Stable Isotope Probing: Methods and Protocols [1st ed. 2019] 978-1-4939-9720-6, 978-1-4939-9721-3

Table of contents :
Front Matter ....Pages i-xi
Experimental Setup and Data Analysis Considerations for DNA- and RNA-SIP Experiments in the Omics Era (Roey Angel)....Pages 1-15
DNA-Based Stable Isotope Probing (Zhongjun Jia, Weiwei Cao, Marcela Hernández García)....Pages 17-29
RNA Stable Isotope Probing (RNA-SIP) (Noor-Ul-Huda Ghori, Benjamin Moreira-Grez, Paton Vuong, Ian Waite, Tim Morald, Michael Wise et al.)....Pages 31-44
Stable Isotope Probing of Microbial Phospholipid Fatty Acids in Environmental Samples (Andrea Watzinger, Rebecca Hood-Nowotny)....Pages 45-55
SIP-Metaproteomics: Linking Microbial Taxonomy, Function, and Activity (Martin Taubert)....Pages 57-69
Chip-SIP: Stable Isotope Probing Analyzed with rRNA-Targeted Microarrays and NanoSIMS (Xavier Mayali, Peter K. Weber, Erin Nuccio, Jory Lietard, Mark Somoza, Steven J. Blazewicz et al.)....Pages 71-87
Quantification of Methanogenic Pathways Using Stable Carbon Isotopic Signatures (Quan Yuan)....Pages 89-94
Stable Isotope-Labeled Single-Cell Raman Spectroscopy Revealing Function and Activity of Environmental Microbes (Li Cui, Kai Yang, Yong-Guan Zhu)....Pages 95-107
Data Analysis for DNA Stable Isotope Probing Experiments Using Multiple Window High-Resolution SIP (Samuel E. Barnett, Nicholas D. Youngblut, Daniel H. Buckley)....Pages 109-128
Stable Isotope Probing of Microorganisms in Environmental Samples with H218O (Egbert Schwartz, Michaela Hayer, Bruce A. Hungate, Rebecca L. Mau)....Pages 129-136
Microbial Taxon-Specific Isotope Incorporation with DNA Quantitative Stable Isotope Probing (Brianna K. Finley, Michaela Hayer, Rebecca L. Mau, Alicia M. Purcell, Benjamin J. Koch, Natasja C. van Gestel et al.)....Pages 137-149
Profiling of Active Microorganisms by Stable Isotope Probing—Metagenomics (Eileen Kröber, Özge Eyice)....Pages 151-161
Targeted Metatranscriptomics of Soil Microbial Communities with Stable Isotope Probing (Ang Hu, Yahai Lu, Marcela Hernández García, Marc G. Dumont)....Pages 163-174
Stable Isotope Probing Techniques and Methodological Considerations Using 15N (Roey Angel)....Pages 175-187
DNA and RNA Stable Isotope Probing of Methylotrophic Methanogenic Archaea (Xiuran Yin, Ajinkya C. Kulkarni, Michael W. Friedrich)....Pages 189-206
Method Development for DNA and Proteome SIP Analysis of Activated Sludge for Anaerobic Dichloromethane Biodegradation (Miao Hu, Matthew Lee, Ling Zhong, Michael J. Manefield)....Pages 207-219
RNA-Based Stable Isotope Probing (RNA-SIP) in the Gut Environment (Severin Weis, Sylvia Schnell, Markus Egert)....Pages 221-231
Stable Isotope Probing of Microbiota Structure and Function in the Plant Rhizosphere (Wafa Achouak, Feth el Zahar Haichar)....Pages 233-243
Back Matter ....Pages 245-247

Citation preview

Methods in Molecular Biology 2046

Marc G. Dumont Marcela Hernández García Editors

Stable Isotope Probing Methods and Protocols

METHODS

IN

MOLECULAR BIOLOGY

Series Editor John M. Walker School of Life and Medical Sciences University of Hertfordshire Hatfield, Hertfordshire AL10 9AB, 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.

Stable Isotope Probing Methods and Protocols

Edited by

Marc G. Dumont School of Biological Sciences, University of Southampton, Southampton, UK

Marcela Hernández García School of Biological Sciences, University of Southampton, Southampton, UK

Editors Marc G. Dumont School of Biological Sciences University of Southampton Southampton, UK

Marcela Herna´ndez Garcı´a School of Biological Sciences University of Southampton Southampton, UK

ISSN 1064-3745 ISSN 1940-6029 (electronic) Methods in Molecular Biology ISBN 978-1-4939-9720-6 ISBN 978-1-4939-9721-3 (eBook) https://doi.org/10.1007/978-1-4939-9721-3 © Springer Science+Business Media, LLC, part of Springer Nature 2019 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, express 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: Image of a CsCl gradient being fractionated after ultracentrifugation. This Humana imprint is published by the registered company Springer Science+Business Media, LLC, part of Springer Nature. The registered company address is: 233 Spring Street, New York, NY 10013, U.S.A.

Preface The term “stable isotope probing” (SIP) and arguably the era of SIP-related approaches to study microbial ecology began with the work in Colin Murrell’s group at the University of Warwick. The initial study, published in 2000, targeted methylotrophs in soil with 13Cmethanol and recovered the labeled DNA from cesium chloride (CsCl) density gradients. In fact, stable isotope labeling of lipids, such as PLFAs, had been used in microbial ecology experiments for some time previously. The SIP approach rapidly developed and soon was being used to target various microbial biomarkers, including RNA and proteins. SIP technologies are now standard tools for investigating microbial communities and have proved their value in more than a thousand published studies. The versatility was not necessarily clear from the outset, with initial speculation that the utility would be restricted to targeting functional groups of microbes with relatively restricted carbon-use profiles, which methylotrophs were a case in point. In fact, SIP has been used to target a wide variety of functional groups of microbes. Another initial concern was that cross-feeding of labeled substrates from the primary consumers to lower trophic levels would be so rapid that it could be impossible to identify the primary consumers. Although cross-feeding does occur, the initial fears seem to be unfounded, and the potential drawbacks tend to be outweighed by the insights it can provide into trophic interactions. The objective of this book is to provide definitive methods to perform SIP experiments. There is some overlap between the chapters describing DNA- and RNA-SIP methods, which we believe is beneficial as they show the varying protocols that researchers have independently developed in their labs. In some cases, these reflect a different availability of equipment. There is not one definitive method but variations that might suit different researchers depending on their research questions and resources. This book covers a wide spectrum of stable isotope methods used in microbial ecology, such as methods to target and analyze labeled DNA, rRNA, mRNA, protein, and PLFA. The protocols to study stable isotope fractionation by microbial pathways and the analysis of labeled communities with Raman microscopy are also covered. Additionally, Chip-SIP, which uses microchips and NanoSIMS to detect stable isotope incorporation into targeted microbial groups, is described. Some of the most exciting recent developments have been the introduction of sophisticated analysis tools to quantify and sensitively detect labeling of DNA and RNA, such as quantitative SIP (qSIP) and high-resolution SIP (HR-SIP). As these protocols rely extensively on in silico analyses, the chapters have step-by-step instructions to implement the bioinformatic pipelines. Also included is a general description of tools that can be used to analyze metagenomics data from DNA-SIP experiments. Although 13C is the most widely used isotope in SIP methods, other elements have been applied in SIP, for instance, 15N, which presents certain challenges. 18O can also be used in traditional SIP experiments. For example, water (H218O) is the most general substrate and is proving to be a powerful means to identify active populations under different conditions. The protocols for 15N-SIP and H218O-SIP are described in detail here.

v

vi

Preface

The case studies provide valuable insights into how SIP technologies can be applied. Included are several chapters describing how researchers have used SIP to answer specific research questions. We believe that this book provides the readers with up-to-date protocols ranging from basic to the most sophisticated applications of SIP and will benefit anyone wishing to use these methods. Southampton, UK

Marc G. Dumont Marcela Herna´ndez Garcı´a

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

1 Experimental Setup and Data Analysis Considerations for DNA- and RNA-SIP Experiments in the Omics Era. . . . . . . . . . . . . . . . . . . . . . Roey Angel 2 DNA-Based Stable Isotope Probing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zhongjun Jia, Weiwei Cao, and Marcela Herna´ndez Garcı´a 3 RNA Stable Isotope Probing (RNA-SIP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Noor-Ul-Huda Ghori, Benjamin Moreira-Grez, Paton Vuong, Ian Waite, Tim Morald, Michael Wise, and Andrew S. Whiteley 4 Stable Isotope Probing of Microbial Phospholipid Fatty Acids in Environmental Samples. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Andrea Watzinger and Rebecca Hood-Nowotny 5 SIP-Metaproteomics: Linking Microbial Taxonomy, Function, and Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Martin Taubert 6 Chip-SIP: Stable Isotope Probing Analyzed with rRNA-Targeted Microarrays and NanoSIMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Xavier Mayali, Peter K. Weber, Erin Nuccio, Jory Lietard, Mark Somoza, Steven J. Blazewicz, and Jennifer Pett-Ridge 7 Quantification of Methanogenic Pathways Using Stable Carbon Isotopic Signatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quan Yuan 8 Stable Isotope-Labeled Single-Cell Raman Spectroscopy Revealing Function and Activity of Environmental Microbes . . . . . . . . . . . . . . . . . Li Cui, Kai Yang, and Yong-Guan Zhu 9 Data Analysis for DNA Stable Isotope Probing Experiments Using Multiple Window High-Resolution SIP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Samuel E. Barnett, Nicholas D. Youngblut, and Daniel H. Buckley 10 Stable Isotope Probing of Microorganisms in Environmental Samples with H218O . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Egbert Schwartz, Michaela Hayer, Bruce A. Hungate, and Rebecca L. Mau 11 Microbial Taxon-Specific Isotope Incorporation with DNA Quantitative Stable Isotope Probing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Brianna K. Finley, Michaela Hayer, Rebecca L. Mau, Alicia M. Purcell, Benjamin J. Koch, Natasja C. van Gestel, Egbert Schwartz, and Bruce A. Hungate 12 Profiling of Active Microorganisms by Stable Isotope Probing—Metagenomics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ¨ zge Eyice Eileen Kro¨ber and O

vii

v ix

1 17 31

45

57

71

89

95

109

129

137

151

viii

13

14

15

16

17 18

Contents

Targeted Metatranscriptomics of Soil Microbial Communities with Stable Isotope Probing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ang Hu, Yahai Lu, Marcela Herna´ndez Garcı´a, and Marc G. Dumont Stable Isotope Probing Techniques and Methodological Considerations Using 15N . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Roey Angel DNA and RNA Stable Isotope Probing of Methylotrophic Methanogenic Archaea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Xiuran Yin, Ajinkya C. Kulkarni, and Michael W. Friedrich Method Development for DNA and Proteome SIP Analysis of Activated Sludge for Anaerobic Dichloromethane Biodegradation . . . . . . . . . . Miao Hu, Matthew Lee, Ling Zhong, and Michael J. Manefield RNA-Based Stable Isotope Probing (RNA-SIP) in the Gut Environment . . . . . . Severin Weis, Sylvia Schnell, and Markus Egert Stable Isotope Probing of Microbiota Structure and Function in the Plant Rhizosphere. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wafa Achouak and Feth el Zahar Haichar

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

163

175

189

207 221

233 245

Contributors WAFA ACHOUAK  Aix Marseille University, CNRS, CEA, UMR 7265 BVME, LEMIRE, ECCOREV FR 3098, Saint-Paul-lez-Durance, France ROEY ANGEL  Soil and Water Research Infrastructure and Institute of Soil Biology, Biology ˇ eske´ Budeˇjovice, Czech Republic Centre CAS, C SAMUEL E. BARNETT  School of Integrative Plant Science, Cornell University, Ithaca, NY, USA STEVEN J. BLAZEWICZ  Physical and Life Sciences Directorate, Lawrence Livermore National Laboratory, Livermore, CA, USA DANIEL H. BUCKLEY  School of Integrative Plant Science, Cornell University, Ithaca, NY, USA WEIWEI CAO  State Key Laboratory of Soil and Sustainable Agriculture, Institute of Soil Science, Chinese Academy of Sciences, Nanjing, People’s Republic of China LI CUI  Key Lab of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen, China MARC G. DUMONT  School of Biological Sciences, University of Southampton, Southampton, UK MARKUS EGERT  Faculty of Medical and Life Sciences, Institute of Precision Medicine, Microbiology and Hygiene Group, Furtwangen University, Villingen-Schwenningen, Germany FETH EL ZAHAR HAICHAR  Laboratoire d’Ecologie Microbienne, University of Lyon, Universite´ Claude Bernard Lyon 1, UMR INRA 1418, UMR CNRS 5557, Villeurbanne Cedex, France ¨ ZGE EYICE  School of Biological and Chemical Sciences, Queen Mary University of London, O London, UK BRIANNA K. FINLEY  Department of Biological Sciences, Center for Ecosystem Science and Society, Northern Arizona University, Flagstaff, AZ, USA MICHAEL W. FRIEDRICH  Microbial Ecophysiology Group, Faculty of Biology/Chemistry, University of Bremen, Bremen, Germany; MARUM—Center for Marine Environmental Sciences, University of Bremen, Bremen, Germany MARCELA HERNA´NDEZ GARCI´A  School of Biological Sciences, University of Southampton, Southampton, UK; Max-Planck-Institute for Terrestrial Microbiology, Marburg, Germany NOOR-UL-HUDA GHORI  Molecular Microbial Ecology Group, The UWA School of Agriculture and Enviornment (SAgE), The University of Western Australia, Crawley, WA, Australia MICHAELA HAYER  Department of Biological Sciences, Center for Ecosystem Science and Society, Northern Arizona University, Flagstaff, AZ, USA REBECCA HOOD-NOWOTNY  Institute of Soil Research, University of Natural Resources and Life Sciences Vienna, Tulln, Austria ANG HU  College of Resources and Environment, Hunan Agricultural University, Changsha, China MIAO HU  UNSW Water Research Centre, School of Civil and Environmental Engineering, University of New South Wales, Sydney, NSW, Australia

ix

x

Contributors

BRUCE A. HUNGATE  Department of Biological Sciences, Center for Ecosystem Science and Society, Northern Arizona University, Flagstaff, AZ, USA ZHONGJUN JIA  State Key Laboratory of Soil and Sustainable Agriculture, Institute of Soil Science, Chinese Academy of Sciences, Nanjing, People’s Republic of China BENJAMIN J. KOCH  Department of Biological Sciences, Center for Ecosystem Science and Society, Northern Arizona University, Flagstaff, AZ, USA EILEEN KRO¨BER  Microbial Biogeochemistry, RA Landscape Functioning, ZALF Leibniz Centre for Landscape Research, Mu¨ncheberg, Germany AJINKYA C. KULKARNI  Microbial Ecophysiology Group, Faculty of Biology/Chemistry, University of Bremen, Bremen, Germany; MARUM—Center for Marine Environmental Sciences, University of Bremen, Bremen, Germany; International Max Planck Research School for Marine Microbiology, Max Planck Institute for Marine Microbiology, Bremen, Germany MATTHEW LEE  School of Civil and Environmental Engineering, UNSW Water Research Centre, University of New South Wales, Sydney, NSW, Australia JORY LIETARD  Institute of Inorganic Chemistry, Faculty of Chemistry, University of Vienna, Vienna, Austria YAHAI LU  College of Urban and Environmental Sciences, Peking University, Beijing, China MICHAEL J. MANEFIELD  UNSW Water Research Centre, School of Civil and Environmental Engineering, University of New South Wales, Sydney, NSW, Australia; School of Chemical Engineering, University of New South Wales, Sydney, NSW, Australia REBECCA L. MAU  Department of Biological Sciences, Center for Ecosystem Science and Society, Northern Arizona University, Flagstaff, AZ, USA; Pathogen and Microbiome Institute, Northern Arizona University, Flagstaff, AZ, USA XAVIER MAYALI  Physical and Life Sciences Directorate, Lawrence Livermore National Laboratory, Livermore, CA, USA TIM MORALD  Molecular Microbial Ecology Group, The UWA School of Agriculture and Enviornment (SAgE), The University of Western Australia, Crawley, WA, Australia BENJAMIN MOREIRA-GREZ  Molecular Microbial Ecology Group, The UWA School of Agriculture and Enviornment (SAgE), The University of Western Australia, Crawley, WA, Australia ERIN NUCCIO  Physical and Life Sciences Directorate, Lawrence Livermore National Laboratory, Livermore, CA, USA JENNIFER PETT-RIDGE  Physical and Life Sciences Directorate, Lawrence Livermore National Laboratory, Livermore, CA, USA ALICIA M. PURCELL  Department of Biological Sciences, Center for Ecosystem Science and Society, Northern Arizona University, Flagstaff, AZ, USA SYLVIA SCHNELL  Institute of Applied Microbiology, Research Center for BioSystems, Land Use, and Nutrition (IFZ), Justus–Liebig–University Giessen, Giessen, Germany EGBERT SCHWARTZ  Department of Biological Sciences, Center for Ecosystem Science and Society, Northern Arizona University, Flagstaff, AZ, USA MARK SOMOZA  Institute of Inorganic Chemistry, Faculty of Chemistry, University of Vienna, Vienna, Austria MARTIN TAUBERT  Faculty of Biological Sciences, Institute of Biodiversity, Friedrich Schiller University Jena, Jena, Germany NATASJA C. VAN GESTEL  Department of Biological Sciences, Texas Tech University, Lubbock, TX, USA

