion (dyes) and cation (interior of the β-annulus capsid) [25]. Similarly, very low concentration (13.8 nM) of M13 phage DNA is enough for encapsulation because this concentration corresponds to 100 μM anion [25]. Relatively low concentration (0.1 μM) of CdTe NPs is used because a suitable concentration range of fluorophore for the FCS measurement is 1–100 nM (see Note 9) [26]. 4. It is important to note that an aqueous solution/dispersion of the guests should be added to the powdered β-annulus peptide for encapsulation. Simple mixing of aqueous solutions of β-annulus peptide and guest molecules results in no encapsulation because the encapsulation proceeds during the capsid formation. 5. The concentration of β-annulus-GGGCG peptide should be below CAC (29 μM in this case) before conjugation with AuNP in order to construct β-annulus–AuNP conjugate (not self-assembled capsid). When AuNPs are added to preformed β-annulus capsid, the AuNPs apparently cooperatively adsorb onto Cys residues on the surface of the capsids, resulting in the relatively uncontrolled assemblies of AuNPs. 6. The efficiency of NCL reaction between β-annulus-Nbz and coiled-coil B peptide is very low, probably because of the fast hydrolysis of β-annulus-Nbz than the thioesterification of the peptide at the C-terminus by MPAA. Use of thioester catalysis to accelerate the thioesterification of the peptide makes no significant effect on the thioester exchange reaction. Thus, we convert β-annulus-Nbz into β-annulus-SBn to proceed the NCL reaction with coiled-coil B peptide.

Nanocapsules from β-Annulus Peptides

119

7. Selection of stain solution is important to get clear TEM images. Although uranyl acetate is good for staining of the β-annulus capsids (Fig. 2) [20] with minimal radiological hazards, uranyl acetate and its contaminated waste should be carefully handled. Sodium phosphotungstate is also good for staining without radiological risks (Figs. 8c and 9c) [30, 31]. In the case of the DNA-encapsulated capsids, DNA and capsid can be stained with cisplatin and uranyl acetate, respectively, to investigate the encapsulation of DNA [25]. In the case of the metal NPs such as ZnO NPs [27] and AuNPs (Fig. 7b) [29], staining is not required. In addition, the staining time and the concentration of the stain solution should be optimized. 8. CAC of the β-annulus peptides can be changed depending on the encapsulated and decorated molecules. For instance, CACs of Ni-NTA-β-annulus peptide [28] and coiled-coil B-β-annulus peptide [31] are low (0.053 μM and 5–10 μM, respectively) compared to that of the original β-annulus peptide (25 μM) [20], indicating the formation of the stable capsids. The stabilizations are possibly due to the interaction of the modified moiety (Ni-NTA and coiled-coil B peptide) as discussed in our published papers [28, 31]. In contrast, ZnO-binding β-annulus peptide at 0.1 mM (high concentration compared to the CMC of the original β-annulus peptide) forms irregular fibrous assemblies whereas the peptide forms the capsid structure at 1.0 mM [27]. The result indicates that the formation of the capsid is inhibited by adding the ZnO-binding peptide (HCVAHR) to the N-terminus of the β-annulus peptide. Interestingly, ZnO-binding β-annulus peptide forms the capsid even at 0.1 mM by incubation with ZnO NPs, indicating the stabilization of the capsids by ZnO NPs. 9. At 10 μM CdTe NPs, any analyzable autocorrelation function curve cannot be obtained by FCS measurement. Generally, suitable fluorophore concentration for FCS measurement is in the range of 1–100 nM. In this concentration range, it is possible to yield a good signal-to-noise ratio because of the large signal fluctuations from fluorophores diffusing in the focal volume.

Acknowledgments This work was supported by KAKENHI (15H03838, 18H02089, 18H04558 for K. M.) from the Japan Society for the Promotion of Science (JSPS).

120

Hiroshi Inaba and Kazunori Matsuura

References 1. Boyle AL, Woolfson DN (2012) Rational design of peptide-based biosupramolecular systems. In: Gale PA, Steed JW (eds) Supramolecular chemistry: from molecules to Nanomaterials. Wiley, Hoboken 2. Castillo J, Sasso L, Svendsen WE (2012) Selfassembled peptide nanostructures: advances and applications in nanobiotechnology. Pan Stanford Publishing 3. De Santis E, Ryadnov MG (2015) Peptide selfassembly for nanomaterials: the old new kid on the block. Chem Soc Rev 44:8288–8300 4. Luo Q, Hou C, Bai Y, Wang R, Liu J (2016) Protein assembly: versatile approaches to construct highly ordered nanostructures. Chem Rev 116:13571–13632 5. Heddle JG, Chakraborti S, Iwasaki K (2017) Natural and artificial protein cages: design, structure and therapeutic applications. Curr Opin Struct Biol 43:148–155 6. Matsuura K (2014) Rational design of selfassembled proteins and peptides for nanoand micro-sized architectures. RSC Adv 4:2942–2953 7. Matsuura K (2018) Synthetic approaches to construct viral capsid-like spherical nanomaterials. Chem Commun 54:8944–8959 8. Inaba H, Matsuura K (2019) Peptide nanomaterials designed from natural supramolecular systems. Chem Rec 19(5):843–858 9. King NP, Sheffler W, Sawaya MR, Vollmar BS, Sumida JP, Andre´ I, Gonen T, Yeates TO, Baker D (2012) Computational design of selfassembling protein nanomaterials with atomic level accuracy. Science 336:1171–1174 10. King NP, Bale JB, Sheffler W, McNamara DE, Gonen S, Gonen T, Yeates TO, Baker D (2014) Accurate design of co-assembling multi-component protein nanomaterials. Nature 510:103–108 11. Bale JB, Gonen S, Liu Y, Sheffler W, Ellis D, Thomas C, Cascio D, Yeates TO, Gonen T, King NP, Baker D (2016) Accurate design of megadalton-scale two-component icosahedral protein complexes. Science 353:389–394 12. Hsia Y, Bale JB, Gonen S, Shi D, Sheffler W, Fong KK, Nattermann U, Xu C, Huang P-S, Ravichandran R, Yi S, Davis TN, Gonen T, King NP, Baker D (2016) Design of a hyperstable 60-subunit protein dodecahedron. Nature 535:136–139 13. Kawakami N, Kondo H, Matsuzawa Y, Hayasaka K, Nasu E, Sasahara K, Arai R, Miyamoto K (2018) Design of hollow protein nanoparticles with modifiable interior and exterior

surfaces. Angew Chem Int Ed 57:12400–12404 14. Ghosh S, Reches M, Gazit E, Verma S (2007) Bioinspired design of nanocages by selfassembling triskelion peptide elements. Angew Chem Int Ed 46:2002–2004 15. Castelletto V, De Santis E, Alkassem H, Lamarre B, Noble JE, Ray S, Bella A, Burns JR, Hoogenboom BW, Ryadnov MG (2016) Structurally plastic peptide capsules for synthetic antimicrobial viruses. Chem Sci 7:1707–1711 16. Fletcher JM, Harniman RL, Barnes FRH, Boyle AL, Collins A, Mantell J, Sharp TH, Antognozzi M, Booth PJ, Linden N, Miles MJ, Sessions RB, Verkade P, Woolfson DN (2013) Self-assembling cages from coiled-coil peptide modules. Science 340:595–599 17. Ross JF, Bridges A, Fletcher JM, Shoemark D, Alibhai D, Bray HEV, Beesley JL, Dawson WM, Hodgson LR, Mantell J, Verkade P, Edge CM, Sessions RB, Tew D, Woolfson DN (2017) Decorating self-assembled peptide cages with proteins. ACS Nano 11:7901–7914 18. Noble JE, De Santis E, Ravi J, Lamarre B, Castelletto V, Mantell J, Ray S, Ryadnov MG (2016) A de novo virus-like topology for synthetic virions. J Am Chem Soc 138:12202–12210 19. De Santis E, Alkassem H, Lamarre B, Faruqui N, Bella A, Noble JE, Micale N, Ray S, Burns JR, Yon AR, Hoogenboom BW, Ryadnov MG (2017) Antimicrobial peptide capsids of de novo design. Nat Commun 8:2263 20. Matsuura K, Watanabe K, Matsuzaki T, Sakurai K, Kimizuka N (2010) Self-assembled synthetic viral capsids from a 24-mer viral peptide fragment. Angew Chem Int Ed 49:9662–9665 21. Matsuura K, Mizuguchi Y, Kimizuka N (2016) Peptide nanospheres self-assembled from a modified β-annulus peptide of Sesbania mosaic virus. Biopolymers 106:470–475 22. Olson AJ, Bricogne G, Harrison SC (1983) Structure of tomato bushy stunt virus IV the virus particle at 2·9 A˚ resolution. J Mol Biol 171:61–93 23. Harrison SC, Olson AJ, Schutt CE, Winkler FK (1978) Tomato bushy stunt virus at 2.9 A˚ resolution. Nature 276:368–373 24. Hsu C, Singh P, Ochoa W, Manayani DJ, Manchester M, Schneemann A, Reddy VS (2006) Characterization of polymorphism

Nanocapsules from β-Annulus Peptides displayed by the coat protein mutants of tomato bushy stunt virus. Virology 349:222–229 25. Matsuura K, Watanabe K, Matsushita Y, Kimizuka N (2013) Guest-binding behavior of peptide nanocapsules self-assembled from viral peptide fragments. Polym J 45:529–534 26. Fujita S, Matsuura K (2016) Encapsulation of CdTe quantum dots into synthetic viral capsids. Chem Lett 45:922–924 27. Fujita S, Matsuura K (2014) Inclusion of zinc oxide nanoparticles into virus-like peptide nanocapsules self-assembled from viral β-annulus peptide. Nanomaterials 4:778–791 28. Matsuura K, Nakamura T, Watanabe K, Noguchi T, Minamihata K, Kamiya N, Kimizuka N (2016) Self-assembly of Ni-NTA-modified β-annulus peptides into artificial viral

121

capsids and encapsulation of His-tagged proteins. Org Biomol Chem 14:7869–7874 29. Matsuura K, Ueno G, Fujita S (2015) Selfassembled artificial viral capsid decorated with gold nanoparticles. Polym J 47:146–151 30. Nakamura Y, Yamada S, Nishikawa S, Matsuura K (2017) DNA-modified artificial viral capsids self-assembled from DNA-conjugated β-annulus peptide. J Pept Sci 23:636–643 31. Fujita S, Matsuura K (2017) Self-assembled artificial viral capsids bearing coiled-coils at the surface. Org Biomol Chem 15:5070–5077 32. Uekawa N, Yamazaki A, Ishii S, Kojima T, Kakegawa K (2010) Synthesis of a stable sol of ZnO nanoparticles by low-temperature heating of Zn(OH)2 in ethylene glycol containing Zn2+ ions. J Ceram Soc Jpn 118:96–101

Chapter 8 Electrostatic Self-Assembly of Protein Cage Arrays Soumyananda Chakraborti, Antti Korpi, Jonathan G. Heddle, and Mauri A. Kostiainen Abstract Protein and peptide cages are nanoscale containers, which are of particular interest in nanoscience due to their well-defined dimensions and enclosed central cavities that can be filled with material that is protected from the outside environment. Ferritin is a typical example of protein cage, formed by 24 polypeptide chains that self-assemble into a hollow, roughly spherical protein cage with external and internal diameters of approximately 12 nm and 8 nm, respectively. The interior cavity of ferritin provides a unique reaction vessel to carry out reactions separated from the exterior environment. In nature, the cavity is utilized for sequestration and biomineralization to render iron inert and safe by shielding from the external environment. Materials scientists have been inspired by this system and exploited a range of ferritin superfamily proteins as supramolecular templates to encapsulate cargoes ranging from cancer drugs to therapeutic proteins. Interesting possibilities arise if such containers can themselves be arranged into even higher-order structures such as crystalline arrays. Here, we describe how crystalline arrays of negatively charged ferritin protein cages can be built by taking advantage of electrostatic interactions with cationic gold nanoparticles. Key words Protein cage, Nanocrystals, Protein engineering, Protein design, Nanocontainers

1

Introduction Protein cages are typically spherical assemblies from under 10 to many 10s of nanometres in diameter and having a hollow central cavity. They are typified by viruses which, in nature, carry a genome cargo and demonstrate particular potential for gene delivery applications. Manipulation of natural and artificial protein cages is an area of growing interest [1, 2], and the idea that empty protein cages can be filled with different cargoes of choice and modified chemically as well as genetically has recently grown in prominence [3]. The ability to organize protein cages into ordered arrays in three-dimensions may result in interesting physical properties or other useful characteristics depending on the nature of the encapsulated cargo [4]. Such arrays can be achieved in a number

Maxim G. Ryadnov (ed.), Polypeptide Materials: Methods and Protocols, Methods in Molecular Biology, vol. 2208, https://doi.org/10.1007/978-1-0716-0928-6_8, © Springer Science+Business Media, LLC, part of Springer Nature 2021

123

124

Soumyananda Chakraborti et al.

of ways, including (a) covalent “crosslinking” of protein cages, e.g., by chemical crosslinkers between appropriate amino acid side chains [5] and (b) complementary binding interactions, e.g., selfassociating peptides can be recombinantly added to the surface of the protein cages which then interact. The disadvantage of covalent systems is the irreversibility, meaning that “incorrect” crosslinks will quickly lead to lattice defects. Recombinantly produced modifications that introduce noncovalent interactions are timeconsuming to produce and have the disadvantage that the modifications are introduced at every equivalent position on the protein cage, which may not be desirable given that protein cages are typically constructed of many 10s of identical protein subunits. An alternative approach is to use electrostatic effects, which allow self-assembly with dynamic and reversible rearrangements. This has been demonstrated using tobacco mosaic virus [6], cowpea chlorotic mottle virus, and ferritin protein cages that produce superlattices when mixed with appropriately modified gold nanoparticles [7, 8]. Subsequently, many other linker molecules, which are able to electrostatically self-assemble with protein cages, have been developed in the past few years [9–11]. Recently, 3D protein cage arrays have been found effective in, for example, enzyme cascade reactions and other applications [3, 12]. Ferritin is of interest as it is a widely characterized bionanotechnological tool and can be easily modified [13, 14]. Here, we give a detailed method for the preparation of electrostatically self-assembled binary lattices consisting of ferritin protein cages and cationic gold nanoparticles (Fig. 1), based on our recently published research [15].