Contributors

xi

PATON VUONG  Molecular Microbial Ecology Group, The UWA School of Agriculture and Enviornment (SAgE), The University of Western Australia, Crawley, WA, Australia IAN WAITE  Molecular Microbial Ecology Group, The UWA School of Agriculture and Enviornment (SAgE), The University of Western Australia, Crawley, WA, Australia ANDREA WATZINGER  Institute of Soil Research, University of Natural Resources and Life Sciences Vienna, Tulln, Austria PETER K. WEBER  Physical and Life Sciences Directorate, Lawrence Livermore National Laboratory, Livermore, CA, USA SEVERIN WEIS  Faculty of Medical and Life Sciences, Institute of Precision Medicine, Microbiology and Hygiene Group, Furtwangen University, Villingen-Schwenningen, Germany; Institute of Applied Microbiology, Research Center for BioSystems, Land Use, and Nutrition (IFZ), Justus–Liebig–University Giessen, Giessen, Germany ANDREW S. WHITELEY  Molecular Microbial Ecology Group, The UWA School of Agriculture and Enviornment (SAgE), The University of Western Australia, Crawley, WA, Australia; Faculty of Science, The University of Western Australia, Crawley, WA, Australia MICHAEL WISE  Department of Computer Science and Engineering, The University of Western Australia, Perth, WA, Australia KAI YANG  Key Lab of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen, China XIURAN YIN  Microbial Ecophysiology Group, Faculty of Biology/Chemistry, University of Bremen, Bremen, Germany; MARUM—Center for Marine Environmental Sciences, University of Bremen, Bremen, Germany; International Max Planck Research School for Marine Microbiology, Max Planck Institute for Marine Microbiology, Bremen, Germany NICHOLAS D. YOUNGBLUT  Department of Microbiome Science, Max Planck Institute for Developmental Biology, Tu¨bingen, Germany QUAN YUAN  State Key Laboratory of Environmental Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences, Guiyang, China LING ZHONG  Bioanalytical Mass Spectrometry Facility, University of New South Wales, Sydney, NSW, Australia YONG-GUAN ZHU  Key Lab of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen, China

Chapter 1 Experimental Setup and Data Analysis Considerations for DNA- and RNA-SIP Experiments in the Omics Era Roey Angel Abstract Careful and thoughtful experimental design is crucial to the success of any SIP experiment. This chapter discusses the essential aspects of designing a SIP experiment, focusing primarily on DNA- and RNA-SIP. The design aspects discussed here begin with considerations for carrying out the incubation, such as, the effect of choosing different stable isotopes and target biomolecules, to what degree should a labeled substrate be enriched, what concentration to use, and how long should the incubation take. Then tips and pitfalls in the technical execution of SIP are listed, including how much nucleic acids should be loaded, how many fractions to collect, and what centrifuge rotor to use. Lastly, a brief overview of the current methods for analyzing SIP data is presented, focusing on high-throughput amplicon sequencing, together with a discussion on how the choice of analysis method might affect the experimental design. Key words DNA-SIP, RNA-SIP, Amplicon sequencing, Omics, Network analysis

1

Introduction The success of any lab experiment hinges on a thoughtful design of the experimental system, careful execution of protocols, and statistically sound data analysis. While SIP protocols have matured and become standardized over the past 20 years since their introduction, what surrounds the gradient generation and fractionation, i.e., the experimental design and data analysis, have been somewhat neglected. Other chapters in this book provide detailed protocols on how to perform SIP in the lab and how to analyze the data using specific methods. This chapter, on the other hand, discusses general considerations in conceptualizing a SIP experiment, designing the experimental set-up and choosing the right analysis method. The focus here is on DNA- and RNA-SIP experiments since these are the most flexible and most widely used forms of SIP. Table 1 summarizes the main points to consider during each of the various steps in designing a SIP experiment.

Marc G. Dumont and Marcela Herna´ndez Garcı´a (eds.), Stable Isotope Probing: Methods and Protocols, Methods in Molecular Biology, vol. 2046, https://doi.org/10.1007/978-1-4939-9721-3_1, © Springer Science+Business Media, LLC, part of Springer Nature 2019

1

2

Roey Angel

Table 1 What should be considered during each of the various steps in the design of a SIP experiment Experimental design step What to consider Which stable isotope to use?

Choice of stable isotope primarily depends on the substrate being used, but different stable isotopes differ in their ability to label nucleic acids and lipids

Prior to incubation

Measure or estimate the turnover rate of the substrate that will be used for labeling

Target molecule

Will dramatically affect what type of data will be produced and what can be learned from it

Incubation duration

Short incubation times might lead to insufficient labeling of the target molecule but long incubation times increase the risk of cross-feeding

Substrate enrichment Substrate should almost always be fully labeled, concentration should be level and concentration within a realistic range for the sample Amount of nucleic acids to load

Varies for DNA- and RNA-SIP. Will also depend on the downstream application

Number of fractions to collect and sequencing depth

More fractions means higher sensitivity but also higher contamination potential and sequencing costs

Unlabeled controls

Should always be included but the exact number will depend on the requirements of the data analysis method

Type of rotor

Traditionally vertical but fixed angle has been recently suggested to be advantageous

Data analysis

Consider how many gradients, fractions, and types of samples (e.g., controls, time series, various concentration levels, etc.) will be needed for the chosen data analysis pipeline

2

Choice of Stable Isotope Every SIP experiment is based on incubating the sample in the presence of a heavy isotope labeled substrate. In theory, every element that is present in the target biomolecule—DNA, RNA, phospholipid-derived fatty acids, or proteins—can be labeled and therefore be used in a SIP experiment. The only exception is, of course, phosphorus for which the common form—31P—is the only stable isotope that exists. In practice, however, SIP experiments almost exclusively use 13C as the isotope of choice, with a tiny minority using 18O and 15N. The choice of substrate and stable isotope in a SIP experiment is of course directly related to the metabolic process or microbial guild of interest. Naturally, in SIP, target microbes can only be isotopically labeled through assimilatory processes. This is somewhat unfortunate because many of the microbially mediated biogeochemical processes of interest are

Experimental Set-up and Data Analysis

3

Table 2 Number of additional neutrons per nucleotide in a DNA or RNA molecule given full labeling (when all respective atoms are replaced by a heavier stable isotope) Carbon-13

Oxygen-18

Nitrogen-15

Deuterium

Adenine

10

10/12

5

9

Guanine

10

12/14

5

10

Cytosine

9

12/14

3

9

Thymine/Uracil

10/9

14/16

2

8

Mean

9.75/9.5

12/14

3.75

9

energy-yielding dissimilatory processes, involving only electron transfer between two compounds and leave no trace in the biomass. In such cases, the microbial guild of interest can only be labeled indirectly through an assimilatory process that is powered by the dissimilatory process of interest (e.g., using 18O-H2O or 13C-CO2 as general substrates for all active organisms and for autotrophs, respectively). Beyond the question of which biological process or microbial target group to study, the different stable isotopes used for SIP differ in their ability to label nucleic acids and therefore lead to buoyant density (BD) changes. Table 2 lists and compares the number of additional neutrons gained per nucleotide in a DNA or RNA molecule by replacing all the atomic positions of a particular element with its heavier stable isotope. Theoretically, the highest mass increase from labeling is achieved by using 18O, with added 12 or 14 neutrons on average for a hypothetical DNA or RNA molecule, respectively. This is, of course, thanks to the fact that labeling with 18O adds two neutrons per atom compared to only one for either 13C, 15N, or D, therefore leading to higher overall mass increase despite the lower number of atoms in the molecule. In contrast, N is, unfortunately, the rarest in nucleic acids compared to C, O, or H and labeling with 15N can lead to a maximum of 3.75 added neutrons per base, on average, or 2.5 times less in mass increase compared to labeling with 13C. This was confirmed experimentally already over 40 years ago when it was shown that fully 15 N-labeled DNA in CsCl has a BD gain of ca. 0.016 g/ml compared to a BD gain of ca. 0.036 g/ml with 13C [1]. Similarly, RNA fully labeled with 15N showed a BD gain of ~0.015 g/ml [2] compared to 0.035 for 13C [3]. The lower maximum mass addition to DNA and RNA through 15N-labeling means a smaller shift of labeled nucleic acids away from unlabeled nucleic acids in an isopycnic gradient compared to 13C-labeling. Still, this more modest shift in BD is nevertheless sufficient to detect labeling in DNA originating from a single organism, as was shown already in the

4

Roey Angel

classical work of Meselson and Stahl [4]. However, for DNA-based SIP this creates a major challenge since double-stranded DNA migrates in a BD gradient not only as a function of its mass but also as a function of its hydration state. The latter is ultimately determined by the G + C content of the DNA and causes an undesired migration of unlabeled high-GC DNA towards the denser regions of the gradient [5]. Already in the first attempts to develop 15N-SIP, it was noticed that due to the relatively small migration of 15N-labeled DNA, unlabeled DNA with high-G + C content could overlap with even fully-labeled DNA of lower G + C content, and obscure the ability to differentiate labeled from unlabeled taxa [6, 7]. This is further intensified by the fact that A-T base pairs contain only seven nitrogen atoms compared to eight in a G-C base pair, resulting in a lower, albeit minor labeling of the A-T base pair [8]. Surprisingly, while 18O labeling should theoretically increase the mass of DNA by 23% and of RNA by 47% compared to labeling with 13C, in practice the observed shifts in BD in 18O-SIP gradients are not much different than in 13C-SIP gradients (0.04 g/ml) [9, 10], indicating that not all positions can be replaced with a heavy isotope. Deuterium has been used in SIP experiments coupled with either Raman microspectroscopy [11] or metabolomics [12], but because of the toxicity of deuterated water (heavy water) at high concentrations, it is probably not suitable for DNA or RNA-SIP. Considering these, it is easy to understand why carbon is the most widely used isotope in SIP. Carbon is abundant enough in biomolecules to allow for easy labeling. In many cases, carbon-based substrates are used for both assimilatory and dissimilatory processes in the cell, so biomass labeling is easily achieved using any of a selection of different substrates. In contrast, many N-transforming processes are dissimilatory, while at the same time many N-assimilation processes are common between different functional groups of microorganisms and therefore provide relatively little differentiating power. Similarly, oxygen is also found abundantly in various terminal electron acceptors used for respiration, which are therefore unsuitable for SIP, or alternatively in water, which is assimilated into the biomass by all known organisms.

3

Setting Up an Experiment SIP experiments are usually relatively complex, laborious, and timeconsuming, and can, therefore, fail because of various reasons and at different stages. Thus, the experimental design of a SIP experiment should be carefully considered in advance and cover all aspects and phases, including preliminary knowledge of the environment

Experimental Set-up and Data Analysis

5

and the targeted process, the nature and duration of the incubation, through possible pitfalls, and down to the desired method of data analysis. Before deciding on a SIP experiment, it is important to gain some preliminary knowledge of the system in question and the microbial guild to be targeted. For SIP to be successful, sufficient substrate needs to be processed and assimilated by the microbes during the incubation period. Therefore, one of the first and most important preliminary tests to perform is to measure the rate and dynamics of the process in question to estimate the length of the incubation period that is needed. Although the relationship between substrate consumption and level of labeling depends on the assimilation efficiency and the size of the active microbial guild and is therefore difficult to establish, some insights and ballpark estimates can nevertheless be made. Also, it is advisable to measure the enrichment level of the total DNA or RNA extracted from the sample to assess if detection of labeled microbes will be feasible [2, 13, 14]. Again, while it is impossible to draw a general direct relation between the level of enrichment of nucleic acids and the outcome of the SIP, because this will depend on whether or not the label is concentrated within a small group of highly labeled microbes or shared among many members, but a qualitative relationship can nevertheless easily be drawn for specific environments and microbial guilds. 3.1 Which Bio-molecule to Target

The term SIP was first attributed to the method for identifying labeled microbes through the incorporation of a stable isotope into their DNA [15]. While this is still the most commonly used ’flavor’ of SIP, other types of SIP are also popular, since in essence nearly every stable bio-molecule in the cell can be used as a target for SIP. Targeting DNA is advantageous because DNA is the gold-standard for taxonomic classification of organisms and for hypothesising about potential functions. It is also popular because DNA amplification and sequencing technologies are affordable and wide spread in most molecular and microbiological labs. The development of a protocol for targeting RNA instead of DNA in a SIP experiment [13] then quickly followed. RNA-SIP offers the same taxonomic resolution power as DNA-SIP but because RNA synthesis is uncoupled to cell replication, it offers higher sensitivity, though at the cost of a somewhat more laborious and sensitive lab work. A further advantage of RNA-SIP is that unlike DNA, RNA does not migrate based on its G + C content in a density gradient, so the potential for detecting false-positives is theoretically lower (see Subheadings 2, 3.6, 4.3, and Chapter 9). Targeting PLFA [16] is another popular way for running SIP that even predates the use of DNA-SIP for detecting active microbes in the environment. Because of the use of an isotope-ratio mass spectrometer (IRMS), which is capable of a much finer mass separation compared to density gradient, PLFA-SIP offers significantly higher sensitivity over DNA- or RNA-SIP, which can be important when studying

6

Roey Angel

organisms with very low specific activity such as deep subsurface microorganisms [17] or bacteria that oxidize atmospheric methane [18]. However, in addition to excluding the use of 15N-labeled substrates, PLFA inherently offers a much more limited capacity for taxonomic affiliation of microbes compared to DNA or RNA and can only differentiate between groups at broad level [19]. Targeting proteins and metabolites is also an option (e.g. [12, 20]), thus providing a direct and unquestionable proof of processing a labeled substrate. However, these methods are very laborious, low throughput, and require significant in-house experience in sample processing, and analysis of the output data. Lastly, identification of isotopically labeled microbes at the single-cell levels is also gaining interest lately using tools such as NanoSIMS [21] and SIP-Raman [22] microspectroscopy, however, their application is still limited because they are costly, low-throughput, and relay on equipment that is found in only a handful of labs around the world. 3.2 Duration of Incubation

As mentioned, incubation length will depend on the one hand on the rate at which the process in question is proceeding and its specific assimilation efficiency. Incubation in the presence of the labeled substrate should allow enough time for the nucleic acids to become sufficiently labeled to be detected above the background. For very fast processes such as water uptake, incubation time can be as short as a few hours [10, 23], while for very slow processes, such as nitrogen fixation, incubation can be as long as several days to weeks [2, 24, 25]. Incubation time should also vary if targeting DNA or RNA. Labeling of RNA can be detected earlier because it does not require cell replication and because its synthesis is not semi-conservative as DNA replication (although this does not preclude a significant dilution of newly synthesized RNA with light isotope as a result of recycling of building blocks within the cell). In general, it is assumed that DNA or RNA molecules should be labeled to at least 30 atom% to differentiate them from unlabeled molecules in a BD gradient [8, 26]. On the other hand, long incubation times bear the risk of labeling community members that do not perform the metabolic activity in question but were labeled through cross-feeding. Because microbes are interlinked through a network of trophic interactions, any labeled element will eventually be spread among many members of the community, regardless of how specific the process in question is. Cross-feeding in isotope-labeling experiments has been acknowledged from the start and has been shown for nitrogen as well as carbon (e.g., [27, 28]). Although typically considered to be an unwanted side effect in SIP experiments, cross-feeding has also been taken advantage of several times to study substrate flow patterns microbial interactions on a temporal scale [29, 30]. Since cross-feeding in a microbial community cannot simply be put to a halt, the typical way of dealing with this issue is to sample at several time points, limit the incubation time to the minimum necessary for labeling, and

Experimental Set-up and Data Analysis

7

combine complementary lines of evidence when concluding that a specific taxon indeed performs the metabolism in question. 3.3 Substrate Enrichment Level and Concentration

Substrates used in SIP experiments are in almost all cases “fully” labeled, i.e., all positions are enriched with the labeled isotope to the highest level possible (>97 atom%). This, of course, stems from the need to achieve high levels of labeling in nucleic acids to detect labeled microbes. However, labeling of carbon only at specific positions could also be employed, for example, to study microbial guilds that would attack the substrate at a specific position of interest, while excluding others. The substrate concentration can also affect the rate and strength of labeling; however, presenting a sample with unrealistically high concentrations can lead to undesired consequences such as drastic community changes or a rapid enrichment of a fast-growing sub-population with low substrate affinity. Therefore, it is best to remain within the range (typically on the higher end) of substrate concentrations that are expected to be found in the environment.

3.4 Amount of Nucleic Acids to Load

Typical DNA-SIP gradients are prepared with 0.5–5 μg of DNA, but there does not seem to be a hard limit for the amount of DNA that can be loaded on a gradient. In RNA-SIP gradients, overloading with RNA will cause aggregation that will prevent efficient separation. The typical recommended amount is around 500 ng for a 5.5-ml gradient [31]. However, this issue was never been studied systematically. For PCR purposes, this amount should be more than enough to target the rRNA or any other functional gene. However, for metagenomic or metatranscriptomic sequencing of the fractions, larger amounts of the template will be needed. This can be achieved either by pooling together several fractions from several different gradients or by multiple displacement amplification (e.g., [32]). In addition, because rRNA accounts for over 80% of the total rRNA in a bacterial cell (with SSU rRNA alone accounting for ca. 27%), enrichment of mRNA might be needed for metatranscriptomic analysis [33].

3.5 Number of Fractions to Collect, and Sequencing Depth

Regardless of which method is used for analyzing the data, success in a SIP experiment is determined by the ability to detect microbial phylotypes that are present in the denser fractions of a labeled gradient and are either absent or have lower abundance in the lighter fractions of the same gradient, or in the denser fraction of a control gradient. The detection limit in SIP experiments is itself not a fixed value but will depend on the sequencing depth, on the number of fractions being collected from each gradient, and on which method is being used to analyze the data (see Subheading 4). Using state of the art sequencing technologies, it is now easy to obtain thousands of sequences per fraction. However, this, of course, comes at a cost, which might not be necessary. It is therefore advisable, if possible, to first obtain an estimate of the size of the microbial guild in question

8

Roey Angel

compared to the total microbial population, using for example qPCR or fluorescent microscopy. The smaller the size of the target community, the harder it will be to detect its labeling above the detection limit. Naturally, this will almost inevitably be an overestimation of the required sequencing depth since the target population is expected to be enriched in the denser fractions of the gradient, but this will at least give a rough estimate for the sequencing depth needed. The number of fractions collected can also affect the detection limit. While a higher number of fractions will most likely increase the sensitivity, it also entails higher sample processing efforts and costs. In addition, more fractions also mean less template per fraction and thus also an increased difficulty to amplify the target and a higher chance of contamination with foreign nucleic acids from the environment. Typically 12–20 fractions are collected, of which about 10–16 end up being analyzed because the lightest and heaviest fractions contain little to no nucleic acids. 3.6 Unlabeled Controls

As in any lab experiment, appropriate controls should be set up in parallel to minimize the detection of false-positives. Many of the older published works included only one or two no-label controls, usually at the last time point or at the highest amendment level. Recently, however, particularly with the growing use of highthroughput sequencing and statistical models to detect labeled OTUs the need to include more no-label controls in the experiment to correctly detect labeled phylotypes has been growing, but on the other hand also became easier to achieve. The exact number and type of no-label controls will depend on the exact statistical method used to analyze the data, but also on the type of SIP being performed since DNA-SIP is more prone to detecting false positives than RNA-SIP because of the effect of the G + C-content on DNA BD (see Subheading 4). Ideally, every labeled sample will have its parallel no-label control. However, this is very laborious and costly, and might not be needed. Since RNA-SIP does not suffer from the bias caused by G + C-based migration as in DNA-SIP, it is possible to compare fractions within a gradient, rather than between gradients, and thus reduce the number of controls (see Subheading 4). Similarly, methods that are only interested in identifying labeling of a phylotype (e.g., differential abundance) but not necessarily quantifying it (e.g., qSIP) remain robust even when some controls are omitted (see Subheading 4 and Chapter 11).

3.7

Traditionally a vertical rotor was preferred over a fixed-angle one for SIP experiments because it provides a shallower gradient and therefore a higher degree of separation between densities. Recent modelling work suggests, however, that this comes at the cost of a higher diffusion of nucleic acids throughout the gradient, thus leading to a higher background [34]. Both rotor types were successfully used in SIP, but to date, no experimental comparison has been published.

Type of Rotor

Experimental Set-up and Data Analysis

4

9

Data Analysis: Analysis of Barcoded Amplicon Data for SIP Arguably, the most significant advancement in the field of DNAand RNA-SIP in recent years came from the introduction of highthroughput sequencing techniques and their adoption to the study of microbial communities using barcoded amplicon sequencing [35–37]. The ability to sequence dozens of samples simultaneously to a very high depth meant that it was now possible to identify rare taxa that were labeled but also taxa that are only partially labeled. Before the adoption of high-throughput sequencing (HT-sequencing), successful labeling of DNA or RNA was done visually, either by detecting a second band of nucleic acids under UV light following ethidium bromide staining or fractionating the gradient into multiple fractions, amplifying the nucleic acids using PCR or qPCR and evaluating the intensity of the bands or copy numbers. The use of fingerprinting techniques such as DGGE and TRFLP enabled not only a more sensitive comparison between fractions but also a direct, albeit qualitative, insight into how many phylotypes were labeled. However, it still suffered from low resolution and a high degree of noise that are inherent to these methods. Moreover, the unequivocal identification of the labeled microbes was still low-throughput, laborious, and costly since it required the construction of clone libraries followed by Sanger sequencing. Barcoded amplicon sequencing allows for robust, semi-quantitative comparison of different fractions along a density gradient, as well as an identification of the identity of which microbes became labeled and which did not. Moreover, the ability to obtain thousands of sequences per sample meant that even labeling of minor members of the community could be detected—something that could not be achieved with standard molecular fingerprinting techniques or Sanger sequencing. The adoption of HT-sequencing technologies also called for new analytical methods that could take advantage of this increase in sensitivity through statistical modelling and enable robust detection of either minor or partially labeled members of the active guild [38, 39]. However, alongside with added sensitivity barcoded amplicon sequencing also presents some challenges for comparing samples because it is difficult to control the number of sequences per sample, also known as the library depth. The problem is not unique to analyzing SIP experiments and poses a major analytical challenge in the field of microbiome studies and comparative transcriptomics (RNA-Seq). In essence, most statistical methods used for comparison assume that across different samples, templates with identical relative abundance should have equal chances of being sequenced and thus any observed differences are an indication that the true abundance of the given sequence differs between the samples. In ecology, the issue is known as “sampling effort.”

10

Roey Angel

Traditionally, the most common way to alleviate the problem of unequal sequencing depths was to randomly sub-sample sequences from each sample down to the smallest sample size so that all samples become equal (a process sometimes called “rarefaction”). This practice, however, came under scrutiny in recent years and sparked some heated polemic papers on how to best handle microbiome data [40]. While the severity of the bias caused by random sub-sampling is debated, it is generally accepted that this is a sub-optimal way to deal with the problem. Another common approach is to convert all abundances to relative abundances and compare the different sequences on a fraction (or percentage) basis. This, however, leads to other problems since it maintains the correlation between the sequencing depth and the number of unique sequences (or OTUs) while at the same time drastically reducing the number of degrees of freedom by coercing the sum of abundance in each sample to 100% [41]. More recent methods try to “eat the cake and leave it whole” by attempting to equalize the variance between samples through a scaling factor while not discarding any data (covered in [42]). Whichever method is chosen, it is important to remember that no statistical trick can solve the inherent problems that stem from large differences in library sizes and these should be handled at the level of sample preparation or sequencing and not data analysis. 4.1 Differential Abundance Analysis and Quantitative Analysis

The most common methods for comparing fractions in SIP experiments were developed for analyzing RNA-Seq datasets. The parallels are apparent; typical RNA-Seq experiments are designed as a case-control study and the analytical challenge is to identify which sequences are differentially expressed (either up-regulated or down-regulated) compared to the control, while overcoming the natural variance and differences in library sizes. Similarly, in SIP experiments one would like to identify which sequences are “differentially abundant” in the fractions where labeled nucleic acids are expected to be present compared to those where unlabeled nucleic acids are present. An important difference to RNA-Seq experiments is, however, that only enriched sequences in the ‘heavy’ fractions are of interest, while depleted sequences should only occur when labeling is strong enough to displace unlabeled sequences from the ‘light’ fractions to a noticeable degree. Nearly all existing data analysis methods should apply to both DNA- and RNA-SIP, albeit with some differences. This book offers two recent and very robust ways to analyze SIP datasets: quantitative SIP (qSIP; see Chapter 11) and high-resolution SIP (HR-SIP; see Chapter 9). Both yield similar results, but they nevertheless differ in some details (discussed in [43]). While high-resolution SIP, like all other differential abundance methods, aims only at detecting labeled phylotypes, qSIP also attempts to quantify the level of enrichment per phylotype, but requires additional quantitative

Experimental Set-up and Data Analysis

11

data from qPCR and also a matching unlabeled control sample for every labeled sample, to reliably detect growth. 4.2 Data Analysis for RNA-SIP Experiments

Since both HR-SIP and qSIP are carefully detailed in this book, repeating the steps here would be redundant. However, because the methods were published for DNA-SIP, some differences to RNA-SIP should be noted. In principle, both methods rely on a comparison of the gradient fractions from labeled samples to those from unlabeled control samples (between-gradient comparison). Moreover, both assume and make use of the fact that while DNA and RNA will concentrate around their theoretical BD, they diffuse throughout the gradient in a Gaussian shape so that amplifiable amounts of nucleic acids are present in every fraction in the gradient [2, 34]. However, because the course of development of a microbial community is controlled by stochastic processes in addition to deterministic ones, parallel incubations from the same parent community often lead to different communities after a while, even if conditions are kept as similar as possible. Consequently, it was demonstrated that these stochastic variations reduce the detection accuracy and it was recommended that the Bray-Curtis dissimilarity between communities of labeled and unlabeled samples that are being compared should ideally be >0.2 [34]. Between-gradient comparisons are crucial for DNA-SIP because as mentioned above, the DNA of different taxa will migrate in the gradient also based on their G + C content. Moreover, the migration based on G + C content is not constant per phylotype. Instead, it will vary based on the size of the DNA fragment surrounding the gene of target, which varies stochastically in most DNA extraction methods [7]. In RNA-SIP, however, the buoyant density of RNA is less affected by G + C content, and one can assume that in a gradient from an unlabeled sample the relative abundance of each taxon should remain relatively constant throughout the different fractions. In contrast, in a gradient from a labeled sample, some taxa will be more abundant in the heavy fractions compared to the lighter ones, while the relative abundance of unlabeled taxa will remain constant throughout the gradient or decline in the heavy fractions if the labeled taxa make up a significant proportion of the entire community. Therefore, since in RNA-SIP differential migration of taxa is only expected as a response of labeling, detection of labeled taxa can also be done in a within-gradient fashion by comparing the relative abundances of taxa in the heavy fractions (i.e., ca. 1.72–1.76 g/ml for DNA-SIP or 1.80–1.84 g/ml for RNA-SIP) with those in the light fractions (i.e., ca. 1.68–1.72 g/ml for DNA-SIP or 1.77–1.80 g/ml for RNA-SIP). However, some label-free controls should nevertheless be set up (e.g., paralleling the beginning and end time points or the highest and lowest treatment extremes) and analyzed because they can help to fine tune the statistical cutoff parameters so that false positives can be avoided [2].

12

Roey Angel

4.3 Network Analysis Using SIP Data

Network analysis—the prediction of microbial associations from presence-absence or abundance data—is gaining popularity in ecological studies in general and microbiome studies in particular [41]. This type of analysis has also been used in concert with SIP to detect, for example, positive and negative correlation between phylotypes of ammonia-oxidizing archaea, nitrite-oxidizing bacteria, and methanotrophs [44], clusters of anaerobic and aerobic bacteria in rewetted biological soil crusts [10], or to identify community members that interact with methane-oxidizing bacteria [45]. However, in contrast to a standard network analysis on microbiome data, the interpretation of the results from a SIP experiment may not be so straightforward. First, most probably only the “heavy” fractions from the labeled gradients should be analyzed because changes in the “light” fractions are either already reflected in the “heavy” fractions (i.e., phylotypes becoming labeled and hence depleted in the “light” fractions), or not directly related to substrate incorporation (e.g., growth and death of phylotypes in the general community). Secondly, while the interpretation of positive correlations in the heavy fractions are relatively easy to interpret (i.e., two phylotypes acquire label under similar conditions), it is not entirely clear what negative interactions mean if anything at all. Thirdly, it is important to bear in mind that network analysis does not reveal the mode of the interaction between two interacting phylotypes and a positive correlating could mean that both use the same substrate, that there is cross-feeding occurring (and thus the interaction is positive-positive or at least positive-neutral), or that one phylotype is praying on another (positive-negative interaction). Lastly, it should be noted that many replicates are required for a network to be stable (at least 25) and that communities should be reasonably similar in all samples [46]. For SIP studies, this probably translates into an analysis of at least 25 “heavy” gradient fractions, coming from both labeled and no-label control incubations. However, when analyzing data from DNA-SIP experiments care should be taken when analyzing multiple fractions from the same gradient since positive correlations could simply be a result of similar G + C contents.