Fig. 1 Schematic representation of superlattice (left) and cryo-TEM image (right) of the crystalline structures formed by ferritin and AuNPs. In the schematic, ferritins are colored blue and purple, AuNP yellow

3D Protein Cage Arrays

2

125

Materials All the reagents used in this study are analytical grade; they were mostly purchased from major suppliers (Sigma, Fisher Scientific, and VWR) and used as received unless otherwise specified. All samples were prepared using ultrapure water produced by a Nanopure® water purification system (Thermo Fisher Scientific). Glassware was purchased from Fisher Scientific. 1. Ferritin preparation buffer: 0.02 M HEPES, pH 7.4: Weigh 4.77 g of HEPES (Sigma catalog number: H3375) and transfer to a glass beaker. Add water to a volume of 900 mL. Mix and adjust pH with NaOH (see Notes 1 and 2). Make up to 1000 mL with water. Filter the prepared buffer through 0.45μM Whatman filter paper and then autoclave. Store at 4
C. 2. 0.05 M Tris–HCl, pH 8.5: Weigh 7.88 g of Tris–HCl (Sigma catalog number: T3252) and transfer to a glass beaker. Add water to a volume of 900 mL. Mix and adjust pH with NaOH. Make up to 1000 mL with water. Filter the prepared buffer through 0.45-μM Whatman filter paper and then autoclave. Store at 4
C. 3. Ferritin cage assembling buffer: 0.1 M MgCl2 in Tris–HCl, pH 8.5: Weigh 20.33 g of MgCl2 · 6 H2O (Sigma catalog number: M2670) and transfer to a glass beaker. Add Tris– HCl, pH 8.5, to a volume of 900 mL. Filter the prepared buffer through 0.45-μM Whatman filter paper and then autoclave. Adjust pH with NaOH. Make up to 1000 mL with TrisHCl. Store at 4
C (see Note 3). 4. Prepare 1 M NaCl by dissolving 0.584 g of NaCl (Sigma catalog number S7653) in 10 mL filtered MilliQ water. Store at room temperature. 5. Gold nanoparticles (AuNPs) can be prepared according to the method developed by Brust and Schiffrin [16, 17]. 6. Expression plasmids, ferritin expression plasmid: Ferritins, in general, are facile to produce and purify. The gene sequence of ferritin from Thermotoga maritima is available from GenBank (Genbank ID: GenBank: AAD36204.1) and can be codon optimized against the model organism E.coli. After optimization, the predicted, translated protein sequence should be checked against the protein sequence database UniProt (UniProt ID: Q9X0L2). Gene synthesis along with cloning of the gene to pET21a (+) plasmid (see Note 4) can be performed using a commercial service (we typically use GenScript, https://www.genscript.com). For our studies, the ferritin from Thermotoga maritima ferritin was most appropriate because of its specific characteristics (see Note 5). 7. Ferritin protein: Highly pure (>95%) recombinant Thermotoga maritima ferritin protein can be produced and purified as

126

Soumyananda Chakraborti et al.

reported earlier [15, 18]. Protein is stored in 20 mM HEPES buffer (pH 7.4) at 4
C (see Note 6). For long-term storage, 15% glycerol is added, and the sample is flash-frozen using liquid nitrogen (180
C) and finally stored in a 80
C freezer (see Note 7). 8. Amicon Ultra- 0.5 Centrifugal Filter Unit (molecular weight cutoff 10,000 Da, Millipore). 9. Eppendorf MiniSpin Centrifuge, rotor F-45-12-11 and Eppendorf centrifuge 5424R, rotor FA-45-24-11. 10. 200-mesh copper grids with lacey carbon support film (Electron Microscopy Sciences). 11. Leica EM GP2 automatic plunge freezer.

3

Methods

3.1 AuNP Surface Modification

Gold nanoparticles prepared using the Brust–Schiffrin method are stabilized by alkanethiols, which are not effective for generating superlattices through electrostatic self-assembly. Therefore, the AuNP surface was further modified with a cationic ligand (Fig. 2). This is first synthesized by a slight modification of the method developed by Rotello et al. [19] (Fig. 3) as follows: 1. Weigh 30 mg of AuNPs using an electrotonic balance. 2. Dissolve AuNPs in 5 mL of dichloromethane. 3. Transfer the AuNP solution to a 50-mL glass beaker (see Note 8), add a magnetic stirrer bar, place on a magnetic stirrer, and stir (e.g., at 500 rpm). 4. Weigh 120 mg (0.28 mmol) of cationic ligand and transfer it into a 25-mL glass beaker. 5. Filter 5 mL of MilliQ water through 0.45-μM Whatman paper and to the beaker containing cationic ligand. Dissolve by simple shaking. 6. Transfer the solution to the glass beaker containing the AuNPs. Incubate for 1 h at room temperature (RT) with continuous stirring. During the process of mixing, gold nanoparticles will spontaneously transfer from the dichloromethane phase to the aqueous phase. 7. Carefully transfer the AuNP-containing the aqueous phase to a 50-mL round-bottom flask. Discard dichloromethane. For handling dichloromethane, refer to the link (http://www. sciencemadness.org/smwiki/index.php/Dichloromethane) for instructions. 8. Fix the round-bottom flask to a rotary evaporator and evaporate the remaining organic solvent until AuNP is completely dry (see Notes 9 and 10).

3D Protein Cage Arrays

Dcore ~ 2.5 nm

127

Dh ~ 8.5 nm S

O

5

4

N

Fig. 2 Cartoon showing gold nanoparticles with a cationic ligand used in the described method. The diameter of gold nanoparticles (Dcore) was determined using TEM, and the hydrodynamic radius (Dh) of gold nanoparticles capped with a cationic ligand is measured using DLS. The respective sizes of gold nanoparticles with and without capping agents are 2.5 and 8.5 nm, respectively

Fig. 3 Schematic showing modification of the AuNPs. Mix Brust–Schiffrin AuNPs in organic solvent and the cationic ligand in aqueous solvent, incubate together, separate the aquatic phase, and evaporate traces of organic solvent. Purify the cationic AuNPs by dialysis against water and collect by lyophilization

9. Prepare a dialysis tank by filling a 1-L glass beaker with 1 L of filtered MilliQ water and a magnetic stirrer bar. 10. Collect dried AuNP and further dissolve in 1 mL filtered MilliQ water and transfer to a cellulose membrane dialysis tube (molecular weight cutoff ¼ 14,000, see Note 11). Seal the bag using dialysis clips. Carefully check the bag for any leaks. 11. Transfer the dialysis bag to the dialysis tank. For better rotational control, a floating device can be added to one end of the dialysis bag. Cover the top of the beaker with aluminum foil to prevent contamination with dust. Place on a magnetic stirrer and stir at a constant speed of 200 rpm at room temperature for 48 h. During dialysis, replace the water in the tank at 6 h, 12 h, and 24 h. 12. After the dialysis is complete, remove the dialysis bag from the beaker and dismantle by removing the dialysis clip. Carefully transfer the AuNP solution to a 10-mL round-bottom flask. 13. Fit the 10-mL round-bottom flask to a lyophilizer and operate in a high vacuum mode. After 30 min, stop the instrument and collect the lyophilized AuNPs (brown solid; see Note 12).

128

Soumyananda Chakraborti et al.

14. The isolated AuNP can be further characterized for size and shape by DLS (dynamic light scattering) and TEM (transmission electron microscopy) and for the presence of cationic ligand on the surface by NMR (nuclear magnetic resonance). These methods are beyond the scope of this article but are explained elsewhere [8]. The synthesized gold nanoparticle should have a core size of approx. 2.5 nm in diameter, with a total diameter for the particle being approx. 8.5 nm once the ligand is taken into account. 15. Prepare AuNP stocks by dissolving dried particles in Milli-Q water to give a final concentration of ~10 mg/mL and store at 4
C (see Note 13). 3.2 Preparation of Ferritin

1. TmFtn requires buffer exchange into a buffer of appropriate pH (50 mM Tris-HCl, pH 8.5) for superlattice formation. Exchange can be achieved by a standard protocol such as follows: Take 100 μL of ferritin and place it into a 0.5 mL Amicon Ultra Centrifugal Filter Unit. Add an appropriate buffer to make up to a total volume to 0.5 mL and centrifuge at 3381  g for 30 min at 4
C, with a proper balance. This should result in a decrease of the volume of the solution retained in the filter unit to approx. one fifth, though this can change on a case-by case basis, so caution should be taken in choosing centrifugation time. The centrifugation unit should be disassembled and the flow-through discarded. Take care not to discard the retained solution. 2. After the fifth wash, the protein sample solution should be carefully collected from the filter unit using a 200-μL pipette and transferred to a 1.5-mL microcentrifuge tube; adjust the volume of protein sample to 100 μL by adding fresh buffer. 3. Determine protein concentration using a UV–Vis spectrometer and the theoretical extinction coefficient of 29,910 M1 cm1 for TmFtn. Remember to measure the buffer alone as a blank and subtract it from the reading obtained with the proteincontaining sample. 4. The protein is expected to be a dimer in solution under these conditions. Assess homogeneity by DLS. In our case, we placed 30 μL of protein sample into a black quartz cuvette and measured with a Malvern Zetasizer Nano ZSP (see Note 14). Perform measurements at 298 K and make each measurement in triplicate. A single peak corresponding to the size of ferritin dimer (approx. 5 nm diameter) should be obtained. 5. For generating cage, mix 100 μL of (20 mg/mL) of ferritin dimer with an equal volume of 100 mM MgCl2 and Tris-HCl (50 mM, pH 8.5) and incubate for approx. 12 h at RT in a 1.5mL microcentrifuge tube. Then, centrifuge at 16,863  g for

3D Protein Cage Arrays

129

10 min at RT to remove any large aggregates. After centrifugation, carefully collect the supernatant and transfer to a new 1.5mL microcentrifuge tube. 6. After cage formation, assess sample quality by again using DLS and parameters as described above. A single peak corresponding to the size of ferritin (approx. 12 nm diameter) should be obtained (see Note 15). 7. TEM is used to confirm cage formation. For this, add four microliters of sample (0.1 mg/mL protein cage) on to a copper-Formvar grid (QUANTIFOIL) prepared by glow discharging (see Note 16). Incubate for 1 min and remove the excess solution by blotting with filter paper. Shortly after soaking, add 4 μL of staining solution, 2% uranyl acetate (see Note 17), incubate for 1 min, and remove the excess stain by blotting with filter paper. Full details of TEM sample preparation and sample acquisitions are beyond the scope of this chapter and should be carried out in conjunction with a TEM expert. 3.3 Superlattice Formation

1. For superlattice formation, add 10 μL of 8 mg/mL ferritin to a 250-μL microcentrifuge tube and mix with 5 μL of Tris-HCl buffer (pH 8.5). 2. Adjust electrolyte concentration so that the final concentration is in the range of 0–100 mM NaCl by adding 1 μL of NaCl to the reaction mixture from the appropriate serially diluted NaCl stock (see Note 18). Optimal NaCl concentration for superlattice formation is ~20 mM for Thermotoga maritima ferritin. 3. Finally, add 4 μL of 20 mg/mL AuNP to the ferritin and mix AuNP with ferritin by pipetting. The final volume of reaction should be kept at 20 μL. 4. Keep ferritin to AuNP ratio at 1:1 (w/w) for the final reaction mixture to avoid excess AuNP. The final Mg2+ concentration in the reaction should be 25 mM. 5. Incubate the reaction mixture for 10–15 min at room temperature (see Note 19). 6. Characterization of superlattice should be first carried out by DLS analysis. Further characterization of superlattice can be made by small-angle X-ray scattering (SAXS) and cryo-TEM.

3.4 Further Characterization

Cryo-TEM images can be obtained. For example, using a JEM 3200FSC field emission microscope (JEOL) operated at 300 kV in bright-field mode with an Omega-type Zero-loss energy filter. Cryo-TEM sample preparation and data collection is complex and should be carried out with an expert in the field. A brief overview of the method as applied to this work is as follows:

130

Soumyananda Chakraborti et al.

1. Prepare 200-mesh copper grids with lacey carbon support film by plasma cleaning. For example, treating in a Gatan Solarus (Model 950) plasma cleaner for 30 s. 2. Place 3 μl of freshly prepared aqueous sample dispersion onto the grids. 3. Plunge-freeze in a 170
C ethane/propane mixture at 100% humidity. 4. Cryo-transfer glassy specimens to the microscope. 5. Capture images, e.g., with Gatan Digital Micrograph® software while maintaining the specimen temperature at 187
C. SAXS can be used to verify superlattice formation. The details of this method are beyond the scope of this chapter but can be found elsewhere. [8–11] SAXS results should show that ferritinAuNP complexes arrange into crystal lattices with an interpenetrating face-centered cubic (fcc) structure.

4

Notes 1. Concentrated NaOH (10 M) can be used at first to achieve a pH close to 8. Then, use a lower concentration of NaOH to avoid a sudden increase in pH above the required pH. Before using a pH meter, it should be calibrated against standard pH solutions available commercially. 2. Use a magnetic stirrer for mixing. 3. It is advisable to use freshly prepared buffer; however, if necessary, buffer can be stored at 4
C for limited periods. 4. Thermotoga maritima ferritin (without any modification) is commercially available from MoLiRom (http://www.mol irom.com). 5. The ferritin from Thermotoga maritima is extremely thermostable [15, 18], and no apparent denaturation of protein was observed when it is boiled at 100
C; furthermore, Thermotoga maritima ferritin shows divalent metal (Mg+2, Ca+2)-mediated assembly behavior, which is unusual among the ferritin family of proteins. 6. All the purified ferritin samples were filtered through a 0.22-μ m filter to avoid bacterial contamination. Ferritin aliquots were always prepared under laminar flow under completely sterile conditions. 7. For long-term protein storage, we used 15% glycerol; this is particularly useful if your sample is to be flash-frozen, as it prevents ice crystal formation. The exact percent glycerol required may need to be optimized depending on the protein.

3D Protein Cage Arrays

131

Handling of liquid N2 must be performed with caution, as contact can cause severe burns. The use of appropriate protective clothing, gloves, and laboratory goggles is advised. 8. Glassware for gold nanoparticle reactions should be thoroughly cleaned before use. In our case, we first cleaned with chromic acid, followed by water (at least 3–4 times). Glassware was then baked at 200
C for 1 h. Try to avoid cleaning of glassware using detergent (to avoid any contamination), trace amounts of which may interfere with the reaction. 9. Rotary evaporation should be carried out using a water bath at 35–40
C and rotation speed greater than 100 rpm to assure stable boiling. To avoid loss of material by splashing of the solvent out of the sample flask, slowly approaching the target pressure is recommended. The boiling point of dichloromethane is approximately 40
C at atmospheric pressure, so no or very weak vacuum is needed for evaporation. However, it is recommended to keep the solution at 500–750 mbar and 40
C for 5–10 min to completely remove the organic solvent, as these conditions exceed the boiling point of dichloromethane but are insufficient for evaporating water. 10. Rotary evaporators can house high vacuums inside glass components, which may break catastrophically if they are not initially intact, and rapid pressure changes can occur if the sample flask is detached incorrectly, which may cause splashing of solvents. Therefore, appropriate safety gear should always be worn when handling the instrument, including safety goggles, laboratory coat, and gloves. 11. Before starting dialysis, the dialysis bag should be cleaned properly; in our case, we cleaned the dialysis bag extensively with filtered MilliQ water (3–4 times) before use. According to our observations, full recovery of samples after dialysis cannot be achieved. 12. It was found in many studies that AuNPs, especially those smaller than 5 nm, are highly toxic to cells [20], so care must be taken during handling them; using personal protection is advisable. 13. Pipette up and down to make sure complete mixing of protein in the solution. 14. The cuvette for DLS should be thoroughly cleaned before use. First, with 70% ethanol (3 times), followed by Milli Q water (4 times), and finally dried by N2 purging. DLS measurements should be performed in a clean and dust-free environment, as atmospheric dust can influence the results heavily. DLS can be measured with very low concentration of protein; in our experimental setup, 0.1 mg/mL ferritin protein cage is sufficient.

132

Soumyananda Chakraborti et al.