Acknowledgments The manuscript for this chapter was written online using authors. R.A. was supported by BC CAS, ISB & SoWa RI (MEYS; projects LM2015075, EF16_013/0001782—SoWa Ecosystems Research).

Experimental Set-up and Data Analysis

13

References 1. Birnie GD, Rickwood D (1978) Isopycnic centrifugation in ionic media. In: Birnie GD, Rickwood D (eds) Centrifugal separations: molecular and cell biology. Butterworth, Boston, pp 169–217 2. Angel R, Panho¨lzl C, Gabriel R et al (2018) Application of stable-isotope labelling techniques for the detection of active diazotrophs. Environ Microbiol 20:44–61. https://doi. org/10.1111/1462-2920.13954 3. Lueders T, Manefield M, Friedrich MW (2004) Enhanced sensitivity of DNA- and rRNA-based stable isotope probing by fractionation and quantitative analysis of isopycnic centrifugation gradients. Environ Microbiol 6:73–78. https://doi.org/10.1046/j.1462-2920.2003. 00536.x 4. Meselson M, Stahl FW (1958) The replication of DNA in Escherichia coli. Proc Nat Acad Sci 44:671–682. https://doi.org/10.1073/pnas. 44.7.671 5. Rolfe R, Meselson M (1959) The relative homogeneity of microbial DNA. Proc Nat Acad Sci 45:1039–1043. https://doi.org/10. 1073/pnas.45.7.1039 6. Cupples AM, Shaffer EA, Chee-Sanford JC, Sims GK (2007) DNA buoyant density shifts during 15N-DNA stable isotope probing. Microbiol Res 162:328–334. https://doi. org/10.1016/j.micres.2006.01.016 7. Youngblut ND, Buckley DH (2014) Intragenomic variation in G+C content and its implications for DNA stable isotope probing. Environ Microbiol Rep 6:767–775. https:// doi.org/10.1111/1758-2229.12201 8. Cadisch G, Espana M, Causey R et al (2005) Technical considerations for the use of 15NDNA stable-isotope probing for functional microbial activity in soils. Rapid Commun Mass Spectrom 19:1424–1428. https://doi. org/10.1002/rcm.1908 9. Aanderud ZT, Lennon JT (2011) Validation of heavy-water stable isotope probing for the characterization of rapidly responding soil bacteria. Appl Environ Microbiol 77:4589–4596. https://doi.org/10.1128/aem.02735-10 10. Angel R, Conrad R (2013) Elucidating the microbial resuscitation cascade in biological soil crusts following a simulated rain event. Environ Microbiol 15:2799–2815. https:// doi.org/10.1111/1462-2920.12140 11. Berry D, Mader E, Lee TK et al (2014) Tracking heavy water (D2O) incorporation for identifying and sorting active microbial cells. Proc Nat Acad Sci 112:E194–E203. https://doi. org/10.1073/pnas.1420406112

12. Baran R, Lau R, Bowen BP et al (2017) Extensive turnover of compatible solutes in cyanobacteria revealed by deuterium oxide (D2O) stable isotope probing. ACS Chem Biol 12:674–681. https://doi.org/10.1021/ acschembio.6b00890 13. Manefield M, Whiteley AS, Ostle N et al (2002) Technical considerations for RNA-based stable isotope probing: an approach to associating microbial diversity with microbial community function. Rapid Commun Mass Spectrom 16:2179–2183. https://doi.org/10.1002/rcm.782 14. Blazewicz SJ, Schwartz E (2011) Dynamics of 18 O incorporation from H218O into soil microbial DNA. Microb Ecol 61:911–916. https:// doi.org/10.1007/s00248-011-9826-7 15. Radajewski S, Ineson P, Parekh NR, Murrell JC (2000) Stable-isotope probing as a tool in microbial ecology. Nature 403:646–649. https://doi.org/10.1038/35001054 16. Boschker HTS, Nold SC, Wellsbury P et al (1998) Direct linking of microbial populations to specific biogeochemical processes by 13Clabeling of biomarkers. Nature 392:801–805. https://doi.org/10.1038/33900 17. Pelz O, Chatzinotas A, Zarda-Hess A et al (2001) Tracing toluene-assimilating sulfatereducing bacteria using 13C-incorporation in fatty acids and whole-cell hybridization. FEMS Microbiol Ecol 38:123–131. https://doi.org/ 10.1111/j.1574-6941.2001.tb00890.x 18. Knief C, Lipski A, Dunfield PF (2003) Diversity and activity of methanotrophic bacteria in different upland soils. Appl Environ Microbiol 69:6703–6714. https://doi.org/10.1128/ aem.69.11.6703-6714.2003 19. Evershed RP, Crossman ZM, Bull ID et al (2006) 13C-labeling of lipids to investigate microbial communities in the environment. Curr Opin Biotechnol 17:72–82. https://doi. org/10.1016/j.copbio.2006.01.003 20. Jehmlich N, Schmidt F, Hartwich M et al (2008) Incorporation of carbon and nitrogen atoms into proteins measured by protein-based stable isotope probing (Protein-SIP). Rapid Commun Mass Spectrom 22:2889–2897. https://doi.org/10.1002/rcm.3684 21. Li T, Wu T-D, Maze´as L et al (2008) Simultaneous analysis of microbial identity and function using NanoSIMS. Environ Microbiol 10:580–588. https://doi.org/10.1111/j. 1462-2920.2007.01478.x 22. Wang Y, Huang WE, Cui L, Wagner M (2016) Single cell stable isotope probing in microbiology using Raman microspectroscopy. Curr

14

Roey Angel

Opin Biotechnol 41:34–42. https://doi.org/ 10.1016/j.copbio.2016.04.018 23. Blazewicz SJ, Schwartz E, Firestone MK (2014) Growth and death of bacteria and fungi underlie rainfall-induced carbon dioxide pulses from seasonally dried soil. Ecology 95:1162–1172. https://doi.org/10.1890/ 13-1031.1 24. Buckley DH, Huangyutitham V, Hsu S-F, Nelson TA (2007) Stable isotope probing with 15 N2 reveals novel noncultivated diazotrophs in soil. Appl Environ Microbiol 73:3196–3204. https://doi.org/10.1128/ aem.02610-06 25. Pepe-Ranney C, Koechli C, Potrafka R et al (2015) Non-cyanobacterial diazotrophs mediate dinitrogen fixation in biological soil crusts during early crust formation. ISME J 10:287–298. https://doi.org/10.1038/ ismej.2015.106 26. Buckley DH, Huangyutitham V, Hsu S-F, Nelson TA (2007) Stable isotope probing with 15 N achieved by disentangling the effects of genome G+C content and isotope enrichment on DNA density. Appl Environ Microbiol 73:3189–3195. https://doi.org/10.1128/ aem.02609-06 27. McDonald IR, Radajewski S, Murrell JC (2005) Stable isotope probing of nucleic acids in methanotrophs and methylotrophs: a review. Org Geochem 36:779–787. https://doi.org/ 10.1016/j.orggeochem.2005.01.005 28. Adam B, Klawonn I, Svede´n JB et al (2015) N2-fixation ammonium release and N-transfer to the microbial and classical food web within a plankton community. ISME J 10:450–459. https://doi.org/10.1038/ismej.2015.126 29. DeRito CM, Pumphrey GM, Madsen EL (2005) Use of field-based stable isotope probing to identify adapted populations and track carbon flow through a phenol-degrading soil microbial community. Appl Environ Microbiol 71:7858–7865. https://doi.org/10.1128/ aem.71.12.7858-7865.2005 30. Pepe-Ranney C, Campbell AN, Koechli CN et al (2016) Unearthing the ecology of soil microorganisms using a high resolution DNA-SIP approach to explore cellulose and xylose metabolism in soil. Front Microbiol. https://doi.org/10.3389/fmicb.2016.00703 31. Lueders T, Manefield M, Friedrich MW (2003) Enhanced sensitivity of DNA- and rRNA-based stable isotope probing by fractionation and quantitative analysis of isopycnic centrifugation gradients. Environ Microbiol 6:73–78. https://doi.org/10.1046/j.1462-2920.2003. 00536.x

32. Chen Y, Dumont MG, Neufeld JD et al (2008) Revealing the uncultivated majority: combining DNA stable-isotope probing multiple displacement amplification and metagenomic analyses of uncultivated Methylocystis in acidic peatlands. Environ Microbiol 10:2609–2622. https://doi.org/10.1111/j.1462-2920.2008. 01683.x 33. Dumont MG, Pommerenke B, Casper P (2013) Using stable isotope probing to obtain a targeted metatranscriptome of aerobic methanotrophs in lake sediment. Environmental Microbiology Reports 13, 757–764. Wiley 34. Youngblut ND, Barnett SE, Buckley DH (2018) SIPSim: a modeling toolkit to predict accuracy and aid design of DNA-SIP experiments. Front Microbiol. https://doi.org/10. 3389/fmicb.2018.00570 35. Xia W, Zhang C, Zeng X et al (2011) Autotrophic growth of nitrifying community in an agricultural soil. ISME J 5:1226–1236. https:// doi.org/10.1038/ismej.2011.5 36. Lee TK, Lee J, Sul WJ et al (2011) Novel biphenyl-oxidizing bacteria and dioxygenase genes from a Korean tidal mudflat. Appl Environ Microbiol 77:3888–3891. https://doi. org/10.1128/aem.00023-11 37. Pilloni G, von NF, Engel M, Lueders T (2011) Electron acceptor-dependent identification of key anaerobic toluene degraders at a tar-oilcontaminated aquifer by Pyro-SIP. FEMS Microbiol Ecol 78:165–175. https://doi.org/ 10.1111/j.1574-6941.2011.01083.x 38. Zemb O, Lee M, Gutierrez-Zamora ML et al (2012) Improvement of RNA-SIP by pyrosequencing to identify putative 4-n-nonylphenol degraders in activated sludge. Water Res 46:601–610. https://doi.org/10.1016/j. watres.2011.10.047 39. Zumsteg A, Schmutz S, Frey B (2013) Identification of biomass utilizing bacteria in a carbon-depleted glacier forefield soil by the use of 13C DNA stable isotope probing. Environ Microbiol Rep 5:424–437. https://doi. org/10.1111/1758-2229.12027 40. McMurdie PJ, Holmes S (2014) Waste not want not: why rarefying microbiome data is inadmissible. PLoS Comp Biol 10:e1003531. https://doi.org/10.1371/journal.pcbi. 1003531 41. Faust K, Raes J (2012) Microbial interactions: from networks to models. Nat Rev Microbiol 10:538–550. https://doi.org/10.1038/ nrmicro2832 42. Weiss S, Xu ZZ, Peddada S et al (2017) Normalization and microbial differential abundance strategies depend upon data characteristics.

Experimental Set-up and Data Analysis Microbiome. https://doi.org/10.1186/ s40168-017-0237-y 43. Youngblut ND, Barnett SE, Buckley DH (2018) HTSSIP: an R package for analysis of high throughput sequencing data from nucleic acid stable isotope probing (SIP) experiments. PLoS One 13:e0189616. https://doi.org/10. 1371/journal.pone.0189616 44. Daebeler A, Bodelier PLE, Yan Z et al (2014) Interactions between Thaumarchaea Nitrospira and methanotrophs modulate autotrophic nitrification in volcanic grassland soil. ISME J

15

8:2397–2410. https://doi.org/10.1038/ ismej.2014.81 45. Ho A, Angel R, Veraart AJ et al (2016) Biotic interactions in microbial communities as modulators of biogeochemical processes: methanotrophy as a model system. Front Microbiol. https://doi.org/10.3389/fmicb.2016.01285 46. Berry D, Widder S (2014) Deciphering microbial interactions and detecting keystone species with co-occurrence networks. Front Microbiol. https://doi.org/10.3389/fmicb.2014.00219

Chapter 2 DNA-Based Stable Isotope Probing Zhongjun Jia, Weiwei Cao, and Marcela Herna´ndez Garcı´a Abstract Microbiomes on Earth are often considered the most heterogeneous biological entities, but their vital roles in driving global biogeochemical cycles often remain elusive. DNA-based stable isotope probing (DNA-SIP) provides a powerful means to establish a direct link between biogeochemical processes and the taxonomic identities of active microorganisms involved in the processes. Combined with highthroughput sequencing, it significantly aids in deciphering ecophysiological functions of active microorganisms at the level of microbial communities. DNA-SIP relies solely on the propagation of targeted microbial communities, during which the entire genomes of daughter cells are synthesized and increasingly 13 C-labeled. This growth on 13C-labeled substrate in association with cell division provides solid evidence for the functional importance and metabolic potential of targeted microorganisms. The essential prerequisite for a successful DNA-SIP experiment is the identification, with confidence, of isotopically enriched 13 C-DNA, of which the amount is generally too low to allow for the direct measurement of 13C atomic percent of nucleic acid. The 13C labeling can be readily identified in the fractionated DNA by quantification of functional genes specific to the known targeted microorganisms, and by high-throughput sequencing of the total microbial communities via 16S rRNA genes without prior knowledge of which microorganisms are 13 C-labeled (i.e., highly enriched in the heavy fractions relative to 12C (natural isotope abundance) control treatments). In this chapter, the protocol for obtaining DNA highly enriched in heavy isotope is presented using diazotrophic methanotrophs in a paddy soil as a case study. Key words Microbial ecophysiology, DNA-SIP, 13C-DNA, Next-generation sequencing

1

Introduction A fingernail-sized patch of soil contains billions of microbial cells, the genetic resource of which is suggestively ~1000 times richer than the human genome. However, the majority of microbial communities have escaped cultivation so far and their functions remain poorly understood. DNA-based stable isotope probing (DNA-SIP) is a powerful means to narrow down the sequence space by targeting the active part of the community, particularly for those uncultured microorganisms. DNA-SIP was invented in 2000 by Colin Murrell and colleagues [1]. The principle of DNA-SIP can be traced back to a historical experiment that

Marc G. Dumont and Marcela Herna´ndez Garcı´a (eds.), Stable Isotope Probing: Methods and Protocols, Methods in Molecular Biology, vol. 2046, https://doi.org/10.1007/978-1-4939-9721-3_2, © Springer Science+Business Media, LLC, part of Springer Nature 2019

17

18

Zhongjun Jia et al.