15. Thermotoga maritima ferritin cage has an octahedral symmetry and can store iron in its 8 nm cavity; the external diameter of ferritin is ~12 nm when measured under TEM and 12–16 nm when measured using DLS. 16. Glow discharge was performed to make TEM grids (copperFormvar) more hydrophilic. Normally, copper-Formvar grids are highly hydrophobic in nature. By performing glow discharge, the grids become more hydrophilic; the surface of grids also becomes more negatively charged on glow discharging, which helps aqueous solutions to spread on the grid surface. 17. For negative staining in TEM, we use uranyl acetate, which is radioactive. In addition, uranium is a toxic heavy metal. Appropriate safety cautions and local regulations must be followed when using it. As an alternative to uranyl acetate, one can use phosphor-tungstic acid (PTA) or platinum blue. These compounds are less hazardous compared to uranyl acetate but, on a case-by-case basis, may not be very effective at staining samples. 18. Self-assembly of ferritin into superlattices required fine adjustment of the interparticle interactions. Normally, very strong interaction between the particles leads to the formation of amorphous aggregates. On the other hand, weak interparticle interaction fails to produce any kind of assemblies [8]. Changing electrolyte concentration of the reaction mixture is the most common way to fine-tune interparticle interaction; another way is changing pH. Both ways were found to be very effective; however, thorough optimization is needed in order to get the best result. 19. We found that freshly prepared ferritin was more effective in superlattice formation, compared to old (aged) protein samples, especially those protein samples which were frozen several times. We also observed that protein stored at 80
C for a month is suitable for superlattice formation.

Acknowledgments SC was supported by the Homing programme of the Foundation for Polish Science co-financed by the European Union under the European Regional Development Fund, grant No. Homing/ 2017-3/22 awarded to SC. JGH was funded by the TEAM Programme of the Foundation for Polish Science co-financed by the European Union under the European Regional Development Fund (TEAM/2016-3/23) awarded to J.G.H. A.K. and M.K. acknowledge the financial support from the Academy of

3D Protein Cage Arrays

133

Finland (Grants 308578, 303804, 273645, 267497, 272579) and Sigrid Juselius Foundation. The research made use of the instrumentation of Aalto University Nano-microscopy Center. References 1. Heddle JG et al (2017) Natural and artificial protein cages: design, structure and therapeutic applications. Curr Opin Struct Biol 43:148–115. https://doi.org/10.1016/j.sbi. 2017.03.007 2. Heddle JG, Tame JRH (2012) Protein nanotubes, channels and cages. The Royal Society of Chemistry, Cambridge 3. Pulsipher KW et al (2016) Ferritin: versatile host, nanoreactor, and delivery agent. Isr J Chem 56:660–670. https://doi.org/10. 1002/ijch.201600017 4. Aumiller WM et al (2018) Protein cage assembly across multiple length scales. Chem Soc Rev 47:3433–3469. https://doi.org/10. 1039/c7cs00818j 5. Abe S et al (2016) Design of a confined environment using protein cages and crystals for the development of biohybrid materials. Chem Commun 52:6496–6512. https://doi. org/10.1039/c6cc01355d 6. Liljestro¨m V et al (2017) Cooperative colloidal self-assembly of metal-protein superlattice wires. Nat Commun 8:671. https://doi.org/ 10.1038/s41467-017-00697-z 7. Hassinen J et al (2015) Rapid cationization of gold nanoparticles by two-step phase transfer. Angew Chem Int Ed Engl 54:7990–7993. https://doi.org/10.1002/anie.201503655 8. Kostiainen MA et al (2013) Electrostatic assembly of binary nanoparticle superlattices using protein cages. Nat Nanotechol 8:52–56. https://doi.org/10.1038/nnano.2012.220 9. Liljestro¨m V et al (2015) Electrostatic selfassembly of soft matter nanoparticle cocrystals with tunable lattice parameters. ACS Nano 9:11278–11285. https://doi.org/10.1021/ acsnano.5b04912 10. Liljestro¨m L et al (2014) Self-assembly and modular functionalization of threedimensional crystals from oppositely charged proteins. Nat Commun 5:4445. https://doi. org/10.1038/ncomms5445 11. Korpi A et al (2018) Self-assembly of electrostatic cocrystals from supercharged fusion peptides and protein cages. ACS Macro Lett

7:318–323. https://doi.org/10.1021/ acsmacrolett.8b00023 12. Uchida M et al (2018) Modular self-assembly of protein cage lattices for multistep catalysis. ACS Nano 12:942–953. https://doi.org/10. 1021/acsnano.7b06049 13. Georgiev IS et al (2018) Two-component ferritin nanoparticles for multimerization of diverse trimeric antigens. ACS Infect Dis 4:788–796. https://doi.org/10.1021/ acsinfecdis.7b00192 14. Tsukamoto R et al (2013) Effect of PEGylation on controllably spaced adsorption of ferritin molecules. Langmuir 29:12737–12743. https://doi.org/10.1021/la4029595 15. Chakraborti S et al (2019) Three-dimensional protein cage array capable of active enzyme capture and artificial chaperone activity. Nano Lett 19:3918–3924. https://doi.org/10. 1021/acs.nanolett.9b01148 16. Kanaras AG et al (2002) Thioalkylated tetraethylene glycol: a new ligand for water soluble monolayer protected gold clusters. Chem Commun 20:2294–2295. https://doi.org/ 10.1039/B207838B 17. Brust M et al (1994) Synthesis of thiolderivatised gold nanoparticles in a two-phase liquid-liquid system. Chem Commun (7):801–802. https://doi.org/10.1039/ C39940000801 18. Ceci P et al (2011) The characterization of Thermotoga maritima ferritin reveals an unusual subunit dissociation behavior and efficient DNA protection from iron-mediated oxidative stress. Extremophiles 15:431–439. https://doi.org/10.1007/s00792-011-03743 19. Miranda OR et al (2010) Enzyme-amplified array sensing of proteins in solution and in biofluids. J Am Chem Soc 132:5285–5289. https://doi.org/10.1021/ja1006756 20. Carnovale C et al (2019) Identifying trends in gold nanoparticle toxicity and uptake: size, shape, capping ligand, and biological corona. ACS Omega 4:242–256. https://doi.org/10. 1021/acsomega.8b03227

Chapter 9 Design and Generation of Self-Assembling Peptide Virus-like Particles with Intrinsic GPCR Inhibitory Activity Sergey G. Tarasov, Marzena Dyba, Joshua Yu, and Nadya Tarasova Abstract Synthetic analogs of the second transmembrane domain (TM) containing a portion of the extracellular loop 1 of G-protein-coupled receptors (GPCR) can serve as biased antagonists of the corresponding receptor. Analogs with negative charges added to the extracellular end self-assemble into round structures. Addition of polyethylene glycol chains of defined length to the C-terminus of the peptides prevents super aggregation and results in highly uniform particles that can fuse with cell membranes spontaneously. Added PEG chains slow down cell fusion, while attachment of receptor ligands to the surface of particles results in receptor-mediated membrane fusion and cell-selective delivery. Critical assembly concentration of TM peptide particles is in the nanomolar range and thus requires nontraditional methods of determination. In this chapter, we outline sequence selection and design of self-assembling GPCR antagonists, methods of the preparation of the nanoparticles, and biophysical methods of particle characterization. The protocols allow for straightforward rational design, generation, and characterization of self-assembling GPCR antagonists for a variety of applications. Key words G-protein-coupled receptors, Microscale thermophoresis, Differential scanning fluorimetry, Dynamic light scattering, Transmembrane peptide, Chemokine receptor

1

Introduction The development of peptide therapeutics is experiencing rapid growth both in industry and academia [1, 2]. Peptides are particularly valuable in filling in for small molecules in inhibition of “undruggable” targets. However, low stability in circulation, rapid metabolism, and poor cell penetration still prevent wider use of peptides as drugs. Peptide inhibitors/activators also present challenges for structure-based design, as the reliability of molecular docking of peptides decreases rapidly with increasing length, particularly, beyond 5 residues [3, 4]. Over a decade ago, we found that analogs of the second transmembrane domain (TM) of Gprotein-coupled receptors can selectively inhibit signaling of the receptor they are derived from [5]. More recent studies revealed

Maxim G. Ryadnov (ed.), Polypeptide Materials: Methods and Protocols, Methods in Molecular Biology, vol. 2208, https://doi.org/10.1007/978-1-0716-0928-6_9, © Springer Science+Business Media, LLC, part of Springer Nature 2021

135

136

Sergey G. Tarasov et al.

that properly derivatized TM antagonists self-assemble into highly homogeneous particles [6]. Nanoparticles fuse with cell membranes spontaneously and inhibit receptors signaling both in vitro and in vivo. Inhibition, at least in the case with the CXCR4 receptor, is due to the formation of a ternary complex with the receptor and the chemokine ligand, CXCL12 [7]. During complex formation, peptide antagonist prevents the portion of CXCL12 critical for receptor-mediated activation of the G-protein from interacting with the receptor. The rest of the chemokine is still capable of stimulating β-arrestin recruitment and receptor trafficking. Consequently, the receptor signaling is inhibited, but receptor downregulation is not. This prevents antagonist tolerance that is typical for regular antagonists and makes TM inhibitors particularly valuable [7]. TM antagonists have been used successfully in several animal studies and were found to be stable to proteolysis despite being made from all L-amino acids [6, 8–10]. It appears that selfassembly protects TM peptides from degradation in circulation [11]. In addition, particles can be targeted to cell surface receptors. Addition of receptor ligands to the particles makes them fuse with cells in receptor-mediated manner similar to viruses, thus allowing for cell-selective delivery. Consequently, TM antagonists present a new paradigm of a drug with multiple biological activities that delivers itself. The strategy is likely to have many applications in the generation of particles with different biological and physicochemical properties.

2

Materials

2.1 Synthesis of Self-Assembling TM Analogs

1. Liberty Blue Corporation).

Microwave

peptide

synthesizer

(CEM

2. Low-loading Rink Amide MBHA resin (Merck). 3. Dimethylformamide, peptide synthesis grade. 4. Methylene chloride. 5. Pierce™ Disposable Columns, 10 ml. 6. Fmoc amino acids (CEM or another supplier). 7. Pseudoproline dipeptides (Merck or ChemImpex). 8. Fmoc-NH-(PEG)27-COOH (Merck or PolyPure). 9. 0.5 M solution of Oxyma Pure (Ethyl cyano(hydroxyimino) acetate) (CEM) in DMF. 10. 1 M solution of N,N0 -diisopropylcarbodiimide (DIC) (ChemImpex) in DMF. 11. 1H-6-Chlorobenzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HCTU, ChemImpex).

Synthetic Virus-like Particles

137

12. N,N-Diisopropylethylamine (DIEA). 13. Deprotection mixture consisting of 7% piperazine and 0.1 M hydroxybenzotriazole (HOBt) hydrate (ChemImpex) in DMF/methanol 9:1 mixture. 14. Freshly prepared resin cleavage cocktail consisting of 90.0% (v/v) trifluoroacetic acid (TFA), 2.5% water, 2.5% triisopropyl-silane, 2.5% 2,20 -(ethylenedioxy)diethanethiol, and 5% thioanisol. All reagents are from Sigma-Aldrich. 2.2 Self-Assembly of Peptides

1. Dimethylsulfoxide. 2. Phosphate-buffered saline (PBS). 3. 4.6% solution of octyl glucoside (n-octyl-β-D-glucopyranoside, Anatrace) in PBS 4. Sonicator bath. 5. Benchtop centrifuge, 5415D (Eppendorf) or similar.

2.3 Dynamic Light Scattering

1. DynaPro NanoStar instrument (Wyatt Technology Corp., Santa Barbara, CA) with a laser wavelength of 658 nm. 2. Dynamics 7.1.7.16 software (Wyatt Technology). 3. Quartz 10 μl cell (Wyatt Technology).

2.4 Microscale Thermophoresis

1. Microscale thermophoresis instrument Monolith NT.115 (Nanotemper GmbH, Germany) with software NT Control v2.1.33 or higher. 2. Standard or Premium MST capillaries (Nanotemper GmbH, Germany). 3. Protein LoBind Eppendorf tubes, 0.5 ml volume.

2.5 Differential Scanning Fluorimetry

1. Differential scanning fluorimetry instrument Prometheus NT.48 (Nanotemper GmbH, Germany) with PR.TemperatureControl software. 2. Standard capillaries for nanoDSF (Nanotemper GmbH).

3

Methods

3.1 Structure-Based Design of TM Antagonists

Structure–activity studies conducted on CXCR4 receptor TM antagonists have shown that X4-2-6 peptide contains an optimal fragment of CXCR4 sequence for the inhibition of receptor signaling. Crystal structure of CXCR4 revealed that the sequence covers about a half of TM2 and entire extracellular loop 1 that is about 5 residues long (Fig. 1a). Although X4-2-6 is predominantly helical in membranes and membrane mimetics, it adopts a hairpin conformation with a short helix at the C-terminus in aqueous solutions (Fig. 1b) and appears to retain it during self-assembly. Impact of

138

Sergey G. Tarasov et al.

Fig. 1 Peptides derived from the second transmembrane domain of G-protein-coupled receptors selfassemble into homogeneous virus-like particles. (a) CXCR4 structure (PDB: 3ODU) with the part used for generation of a self-assembling receptor antagonist highlighted in blue. (b) In solution, TM2-derived peptide adopts predominantly a hairpin conformation as determined by solution NMR. (c) Transmission electron microscopy revealed a round shape of the particles. (d) 3D reconstitution of 2D TEM images confirmed overall spherical shape of the particles

amino acid residue substitutions on self-assembly revealed a critical role of TM2 proline (Pro92 in CXCR4) that is highly conserved in GPCRs and appears to provide for the turn in the hairpin fold of the peptide. Deletion of N-terminal leucine residues results in the formation of amyloid-type fibers rather than rounded structures [11]. Other essential structural features required for precise selfassembly are the free N-terminus and at least two aspartic acid

Synthetic Virus-like Particles

X4-2-6: R3-2-1: R5-2-1: R4-2-1: SP|P61073|CXCR4_HUMAN SP|P51677|CCR3_HUMAN SP|P51681|CCR5_HUMAN SP|P51679|CCR4_HUMAN SP|P32239|GASR_HUMAN SP|P32246|CCR1_HUMAN SP|P51686|CCR9_HUMAN SP|P41597|CCR2_HUMAN SP|P51684|CCR6_HUMAN SP|P32248|CCR7_HUMAN SP|P51685|CCR8_HUMAN SP|P25100|ADA1D_HUMAN SP|P49682|CXCR3_HUMAN SP|P50406|5HT6R_HUMAN :

139

NH2-LLFVI-TLPFWAVDA-VANWYFGNDD-PEG-27-OH NH2-LLFLV-TLPFWIHYVRGHNWVFGHDDD-PEG27-OH NH2-LFFLL-TVPFWAHYA-AAQWDFGNDD-PEG27-OH NH2-LLFVF-SLPFWGYYA-ADQWVFGLGDDD-PEG27-OH VADLLFVI-TLPFWAVDA-VANWYFGNFLC ISDLLFLV-TLPFWIHYVRGHNWVFGHGMC ISDLFFLL-TVPFWAHYA-AAQWDFGNTMC ISD-LLFVFSLPFWGYYA-ADQWVFGLGLC VSDLLLAVACMPFTLLPNLMGTFIFGTVIC ISDLLFLF-TLPFWIDYKLKAQWDFGNTMC IADLLFLV-TLPFWAIAA-ADQWKFQTFMC ISDLLFLI-TLPLWAHSA-ANEWVFGNAMC IADILFVL-TLPFWAVSHATGAWVFSNATC VADILFLL-TLPFWAYSA-AKSWVFGVHFC LSDLLFVF-SFPFQTYYL-LDQWVFGTVMC VADLLLSATVLPFSATMEVLGFWAFGRAFC VADTLLVL-TLPLWAVDA-AVQWVFGSGLC TSDLMVGLVVMPPAMLNALYGRWVLARGLC :* :. .* *

Fig. 2 Examples of TM antagonist sequences and receptor fragments they are derived from. Parts of receptors used for generation of TM antagonists are underlined. Dipeptides substituted with pseudoproline dipeptides during the synthesis are double underlined

residues added to the C-terminus of the receptor fragment. Positively charged N-terminus stabilizes the hairpin by interacting with negatively charged aspartates, while the repulsion of the Asp negative charges is likely to allow for the formation of finite structures. The peptide must be of a certain length for the formation of the stabilizing ion pair and have no negative charges on the N-terminus. Polyethylene glycol must be added to the C-terminus to prevent aggregation of the particles. We found the optimal number of monomeric units in the PEG chain to be 27. Consequently, to design a sequence of self-assembling GPCR antagonists, one follows the following rules. 1. The N-terminal residue of the peptide is the one right after the conserved aspartate of TM2 of the targeted receptor (Fig. 2). 2. The C-terminus is 14–15 residues downstream from the conserved Pro of TM2, and two sequences may need to be tested. In most cases, terminating two residues upstream from conserved cysteine worked the best. 3. To the C-terminus, 2–3 residues of aspartic acid are added, followed by PEG27.

140

Sergey G. Tarasov et al.

3.2 TM Analog Synthesis

Hydrophobic transmembrane peptides have a strong tendency to aggregate during the synthesis, which can lead to deletions and poor yields. Synthesis of self-assembling antagonists is somewhat easier because it starts from a long PEG chain that works as a spacer and insulates the peptide from polystyrene beads. Invariant Pro residue in the middle of the sequence, the use of microwave irradiation, low-loading resins, and pseudoproline dipeptides wherever possible allow counteracting the aggregation propensity and achieving over 80% purity of crude peptides. Peptides have been synthesized on a Liberty Blue Microwave peptide synthesizer (CEM Corporation) using Fmoc chemistry at a 0.1-mmol scale. 1. Due to the high cost of Fmoc-NH-(PEG)27-COOH, it is attached to low-loading Rink Amide MBHA resin (Merck) manually overnight using 1.2-fold excess of the amino acid, HCTU as an activating agent, and an equimolar amount of DIEA. 2. After coupling, wash the resin in a disposable polypropylene column with DMF and methylene chloride and dry under a fume hood. 3. Weigh out 0.1 mmol Fmoc-NH-(PEG)27-resin (see Note 1). Load it into a 50-ml centrifuge tube of resin loader of the peptide synthesizer. 4. Run the synthesis with the following modifications to the published protocol of high-efficiency peptide synthesis [12]. The coupling with DIC/Oxyma is performed for 4 min at 90  C for all residue except for His, for which the reaction is carried out for 10 min at 50  C. Fivefold amino acid excess is used on all cycles, and all residues are double-coupled. 8% piperazine in DMF/methanol 9:1 mixture containing 0.1 M HOBt is used for the deprotection. All deprotection cycles are conducted at room temperature to avoid racemization and aspartimide formation. Pseudoproline analogs used for dipeptides are double underlined in Fig. 2. 5. Filter the resin with peptide through disposable polypropylene column, wash with DCM, and dry under the fume hood. 6. Transfer dry resin into a 10-ml glass vial, add a stir bar, and chill to 80  C on dry ice. Chill 10 ml of deprotection mixture per 1 g of resin to 20  C. After mixing the resin and the mixture, run the cleavage with stirring for 30 min at 5  C and 1.5 h at room temperature. 7. Filter the mixture into a 50-ml centrifuge tube containing 30 ml of ether through a disposable filter column. TM peptides tend to stick to the resin, so wash the resin on filter twice with 0.2 ml of neat TFA. Top the centrifuge tube to 50 ml with

Synthetic Virus-like Particles

141

ether. Use at least tenfold excess of ether for precipitating out the peptide. 8. Separate the precipitate by centrifugation and wash four times with neat ether. After drying under the hood, dissolve 80–100 mg of crude peptide in 2–3 ml of DMSO and purify on a preparative (25 mm  250 mm) Atlantis C3 reverse phase column (Waters) in a 90-min gradient of 0.1% (v/v) trifluoroacetic acid in water and 0.1% trifluoroacetic acid in acetonitrile, with a 10 ml/min flow rate. Analyze the fractions containing peptides on Agilent 6100 LC/MS spectrometer with the use of a Poroshell 300SB-C3 column. Combine the fractions that are more than 95% pure. To remove trifluoroacetic acid, add acetic acid to peptide solution to a final concentration of 5% and freeze-dry. 3.3 Generation of Fluorescent TM Peptides

1. For generation of fluorescent peptides, start the synthesis from manual attachment of Fmoc-homocysteine to the resin (see Note 2). Homocysteine coupling is followed by manual coupling of Fmo-PEG27 and peptide assembly as described in Subheading 3.2. 2. After cleavage and deprotection, freeze-dry peptide at least twice from 5% acetic acid to get rid of TFA that can interfere with subsequent Michael reaction by acidification of solutions. 3. Dissolve 5 mg of fluorescein maleimide (11.5 μmol) in 200 ml DMF and add to 7 μmol of peptide (about 30 mg). Sonicate if peptide would not dissolve. Incubate on a shaker overnight. 4. Purify by HPLC as described in Subheading 3.2.

3.4 Peptide Self-Assembly and Generation of the Particles

1. Prepare 32 mg/ml stock solution of the peptide in DMSO or 4.6% octyl glucoside (see Note 3). Sonicate in a bath sonicator if needed to achieve a transparent solution. 2. Dilute stock solution into phosphate-buffered saline or other buffers with pH 7.0–7.5 to achieve 0.4–0.05 mg/ml final concentrations of the peptide. 3. Sonicate the samples for 10 min in a sonicator bath and leave at ambient temperature overnight (about 20 h) before measurements. 4. Centrifuge solutions for 10 min at maximal speed in a table-top centrifuge. 5. Remove supernatant carefully and use for analysis and other applications. We have found solutions containing antiseptic (NaN3) to be stable for at least 3 weeks at room temperature. Do not refrigerate to avoid aggregation/precipitation.

142

Sergey G. Tarasov et al.

3.5 Dynamic Light Scattering (DLS)

DLS is a method of choice for monitoring self-assembly due to its relatively low cost, speed, and ability to determine both the size and homogeneity of particles. It allows conducting experiments in a wide range of sample buffers, temperatures, and concentrations. DLS is also a noninvasive technique and requires comparatively low amounts of self-assembling material [13]. 1. Load about 5 μl of sample solution into the cell, making sure there are no bubbles. 2. Perform DLS measurements at 20  C. 3. Collect thirty 5-s acquisitions per each measurement. 4. Inspect your acquisitions in Datalog Grid table, Datalog Graph, and Autocorrelation Graph. Mark selected outlying data points in the Datalog Grid. 5. Display data in Regularization Graph (Fig. 3). Use the regularization algorithm to determine the hydrodynamic radii and size distribution of particles. Select the distribution as a percentage of Mass (%Mass) and Rayleigh Spheres for a Model (for particles with 16,000  g) (see Note 4). 3. Dispense 0.5–2 μL of the sample solution (protein sample mixed with the different detergent stocks) into a 96-well Hydrophobic Douglas Instruments Vapor Batch Plate (DI_vbatch) using a Douglas Instruments Oryx8 system (see Notes 5, 6 and 9). If possible, at least three drops per sample should be measured to assess reproducibility.

196

Tristan O. C. Kwan et al.

Fig. 4 DLS measurements of 1 mg/mL AcrB. The sample was centrifuged at 21,130  g for 10 min at 4  C prior to loading three (for reproducibility) drops of 2 μL onto a DI_vbatch plate covered in paraffin oil. Measurements were set to run for 25 scans (10 s each) with 10 repeated measurements per scan (total time ~22 h). The refractive index of water (1.33) and a viscosity value of 1.13 (based on the sample 5% (v/v) glycerol) were used in all calculations performed by the manufacturer’s software. The left panel shows the size distribution plot in the form of signal heat map (blue ¼ low particle concentration, red ¼ high particle concentration). On the right-hand side panel, DLS analysis is shown through the graphs of the autocorrelation function, radius distribution, and radial distribution plot (the blue spot diameter represents the relative scattered light intensity of the detected particles in arbitrary units)

4. The Oryx8 system automatically adds oil directly after dispensing the drop. However, it is recommended to add an extra 1–2 mL of paraffin oil on top (ensuring that drops are not disturbed) to have an even surface of oil (see Note 7). 5. Insert the plate into the DLS SpectroLight 610 system, defining the correct plate type when setting up the experiment. Through the user interface software, select the relevant plate wells for measurement. 6. Ensure that the drops are centered for detection using the Dropsearch function and that a good measurement position can be found using the Finesearch function. 7. Set the number of scans and the number of repeated measurements on the software for the DLS measurements. Figures 5 and 6 show an example where two different IMPs stability were assessed in the presence of different detergents by in situ DLS over time.

In Situ Measurements of Polypeptide Samples by Dynamic Light Scattering. . .

197

Fig. 5 Size distribution plots (signal heat maps: blue ¼ low particle concentration, red ¼ high particle concentration) for the pure detergent micelles (a) and for the protein in the presence of several detergents with the mean autocorrelation function and mean distribution plots (b). The protein (AcrB) was reconstituted in buffer B (10 mM Tris, pH 7.5, 300 mM NaCl, 5% (v/v) glycerol, and 1% detergent) to a final concentration of 1 mg/mL. Detergents stocks (1%) in buffer B were also prepared. Samples were centrifuged at 21,130  g for 10 min at 4  C prior to loading 2 μL of each sample onto a DI_vbatch plate with paraffin oil. DLS measurements were set to run at 20  C for 25 scans (10 s each) with 10 repeated measurements per scan (total time ~22 h). Distribution maps and plots were automatically generated by the available software and exported as images 3.5 Monitoring Nucleation/ Precipitation of Proteins During Crystallization

1. Switch the DLS instrument on and set up the chosen working temperature using the instrument’s temperature control unit. It takes around 15 min for the instrument sample holder to warm up (depending on the exterior temperature). 2. Centrifuge the purified protein sample for 10 min at 4  C in a bench-top centrifuge (>16,000  g) (see Note 4). 3. Dispense 0.5–1μL of the protein sample and 0.5–1 μL of the crystallization buffer (see Note 10) into a 96-well Hydrophobic Douglas Instruments Vapor Batch Plate (DI_vbatch) using a Douglas Instruments Oryx8 system (see Notes 5, 6 and 9).

198

Tristan O. C. Kwan et al.

Fig. 6 Size distribution plots (signal heat maps: blue ¼ low particle concentration, red ¼ high particle concentration) for the pure detergent micelles (a) and for the protein in the presence of several detergents with the mean autocorrelation function and mean distribution plots (b). The protein (LacY) was reconstituted in buffer D (20 mM Tris, pH 7.5, 150 mM NaCl, and 3 critical micelle concentration (CMC) of each detergent) to a final concentration of 1 mg/mL. Detergents stocks (1%) in buffer D were also prepared. Samples were centrifuged at 21,130  g for 10 min at 4  C prior to loading 2 μL of each sample onto a DI_vbatch plate with paraffin oil. DLS measurements were set to run at 20  C for 25 scans (10 s each) with 10 repeated measurements per scan (total time ~29 h). Distribution maps and plots were automatically generated by the available software and exported as images

4. The Oryx8 system automatically adds oil directly after dispensing the drop. However, it is recommended to add an extra 1–2 mL of paraffin oil on top (ensuring that drops are not disturbed) to have an even surface of oil (see Note 7). 5. Insert the plate into the DLS SpectroLight 610 system, defining the correct plate type when setting up the experiment. Through the user interface software, select the relevant plate wells for measurement.

In Situ Measurements of Polypeptide Samples by Dynamic Light Scattering. . .

199

Fig. 7 Microbatch crystallization of AcrB. The protein (1 mg/mL) was mixed with 96 different crystallization solutions at 1:1 ratio using the Oryx System (Douglas Instruments) into a 96-well Douglas Instruments Vapor Batch Plate and covered with paraffin oil. Crystallization was monitored by in situ DLS for all 96 wells (50 scans for 10 s with 10 repeated measurements—total time ~5 days). Crystals appeared in 0.1 M MES, pH 6.7, 24% PEG-400, and 0.3 M NaCl after 4 days

6. Ensure that the drops are centered for detection using the Dropsearch function and that a good measurement position can be found using the Finesearch function. 7. Set the number of scans and the number of repeated measurements on the software for the DLS measurements. 8. Plate wells can be imaged using the Image function with the system’s built-in microscope. Images should be acquired at relevant intervals for visual monitoring of crystal growth. Figure 7 shows an example where the precipitation/nucleation/crystallization process of an IMP is monitored by in situ DLS (see Note 11).

4

Notes 1. E. coli AcrB (Uniprot: P31224) is recombinantly expressed in C43 (DE3) cells, as reported in [11]. The purification scheme was adapted from [12]. Briefly, lyse cells in phosphate-buffered saline (PBS) buffer and collect membranes by ultracentrifugation. Solubilize the protein in PBS, 150 mM NaCl, and 1%

200

Tristan O. C. Kwan et al.

(w/v) n-dodecyl β-D-maltoside (DDM) for 1 h at 4  C and leave to bind to Ni-NTA resin. Wash the resin with buffer A (50 mM Tris, pH 7.5, 300 mM NaCl, 0.1% (w/v) DDM, 5% (v/v) glycerol) and 30 mM imidazole for 10 CV followed by buffer A and 50 mM imidazole for further 10 CV. Collect fractions with buffer A and 250 mM imidazole, supplement with an equimolar amount of Tobacco Etch Virus (TEV) protease and perform dialysis overnight against buffer B (10 mM Tris, pH 7.5, 300 mM NaCl, 0.03% (w/v) DDM, and 5% (v/v) glycerol). Proceed to gel filtration using a Superdex S200 Increase 10/300 GL (GE Healthcare Life Sciences) equilibrated in buffer B and analyze fractions by SDS-PAGE. Pool the desired fractions, concentrate the protein to ~28 mg/mL, and store at 80  C. 2. E. coli LacY (UniProt: P02920) is recombinantly expressed in C43 (DE3) cells. Grow cells at 37  C and induce with 0.4 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) upon reaching an OD600 of 0.5. Reduce the temperature to 25  C and harvest cells after overnight incubation. Lyse cells in PBS and collect membranes by ultracentrifugation. Solubilize the protein in PBS, 150 mM NaCl, and 1% (w/v) DDM for 1 h at 4  C and leave to bind to Ni-NTA resin. Wash the resin with buffer C (PBS, 150 mM NaCl, 0.1% (w/v) DDM) and 10 mM imidazole for 10 CV, followed by buffer C and 20 mM imidazole for 10 CV, then buffer C and 35 mM imidazole for further 10 CV. Collect fractions with buffer C and 250 mM imidazole, supplement with an equimolar amount of TEV protease, and dialyze overnight against buffer D (20 mM Tris pH 7.5, 150 mM NaCl, and 0.03% (w/v) DDM). Perform gel filtration using a Superdex S200 Increase 10/300 GL (GE Healthcare Life Sciences) equilibrated in buffer D. Analyze fractions by SDS-PAGE. Pool the desired fractions, concentrate the protein to ~15 mg/mL, and store at 80  C. 3. All the buffers should be prepared using HPLC-grade water or ultrapure water from a purification system. It is also important to filter the buffers through a 0.2-μm filter to remove any large impurities. 4. Protein samples should be filtered through a 0.2-μm filter or centrifuged at >16,000  g for 10 min prior to DLS measurements in order to remove precipitates or any other larger insoluble particles. 5. Automatic dispensing of the drops is advisable for accurate drop size and position. In addition, drops will be free from air bubbles, which often occurs during manual pipetting. Air bubbles interfere with the DLS measurements.