demonstrated the semiconservation mechanisms of DNA replication in 1958 [2]. Instead of using a single bacterium in the classic Meselson–Stahl experiment [2], DNA-SIP targets a microbial community grown on the 13C-labeled substrate within natural environments or samples. By feeding environmental samples with 13 C-labeled substrates, the genomes of microorganisms assimilating 13C-substrate become increasingly labeled, and the success of DNA-SIP depends on the separation of isotopically labeled DNA by isopycnic density gradient centrifugation of total genomic DNA isolated from SIP microcosms at the end of incubation. The first DNA-SIP experiments used ethidium bromide incorporated in the CsCl gradient to visualize under ultraviolet light the position of the labeled DNA, which could be retrieved with a needle and syringe [1]. It has been shown that 0.1 μg 13 C-DNA is too low to be visualized, suggesting the inherent difficulty of visualizing the labeled DNA by ultraviolet light after ultracentrifugation [3, 4]. To circumvent this problem, Friedrich and Lueders in 2004 developed a more sensitive protocol for the detection of “light” and “heavy” DNA in fractions of centrifugation gradients [5]. It involves the collection of DNA across the entire density range of the SIP gradient, usually separating the gradient into about 15 fractions. The distribution of DNA from the total microbial community throughout the gradient was then quantified by domain-specific PCR of 16S rRNA genes. Unless a large amount of 13C-substrate is consumed resulting in a heavy enrichment of the target population, usually the 13C-DNA cannot be detected on the basis of domain-specific PCR quantification of 16S rRNA genes in “heavy” fractions from 13C-treatment relative to 12C-control. This is particularly evident for microbially mediated processes by phylogenetically restricted groups, and can be circumvented by PCR quantification of biomarker genes specific to the known targeted microorganisms. For instance, Jia and Conrad quantified the distribution patterns of amoA genes specific for microbial ammonia oxidizers, and clearly identified amoA genecontaining 13C-DNA across the density gradient [6]. The resolution of labeled communities was further enhanced by highthroughput sequencing of 16S rRNA genes at the domain-level without prior knowledge of targeted microorganisms [7]. The community analyses of “light” and “heavy” DNA fractions can thus reduce the bias associated with the constant low background of unspecific nucleic acids in heavy DNA fractions. Almost all biologically important elements, except phosphorus, have two or more stable isotopes. Anabolism is the key feature of microbial life, and the difference of buoyant density in DNA relies on the element proportion. For example, lower relative abundance of nitrogen compared with carbon atoms in nucleotide molecules results in smaller differences in buoyant density between 13C and 15 N-fully labeled DNA. To resolve the heavy isotope-labeled from

DNA-SIP

19

Table 1 Elements that constitute nucleic acids and some properties that are important for SIP Nucleotide unit Deoxyribonucleotide

C

H

N

O

P

MW 12 14 C N

A: Deoxyadenylate, dAMP

10

14

5

6

1

G: Deoxyguanylate, dGMP

10

14

5

7

T: Deoxythymidylate, dTMP

10

15

2

C: Deoxycytidylate, dCMP

9

14

3

Increase in MW, % 13

15

331.2

3.019

1.510

1

347.2

2.880

1.440

8

1

322.2

3.104

0.621

7

1

307.2

2.930

0.977

C

N

Natural abundance of isotope (atom %) Light isotope

12

Natural abundance, %

98.93

Heavy isotope

13

Natural abundance, %

1.07

C

C

1

14

16

31

99.99

99.63

99.76

100

2

15

18

0.37

0.20

H

N

H

N

0.01

O

P

O

DNA buoyant density (BD) of genomic DNA with different GC content BD increase g/mL

GC content (%)

BD, g/mL 12 14 C N

30

1.6894

0.0509

0.019

50

1.7090

0.0510

0.020

70

1.7286

0.0510

0.020

13

C

15

N

nonlabeled DNA by ultracentrifugation, other technological considerations are also required including (1) the degree to which DNA is labeled by heavy isotopes and (2) the G+C content of microbial genome. As shown in Table 1, in theory, compared with deoxyribonucleotides composed of 12C14N, 13C-labeled deoxyribonucleotides are increased by 2.8%–3.1% in molecular weight and a corresponding 0.051 g/mL increase in buoyant density. In contrast, 15N-labeled deoxyribonucleotides are increased by about 0.62–1.51% in molecular weight and 0.02 g/mL in buoyant density. It should be noted that the observed buoyant density varies across different labs [2, 3, 5, 8]. It should also be emphasized that G + C content of microbial genomes play a crucial role in 13C-DNA separation [5]. The G + C content of microbial taxa inhabiting in natural environments may vary from 30% to 70%, or even more, and buoyant density of DNA in CsCl medium is proportional to its G + C content [9]. The deviation of DNA buoyant density between 30% and 70% G + C content can reach 0.0392 g/mL as shown in Table 1, almost twice by which 15N-labeled deoxyribonucleotides

20

Zhongjun Jia et al.

increase. The secondary centrifugation after combining firstly isolated heavy fraction DNA with bis-benzimide is also useful to disentangle the effects of genome G + C content [10, 11]. Without prior knowledge of which SIP gradient fractions are highly enriched with the labeled DNA of active microorganisms following isopycnic density gradient ultracentrifugation [5], high-throughput sequencing of universal 16S rRNA genes targeting the total microbial communities can greatly facilitate the interpretation of isotopically enriched DNA from less abundant functional groups [7]. The higher proportion of certain phylotypes in heavy DNA fractions indicates labeling of these microorganisms during 13C-SIP incubation when compared to 12C (i.e., natural 13C abundance) control treatment. For instance, the 16S rRNA genes of nitrite-oxidizing bacteria accounted for 42.1% of the total sequence reads in the “heavy” fraction from the labeled treatment, while only 0.07% was observed in the corresponding “heavy” fraction from the 12C control [7]. Furthermore, the relative abundance of 16S rRNA gene sequences of methanotrophs could be as high as 90% of the total 16S rRNA gene sequences in the “heavy” DNA fractions in labeled treatment, while only 0.21% can be observed in 12C-control treatment [12]. Amplicon sequencing of the total 16S rRNA gene pool allows for the identification of putatively labeled microorganisms with an unprecedented level of detail. Furthermore, metagenomic analysis will allow for the possible discovery of novel genes and reconstruction of novel metabolisms in natural environments [3, 13, 14].

2

Materials

2.1 Reagent Preparation and Setup

1. 0.5 M EDTA solution (pH 8.0): Dissolve 186.1 g of disodium ethylenediamine tetraacetate dihydrate (EDTA) in 900 mL of deionized water, add 2 M NaOH to adjust the pH to 8.0 and make up to 1000 mL with water. Sterilize in an autoclave. 2. 1 M Tris–HCl (pH 8.0): Dissolve 121.1 g Tris base in 800 mL deionized water, add HCl (%) to adjust pH to 8.0, and make up to 1000 mL with deionized water. 3. Tris–EDTA buffer (10 mM Tris–HCl and 1 mM EDTA, pH 8.0): Add 2 mL of 1 M Tris–HCl (pH 8.0) and 0.4 mL of 0.5 M EDTA (pH 8.0) to 197.6 mL deionized water, autoclave. 4. Gradient Buffer (GB, 0.1 M Tris, 0.1 M KCl, and 1 mM EDTA): Mix 50 mL of 1 M Tris–HCl, 3.75 g KCl, and 1.0 mL of 0.5 M EDTA with 400 mL deionized water, dissolve KCl and then add to deionized water to a total volume of 500 mL. Filter-sterilize (0.22 μm) or autoclave. 5. 70% ethanol: Add 370 mL of 95%(v/v) ethanol to 130 mL of deionized water.

DNA-SIP

21

6. CsCl solution: Dissolve 603 g CsCl with GB to a total volume of 500 mL. The final density is usually 1.88–1.89 g/mL at 20
C (see Note 1). 7. Polyethylene Glycol 6000 (30% PEG 6000 and 1.6 M NaCl) solution: Dissolve 150 g of PEG 6000 with 46.8 g NaCl in deionized water to a total volume of 500 mL. Autoclave and then filter-sterilize (0.2 μm). 8. GelRed® Nucleic Acid Gel Stain. 9. Luria–Bertani (LB) broth. 10. High-Profile 96-Well PCR Plates, Bio-Rad. 11. Fast DNA Spin Kit for Soil, MP Bio. 12.

13

C-methane (>98% purity).

13. pmoA primers for qPCR: A189F: 50 -GGNGACTGGGA CTTCTGG-30 , and mb661r: 50 -CCGGMGCAACGTCYTTA CC-30 , 14. SYBR Green Jump-Start™, Taq ReadyMix™. 15. 515F primer for MiSeq Illumina: 50 -TCG TCG GCA GCG TCA GAT GTG TAT AAG AGA CAG XXXXXXXX GTG CCA GCM GCC GCG G-30 (the adapter is underlined, and the barcode is indicated with X’s). 16. 907R primer for MiSeq Illumina: 50 -GTC TCG TGG GCT CGG AGA TGT GTA TAA GAG ACA GCC GTC AAT TCM TTT RAG TTT-30 (the adapter is underlined). 17. Bovine serum albumin (BSA). 18. 2 Premix (TaKaRa Biotech, Tokyo, Japan). 2.2

Equipment

1. Ultracentrifuge: Beckman Coulter, Optima XPN-80. 2. Ultracentrifuge rotor: VTi 65.2 rotor, 16  5.1 mL. 3. Sieve (2 mm mesh). 4. 15-mL screw cap tubes. 5. Ultracentrifuge tubes: 5.1-mL polyallomer (Beckman). 6. Tube sealer: Beckman Coulter. 7. Microfluidic programmable syringe pump. 8. Laboratory clamp and stand. 9. 10-mL and 20-mL syringe, compatible with syringe pump. 10. Rubber tubing (~30 cm, 1.5-mm inner diameter). 11. 26-gauge and 23-gauge syringe needles. 12. Vortex. 13. Refractometer: AR200 digital (Reichert). 14. Microcentrifuge (14,000 rcf).

22

Zhongjun Jia et al.

15. 1.5-mL Eppendorf tubes. 16. 120-mL serum bottles. 17. Butyl rubber stoppers, aluminum caps, crimp sealer. 18. CFX96 Touch Real-Time PCR Detection System, Bio-Rad. 19. Gas chromatograph (GC) with detector for methane (e.g., flame ionizing detector). A CO2 detector is optional (e.g., a methanizer on the GC). 20. NanoDrop (ND-1000). 21. Thermal cycler.

3

Methods

3.1 Microcosm Construction (CH4 Oxidation)

1. Homogenize fresh soil and pass it through 2-mm meshed sieve.

3.1.1 Preparatory Experiment

3. Determine maximum water holding capacity (WHC) and other physiochemical properties of the soil.

2. Place ~2 g fresh soil into 2-mL Eppendorf tube as zero-time control, store at 20
C.

4. Take a part of fresh soils for microcosm construction and preserve the remainder by air-drying. 3.1.2 Incubation Experiment

1. Incubate fresh soil (equivalent to 6.0-g dry weight soil, d.w.s.) at approximately 60% maximum water-holding capacity and 28
C in the dark in a 120-mL serum bottle sealed with butyl stoppers. Inject 1.2 mL of 13CH4 into each bottle after extracting 1.2 mL of air (i.e., the initial methane concentration in bottle is 10,000 ppm). Natural isotope abundance methane (12CH4) treatments should be set up as control. 2. Determine the concentration of CH4, and CO2 if possible, in incubation bottles by gas chromatography at the beginning of incubation and every 8–16 h during incubation until methane is exhausted. 3. Take out soil from bottles and store it at 20
C for total DNA extraction.

3.1.3 DNA Extraction ( See Note 2)

1. Extract total DNA from soil (including zero-time samples and incubation samples. 2. Determine the total DNA concentration and purity (OD260/ OD280 and OD260/OD230) using an ultraviolet spectrophotometer NanoDrop (ND-1000). 3. Quantify the copy numbers of pmoA and 16S rRNA genes by real-time qPCR of the total DNA (see Subheading 3.4.2).

DNA-SIP

3.2 Ultracentrifugation

23

1. Mix 2.0 μg of total DNA from the 13CH4 labeled and control treatments with Gradient Buffer (GB) to a volume of 1.0 mL. 2. Prepare the centrifugation CsCl solution as follows: add 4.9 mL of CsCl, 0.9 mL of GB, and 1.0 mL of GB containing 2.0 μg total DNA to 15-mL screw cap tube. Mix by vortexing. 3. Measure the refractive index of solution using a refractometer. The target value is 1.4029 + 0.0002 (corresponding buoyant density is 1.725 g/mL). If it is too high, add GB (20 μL at a time). If it is too low, add CsCl (20 μL at a time). 4. Transfer 5.1 mL of the solution to an ultracentrifuge tube using a 10-mL sterile syringe (see Note 3). 5. Balance the tubes to within 0.01 g. If necessary, add, or remove, a small amount of the solution to adjust the weight. Seal the ultracentrifuge tubes (a special cap of the centrifuge tube is fastened to the neck of the tube, and it is heat sealed with a tube sealer). Place the balanced tubes symmetrically in pairs into the ultracentrifuge rotor and screw on the nut (note, nut also should be balanced). Ultracentrifuge parameters: 44 h, 20
C, 45,000 rpm (190,000  g), time: hold, Accel: 9; Decel: no break (see Note 4). 6. Take out the rotor and place it in the corresponding groove on the lab table at the end of ultracentrifuge, carefully unscrew the nut, take out the tube vertically and put it on the tube holder (see Note 5). 7. In this step the gradient is fractionated into 15 fractions by displacing the ultracentrifuge tube contents. First, place 15 of 1.5 mL Eppendorf tubes below the ultracentrifuge tube, which is secured with a laboratory clamp and stand (see Notes 6 and 7). Fill a 20-mL syringe with sterile water, and attach the rubber tubing to the end. The 23-gauge needle should be attached to the other end of the needle. The tubing and needle should be filled with water, without air bubbles. Insert the 23-gauge needle into the tube at the shoulder of the ultracentrifuge tube, and fasten the syringe to the pump. Second, pierce a small hole at the bottom of tube with 26-gauge needle, and then discard this needle. Finally, start the syringe pump at a rate of 320-μL/min and collect a new fraction every minute for a total of 15 fractions. This is done by collecting the drops into the 1.5-mL Eppendorf tubes, moving between the tubes so that the gradient is distributed evenly across 15 fractions. 8. Measure the refractive index of the 15 fractions using the refractometer. The buoyant density of each fraction is calculated by the following empirical formula: ρ ¼ 75:9318 þ 99:2031x  31:2551x 2 , where ρ denotes buoyant density (g/mL) and x denotes refractive index (see Note 8).

24

3.3

Zhongjun Jia et al.

DNA Purification

1. Add 550 μL of PEG6000 solution to each DNA fraction, invert by mixing the solution several times, and incubate at room temperature for 2 h or heated for 1 h at 37
C. 2. Centrifuge at 13,000  g for 30 min, remove the supernatant. 3. Rinse the DNA precipitate with 500 μL of ethanol (70%, v/v), centrifuge for 10 min, and remove the supernatant. 4. Repeat step 3 to remove residual CsCl and PEG6000. 5. Keep the Eppendorf tube open at room temperature for 15 min to dry the residual liquid. 6. Dissolve the DNA in 30 μL of TE buffer and then store at 20
C.