In Situ Measurements of Polypeptide Samples by Dynamic Light Scattering. . .

201

6. 72-well Terasaki microbatch plates (Molecular Dimensions) or any other SBS standard 96-well microbatch plates can also be used instead of the hydrophobic Douglas Instruments Vapor Batch Plates (DI_vbatch). 7. For in situ DLS experiments, sample drops should always be covered with inert oil (paraffin, silicone, or any other type of mineral oil). This prevents sample dehydration as the drop volumes are very small. Since these oils are inert and immiscible with water, they do not interact with the protein sample or with the crystallization solution. 8. The viscosity is influenced by protein concentration and components of the buffer such as alcohols and glycerol. For accurate measurements of Rh, the viscosity and refractive index can be measured with a viscometer and a refractive index detector, and then entered manually into the system software. 9. The plates should be very clean and dust-free to avoid errors during the measurements. Hence, they should be undusted with compressed air prior to its use. 10. In our laboratory, we prepare custom crystallization screens using the Microlab STAR Liquid Handling System (Hamilton) to ensure accuracy and reproducibility. 11. The user should be familiar with the difference between aggregation/precipitation and amorphous crystallization and how these different processes differ on the DLS size distribution heat maps (see Fig. 8). a

Tempearture (deg. C) 20 30 40

b 50

60

Tempearture (deg. C) 0

104760 104614 99328 94052 88776 83509 83365 78081 72808 67533 62259 62114 56853 51580 46300 41033 40888 35623 30350 25070 19806 19659 14388 9121 3865 0

10

20

30

40

50

60

Time (s)

10

Time (s)

106065 105910 100629 95352 90074 84800 84657 79385 74101 68828 63551 63404 58148 52879 47598 42329 42184 36921 31647 26365 21097 20950 15684 10413 5153 0

0

1nm

10nm

100nm 1um Radius

10um

0.1mm

1nm

10nm

100nm 1um Radius

10um

0.1mm

Fig. 8 Difference between protein aggregation/precipitation and amorphous crystallization given by size distribution heat maps (blue ¼ low particle concentration, red ¼ high particle concentration). In (a), a conventional profile for protein aggregation/precipitation over time is shown. Here, the hydrodynamic radius of the protein increases over time with the total loss of the monodisperse pure protein. (b) shows a conventional protein nucleation/crystallization profile over time. Nucleation (second distribution peak) is appearing over time without the aggregation of the protein (first distribution peak)

202

Tristan O. C. Kwan et al.

Acknowledgments We acknowledge funding from the United Kingdom’s Department of Business, Energy and Industrial Strategy (BEIS). References 1. Sanders CR, Myers JK (2004) Disease-related misassembly of membrane proteins. Annu Rev Biophys Biomol Struct 33:25–51 2. Reis R, Moraes I (2019) Structural biology and structure–function relationships of membrane proteins. Biochem Soc Trans 47:47–61 3. Kwan TO, Reis R, Siligardi G, Hussain R, Cheruvara H, Moraes I (2019) Selection of biophysical methods for characterisation of membrane proteins. Int J Mol 20:2605 4. Birch J, Axford D, Foadi J, Meyer A, Eckardt A, Thielmann Y, Moraes I (2018) The fine art of integral membrane protein crystallisation. Methods 147:150–162 5. Meyer A, Dierks K, Hussein R, Brillet K, Brognaro H, Betzel C (2015) Systematic analysis of protein–detergent complexes applying dynamic light scattering to optimize solutions for crystallization trials. Acta Crystallogr F Struct Biol Cryst Commun 71:75–81 6. Stetefeld J, McKenna SA, Patel TR (2016) Dynamic light scattering: a practical guide and applications in biomedical sciences. Biophys Rev 8:409–427 7. Lorber B, Fischer F, Bailly M, Roy H, Kern D (2012) Protein analysis by dynamic light

scattering: methods and techniques for students. Biochem Mol Biol Educ 40:372–382 8. Jachimska B, Wasilewska M, Adamczyk Z (2008) Characterization of globular protein solutions by dynamic light scattering, electrophoretic mobility, and viscosity measurements. Langmuir 24:6866–6872 9. Murphy RM (1997) Static and dynamic light scattering of biological macromolecules: what can we learn? Curr Opin Biotecnhol 8:25–30 10. Aivaliotis M, Samolis P, Neofotistou E, Remigy H, Rizos AK, Tsiotis G (2003) Molecular size determination of a membrane protein in surfactants by light scattering. Biochim Biophys Acta Biomembr 1615:69–76 11. Pos KM, Diederichs K (2002) Purification, crystallization and preliminary diffraction studies of AcrB, an inner-membrane multi-drug efflux protein. Acta Crystallogr D Biol Crystallogr 58:1865–1867 12. Veesler D, Blangy S, Cambillau C, Sciara G (2008) There is a baby in the bath water: AcrB contamination is a major problem in membrane-protein crystallization. Acta Crystallogr F Biol Crystallogr Commun 64:880–885

Chapter 14 Measurement of Peptide Coating Thickness and Chemical Composition Using XPS David J. H. Cant, Alexander G. Shard, and Caterina Minelli Abstract X-ray photoelectron spectroscopy is a highly surface-sensitive analytical technique, capable of providing quantitative information on the chemical composition of materials within the top
10 nm of their surface. For samples consisting of distinct underlayer and overlayer materials, the thickness of the coating can also be determined if it falls within this
10 nm information depth, which is often the case for peptide layers. Such measurements are simple to perform for flat samples and can also be performed on nanoparticulate samples provided that either the core radius or total particle radius are known. Here, we describe a straightforward protocol for obtaining such measurements from peptide coatings on both flat surfaces and nanoparticles, including preparation of nanoparticle samples from suspension, data acquisition, and analysis. Key words XPS, Nanoparticle, Peptide, Coating, Thickness, Composition, X-ray, Photoelectron, Spectroscopy

1

Introduction X-ray photoelectron spectroscopy (XPS) is a powerful technique for quantitatively measuring the composition of surfaces. By utilizing the photoelectric effect with an X-ray beam of known wavelength, electrons from atomic core-levels within a sample may be emitted, detected, and their kinetic energies measured. It is then possible to determine their binding energies and therefore the elemental atoms they originate from and, in many cases, their oxidation state. Electrons that are not scattered, or only scattered elastically from the material, (i.e., those that do not lose energy through interaction with other atoms) contribute to the peak features of the XPS spectra, while those that are scattered inelastically contribute to the spectral background. Due to the attenuation of the number of electrons traveling through a material, XPS is highly surface sensitive, and with appropriate interpretation, it can also be used to measure the thicknesses of coatings, which are less than 10 nm thick on surfaces and particles. Measurement of coatings on flat

Maxim G. Ryadnov (ed.), Polypeptide Materials: Methods and Protocols, Methods in Molecular Biology, vol. 2208, https://doi.org/10.1007/978-1-0716-0928-6_14, © Springer Science+Business Media, LLC, part of Springer Nature 2021

203

204

David J. H. Cant et al.

surfaces using XPS has been known for some time [1–4], while measurement of nanoparticle coatings is an area of significant recent interest and development [5–10]. Here, we provide protocols for the acquisition of data from peptide-coated samples and for analyzing such data in order to obtain measurements of the chemical composition and thickness of the peptide coating. For nanoparticle samples, a protocol for drop-casting samples from solution is provided. Typical samples that this protocol applies to are selfassembled peptide mono- and multilayer(s) and peptide functional coatings on nanoparticles.

2 2.1

Materials Flat Surfaces

1. A clean sample of a peptide layer bound to an underlayer material (see Note 1). 2. A sample of the underlayer material undergoing the same preparation steps that are used for the peptide layer attachment, without any peptide layer. 3. A sample of the peptide or a similar organic material of thickness much greater than the XPS information depth; greater than 100 nm is appropriate.

2.2 Nanoparticle Samples

1. A sample of peptide-coated nanoparticles in suspension. 2. A bulk, uniform sample of the nanoparticle core material, i.e., of thickness much greater than the XPS information depth; greater than 100 nm is appropriate. 3. A sample of the peptide or a similar organic material of thickness much greater than the XPS information depth; greater than 100 nm is appropriate. 4. A flat, vacuum-compatible substrate suitable for the sample (see Notes 1 and 2). 5. A vacuum desiccator (see Note 3). 6. A vacuum pump for use with the selected vacuum desiccator (see Note 3). 7. Pipettes for handling liquids. 8. A laboratory centrifuge (see Note 4). 9. Appropriate centrifuge tubes. 10. Ultrapure water. 11. Isopropyl alcohol (IPA). 12. A sonicator bath. 13. Laboratory tissue (see Note 5). 14. A basic optical microscope (optional).

Measurement of Peptide Thickness and Composition Using XPS

2.3 XPS Measurements and Data Analysis

3

205

1. An intensity-calibrated XPS instrument (see Notes 6 and 19). 2. Appropriate XPS data analysis software (see Note 7).

Methods

3.1 Drop-Casting of Nanoparticle Samples from Suspension

1. Decant the nanoparticle suspension into centrifuge tubes. 2. Centrifuge the suspension to separate the nanoparticles of interest from any unbound peptide, stabilizing ligands, and other material that may be present in dispersion (see Note 4). Once the particles are pelleted, carefully remove the supernatant and redisperse the concentrated nanoparticle suspension in ultrapure water. Perform this washing cycle at least three times to ensure all material not bound to the nanoparticles is removed. Care must be taken at this stage, as the nanoparticle suspension is likely to precipitate. Minor agglomeration/ aggregation of nanoparticles will not typically affect an XPS measurement, but significant precipitation may prevent the subsequent steps from being carried out effectively. 3. Clean the substrate onto which the nanoparticle sample is to be deposited prior to the preparation process. Initially, remove large contaminant particles by wiping with laboratory or clean-room-grade tissue (see Note 5) lightly damped with organic solvents such as IPA. Then, thoroughly clean any remaining adhered contaminants from the substrate, for example, by sonication for 30 min in IPA, followed by sonication in ultrapure water for 30 min and then by a final phase of sonication for 30 min in IPA. Either allow substrates to air-dry or (to minimize exposure to dust and other contaminants) dry them under vacuum in a vacuum desiccator, making sure not to use vacuum grease, as this may easily contaminate the sample. 4. Place the substrate(s) in suitable containers within a vacuum desiccator (see Note 8) and carefully pipette about 3 μL (see Note 9) of the washed nanoparticle suspension onto the center of the substrate (Fig. 1). 5. Close the lid of the desiccator, making sure not to use vacuum grease, as this may easily contaminate the sample. 6. Attach the vacuum pump with the desiccator valve closed and then turn the pump on. Gradually open the desiccator valve to pump out the air. Once a vacuum has been achieved, seal the lid such that it cannot be easily removed from the desiccator. 7. Wait for the droplet(s) of suspension to dry within the desiccator—depending on the amount of suspension, the degree of

206

David J. H. Cant et al.

Fig. 1 Flow diagram showing the procedure for drop-casting a suspension of nanoparticles onto a substrate. This procedure should be repeated until a thick, even deposit has been obtained

droplet spreading, and the type of material suspended; this may take 10–20 min (see Note 10). 8. Once the droplet(s) have dried, close the desiccator valve, turn off the vacuum pump, and disconnect it from the desiccator. 9. Gradually open the desiccator valve to allow air to return to the desiccator. Opening the valve rapidly will cause a rush of air, which is likely to disturb, and possibly flip, the substrates. If the surface of the substrate ever contacts another surface, this may lead to contamination, which may affect the quality of any measurements. 10. Observe the deposited nanoparticles on the substrate. Ideally, this deposit should cover an area of the substrate that is comfortably larger than the area illuminated by the XPS X-ray beam, and have a thickness greater than 10 nm. For most sample types, a visual inspection can give a good indication of the degree of coverage and thickness of the deposit; the sample should appear as an opaque, uniform, circular mark (see Fig. 2 for an example of a deposit of gold-core nanoparticles). Repeat steps 4–10 until this is the case, placing droplets centrally upon the existing deposit. Take care to watch for issues such as “coffee-ring” drying, mismatching of droplet position during repeated steps, and sample detachment from the substrate (see Notes 11–13). 11. A final check for the uniformity and thickness of the deposited sample may be performed using an optical microscope to check for gaps in the nanoparticle film; however, this should rarely be necessary.

Measurement of Peptide Thickness and Composition Using XPS

207

Fig. 2 Example of a drop-cast deposit of gold nanoparticles on a section of silicon wafer 3.2 XPS Measurements

1. Samples for XPS measurement should consist of either a flat surface bearing a clean peptide coating (see Note 14) or a dry, uniform deposit of peptide-coated nanoparticles on a substrate, which may be prepared by the method described in Subheading 3.1. Once prepared, at no point should the surface of these samples have come into contact with any other surface, to avoid contamination. Handle the samples using clean stainless-steel (or similar noncontaminating material) tweezers, taking care to handle the edges only, and avoiding touching the areas to be analyzed. 2. In order to facilitate accurate analysis, during the same experiment and using the same instrument settings, analyze samples of pure, flat, “bulk” (e.g., greater than 100 nm in thickness) reference material. This should include a sample of the clean underlayer material (for flat samples) or a bulk sample of the nanoparticle core material (for nanoparticulate samples), as well as a thick flat sample of the coating peptide (e.g., spincoated from solution) or a similar organic material [11, 12]. 3. Mount the samples on the appropriate sample holder for the instrument, taking care to ensure they are securely held, with a suitable region of the sample exposed and available for analysis (see Note 15). Once the samples are mounted, place the sample holder into the sample-entry system for the XPS instrument. It is important that there are no loose pieces of material or samples present, as these may be shaken off during movement and will at best be lost within the instrument, or at worst cause damage to vacuum pumps or the instrument itself.

208

David J. H. Cant et al.

4. Once the samples have been safely entered into the system, move them away from the trajectory of the X-ray beam, then turn on the X-ray source and allow it to equilibrate for a minimum of 20 min (see Note 16). When not actively acquiring data, position samples such that they are not exposed to the X-ray beam, as prolonged exposure may result in damage to the sample. 5. Once the X-ray source has equilibrated, select the areas of the sample to be analyzed. The procedure for doing this will vary with the instrument used and will normally be a standard laboratory procedure. Make sure to select multiple nonoverlapping analysis areas in order to obtain data that are representative of the variation across the sample—3 areas will typically suffice. Align the positions of the areas to be analyzed with the intersection between the X-ray source and analyzer axis and optimize to be within the focus of any electron focusing optics; i.e., the distance between the analyzer and the sample should be optimized to detect the maximum counts from the sample (see Note 17). 6. For some samples, particularly those deposited on nonconducting substrates, some distortion of the spectra may be observed due to the accumulation of charge under X-ray irradiation. Many instruments possess mechanisms by which this may be mitigated, such as low-energy electron flood guns. It is useful to optimize the parameters of any such charge neutralizer prior to collecting data; this may be performed by taking a series of spectra (typically a high-resolution spectrum of the carbon 1 s peak at approximately 285 eV binding energy is suited for this purpose) while varying the parameters of the charge neutralizer. The ideal parameters can then be identified by discerning the “best” spectrum with the sharpest peaks. 7. Acquire a low-resolution survey spectrum from each analysis position (Fig. 3). This should typically be performed with settings as follows: analyzer pass energy in the range 100–200 eV; energy scale step-size of 0.5–1 eV; an acquisition time per step of 500–1000 ms (see Note 18); and a binding energy range from approx. 1200 to 10 eV binding energy (for an aluminum Kα X-ray source with photon energy 1486.6 eV, this corresponds to a kinetic energy of 286.6–1496.6 eV). Prior to acquiring data, it is important to ensure that an intensity scale calibration has been performed at the pass energy and analyzer settings selected (see Note 19). 8. Identify the peaks observed within the spectra. These should, generally, be known in advance based on the expected chemistry of the sample. For any peaks that do not belong to an element expected to be present in the sample, take care to

Measurement of Peptide Thickness and Composition Using XPS

209

Fig. 3 (a) A survey spectrum from a flat gold surface coated with a self-assembled peptide layer and (b) a survey spectrum taken from a sample of gold nanoparticles of radius 30 nm coated with the same peptide layer. Peaks used for quantification are labeled. The peptide sequence is CGGGNPSSLFRYLPSD

correctly identify their origin as these are often indicative of contaminants, which will invalidate the analysis (see Note 20). 9. Acquire high-resolution spectra for each element present within the sample at every analysis position (Fig. 4). Acquire spectra from the most intense peak available for each element, excluding peaks which overlap with others from another element. This should typically be performed with settings as follows: analyzer pass energy in the range 10–50 eV; energy scale step-size of 0.1 eV; an acquisition time per step of at least 500 ms, with considerably more for less intense peaks (see Note 18); and an energy scale range which covers the entire peak width plus a reasonable area of background on either side—for most peaks, approximately 30 eV will be a suitable range; however, larger peaks or doublets with large separation may require a considerably broader energy range. 10. Once high-resolution spectra have been acquired, repeat step 7 so that any change in the sample during the experiment can be noted and, if necessary, accounted for. 11. Review all spectra to ensure they were acquired successfully and with no issues (e.g., evidence of charging effects, evidence of significant sample damage, and lack of sufficient width in the energy scale). Significant issues should be resolved by

210

David J. H. Cant et al.

Fig. 4 High-resolution spectra for elements present in a peptide-coated gold sample, highlighting common background choices—(a) Au 4f peak shown with a Shirley background, (b) the same Au 4f peak with a Tougaard background, (c) a carbon 1s with a Tougaard background, and the main contributing chemical species labeled, (d) an oxygen 1s spectrum with a Tougaard background, and (e) a nitrogen 1s peak with a linear background

reacquisition of the spectra; in the case of sample damage, this may require setting up a new position (step 5) and repeating the measurements with significantly reduced X-ray exposure (e.g., by reducing the total acquisition time or X-ray intensity, which may then be compensated for by using a higher pass energy, where possible). 12. Once satisfactory data have been acquired, turn off the X-ray source, analyzer, and any charge compensation source and follow the appropriate procedure for removal of the samples from the instrument. 3.3 Analysis of XPS Data from Peptide-Coated Samples

1. Export spectra acquired from the samples and any reference materials from the instrument into a format readable by the analysis software (see Note 21). 2. Export the appropriate intensity calibration data for the analyzer modes used; it is good practice to always store and transfer the relevant intensity calibration data alongside each batch of experimental data.

Measurement of Peptide Thickness and Composition Using XPS

211

3. Open the data in the analysis software and apply the appropriate intensity calibration to each scan. The method for this will vary depending on the software but should be relatively simple on most. If this cannot easily be performed within the software available, an alternative method is to apply the correction manually (see Note 22). 4. Identify the elements present by manually analyzing the survey spectra or by means of the analysis software. Verify that the detected elements are those expected, i.e., carbon, nitrogen, oxygen, and any element specific to the amino acids forming the peptide layer. Depending on the elemental concentration, depth distribution, photoionization cross section, and electron attenuation length, some expected elements may not be detected. Typical detection limits for XPS fall in the range 0.1–1 at% (atomic%, i.e., relative proportion by number of atoms of a specific element; note that hydrogen atoms are not counted) of a homogeneous-equivalent material [13]; this means that some elements which may be present in low concentrations within a peptide are likely to be difficult to detect; a typical example would be sulfur within the cysteine group used to bind to gold, which will have a detection limit of around 1%. Elements that are not detected due to low concentration may usually be neglected in the analysis without a significant impact on the result. Electrons from the substrate may or may not be detected, depending on the thickness of the sample. Investigate the presence of any unexpected element in the XPS spectra, as this may indicate the presence of impurities in the samples or contaminations occurring during sample preparation or mounting (see Note 20). 5. For each identified element, select a photoelectron peak to use for quantification. This should typically be the most intense sharp peak belonging to a given element, which does not overlap with peaks of other elements. For cases in which both the substrate and peptide contain the same element in different chemical states, the acquired high-resolution spectra must be used (see Note 23). 6. In order to measure the intensity of each peak, the background signal must be excluded from the quantification. For most analysis software, this will initially consist of selecting a region around each peak, within which a model for the subtraction of the inelastic electron background is applied. The most commonly used background models are the Shirley, Tougaard, and Linear forms; examples of these are shown in Fig. 4a, b, e. As an example system, the survey spectra shown in Fig. 3 are from gold (nanoparticles and a flat surface) coated with a selfassembled peptide layer; in this case, we recommend using the Tougaard model for the gold 4f, carbon 1s, and oxygen

212

David J. H. Cant et al.

1s peaks. For the nitrogen 1s peak, a linear model is used due to its relatively low intensity and its position lying on the slope of the inelastic background from the gold 4d peaks. A full discussion of the advantages and disadvantages of each background model is beyond the scope of this protocol; however, some basic points are discussed in Note 24. 7. Once a background for each peak has been defined, obtain the resulting intensity-calibrated areas from the software. Then, by dividing by the appropriate sensitivity factors (see Note 25), correct the peak areas to account for differences in photoionisation cross section and electron attenuation. The NPL Average Matrix Sensitivity Factors (AMRSFs) are suitable for use with intensity-scale-calibrated instruments using aluminum Kα or magnesium Kα X-rays and are available on the NPL website [14, 15]. 8. Using the corrected areas, calculate the “homogeneous-equivalent atomic concentrations” (commonly reported as “atomic %”) by dividing each corrected peak area by the sum of all corrected peak areas (note that only one peak per element should be used for these calculations). See Table 2 and Note 28 for measured concentrations and their use in calculations. Figure 5 shows a comparison between the expected and calculated compositions for this example system. 9. In order to measure the peptide coating thickness for both flat and nanoparticulate samples, the first step is to calculate the intensity ratio between one element in the peptide and one in the underlayer, normalized by the intensities obtained from samples of the pure material (see Note 26). This is described by Eq. (1) (adapted from the supporting information of Ray et al. [16]). A p,u ¼

Ip ½pχ u Iu 1 1 ¼ f Ip Iu ½uχ p

ð1Þ

Ap, u is the normalized intensity ratio between a peak arising from an element, p, in the peptide, and a peak arising from an element, u, in the underlayer. Ip is the intensity of a peak arising from an element, p, in the peptide. Iu is the intensity of a peak arising from an element, u, in the underlayer. I1 p is the intensity of a peak arising from an element, p, in a pure, homogeneous sample of the peptide. I1 u is the intensity of a peak arising from an element, u, in a pure, homogeneous sample of the underlayer. χ p is the atomic fraction of p in the pure peptide. χ u is the atomic fraction of u in the pure underlayer.

Measurement of Peptide Thickness and Composition Using XPS

213

Fig. 5 Bar chart showing the expected elemental composition of the peptide layer from the peptide structure, compared to the final measured compositions for the flat and nanoparticle samples. The increased carbon and corresponding decreases in other elements is likely due to a combination of the orientation of the peptide and distribution of elements within it, as well as a small contribution from adventitious carbon. In the case of the nanoparticle samples, the carbon may also originate from the nanoparticle coating prior to functionalization with the peptide layer

[p] is the measured “atomic%” of p (arising from the peptide) from the coated sample. [u] is the measured “atomic%” of u (arising from the underlayer) from the coated sample. f is a factor which comprises the various differences in signal between the underlayer material and peptide due to aspects such as density and intrinsic loss processes [16] (see Note 26). 10. In order to measure the coating thickness from this intensity ratio, the other required pieces of information are the effective attenuation lengths (EAL) through the underlayer and coating materials of electrons contributing to the chosen peaks. These may either be estimated [17] or looked up in an appropriate database [18]. Here, we use the notation Li, J to refer to an EAL for an electron arising from material i and travelling through material J, as used by Shard [9]. In this instance, p and u refer to electrons from the peptide and the underlayer respectively, while P and U refer to the peptide and underlayer

214

David J. H. Cant et al.

Table 1 Effective attenuation lengths in nm for typical elements within a peptide (carbon, oxygen, and nitrogen) and two example underlayers (gold and silicon) calculated using the method by Seah [17] Effective attenuation lengths C 1s (nm) electrons

O 1s electrons

N 1s electrons

Au 4f electrons

Si 2p electrons

Through an organic

3.32

2.74

3.05

3.79

3.76

Through gold

1.17

0.99

1.09

1.32

–

Through silicon

2.07

1.71

1.90

–

2.34

materials themselves. Typical EAL values are given in Table 1— for example, Lu, P for a system with a gold underlayer using the gold 4f peak would be 3.79 nm, while Lp, P would be different for each element in the peptide, having values for carbon, oxygen, and nitrogen of 3.32, 2.74, and 3.05 nm respectively. These attenuation lengths are used in the example calculation given in Note 28. 11. Once effective attenuation lengths for electrons from every relevant peak traveling through both underlayer and coating materials have been obtained, use Eq. (2) (to obtain tflat for flat samples) or Eqs. 3–6 (to obtain t NP for nanoparticle samples) to estimate the coating thickness. For the nanoparticle equations, r denotes the radius of the nanoparticle core. All equations are adapted from Shard [9]. !
 0:05 0:42 0:58 A p,u 2:2 ln A p,u L 0:95 u,P L p,P þ 2A p,u L u,P L p,P ð2Þ t flat ¼ L p,P A p,u 2:2 þ 1:9
 t 1 L p,P þ γ 3:5 þ 0:565γ 2:5 rt 0 t NP ¼ ð3Þ ð1 þ 1:8γ ÞL p,P þ ðγ 3:5 þ 0:565γ 2:5 Þr 2 3 !1=3 A p,u L 2p,P ð4Þ þ1  15 t 0 ¼ r4 L u,P L p,U t1 ¼


 0:9 0:1 0:41 0:59 0:74A 3:6 p,u ln A p,u L u,P L p,P þ 4:2A p,u L u,P L p,P A 3:6 p,u þ 8:9 γ¼

0:4 0:1 L 0:5 u,P L p,U L p,P

A 0:1 p,u r

ð5Þ

ð6Þ

12. Perform this calculation for each element pair within the underlayer and coating to obtain multiple estimates for the coating thickness. Significant variation in these estimates may indicate an excess of one or more elements arising from

734

42.6

N 1s

Au 4f

92.6

135

O 1s

C 1s

Nanoparticle sample

1700

46.3

N 1s

Au 4f

89.5

114

O 1s

C 1s

Flat surface

Element Peak intensity and peak (calibrated)

16.1

10.8

16.0

57.1

33.1

10.5

13.7

42.7

–

0.168

0.210

0.622

–

0.168

0.210

0.622

–

2.24

2.64

3.19

–

1.88

1.97

2.07

–

2.3

2.5

3.0

–

2.4

2.3

2.7

3.03 2.95 –

0.128 –

–

–

0.183

1.16

0.153

2.88

1.24

0.186

0.689

1.09

0.661

–

2.8

2.8

2.8

–

2.6

2.6

2.6

Expected peptide composition Measured peptide Ap, u, 0 Measured Measured atomic χ p, 0 (initial Initial thickness layer composition χ p Ap, u thickness concentration (%) (initial estimate) estimate) estimate (nm) (iterated) (iterated) (nm)

Table 2 Measured chemical composition and layer thickness for self-assembled peptide monolayers on a flat gold surface and on gold nanoparticles. Relevant XPS spectra are shown in Figs. 3 and 4 Measurement of Peptide Thickness and Composition Using XPS 215

216

David J. H. Cant et al.

contamination—otherwise, in the absence of any composition gradients or peculiarities, the average across all element pairs represents the best estimate of the coating thickness (see Note 27). The variation may be reduced by iteratively recalculating thickness using different values for the pure material compositions (χ k and χ u from Eq. 1) in order to obtain more representative estimates for both the coating thickness and the effective composition of the material (e.g., accounting for effects such as orientation and ordering). A comparison of final calculated compositions compared to the expected composition from the peptide structure is given in Note 28 and shown in Figure 5. The uncertainty in such measurements is typically expected to be dominated by the 10% uncertainty in the estimation of the EAL. 13. The measurement of elemental composition and thickness of peptide coatings by XPS can also be performed by using the high-resolution spectra instead of the survey spectra. Representative high-resolution spectra are shown in Fig. 4. The main advantage is that the measurement of the areas under the peaks is more accurate. However, the acquisition of high-resolution spectra increases significantly the experiment time. Highresolution spectra are also useful to quantitatively evaluate the chemical state of each element in the peptide, which is useful to verify, for example, the peptide structure or the occurrence of specific chemical reactions (see Fig. 4 and Note 23).