3.4 Identification of 13C-DNA 3.4.1 Method 1: Quantification of Total DNA

3.4.2 Method 2: Quantification of Biomarker Genes Specific to Target Microorganisms

1. Measure the DNA concentration in each fraction using a NanoDrop spectrophotometer. Compared with the 12CH4 control, 13CH4 labeled samples may have a higher concentration in the denser gradient fractions. 2. Alternatively, load 5 μL from each fraction onto a 1% agarose gel, and run the electrophoresis. Stain the gel with GelRed® Nucleic Acid Gel Stain and take an image on a UV transilluminator. Brighter bands in the heavy fraction suggest that genomic DNA is 13C-labeled (see Note 9). The identification of 13C-DNA can be performed by quantification of methanotroph functional genes (pmoA), universal 16S rRNA genes, and by high-throughput sequencing of universal 16S rRNA genes (Fig. 1). When compared with the 12CH4 treatments, pmoA copies in 13CH4 treatments peak in heavy fractions, indicating that pmoA gene-carrying methanotrophs assimilated the 13 CH4. Similarly, if the sample was labeled with 15N2 instead of 13 CH4, quantification of nifH genes will indicate if diazotrophic nitrogen fixation played an important role in the microcosm. In this example we describe the protocol for pmoA qPCR. 1. Prepare qPCR standards for pmoA. The genomic DNA of a pure culture of a microorganism containing the target biomarker gene can be used for this. Alternatively, a PCR product or clone of the gene can also be used. For each assay prepare a fresh dilution series from 101 to 107 copies/μL. 2. Prepare SYBR Green quantitative PCR master mix for a final volume of 25 μL. The 25-μL master mix should include 12.5 μL of SYBR Green Jump-Start™ Taq ReadyMix™, 0.5 μM of each primer, 200 ng BSA/μL, 2.5 mM MgCl2, and 2.0 μL template DNA. A volume of 23-μL per sample should be prepared, plus slightly extra to account for pipetting errors. 3. Add 2 μL of the standards in triplicate at different concentrations to a 96-well qPCR plate. Add 2 μL of DNA from each

DNA-SIP

Copy numbers of pmoA, nifH and 16S rRNA genes 0

Relative abundance of MOB-like 16S rRNA reads by MiSeq amplicon sequencing(%)

4x105 8x105 0 5x104 1x105 0 6x106 1x107

1.70 pmoA

nifH

16S rRNA

25

0 15th fraction

20

40

MOB-like

0

20

40

0

Type I

20

40

Type II

14th fraction 13th fraction

CsCl Buoyant Density, g ml-1

1.71

12th fraction 11th fraction

1.72

10th fraction 9th fraction 8th fraction

1.73

7th fraction 6th fraction 5th fraction

1.74

4th fraction 3th fraction

1.75

2th fraction

0.0 0.1

Ratio of pmoA /16S rRNA Ratio of nifH /16S rRNA

0.2

1th fraction 12CH + 14N 4 2

13CH

4+

15N

2

Fig. 1 Distribution of specific functional pmoA genes, nifH genes, universal 16S rRNA genes, relative frequency of the 16S rRNA gene sequences affiliated with methane-oxidizing bacteria (MOB, methanotrophs), type I and type II MOBs in all the fractionated DNA across the entire CsCl buoyant density range of SIP gradients from the 13 CH4- and 12CH4-amended microcosms

gradient fraction into the plate. Add 23 μL of the SYBR Green master mix to the templates. A control is always run with water as template instead of DNA extract. 4. Run the real-time PCR according to the cycling conditions as follows: 95
C for 3.0 min; 40 cycles of 95
C for 10 s, 55
C for 30 s; 72
C for 30 s; 80
C for 5 s with plate read; then melt curve at 65
C to 95
C, incrementing 0.5
C per step including 5 s for plate read. 3.4.3 Method 3: Amplicon Sequencing of Universal 16S rRNA Genes

1. Prepare conventional PCR master mix for a final volume of 25 μL. The master mix includes 12.5 μL of 2 Premix (TaKaRa Biotech), 2.0 μL of forward and reverse primers (each at 0.5 μM final concentration), and 6.5 μL of molecular-grade water is added to reach a 23-μL volume per reaction. 2. Add 2 μL of DNA from each gradient fraction into the plate; prepare triplicates for each sample. A control is always run with water as template instead of DNA extract. 3. Add 23 μL of the PCR master mix to each template. 4. Perform PCR on a thermal cycler using the following program:

26

Zhongjun Jia et al.

95
C for 3 min, 35 cycles of 95
C for 10 s, 55
C for 30 s, 72
C for 30 s, and a final elongation at 72
C for 8 min; hold at 4
C. 5. Pool the triplicate PCR amplicons. 6. Combine the amplicons in an equimolar concentration and send to a sequencing facility. 3.5 Case Study with Soil Labeled with 13CH4 and 15N2

In this example we labeled methanotrophs in soil with 13CH4, and also included 15N2. The SIP gradients were analyzed by qPCR of 16S rRNA, pmoA, and nifH genes. Also, amplicon sequencing of 16S rRNA genes was performed. Quantitative analysis of universal 16S rRNA genes provided no clear evidence for labeling because the majority of 16S rRNA genes occurred in the light fraction, while only a slightly higher peak could be observed in heavy fraction (Fig. 1). However, the highest ratio of pmoA to 16S rRNA genes appeared in the “heavy” fraction, suggesting pmoA gene-carrying methanotrophs were highly enriched in the heavy gradient fractions. Meanwhile the ratios of nifH genes to 16S rRNA genes remained constantly low in the fractionated DNA across the entire buoyant density of the SIP gradients, suggesting that only a small fraction of methanotrophs contain nifH genes. It thus seems plausible that quantification of domain-specific or universal 16S rRNA genes is not applicable for DNA-based SIP of phylogenetically restricted microbial specialists. However, when microbial generalists are targeted for SIP studies, a wide variety of microorganisms are expected to be labeled, and quantitative analysis of domainspecific 16S rRNA genes could be applicable. For example, our unpublished data showed that the majority of soil genomic DNA was labeled and spun down to the “heavy” fraction in SIP microcosms amended with 13C-labeled glucose. In this case, it is difficult to use a single functional biomarker gene or 16S rRNA genespecific primers to infer the distribution of phylogenetically distinct phylotypes that were involved in carbon-assimilation at different trophic levels of glucose decomposition. High-throughput sequencing of universal 16S rRNA genes was performed in all of the fractionated DNA across the entire buoyant density of the SIP gradients (Fig. 1). The results showed clearly that the highest relative abundance of methane-oxidizing bacteria (MOB, methanotrophs) occurred in the fourth to sixth fractions, being similar to quantitative distribution pattern of pmoA gene copies and the ratio of pmoA genes to 16S rRNA genes. A phylogenetic analysis further suggested that both type I and II methanotrophs were labeled, although they were concentrated at different “heavy” fractions. This could be in part explained by the higher G + C content of type II methanotrophs than their type I counterparts. Amplicon sequencing can serve as a standard protocol for identification of 13C-DNA in future [6]. This is most important

DNA-SIP

27

for active microorganisms that often constitute a small fraction of the total microbial communities in natural environments and those who have a slow rate of substrate utilization. For instance, under field conditions, methanotrophs often comprise less than 0.5% of the total microbial communities, and ammonia-oxidizing bacteria (AOB) and archaea (AOA) are notoriously slow growing. The key reasoning is that in complex environmental samples only small amount of the soil genomic DNA can be labeled and migrate to the “heavy” DNA fractions, while the majority of soil genomic DNA stays in the “light” DNA fraction at the top of the ultracentrifugation tube after isopycnic centrifugation regardless of either 13 C- or 12C-treatment. This is generally acknowledged for most DNA/RNA-SIP experiments approaching in situ metabolisms of microorganisms in complex environment because the specifically targeted group of active microorganism usually has low abundance in the natural environment. Thus, the targeted microorganisms of labeled would have high proportions given the small population size of the total microbial community resolved in the presumably 13 C-labeled “heavy” DNA fraction, but not in the 12C-DNA fraction. For instance, as shown in Fig. 1, (1) The amount of total genomic DNA in the presumably 13C-labeled “heavy” DNA fraction was low for both 12C and 13C treatment; (2) 13C-DNA from the labeled methanotrophic communities is low but can be clearly demonstrated against the small amount of total genomic DNA in the “heavy” DNA fraction after ultracentrifugation by amplicon sequencing of total 16S rRNA genes in the heavy fraction using universal primer; (3) However, this is not the case for the “heavy” fractions from 12C-treatments since methanotrophic community stayed only in the “light” fractions where majority of soil genomic DNA were resolved. In summary, high-throughput sequencing of total 16S rRNA genes enables accurate analysis of the relative abundance for each individual phylotypes in the “heavy” fractions, and the subsequent pairwise comparison of each individual phylotypes can help identifying phylotypes whose relative abundances are significantly higher in the heavy DNA fractions from 13C-labeled treatment than those from 12C-labeled treatment. The “heavy” DNA fractions containing these highly enriched phylotypes can be considered 13C-labeled for further analysis such as metagenome sequencing.

4

Notes 1. CsCl is more easily dissolved at 30
C. 2. Ideally, the 13C-substrate concentration should be as close to the in situ concentration as possible; however, to ensure that the background concentration of the unlabeled substrate in the

28

Zhongjun Jia et al.

sample does not excessively dilute the label. Generally, 5–500 μmol of labeled substrate should be assimilated by microorganisms per gram of soil, or 1–100 μmol per liter of seawater, to obtain sufficient 13C-DNA to be detected. 3. A longer (>6 cm) needle with 10-mL syringe can be used to gently inject the CsCl solution into the ultracentrifuge tube. To prevent the tube from bursting during ultracentrifugation, tubes should be filled at least to its neck. 4. Firstly, it is suggested to check whether the speed reaches 45,000 rpm (190,000  g) at the start of the centrifugation. The ultracentrifuge should be stopped if the speed does not reach the set point, as it may indicate that a tube has burst or the rotor is unbalanced. 5. The ultracentrifuge tube should be handled as gently as possible to minimize disturbing the gradient. 6. Pay attention not too clamp the ultracentrifuge too tight or loose. If overly tight, excessive pressure may cause the CsCl solution to flow out of the tube as soon as it is pierced. If the tube is not held firmly, it can fall. 7. The buoyant density of fractions ranges between 1.690 and 1.770 g/mL. The formula for ρ–x needs to be determined in advance: prepare different densities of CsCl and GB mixtures within the range of 1.650–1.850 g/mL. Measure its refractive index (nD-TC mode), plot the data and the regression to determine the relationship between refractive index and density. When a new CsCl solution is used, the formula may need to be corrected. 8. Labeled microorganisms usually account for a small proportion of the total microbial communities. Therefore, even if the labeled DNA is successfully separated under ultracentrifugation, it will not be visible in an agarose gel. We find that a heavy fraction with 0.7 ng 13C-DNA/μL in 30 μL of TE buffer does not generate a distinct band in a gel. In addition, due to the large amount of unlabeled DNA and the constant low background in ultracentrifugation tube, the visualization of 13 C-DNA in heavy fractions is extremely difficult. Methods such as HRSIP (see Chapter 9) are ideal for detecting labeled DNA. References 1. Radajewski S, Ineson P, Parekh NR, Murrell JC (2000) Stable-isotope probing as a tool in microbial ecology. Nature 403(6770):646–649 2. Meselson M, Stahl FW (1958) The replication of DNA in Escherichia coli. Proc Natl Acad Sci U S A 44(7):671–682

3. Dumont MG, Radajewski SM, Miguez CB, McDonald IR, Murrell JC (2006) Identification of a complete methane monooxygenase operon from soil by combining stable isotope probing and metagenomic analysis. Environ Microbiol 8(7):1240–1250

DNA-SIP 4. Neufeld JD, Vohra J, Dumont MG, Lueders T, Manefield M, Friedrich MW et al (2007) DNA stable-isotope probing. Nat Protoc 2 (4):860–866 5. Lueders T, Manefield M, Friedrich MW (2004) Enhanced sensitivity of DNA- and rRNA-based stable isotope probing by fractionation and quantitative analysis of isopycnic centrifugation gradients. Environ Microbiol 6(1):73–78 6. Jia Z, Conrad R (2009) Bacteria rather than Archaea dominate microbial ammonia oxidation in an agricultural soil. Environ Microbiol 11(7):1658–1671 7. Xia W, Zhang C, Zeng X, Feng Y, Weng J, Lin X et al (2011) Autotrophic growth of nitrifying community in an agricultural soil. ISME J 5 (7):1226–1236 8. Cupples AM, Shaffer EA, Chee-Sanford JC, Sims GK (2007) DNA buoyant density shifts during 15N-DNA stable isotope probing. Microbiol Res 162(4):328–334 9. Schildkraut CL, Marmur J, Doty P (1962) Determination of the base composition of deoxyribonucleic acid from its buoyant density in CsCl. J Mol Biol 4:430–443

29

10. Buckley DH, Huangyutitham V, Hsu SF, Nelson TA (2007) Stable isotope probing with 15 N achieved by disentangling the effects of genome G+C content and isotope enrichment on DNA density. Appl Environ Microbiol 73 (10):3189–3195 11. Karlovsky P, de Cock AW (1991) Buoyant density of DNA-Hoechst 33258 (bisbenzimide) complexes in CsCl gradients: Hoechst 33258 binds to single AT base pairs. Anal Biochem 194(1):192–197 12. Zheng Y, Huang R, Wang BZ, Bodelier PLE, Jia ZJ (2014) Competitive interactions between methane- and ammonia-oxidizing bacteria modulate carbon and nitrogen cycling in paddy soil. Biogeosciences 11:3353–3368 13. Chen Y, Murrell JC (2010) When metagenomics meets stable-isotope probing: progress and perspectives. Trends Microbiol 18 (4):157–163 14. Kalyuzhnaya MG, Lapidus A, Ivanova N, Copeland AC, AC MH, Szeto E et al (2008) High-resolution metagenomics targets specific functional types in complex microbial communities. Nat Biotechnol 26(9):1029–1034

Chapter 3 RNA Stable Isotope Probing (RNA-SIP) Noor-Ul-Huda Ghori, Benjamin Moreira-Grez, Paton Vuong, Ian Waite, Tim Morald, Michael Wise, and Andrew S. Whiteley Abstract Stable isotope probing is a combined molecular and isotopic technique used to probe the identity and function of uncultivated microorganisms within environmental samples. Employing stable isotopes of common elements such as carbon and nitrogen, RNA-SIP exploits an increase in the buoyant density of RNA caused by the active metabolism and incorporation of heavier mass isotopes into the RNA after cellular utilization of labeled substrates pulsed into the community. Labeled RNAs are subsequently separated from unlabeled RNAs by density gradient centrifugation followed by identification of the RNAs by sequencing. Therefore, RNA stable isotope probing is a culture-independent technique that provides simultaneous information about microbiome community, composition and function. This chapter presents the detailed protocol for performing an RNA-SIP experiment, including the formation, ultracentrifugation, and fractional analyses of stable isotope-labeled RNAs extracted from environmental samples. Key words RNA-SIP, 16S rRNA, Gradient centrifugation, Community diversity, Community function, Isotope-labeled substrate

1

Introduction One of the fundamental questions in microbial ecology still remains “who is doing what?” [1]. The Golden Age of 16S rRNA sequencing, commencing with Carl Woese’s seminal work [2], revealed the true extent of microbial diversity on Earth but raised a significant issue in that, in most instances, functions could not be assigned to many of the taxa identified by environmental sequence surveys [3]. This was recognized as a major pitfall for all but the most well-known taxa whose aligned sequences defined unique functional groups, such as methanotrophs [4] and methanogens [5]. Prior to the advent of rRNA gene surveys, functional characterization of microbial populations was addressed through the laborious task of isolating microbial strains into the laboratory followed by characterizing the isolates at the physiological, biochemical and genetic levels. The metabolic properties and cellular

Marc G. Dumont and Marcela Herna´ndez Garcı´a (eds.), Stable Isotope Probing: Methods and Protocols, Methods in Molecular Biology, vol. 2046, https://doi.org/10.1007/978-1-4939-9721-3_3, © Springer Science+Business Media, LLC, part of Springer Nature 2019

31

32

Noor-Ul-Huda Ghori et al.

interactions data acquired this way could be used to infer potential functions of these microbes and their close relatives [6]. This approach, however, was limited in terms of isolating only a small portion of the community which was cultivable [3, 7] and when compared to the sequences known to be present, clearly provided a highly biased view of in situ diversity and function. In the first years of the new millennium, a significant development to resolve this issue was provided by Colin Murrell’s laboratory in the UK, with the advent of 16S rRNA sequencing linked to specific substrate mediated isotopic labeling [8], a technique we now know as stable isotope probing (SIP). Stable isotope probing is both simple in execution and elegant as a concept, the workflow being represented in Fig. 1. It is based upon the premise “you are what you eat.” Substrates synthesized using nonradioactive stable isotopes (such as 13C, and later 15N and 18 O) can be pulsed into natural samples and any organisms utilizing and incorporating the substrate will become labeled with the stable isotope signature [9]. Originally the method was used to trace single carbon compounds into polar lipid derived fatty acids of active sulfate reducers [10] and methylotrophs [10, 11]. The resolution of the method was increased substantially as it moved from PLFA to DNA based sequencing of heavy DNA, recovered by density gradient centrifugation, bringing it directly into line with the significant 16S rRNA sequencing efforts being performed [8]. Several years later, density gradient separation of RNA, termed RNA-SIP, was developed, which targeted RNA as the biomarker to identify active bacteria [12]. While complementary to DNA-SIP, targeting the heavy-labeled RNA directly made sense for several reasons, including a relatively higher abundance of RNA within a cell when compared to DNA content, faster turnover during activity periods and the ability to target gene expression through mRNA [13]. In fact, the various cellular RNA molecules can be turned over (and hence labeled) without the need for DNA synthesis or replication of the organism itself [14], an important consideration in many low growth environments, where cells may be actively turning over RNA but not replicating their DNA to a significant degree [15]. Finally, in common with DNA-SIP, RNA-SIP allows for identification of microorganisms without any prior knowledge of their identity, owing to the sequence-based resolution provided by SSU rRNA [14]. Since the RNA-SIP protocol was first published [16] it has helped answer many fascinating questions including assessing the relationship between function and diversity at both the identity [14] and specific functional gene level [13]. Studies that have applied the RNA-SIP technique are widespread and include microorganisms in sandy soils [17], sulfate reducing bacteria in paddy soils [18], methylotrophs in rice field soils [19, 20] and novel applications of food web tracing by identification of bacterial micropredators in a soil trophic network [21].