4

Notes 1. For clarity, we use the terms “substrate” and “underlayer” throughout this protocol with specific meanings: “underlayer” refers to the material to which a peptide is bound and from which an XPS signal must be measured—for nanoparticle samples, this is the nanoparticle core; “substrate” is used only to refer to the material upon which a sample of nanoparticles is deposited. In the ideal scenario, no signal from the substrate should be observed in the XPS measurements. 2. A selected substrate should ideally not contain any elements likely to also occur in the sample so that the presence of peaks from the substrate within the spectra do not interfere with quantification of the signal from the sample. Likewise, the substrate should not contain elements whose peaks overlap those from the sample in a manner that prevents quantification. Typical substrates for peptide layers are silicon wafer and gold. Polytetrafluoroethylene (PTFE) may also be used; however, this will necessitate peak fitting a high-resolution carbon 1s spectrum to ensure that no carbon intensity from the substrate

Measurement of Peptide Thickness and Composition Using XPS

217

contributes to the values used for thickness calculations. Substrate size and shape should be considered in relation to the sample mounting system used with the intended XPS instrument. In many cases, this will likely necessitate cutting of the substrate to size. Typically, 1 cm  1 cm substrates are suitable for most instruments and are often suitable for a number of other analytical techniques. 3. A vacuum desiccator and associated pump are not a compulsory requirement, as sample drying may also be performed in air. However, air-drying during the sample preparation process may take an excessive amount of time if several drying cycles are to be performed, and may also lead to a greater degree of adventitious carbon contamination from the significantly increased exposure to air [8]. Drying of any materials used as part of preparation for a surface analysis experiment should never be performed using a compressed air line, as these may contain contaminants that could end up on the sample. Inert gas lines such as nitrogen or argon from a cylinder are typically fine. 4. Centrifugation is one method available for purification of nanoparticle suspensions and is typically one of the most efficient methods [19]. If centrifugation is not available, dialysis may be performed instead. 5. Clean-room-grade tissue is optimal; however, any standard lint-free laboratory tissue is appropriate. 6. Modern XPS instruments are available with a number of different anode materials—the primary difference between these is in the energy of the photons produced. The most common anode type used is aluminum, monochromated to the Kα emission line. This is also typically the most suitable for analysis of peptide coatings in terms of depth sensitivity and intensity. Several XPS instruments also now provide the ability to micro-focus the X-ray beam, resulting in greater intensity from a smaller analysis area. However, as a high intensity of X-rays (and the resultant emitted electrons) can very easily and rapidly damage organic materials, this is not recommended for use in the analysis of organic coatings such as peptides unless the X-ray intensity is significantly reduced. 7. A variety of XPS analysis software is available [20–22], and many instruments are provided with a license for the manufacturer’s own analysis application. For the general analyst, CasaXPS [20] is a user-friendly option, with an extremely large amount of documentation and support available online. 8. Suitable containers will vary depending on the substrate, but for most, suitably sized semiconductor wafer carrier trays are ideal: these typically possess a curved surface on the “face-up”

218

David J. H. Cant et al.

side, such that when jostled any flat sample within will likely only contact the container with its edges, minimizing contamination from the carrier itself. It is still advisable to clean (e.g., by wiping with IPA and drying) the interior of any vessel before it is used to carry samples for surface analysis. It is strongly recommended that containers using adhesives to secure the sample (such as gel-pack boxes) are likewise avoided, as these are highly likely to cause contamination of the sample. 9. Depositing 3 μL of nanoparticle suspension at a time is suitable for most samples but may take a considerable amount of time to form a good deposit, particularly for samples with low concentration. Likewise, for highly hydrophobic substrates, the resulting deposit may be too small. Larger volumes up to around 20 μL may be used; however, these will increase the area the droplet covers and may also increase the likelihood of “coffee-ring” drying effects. To avoid this, it is always preferable to use the highest concentration of nanoparticles that will remain stable in suspension, along with the lowest practicable droplet size. 10. If the droplet drying takes significantly longer than 20 min without a clear reason (such as a large droplet, or a sample that is suspended in a solvent less volatile than water), this may indicate a leak or improper seal on the desiccator lid. In this case, the desiccator valve should be closed, the pump switched off and disconnected from the desiccator, and the desiccator gradually allowed to come back up to atmospheric pressure, before starting the drying process again. If the drying process is consistently taking a significantly long time, it is advisable to empty the desiccator and then perform a leak check on both the desiccator and the tubing between it and the pump. 11. The coffee-ring effect during drying is a common issue which affects drop-casting of nanoparticle suspensions. There are several possible ways this may be mitigated. Using a larger number of smaller droplets may help minimize the “uncovered” area at the center of the ring. Similarly, depositing droplets in gradually reducing volumes each cycle may result in a series of overlapping, concentric “coffee-rings” and allow an almost even deposit to be achieved. There has been some discussion [23, 24] of drying in the proximity of ethanol vapor in order to prevent coffee-ring effects—this can be performed by introducing a broad, open container of ethanol into the vacuum desiccator; however, some trial-and-error may be required to determine the ideal conditions for a given setup, as this may also cause the droplets to spread and cover a significantly larger area of the substrate.

Measurement of Peptide Thickness and Composition Using XPS

219

12. Ideally, sequential deposits should be in the same location as each other. However, for some samples, this may be difficult to achieve. This often occurs with a sample that is of higher hydrophobicity than the substrate onto which it is being deposited. This may only easily be mitigated by selecting a more hydrophobic substrate, or (in less severe cases) taking greater care when placing the droplets. 13. Detachment, or “flaking” of the nanoparticle deposit should rarely happen, and likely indicates that too much sample has been deposited. If this occurs before a satisfactory deposit has been achieved, it likely indicates uneven film formation due to one of the issues discussed in Notes 11 and 12. 14. Samples of peptide coatings on flat, macroscopic surfaces (as opposed to nanoparticles) should consist of a substrate of appropriate size for the sample holder of the XPS instrument, coated with a uniform, complete (i.e., without holes) coating of the desired peptide covering an area significantly greater than the analysis area of the XPS instrument. Such samples should be thoroughly rinsed to remove excess unbound molecules and other contaminants from the surface, as these will falsely contribute to the signal observed. 15. Typical sample mounting methods for XPS include the use of metal clips—which should be positioned such that they do not obscure the area to be analysed—or often conductive, vacuumcompatible, adhesive tapes—which should be used carefully and sparingly—to minimize sample contamination, impact on vacuum quality, and difficulty of removal postanalysis. 16. Some instruments may require a longer equilibration time. This can be tested by taking repeated survey spectra under identical conditions from a reference sample unlikely to be damaged by X-ray exposure (e.g. gold foil) until no variation between several sequential spectra is observed. 17. For flat samples, thickness calculations are simplest when the sample is oriented such that the surface normal is parallel to the analyzer axis; i.e., electrons traveling from the sample to the analyzer are emitted perpendicular to the surface, through the minimum amount of coating material. For nanoparticle samples however, this will not have a significant impact unless the angle between the surface normal and analyzer axis is greater than about
50 [25]. 18. The total acquisition time required for a single spectrum may depend on a number of instrument-specific factors such as X-ray source intensity, detector efficiency, and analyzer pass energy, which can all have an impact on the final count-rate. As such, the parameters given here should be considered a rule of thumb and may be freely adjusted in order to achieve a

220

David J. H. Cant et al.

reasonable signal to noise within the obtained spectra. The spectra given in Fig. 3 may be seen as an example of good signal to noise. Data with a signal to noise of even 2–4 times worse than these may still provide reasonable thickness measurements. For high-resolution spectra, the acquisition time required may vary drastically depending on the intensity of the peak—while for many instruments an acquisition time per step of around 500 ms may produce a good signal to noise for the most intense peak in a spectrum, a peak half the intensity will require an acquisition time four times longer to achieve the same signal to noise ratio. For all spectra, acquisition times may often be manipulated by changing the dwell time per step or by performing multiple scans of the same region and summing the spectra. On some systems, the latter method will include the ability to view the separate scans individually; this is recommended for samples that are particularly likely to become damaged by X-ray exposure so that any changes in the spectrum can be noted. 19. Any quantitative measurements to be performed using XPS require that the instrument has undergone intensity scale calibration for the specific set of analyzer parameters being used. Primarily, this refers to the pass energy, the settings of any apertures used within the analyzer, and the settings of any electron optics involved in the collection of photoelectrons emitted from the sample. Several methods exist for performing intensity calibrations [26, 27], which should typically be carried out by the individual responsible for the day-to-day operation of the instrument. Such calibrations should ideally be performed on at least a yearly basis. Should a calibration not be available for a desired analyzer mode, it is reasonable to acquire data with the aim of obtaining calibration data shortly afterward. Ideally, this should be done as soon as possible, as any significant change to analyzer characteristics, e.g., due to a fault, repair, or otherwise in the meantime, may significantly impact the validity of the calibration. 20. Common contaminants include sodium, from citrate stabilizers or contamination from improper handling; silicon, from PDMS grease; chlorine, from improper handling (with sodium); and adventitious carbon, although this is often indistinguishable from carbon within the sample. Identification of peaks can be performed by reference to a number of resources [28, 29], and often an appropriate tool will be included in the software used to run the XPS instrument. 21. For analysis software associated with the instrument used, file type and format should not be an issue. For other XPS analysis software, a common format used is the “VAMAS” format (*. vms), which is a standard format designed to be able to contain

Measurement of Peptide Thickness and Composition Using XPS

221

all the information that may need to be logged for XPS experiments. If this format is unavailable, either to export from the instrument or to read into the analysis software, basic text formats (e.g., comma-separated values, *.csv or *.txt) may be used, but care should be taken to understand exactly how such data are interpreted by the software. 22. In order to perform an intensity calibration directly to the raw data of a spectrum, both the intensity calibration data and experimental data must first be opened in suitable spreadsheet manipulation software. Then, the intensity of the experimental data spectrum at each energy must be divided by the intensity calibration data for the same energy—for various reasons, the energy steps for these two data sets may not always match up directly, in which case some basic interpolation of the intensity calibration data may be necessary. 23. As long as the difference in the chemical state can be resolved in the high-resolution spectra, the signal from both the underlayer and peptide for a shared element can be quantified separately by the use of peak fitting. Typical binding energy shifts for specific chemical states within a given XPS peak can be found from a number of sources [28–30], and there are several guides available to guide the user through the process of performing peak fitting, particularly if using CasaXPS analysis software [20]. 24. Background choice is not a trivial subject within XPS analysis, and the background selected for a peak can be one of the largest contributing factors to the uncertainty in the quantification of the peak intensity. Linear backgrounds, the simplest form of background used, are rarely appropriate. On some occasions however, they may be suitable—the most common of which is the quantification of a relatively low-intensity peak, which is located on a downward sloping inelastic background (i.e., at a higher binding energy) from a more intense peak, as shown in Fig. 4c in the case of N 1s. Crucially, to appropriately use a linear background, the background slope on either side of the peak must be essentially identical, with no offset or change in gradient. The Shirley and Tougaard methods are both more appropriate as general models for background subtraction and are fairly commonly used, with some discussion in the literature of their differences and applicability [31, 32]. For the majority of cases, where data from only one instrument is to be considered, the use of either model is acceptable, so long it is consistently applied. Regardless of the background subtraction model selected, it is important that the endpoints of the region in which the background is defined lie significantly beyond the peak and are not unduly skewed by any noise—averaging over at least 5 data points at each end is recommended.

222

David J. H. Cant et al.

25. Often, XPS analysis software includes a default set of sensitivity factors that may be applied—for software provided with an instrument, this may even include some form of built-in intensity scale correction (these are typically referred to as “instrumental sensitivity factors”); intensity scale corrections of this nature may be inaccurate due to inter-instrument variability and drift over time, so it is advisable to perform intensity scale calibration separately and use sensitivity factors that do not account for instrumental aspects. 26. The factor f in Eq. (1) may be determined from the XPS measurements of the pure reference materials for the underlayer and the peptide coating. These should be measured under identical instrument conditions and parameters to those used for the coated sample, and ideally on the same day. The intensity and sensitivity-corrected intensities for all the elements present in a material (utilizing the same peaks as used for measurement of the coated sample) should be determined. These should then be summed separately for each material, and the sum of intensities from the underlayer material should be divided by those for the peptide to give a value for f. For the common case of a typical organic material (such as a peptide) on a gold underlayer, a value of f ¼ 0.56 is reasonable [16, 33]. 27. Estimates of coating thickness performed using XPS measure the peptide in a dry state under ultra-high vacuum. As such, they are expected to be lower than the thickness of the same peptide layer in a hydrated state. For example, previous comparisons of measurement methods have obtained estimates for the average peptide thickness on a range of nanoparticles of (4.3  0.4) nm, with solution-based measurements such as dynamic light scattering giving an estimate of (4.8  0.5) nm; the expected length of this peptide from the structure would be approximately 5.3 nm [33]. 28. An example of a full calculation of peptide thickness using this method, applied to the data shown in Fig. 3, is as follows: intensities are measured from the survey spectra for one peak from each element present: gold 4f, carbon 1s, oxygen 1s, and nitrogen 1s. These intensities are corrected using the appropriate intensity calibration for the instrument, adjusted using the NPL AMRSFs [14, 15], and the atomic percentages are calculated; Table 2 provides these values and others used throughout this calculation. As this sample consists of an organic layer on gold, f ¼ 0.56 can be used [33], and a value for Ap, u can be obtained for each peptide–underlayer element pair—in this case, AC, Au, AO, Au, and AN, Au. The chemical structure of the peptide is CGGGNPSSLFRYLPSD, i.e., C74H110N20O25S1. As hydrogen is not detected by XPS, this would give expected carbon, oxygen, nitrogen, and sulfur

Measurement of Peptide Thickness and Composition Using XPS

223

atomic concentrations of 61.7%, 20.8%, 16.7%, and 0.8% respectively. As it is below the detection limit, we neglect sulfur for the calculations and normalize the carbon, oxygen, and nitrogen to sum to 100%. This provides starting estimates for the atomic fraction of χ C ¼ 0.622, χ O ¼ 0.210, and χ N ¼ 0.168. Attenuation lengths Li, J are calculated using the method given by Seah [17] and are given in Table 1. At this point, thickness estimates can be obtained using Eq. (2) for the flat surface and Eqs. (3–6) for the nanoparticle sample, with a nanoparticle radius of 30 nm. With the initial composition assumption for the peptide, the thickness values calculated for each element are notably different. We then iteratively change the assumed composition until a minimum deviation between these thickness values is obtained, and the resulting composition and average thickness across the elements are considered the best estimates achievable. The results in Table 2 for peptide layers on flat substrates and nanoparticles both give a similar estimate for the thickness of the peptide within the
10% uncertainty that arises from the calculation of the attenuation lengths [17]. In the calculated compositions, the carbon concentration is higher than expected, likely due to the presence of adventitious carbon contamination; otherwise, the composition estimates provide reasonable agreement with the peptide formula, within the expected uncertainty. Figure 5 shows a bar chart of the expected compositions from the peptide structure, compared to these calculated values for both the flat and nanoparticle samples.

Acknowledgments This work was supported by the National Measurement System of the UK Department of Business, Energy and Industrial Strategy (BEIS). References 1. Seah MP, Spencer SJ (2003) Ultrathin SiO2 on Si IV. Intensity measurement in XPS and deduced thickness linearity. Surf Interface Anal 35:515–524 2. Cumpson PJ (2000) The Thickogram: a method for easy film thickness measurement in XPS. Surf Interface Anal 29:403–406 3. Hill JM, Royce DG, Fadley CS et al (1976) Properties of oxidized silicon as determined by angular-dependent X-ray photoelectron spectroscopy. Chem Phys Lett 44:225–231

4. Fadley CS, Baird RJ, Siekhaus W et al (1974) Surface analysis and angular distributions in X-ray photoelectron spectroscopy. J Electron Spectros Relat Phenomena 4(2):93–137 5. Shard AG, Wang J, Spencer SJ (2009) XPS topofactors: determining overlayer thickness on particles and fibres. Surf Interface Anal 41:541–548 6. Gaspar DJ, Baer DR, Castner DG et al (2010) Application of surface chemical analysis tools for characterization of nanoparticles. Anal Bioanal Chem 396:983–1002

224

David J. H. Cant et al.

7. Chudzicki M, Werner WSM, Shard AG et al (2015) Evaluating the internal structure of core-shell nanoparticles using X-ray photoelectron intensities and simulated spectra. J Phys Chem C 119:17687–17696 8. Belsey NA, Cant DJH, Minelli C et al (2016) Versailles project on advanced materials and standards Interlaboratory study on measuring the thickness and chemistry of nanoparticle coatings using XPS and LEIS. J Phys Chem C 120:24070–24079 9. Shard AG (2012) A straightforward method for interpreting XPS data from core-shell nanoparticles. J Phys Chem C 116:16806–16813 10. Cant DJH, Wang Y-C, Castner DG et al (2016) A technique for calculation of shell thicknesses for core-shell-shell nanoparticles from XPS data. Surf Interface Anal 48:274–282 11. Baer DR, Karakoti AS, Clifford CA et al (2018) Importance of sample preparation on reliable surface characterisation of nano-objects: ISO standard 20579-4. Surf Interface Anal 50:902–906 12. ISO 20579-4:2018 - Surface chemical analysis -- Guidelines to sample handling, preparation and mounting -- Part 4: Reporting information related to the history, preparation, handling and mounting of nano-objects prior to surface analysis, https://www.iso.org/standard/ 68833.html 13. Shard AG (2014) Detection limits in XPS for more than 6000 binary systems using Al and mg Kα X-rays. Surf Interface Anal 46:175–185 14. Surface technology—NPL, https://www.npl. co.uk/surface-technology 15. Seah MP, Gilmore IS, Spencer SJ (2001) Quantitative XPS: I. analysis of X-ray photoelectron intensities from elemental data in a digital photoelectron database. J Electron Spectros Relat Phenomena 120:93–111 16. Ray S, Steven RT, Green FM et al (2015) Neutralized Chimeric Avidin Binding at a Reference Biosensor Surface. Langmuir 31:1921–1930 17. Seah MP (2012) Simple universal curve for the energy-dependent electron attenuation length for all materials. Surf Interface Anal 44:1353–1359 18. NIST Standard Reference Database 82 | NIST Electron Effective-Attenuation-Length Database. https://www.nist.gov/srd/nist-stan dard-reference-database-82 19. Gilliland D, Ceccone G, Spampinato V et al (2017) Influence of different cleaning

processes on the surface chemistry of gold nanoparticles. Biointerphases 12:031003 20. Casa Software Ltd. http://www.casaxps.com/ 21. AAnalyzer. http://rdataa.com/aanalyzer/ aanaHome.htm 22. Hesse R, Chasse´ T, Szargan R (2003) Unifit 2002-universal analysis software for photoelectron spectra. Anal Bioanal Chem 375:856–863 23. Majumder M, Rendall CS, Eukel JA et al (2012) Overcoming the “coffee-stain” effect by compositional Marangoni-flow-assisted drop-drying. J Phys Chem B 116:6536–6542 24. Hu H, Larson RG (2006) Marangoni effect reverses coffee-ring depositions. J Phys Chem B 110:7090–7094 25. Werner WSM, Chudzicki M, Smekal W et al (2014) Interpretation of nanoparticle X-ray photoelectron intensities. Appl Phys Lett 104:243106 26. Shard AG, Spencer SJ (2019) Intensity calibration for monochromated Al Kα XPS instruments using polyethylene. Surf Interface Anal 51:618–626 27. Seah MP (1993) XPS reference procedure for the accurate intensity calibration of electron spectrometers? Results of a BCR intercomparison co-sponsored by the VAMAS SCA TWA. Surf Interface Anal 20:243–266 28. Moulder JF, Chastain J (1992) Handbook of x-ray photoelectron spectroscopy : a reference book of standard spectra for identification and interpretation of XPS data. Physical Electronics Division, Perkin-Elmer Corp 29. Naumkin A V., Kraut-Vass A, Powell CJ, et al NIST X-ray photoelectron spectroscopy database. http://books.google.com.ar/books? id¼I6LhMgEACAAJ 30. Beamson G, Briggs D (1993) High resolution XPS of organic polymers: the Scienta ESCA300 database. J Chem Educ 70:A25 31. Castle JE, Chapman-Kpodo H, Proctor A et al (2000) Curve-fitting in XPS using extrinsic and intrinsic background structure. J Electron Spectros Relat Phenomena 106:65–80 32. Seah MP (1999) Background subtraction: I. general behaviour of Tougaard-style backgrounds in AES and XPS. Surf Sci 420:285–294 33. Belsey NA, Shard AG, Minelli C (2015) Analysis of protein coatings on gold nanoparticles by XPS and liquid-based particle sizing techniques. Biointerphases 10:019012

Chapter 15 Imaging the Effects of Peptide Materials on Phospholipid Membranes by Atomic Force Microscopy Katharine Hammond, Georgina Benn, Isabel Bennett, Edward S. Parsons, Maxim G. Ryadnov, Bart W. Hoogenboom, and Alice L. B. Pyne Abstract Recent advances in biomolecular design require accurate measurements performed in native or near-native environments in real time. Atomic force microscopy (AFM) is a powerful tool to observe the dynamics of biologically relevant processes at aqueous interfaces with high spatial resolution. Here, we describe imaging protocols to characterize the effects of peptide materials on phospholipid membranes in solution by AFM. These protocols can be used to determine the mechanism and kinetics of membrane-associated activities at the nanoscale. Key words Atomic force microscopy, Antimicrobial peptides, Peptide materials, Supported lipid bilayers, Phospholipid membranes

1

Introduction The plasma membrane represents the physical barrier between a cell and its environment. In its most basic form, it comprises a bilayer motif with a thickness of only a few tens of A˚ngstro¨ms [1]. The membrane has evolved to selectively allow the passage of ions, small molecules, and larger macromolecules in and out the cell. The plasma membrane is also vulnerable to attack, e.g., by pore-forming proteins and membrane-targeting peptide assemblies [2, 3]. Particularly notable are antimicrobial peptides that disrupt bacterial membranes, often leading to lysis and cell death [4]. An in-depth understanding of how these peptides interact with membranes can improve the design of more effective antimicrobials demonstrating higher activity and specificity [5]. Recent progress in atomic force microscopy (AFM) has enabled the study of the plasma membrane interface, and its disruption by

Katharine Hammond and Georgina Benn contributed equally to this work. Maxim G. Ryadnov (ed.), Polypeptide Materials: Methods and Protocols, Methods in Molecular Biology, vol. 2208, https://doi.org/10.1007/978-1-0716-0928-6_15, © Springer Science+Business Media, LLC, part of Springer Nature 2021

225

226

Katharine Hammond et al.

antimicrobial peptides and peptide nanoscale assemblies in aqueous solution (e.g., [3, 6–11]). A straightforward approach to probe the impact of peptide treatments on the plasma membrane is to use supported lipid bilayers (SLBs) assembled on a flat mica surface [12]. The lipid composition of SLBs can be tuned to mimic eukaryotic or prokaryotic cell membranes. Bacterial membranes are naturally rich in anionic lipids, phosphatidylethanolamine (PE), and cardiolipin [13]; the most common bacterial membrane mimics consist of mixtures of various zwitterionic and anionic phospholipids. Physiological SLBs form fluid phases and can be prepared from phospholipids with gel-to-fluid transition temperatures that are significantly lower than room temperature [14]. SLBs can be flat to within a few A˚ngstro¨ms, allowing the visualization of changes to the bilayer surface with nanometre resolution by AFM in aqueous solution [15]. Antimicrobial peptides can interact with phospholipid bilayers in many ways, inducing poration, localized thinning effects, and surface roughening [16]. These interactions induce topographical changes, which can be distinguished by AFM, as shown in Fig. 1. AFM imaging of membranes treated with peptide should be carried out in fluid to directly observe surface alterations with temporal resolution and minimize potential artifacts due to uncontrolled drying effects. Here, we provide a protocol using the rapid force-distance (or off-resonance tapping) imaging mode as implemented by Bruker (Santa Barbara, CA), denoted as PeakForce Tapping®. The protocol also applies to experiments carried out using the more common intermittent contact, or tapping mode, and has also been performed in contact mode (e.g., [11]).

2

Materials Prepare and store all reagents at room temperature, unless indicated otherwise.

2.1 General Materials

1. 10 mm mica substrates (Agar Scientific). 2. 15 mm magnetic stainless-steel discs (Agar Scientific). 3. Adhesive backed PTFE (Bytac® surface protection laminate). 4. Scalpel or punch. 5. Scotch tape. 6. Araldite® 2-part epoxy resin. 7. Stainless-steel tweezers. 8. Plasma cleaner (Zepto, Deiner Electronics).

Imaging Peptide Materials on Membranes

227

Fig. 1 AFM observation of peptide material effects on supported lipid bilayers (SLBs). (a) Topography of an untreated SLB. The surface is featureless and flat to within 300–700 nm).

Imaging Peptide Materials on Membranes

229

Fig. 2 (a) Magnetic steel sample disc, (b) PTFE adhered to magnetic steel sample disc, and (c) mica disc glued to PTFE surface using Araldite® 2-part epoxy resin

5. Sonicate the vesicle suspension in a bath sonicator at 37 kHz, at a temperature above the gel to fluid transition temperature of the lipids (see Note 5) until the solution becomes clear. This should take between 5 min and 1 h, depending on the lipid species, the concentration, and the position in the bath sonicator (whether the suspension is placed in a sonic ‘hotspot’). 6. Filter the vesicle suspension using an Avanti mini-extruder through 50-nm polycarbonate membranes a minimum of 20 times to ensure vesicles of uniform size. This should also be done above the gel to fluid transition temperature of the lipids. The final extruded suspension should be taken from the opposite side to the initial insertion to ensure all vesicles in solution have been passed through the membrane. 3.3 Formation of a Supported Lipid Bilayer

Prepare SLBs via the vesicle fusion method (Fig. 3) (see Note 6). 1. Add 80 μL of buffer solution to a freshly cleaved mica substrate surface, followed by 10 μL of the SUV suspension (1 mg/mL). Finally, add 10 μL calcium chloride solution to give a final calcium concentration of 10 mM. Calcium aids the adsorption and rupture of the SUVs to the mica surface. 2. Incubate the SUV solution on the mica substrate at room temperature (see Note 7) for 30 min to allow adsorption and rupture of the vesicles onto the mica surface. Incubation should be performed in a closed or humid chamber, to ensure the surface is kept hydrated throughout (see Note 8). 3. Wash the mica surface thoroughly by adding and removing 50 μL of buffer solution ten times. Take care to ensure the surface is kept hydrated at all times. This will remove excess vesicles from solution, yielding a uniform bilayer free of adsorbed vesicles (as assessed by AFM imaging).

3.4 AFM Imaging of Supported Lipid Bilayers

1. Prior to use, soak the chosen cantilever in isopropanol:ethanol (1:1) and dry. Then, plasma clean in air for 30 s at 10% power.

230

Katharine Hammond et al.

Fig. 3 Using the vesicle fusion method, small unilamellar vesicles are adsorbed onto a flat substrate using divalent cations; they then flatten and rupture to form a continuous bilayer Table 1 Typical parameters used for an MSNL-E cantilever on a Multimode 8 AFM system operated in PeakForce Tapping mode in liquid Parameter

Typical value

Scan size

500 nm to 2 μm

Aspect ratio

1

Pixel density

256  256–512  512

Line rate

1–1.5 Hz

Gain

30–40

Set point

0.01–0.03 V

Sync distance

95–120 μs

PeakForce frequency

2–4 Hz

Lift height

7–12 nm

Z range

1–1.5 μm

Deflection limit

12.24 V

2. Mount the chosen cantilever in the AFM and align the laser. Leave the AFM to equilibrate in buffer solution, using a clean freshly cleaved mica disc during sample preparation. 3. Exchange the blank, equilibration mica disc for the mica disc with supported lipid bilayer on top. 4. Engage the cantilever. 5. Imaging parameters will vary, depending on the chosen cantilever. Generally, imaging is carried out at 2, 4, or 8 kHz, with PeakForce amplitudes of 10–20 nm, set points of 0.05–0.2 V (100 pN), and a pixel density of 512 by 512. More detailed, typical parameters for an MSNL-E are shown in Table 1.

Imaging Peptide Materials on Membranes

3.5 AFM Force Spectroscopy of SLBs

231

The flatness of a bilayer, and its similarity to the mica it is deposited on, means that a force curve must be used to confirm its presence. Once the cantilever is engaged, a force curve is performed. 1. To begin with, the tip is far enough away from the sample surface (0–75 nm Z distance) that they do not interact. 2. When the tip first comes into contact with the bilayer surface, there is an increase in force as the tip pushes on the bilayer surface, elastically deforming it. 3. When a high enough force is exerted to push the tip through the bilayer, the tip then moves the distance of the bilayer to the mica surface below. This required force, called the breakthrough force, provides information about the stiffness of the bilayer. In addition, the Z distance of this peak can be used to calculate the width of the bilayer. 4. The tip then pushes on the mica substrate, rapidly increasing the force on, and deflecting, the cantilever. 5. The appearance of this characteristic breakthrough force curve allows us to confirm the presence of a bilayer on the mica surface, as this feature is not present in force curves on bare mica. Figure 4 shows the corresponding points on a force curve and movement of the cantilever.

Fig. 4 (a) Movement of the cantilever as a force curve is performed on a bilayer (pink) on mica (brown). (b) Force curve showing typical bilayer approach and retract curves for SLB on mica. During approach (blue), the tip breaks through the bilayer to the mica surface, giving rise to a characteristic breakthrough force peak. During retraction (red), there is adhesion of the tip to the bilayer, followed by “snapping off” and retraction. Numbers correspond to the numbered points in (a)

232

Katharine Hammond et al.

3.6 AFM Imaging of Peptide Mechanism of Action

1. After forming an SLB, image to confirm it is free from defects. 2. Withdraw the AFM tip from the bilayer surface (see Note 9), taking care that the sample remains hydrated. 3. Prepare a 3 mM solution of antimicrobial peptide in buffer. 4. Inject 10 μL peptide solution to the lipid bilayer. This will give a final peptide concentration of around 0.3 μM (see Note 10). 5. Re-engage the cantilever. 6. Image the peptide treated bilayer until no further changes are observed. This can take between 5 min and 2 h. 7. Imaging parameters will depend on the cantilever used and the mechanism of membrane disruption but are typically similar to the imaging parameters used before peptide addition (see Table 1). Optimal scan size will depend on the mechanism observed, but is generally 1–4 μm, with 512 pixel density, PeakForce amplitudes of 10–20 nm, and set points of 0.05–0.2 V (100 pN). 6. Once the surface has stopped changing, a larger scan size in the same area should be taken. This is to confirm that the membrane disruption observed is not an artifact due to the forces exerted by the continuously scanning AFM tip. A scan in a different sample location should also be taken (see Note 11). 7. In some cases, peptide can stick to the cantilever and interfere with imaging. Furthermore, peptide–lipid vesicles can form and re-bind to the bilayer surface. To reduce these effects, wait for 15 min after peptide injection (step 4 in Subheading 3.6), before washing the sample surface by adding and removing 50 μL of buffer solution ten times. This allows the peptide to bind to the membrane surface but removes excess peptide from the imaging solution, enabling cleaner imaging.

4

Notes 1. Other zwitterionic phospholipids such as 1,2-doileoyl-sn-glycero-3-phosphoethanolamine (DOPE) or 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) may be used instead of, or in combination with POPC. Other anionic phospholipids could also be used, such as 1,2-dioleoyl-sn-glycero-3-phospho-(10 -rac-glycerol) (DOPG). 2. It is important to prepare all solutions using ultrapure water (for example, MilliQ™, which is prepared by purifying

Imaging Peptide Materials on Membranes

233

deionized water to a resistivity of 18 MΩ cm and TOC < 10 ppb at 25  C) and analytical-grade reagents. 3. The use of an appropriate cantilever is important to obtain adequate resolution and to minimize damage to the sample. The choice of cantilever depends on the mode of AFM being employed and the type of sample. For imaging of peptide/ protein-membrane interactions in fluid, PEAKFORCEHIRS-F-B, FastScan D, Biolever Mini, MSNL-E, and MSNLF cantilevers were selected due to their low spring constants (