RNA-SIP

33

Fig. 1 Graphical workflow of the stages involved in RNA-SIP. (a) Natural samples are pulsed in parallel with unlabeled control substrate (e.g., 12C6-Glucose) and equivalent stable isotope-labeled substrate (e.g., 13 C6-Glucose). (b) Both control and labeled RNAs are extracted and subject to isopycnic ultracentrifugation which bands the RNA according to buoyant density, the RNA derived from cells taking up the labeled isotope banding lower in the gradient (c1) due to increased mass afforded by the heavier 13C isotope (orange band). (c2). Each gradient is fractionated to recover the RNA in a series of tubes to isolate the fraction containing the labeled RNAs followed by reverse transcription (d). Subsequently, downstream sequencing/analysis protocols are performed to provide either (e) phylogenetic information if targeting rRNA with specific phylogenetic primers, functional gene information if using specific cDNA primers to gene-specific mRNAs, or, new generation transcriptomic analyses

This chapter discusses the protocol to perform a basic RNA-SIP experiment. It outlines the formation, ultracentrifugation, and fractional analyses of gradients for the separation of 13C stable isotope-labeled RNA extracted from the environment. The protocol discussed here has been used for both rRNA and mRNA analyses and the protocol has been updated from the original publication of RNA-SIP [12], which relied upon reverse

34

Noor-Ul-Huda Ghori et al.

transcription PCR of whole fractionated gradients followed by denaturant gradient gel electrophoresis (DGGE) analysis. Before starting the protocol it should be ensured that the 13C (or 15N, 18O) substrate used should ideally be fully substituted and be of 99 atom% purity. We will not discuss the detailed protocol of RNA extraction as there are many well documented protocols already available (e.g., see Chapter 13), but we use our in house method which has been proven in many environments previously [22]. Following extraction, we further purify RNAs using commercially available kits (see Subheading 2). This combination of methods yields sufficient quality of rRNA and mRNA for RNA-SIP downstream processing and have been successfully applied to soil, water, lithobiont, and bioreactor samples in the past.

2

Materials

2.1 Equipment Required

The protocol described here is optimized for 2.2-mL tubes centrifuged in a fixed-angle tabletop ultracentrifuge rotor. Swing-out rotors are not compatible with isopycnic centrifugation used in RNA or DNA-SIP. For RNA-SIP, the use of other fixed angle rotors and tubes (5 mL, vertical rotors, etc.), requires a straightforward adaptation of the presented protocol through rotor conversions (e.g., by k factor), which can be supplied by many centrifuge manufacturers. A detailed list of the required equipment is as follows: 1. Beckman TLX benchtop ultracentrifuge (Beckman Coulter). 2. Beckman tube topper (Beckman Coulter). 3. Beckman Fraction Recovery System (Beckman Coulter). 4. Razel syringe pump (Semart Technical Ltd.). 5. GeneQuantPro RNA/DNA calculator (Biochrom) or QuBit. 6. Certified DNA/RNase-free filter tips. 7. Fully calibrated pipettes. 8. 1.5-mL microfuge tubes. 9. 2.0-mL microfuge tubes. 10. 2.2-mL polyallomer sealable no. 344625; Beckman Coulter).

centrifuge

tubes

(cat.

11. 2-mL Plastipak syringe. 12. 5-mL Plastipak syringe. 13. 23-gauge Luer lock needles (cat. no. SZR 175525K; Fisher). 14. Nitrile gloves. 15. Three-figure milligram balance. 16. Real-time PCR thermal cycler and plates.

RNA-SIP

35

17. UV sterilizing PCR workstation. 18. PCR tubes. 19. 8-Channel pipettes. 2.2 Reagents Required

1. Agarose: 1.5% agarose gel containing 200 ng/mL ethidium bromide. 2. Tris–Borate EDTA buffer. 3. 10 mg/mL Ethidium bromide. 4. 2.0 g/mL cesium trifluoroacetate: prepare cesium trifluoroacetate (CsTFA) gradient with a starting density of 1.8 g/mL for 2.2 mL volume gradient by mixing 1.761 mL of 2.0 g/mL CsTFA with 75 μL deionized formamide and 344 μL nucleasefree water. This provides 2.180 mL of gradient to which 20 μL of sample is added for a final volume of 2.2 mL. Volumes can be scaled by taking account of the increased volume of the tubes relative to 2.2 mL. For consistency, make a “master mix” containing all the reagents needed for n + 1 gradients, where n is the number of samples to be run. Subsequently, aliquot master mix minus sample volume per gradient. For example, for a 2.2 mL gradient aliquot 2.180 mL of above master mix into a 2-mL microfuge tube, which contains 20 μL of RNA sample. This avoids gradient to gradient variation due to pipetting errors within a centrifugation batch. Deionized formamide is best aliquot in small volumes and frozen at 20  C to avoid freeze–thaw cycles. Once defrosted do not refreeze. 5. Nuclease-free water. 6. Deionized formamide. 7. Nuclease-free 10 Tris–EDTA. 8. Isopropanol (ice cold). 9. Ethanol. 10. Parafilm M. 11. 20 mg/mL bovine serum albumin. 12. AccessQuick RT-PCR System. 13. SYBR Green 10,000 concentrate. To prepare 20 SYBR Green working solution, dilute SYBR Green stock 10,000 in a 1:500 volume ratio in nuclease-free water. Make small aliquots and store at 20  C and thaw only once. 14. Primers at 50 μM stock solutions. 15. 16S and 23S ribosomal RNA standard (from E. coli), 100 μg/ mL in TE buffer, as contained in the RiboGreen RNA Quantitation Kit (Invitrogen).

36

3

Noor-Ul-Huda Ghori et al.

Methods

3.1 13C-Labeled Substrate Incubations, RNA Extraction and Purification

1. It is recommended that initial investigations using light isotopic (12C, 14N etc.) version of substrates backed by chemical analyses are performed prior to the labeled pulsing in order to assess a time course for incorporation of substrate prior to applying the expensive stable isotope substrate. The pulse can be applied once the incorporation rate of the substrate is established, using chemical or isotopic ratio mass spectrometry based analyses to monitor substrate utilization. Always perform parallel control unlabeled (e.g., 12C) and test labeled (e.g., 13 C) substrate pulses to account for where “native” RNA should band versus the labeled RNA. It is also advisable to use a time course of SIP incubation analyses to confirm the entry of the pulse into the microbial community and to avoid “single sample” experiments. This gives confidence that the target organisms are labeling and reduces confounding effects of “cross feeding” if pulses are left too long and a single sample is taken. Further, isotopic ratio mass spectrometry of extracted and purified RNA is also advisable to assess the degree of isotopic enrichments since successful RNA-SIP requires a minimum of 15 atomic % 13C substitution. A range of pulse options for different environments with various substrates can be found in Murrell and Whiteley [23]. Samples obtained from pulses can be stored at 20  C for up to a month or indefinitely at 70  C. 2. Extract RNA or total nucleic acids using an optimized protocol and purify RNA further (e.g., RNA AllPrep Qiagen Kit) if a total DNA/RNA extract is obtained from the primary extraction. 3. Elute pure RNA in 50 μL of nuclease-free water and run 5 μL on a 1.5% agarose gel containing 200 ng/mL ethidium bromide in 1 TBE at 70 V for 20 min to observe intact 16S and 23S rRNA. Once checked for quality of RNA (using the 16S and 23S rRNA bands as a proxy), quantify absolute concentration of RNA by spectrophotometry (e.g., GeneQuantPro RNA/DNA calculator) or fluorimetry (e.g., Qubit) and dilute RNA extract to 100 ng/μL with nuclease-free water. Samples can now be stored at 20  C for a month or indefinitely at 70  C.

3.2 RNA Gradient Preparation and Centrifugation

1. Prepare a sufficient amount of gradient medium (see Subheading 2.2, item 4) to allow an even number of centrifuge tubes (typically up to 10 for the TLA120.2 rotor) depending upon your rotor. Include at least one blank gradient (no RNA

RNA-SIP

37

sample). The blank gradient acts as a reference gradient and their density profiles are calculated after fractionation to assess the efficiency of ultracentrifugation. Always run pairs of gradients, consisting of control (e.g., 12C substrate pulse) gradients with their corresponding parallel labeled samples (e.g., 13C substrate pulse). The control pulsed samples aid in locating the unlabeled community RNA banding point and act as a key reference to establish whether the stable isotope-labeled test samples have incorporated isotope and banded further down the gradient. Further, artificially generated controls are highly useful and can be generated by centrifuging RNA extracted from E. coli grown in unlabeled or labeled substrates within defined M9 medium. For example, for carbon, E. coli grown on 13C fully C6-labeled glucose as the sole source of carbon within the M9 in parallel to RNA extracted from cells grown on normal 12C glucose. For other isotopes, labeled positive control RNA can be generated by growing cells on a single 15N-labeled N source (versus the equivalent 14N) within the M9 medium, or, using heavy 18O water and normal 16O water for 18O SIP applications. These artificial controls are useful to generate and store as a resource to develop and troubleshoot all SIP applications. We routinely grow cells and extract RNA and store as a reference resource for organisms of different GC contents and even defined atomic % RNAs to aid in method development and calibration. For routine use, a minimum of equivalent control and labeled samples together with blank gradients should be analyzed on every run, while positive and negative labeled artificial RNA controls (e.g., derived from E. coli) are useful to ensure run consistency and accurately monitor the position of “heavy” and “light” fraction locations. 2. Add 20 μL of RNA sample containing approximately 500 ng of RNA to a clean and sterile 2-mL microfuge tube. Optimum loading of the gradient is ca. 250 ng RNA per mL of gradient. For example, 5 μL of a 100 ng/μL RNA sample would be added to 15 μL of nuclease-free water in a clean 2-mL microfuge tube to account for the 20 μL sample volume. For the blank gradient simply add 20 μL of nuclease-free water instead. 3. Add 2.180 mL of premixed gradient medium to the microfuge tube to generate 2.2 mL of gradient and take care to avoid bubbles when drawing the 2.2 mL volume into a 2 mL-syringe fitted with a 24-gauge needle. Once aspirated, invert the syringe, tap to dislodge any bubbles to the neck of the syringe and carefully move the plunger to remove any residual air in the syringe without losing any gradient medium. Once primed, deliver the gradient medium into a polyallomer QuickSeal tube by placing the syringe needle through the open tube

38

Noor-Ul-Huda Ghori et al.

neck, tilt the tube and slowly fill from the bottom of the tube to avoid air bubbles while carefully withdrawing the needle as the tube fills. 4. Seal the polyallomer QuickSeal tube with a heat sealer while wearing safety glasses. This avoids the possibility of eye contamination, as occasionally a small amount of gradient medium may be ejected due to reduction in the neck volume as the tube seals in tandem with expansion of the medium due to heating (see Note 1). Place the fully sealed tubes in the rotor (note the tube position) and place a shoulder cap on the tubes to support the tube tops to avoid tube crushing during centrifugation. 5. Spin at 128,000  g for 42–65 h at 20  C, with max acceleration and max deceleration (see Note 2). 3.3 Gradient Fractionation

1. Carefully remove the tubes from the centrifuge rotor with forceps. 2. To optimize the fractionation and to obtain best results, use a fraction recovery system that allows fraction collection from the base of the tube via water displacement at the top of the tube. For example, for the Beckman Fraction Recovery System, prepare the gradient fractionator by connecting a 5-mL syringe filled with 5 mL of nuclease-free water containing 5 μL of any DNA loading buffer to color the water to visualize the progress of fractionation. Prime the line by pressurizing the syringe to force water through the line until a single drop emerges out of the fractionator hood. 3. Place the centrifuge tube in the fraction recovery system and carefully remove the tube top. Squeezing the tube is best avoided and removal of the top can be achieved by using a heated scalpel blade to minimize tube disturbance. Once the top is removed lower the fraction recovery hood onto the top of the open tube and ensure a tight seal is obtained. 4. Pierce the bottom of the tube by inserting the fraction collection needle, ensuring a tight seal is maintained. The seal can be enhanced by stretching a small piece of Parafilm around the base of the tube prior to placing it in the fraction collector. 5. Set a flow rate of 200 μL/min on the syringe pump and simultaneously start the pump and a stop watch. In a series of sterile microfuge tubes collect a fraction every 30 s, amounting to ~100 μL per fraction and 20 fractions per gradient. The last fraction tubes will contain traces of color if DNA loading dye is used in the displacement water as a visual aid, these tubes representing the location of the end of the fractionation series. 6. Calculate the absolute density by refractometry or weighing a known volume of the fractions into a clean and sterile microfuge tube on a three-figure balance. Plot the gradient shape

RNA-SIP Control

SIP

Buoyant Density (g mL)

Fraction Number ~ BD

39

16S Q-PCR Gene copy number

1.75 1.77 12C

1.79 1.81 1.83 1.85

13C

1.87 1.89 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Fig. 2 Typical 2.2-mL volume gradient fraction analysis by density and qRT-PCR of RNA locations for control and labeled isotope SIP tubes

using density in g/mL as a function of fraction number (Fig. 2). Remember that fraction 1 is the bottom of the gradient and last fraction is the top of the gradient. 7. Clean the fractionator after each gradient is fractionated by removing the fractionator hood and pipetting 2 mL of 0.1 M NaOH into the empty centrifuge tube. Allow it to run out of the collection needle. Repeat the process with 2 mL of absolute ethanol. 3.4 RNA Precipitation

1. To the 100 μL fractions add two volumes of ice-cold isopropanol and incubate the tubes at 20  C for 30 min. Samples can be stored at 20  C at this point for later processing. 2. Centrifuge tubes for 20 min at 14,000  g in a prechilled microtube centrifuge at 4  C, carefully remove the supernatant and add 150 μL of ice-cold isopropanol to wash away residual CsTFA. Exercise care when pipetting supernatants and washing as RNA pellets are not usually visible due to low concentrations of RNA distributed over multiple fractions. 3. Spin the tubes at 14,000  g for 5 min at 4  C in prechilled microfuge and remove supernatant. Perform a final spin of the tube for 1 min at 14,000  g and remove any remaining isopropanol using a 20-μL volume pipette tip. 4. Air-dry the pellet and resuspend in 10 μL of RNase-free TE buffer. Samples can be stored at 20  C or optimally at 80  C at this point before further processing.

3.5 Quantification of Bacterial RNA in Gradient Fractions

The following protocol is that previously published for assessment of RNA separation based upon qRT-PCR of 16S rRNA within gradient fractions [12]. However, examples of mRNA-based

40

Noor-Ul-Huda Ghori et al.

Table 1 qRT-PCR reaction mixture qRT-PCR master mix

Volume (μL) Concentration in PCR

Nuclease-free water

16.4

AccessQuick 2 master mix

20

1

BSA (20 mg/mL)

0.4

0.2 μg/μL

SYBR Green working solution (20)

0.2

0.1

519f-primer (50 μM)

0.2

0.25 μM

907r-primer (50 μM)

0.2

0.25 μM

AMV (5 U/μL)

0.6

3U

gene-specific analysis protocols can now be found in [13] or more recent advanced protocols for transcriptome analyses of heavy fractions [24]. This 16S rRNA-based protocol utilizes short amplicons generated with universal primers targeting domain Bacteria but can be adapted for Archaea and Eukarya with the relevant group-specific primer sets within the literature such as [21, 25–27]. 1. Prepare 40-μL qRT-PCR reaction mixtures according to Table 1. 2. Label a set of eight sterile microfuge tubes in a rack with a tenfold dilution series from 101 down to 10 6 and pipette 5 μL of E. coli standard rRNA (100 ng/μL) into the first tube. Dilute this tube 1:10 by adding 45 μL of nuclease-free water. Mix by pipetting a few times. Repeat the process by removing 5 μL of the 1:10 dilution and place in the next tube, adding 45 μL of nuclease-free water and mixing to generate a 1:100 (100) dilution. Repeat this process down to the eighth tube to yield a standard set ranging from 10 ng/μL down to 10 6 ng/μL (see Note 3). 3. Within a 96-well PCR plate (see Note 4), pipette 2 μL of each gradient fraction in triplicate and add 38 μL of qRT-PCR master mix (Subheading 3.5, step 1) to each template and pipette a few times to mix (see Note 5). 4. Pipette triplicate negative controls within the plate by pipetting 2 μL of nuclease-free water into 3 wells and add 38 μL of qRT-PCR mix as in step 4. 5. For the standard curve pipette 2 μL of each of the 8 E. coli rRNA dilution standards (step 2) into adjacent wells, replicating as required, and add 38 μL of qRT-PCR master mix. 6. Amplify in a SYBR Green enabled qPCR thermocycler under the thermal conditions below (Table 2).

RNA-SIP

41

Table 2 qPCR thermal cycler conditions Step

Temperature 

Duration

Reverse transcription

45 C

20 min

Initial denaturation

95  C

5 min

95  C

30 s

Fluorescence

35 cycles of: Denaturation



Annealing

52 C

30 s

Elongation

68  C

30 s

Final elongation

68  C

5 min

Final denaturation



1 min



95 C

Reassociation

55 C

30 s

Dissociation ramp

55–95  C

30 min

Final hold

25  C

hold

SYBR

SYBR

7. Finally, quantify the bacterial rRNA in each gradient fraction to arbitrary E. coli 16S and 23S rRNA in ng/μL units via the collected SYBR Green fluorescence threshold cycles (Ct) data in each qRT-PCR reaction (see Note 6). If all steps have been performed correctly, the gradient shape (density versus fraction number) and location of 16S rRNA fluorescence signals equating to “light” and “heavy” RNAs should be obtained as shown in Fig. 2. Typically, “light” unlabeled RNA bands at a buoyant density of around 1.77–1.80 g/mL corresponding to fractions in the region 11–13. The median density (mid-point) within the tube should reflect the initial gradient density of 1.8 g/mL while the “heavy” RNA bands in the region of fractions 4–7, corresponding to a density of around 1.84–1.87 g/mL. Once located, these fractions can then be subject to sequencing to determine the active organisms [14, 28], RT-PCR of key functional genes [29] or recently described transcriptome analyses using linear amplification [24]. Identified sequences linked to functionally active organisms which are derived from RNA-SIP analyses can be further used to design specific rRNA targeted probes to perform “full cycle” analyses using stable isotope enabled imaging technologies such as Raman-FISH [29]. After probe development, the RNA-SIP functionally defined cells can be identified by FISH probing to visualize their abundance, habitat and ecology as well as simultaneously performing single cell measures of isotopic content through Raman spectroscopy to assess isotope uptake characteristics

42

Noor-Ul-Huda Ghori et al.

[13]. The Raman-FISH spectroscopy based method of isotope measurement is particularly suitable to these applications due to it being microscopy based, the ease of sample processing of previously isotope-labeled samples, and its ability to detect most stable isotopes which are commonly used in SIP based studies [30].

4

Notes 1. Critically, make sure the seal is straight and secure, as defects in the seal will cause tube failure. Check seal integrity by gently squeezing the tube to ensure a watertight neck seal. If liquid emerges from around the neck seal, discard the tube and reform a new gradient. 2. Due to the density of CsTFA, rotors need to be de-rated to a maximum run speed equivalent to 80% of their normal maximum speed. 3. Freshly prepare the standards for each qRT-PCR experiment and do not let them stand for long as they are extremely unstable and cannot be stored for more than 1 h at 4  C. 4. Throughput can also be increased by scaling the volumes to 192 and 384 well PCR systems now commonly used. 5. When first calibrating RNA-SIP, run all gradient fractions to visualize RNA horizons. However, as confidence builds and gradient formation becomes routine, only a subset of fractions (ca. 8–10 bracketing 12C and 13C horizons) will need to be included within the plate. 6. Care must be taken to omit false positive Ct values that may be caused by the formation of primer dimers in samples containing no or extremely low amounts of template from analysis. These false positives can be identified by their PCR products melting profiles recorded during the dissociation ramp.

Acknowledgments The authors wish to thank all the investigators over the years who have played a role in developing RNA-SIP as a method, especially professors Mike Manefield, Tillmann Lu¨eders, and Dr. Ian Douglas of Beckman Coulter. Particular thanks go to Professor Colin Murrell for stimulating discussions over the years on stable isotope enhanced microbial ecology. The Molecular Microbial Ecology Laboratory at UWA is funded by a range of sources, including the Australian Research Council Linkage Program (LP150101111 to A.S.W.), The Australia China Joint Research Centre (A.S.W., I.W.), a Chilean BECAS scholarship (B.M.G.), and a UWA IPRS scholarship (N.H.G.).

RNA-SIP

43

References 1. Dumont MG, Murrell JC (2005) Innovation: stable isotope probing—linking microbial identity to function. Nat Rev Microbiol 3:499–504. https://doi.org/10.1038/ nrmicro1162 2. Woese CR, Fox GE (1977) Phylogenetic structure of the prokaryotic domain: the primary kingdoms. Proc Natl Acad Sci U S A 74:5088–5090. https://doi.org/10.1073/ PNAS.74.11.5088 3. Amann RI, Ludwig WSK (1995) Phylogenetic identification and in situ detection of individual microbial cells without cultivation. Microbiol Rev 59:143–169 4. Whittenbury R, Phillips KC, Wilkinson JF (1970) Enrichment, isolation and some properties of methane-utilizing bacteria. J Gen Microbiol 61:205–218. https://doi.org/10. 1099/00221287-61-2-205 5. Huber H, Thomm M, Ko¨nigw H, Thies G, Stetter KO (1982) Methanococcus thermolithotrophicus, a novel thermophilic lithotrophic methanogen. Arch Microbiol 132:47–50. https://doi.org/10.1007/BF00690816 6. Radajewski S, McDonald IR, Murrell JC (2003) Stable-isotope probing of nucleic acids: a window to the function of uncultured microorganisms. Curr Opin Biotechnol 14:296–302. https://doi.org/10.1016/ S0958-1669(03)00064-8 7. Hugenholtz P, Goebel BMPN (1998) Impact of culture independent studies on the emerging phylogenetic view of bacterial diversity. J Bacteriol 180:4765–4774 8. Radajewski S, Ineson P, Parekh NR, Murrell JC (2000) Stable-isotope probing as a tool in microbial ecology. Nature 403:646–649. https://doi.org/10.1038/35001054 9. Neufeld JD, Dumont MG, Vohra J, Murrell JC (2007) Methodological considerations for the use of stable isotope probing in microbial ecology. Microb Ecol 53:435–442. https://doi. org/10.1007/s00248-006-9125-x 10. Boschker HTS, Nold SC, Wellsbury P, Bos D de GW, Pel R, Parkes RJCTE (1998) Direct linking of microbial populations to specific biogeochemical processes by C-13-labelling of biomarkers. Nature 392:801–805 11. Whiteley AS, Manefield M, Lueders T (2006) Unlocking the ‘microbial black box’ using RNA-based stable isotope probing technologies. Curr Opin Biotechnol 17:67–71. https://doi.org/10.1016/J.COPBIO.2005. 11.002

12. Whiteley AS, Thomson B, Lueders T, Manefield M (2007) RNA stable-isotope probing. Nat Protoc 2:838–844. https://doi.org/10. 1038/nprot.2007.115 13. Huang WE, Ferguson A, Singer AC, Lawson K, Thompson IP, Kalin RM, Larkin MJ, Bailey MJ, Whiteley AS (2009) Resolving genetic functions within microbial populations: in situ analyses using rRNA and mRNA stable isotope probing coupled with single-cell raman-fluorescence in situ hybridization. Appl Environ Microbiol 75:234–241. https://doi. org/10.1128/AEM.01861-08 14. Manefield M, Whiteley AS, Griffiths RI, Bailey MJ (2002) RNA stable isotope probing, a novel means of linking microbial community function to phylogeny. Appl Environ Microbiol 68:5367–5373. https://doi.org/10.1128/ AEM.68.11.5367-5373.2002 15. Ostle N, Whiteley AS, Bailey MJ, Sleep D, Ineson P, Manefield M (2003) Active microbial RNA turnover in a grassland soil estimated using a 13CO2 spike. Soil Biol Biochem 35:877–885. https://doi.org/10.1016/ S0038-0717(03)00117-2 16. Manefield M, Griffiths RI, Leigh MB, Fisher R, Whiteley AS (2005) Functional and compositional comparison of two activated sludge communities remediating coking effluent. Environ Microbiol 7:715–722. https://doi.org/10. 1111/j.1462-2920.2004.00746.x 17. Schwarz A, Adetutu EM, Juhasz AL, AburtoMedina A, Ball AS, Shahsavari E (2017) Microbial degradation of phenanthrene in pristine and contaminated sandy soil. Microb Ecol 75 (4):888–902 18. Liu P, Pommerenke B, Conrad R (2018) Identification of Syntrophobacteraceae as major acetate-degrading sulfate reducing bacteria in Italian paddy soil. Environ Microbiol 20:337–354. https://doi.org/10.1111/ 1462-2920.14001 19. Lueders T, Pommerenke B, Friedrich MW (2004) Stable-isotope probing of microorganisms thriving at thermodynamic limits: syntrophic propionate oxidation in flooded soil. Appl Environ Microbiol 70(10):5778–5786 20. Lueders T, Manefield M, Friedrich MW (2003) Enhanced sensitivity of DNA- and rRNA-based stable isotope probing by fractionation and quantitative analysis of isopycnic centrifugation gradients. Environ Microbiol 6:73–78. https://doi.org/10.1046/j.1462-2920.2003. 00536.x

44

Noor-Ul-Huda Ghori et al.

21. Lueders T, Kindler R, Miltner A, Friedrich MW, Kaestner M (2006) Identification of bacterial micropredators distinctively active in a soil microbial food web. Appl Environ Microbiol 72:5342–5348 22. Griffiths RI, Whiteley AS, O’Donnell AG, Bailey MJ (2000) Rapid method for coextraction of DNA and RNA from natural environments for analysis of ribosomal DNA- and rRNAbased microbial community composition. Appl Environ Microbiol 66:5488–5491. https://doi.org/10.1128/AEM.66.12.54885491.2000 23. Murrel JC, Whiteley AS (2010) Stable isotope probing and related technologies. American Society for Microbiology Press, Washington, DC 24. Bradford LM, Vestergaard G, Ta´ncsics A, Zhu B, Schloter M, Lueders T (2018) Transcriptome-stable isotope probing provides targeted functional and taxonomic insights into microaerobic pollutant-degrading aquifer microbiota. Front Microbiol 9:2696. https:// doi.org/10.3389/fmicb.2018.02696 25. Weisburg WG, Barns SM, Pelletier DA, Lane DJ (1991) 16S ribosomal DNA amplification for phylogenetic study. J Bacteriol 173:697–703. https://doi.org/10.1128/JB. 173.2.697-703.1991

26. Stubner S (2002) Enumeration of 16S rDNA of desulfotomaculum lineage 1 in rice field soil by real-time PCR with SybrGreen™ detection. J Microbiol Methods 50:155–164. https:// doi.org/10.1016/S0167-7012(02)00024-6 27. Lueders T, Wagner B, Claus P, Friedrich MW (2004) Stable isotope probing of rRNA and DNA reveals a dynamic methylotroph community and trophic interactions with fungi and protozoa in oxic rice field soil. Environ Microbiol 6(1):60–72 28. Manefield M, Whiteley AS, Ostle N, Ineson P, Bailey MJ (2002) Technical considerations for RNA-based stable isotope probing: an approach to associating microbial diversity with microbial community function. Rapid Commun Mass Spectrom 16:2179–2183. https://doi.org/10.1002/rcm.782 29. Huang WE, Stoecker K, Griffiths R, Newbold L, Daims H, Whiteley AS, Wagner M (2007) Raman-FISH: combining stableisotope Raman spectroscopy and fluorescence in situ hybridization for the single cell analysis of identity and function. Environ Microbiol 9:1878–1889. https://doi.org/10.1111/j. 1462-2920.2007.01352.x 30. Read D, Huang WE, Whiteley AS (2015) Single cell microbial ecophysiology with RamanFISH. Springer, Berlin, pp 65–76

Chapter 4 Stable Isotope Probing of Microbial Phospholipid Fatty Acids in Environmental Samples Andrea Watzinger and Rebecca Hood-Nowotny Abstract Phospholipid fatty acid (PLFA) extracted from environmental samples describe the microbial community pattern and are sensitive to monitor and quantify shifts in the microbial community. Linkage with the stable isotope technique adds a functional perspective and is frequently used to quantify carbon turnover in microbial communities and detect physiological changes. Here we present a PLFA extraction method by using an organic solvent water mixture, followed by lipid separation based on solid-phase extraction and an alkaline methylation. Finally, we provide a protocol for the carbon stable isotope measurements of the extracted fatty acid methyl esters (FAMEs) by gas chromatograph–isotope ratio mass spectrometer (GC-IRMS) and calculation of concentration and δ13CVPDB values. Key words Phospholipid fatty acid, Stable isotope probing, Lipid extraction, Separation, Solid-phase extraction, Alkaline methylation, Fatty acid methyl ester, Microorganisms, Environmental samples, GC-IRMS

1

Introduction The first method for lipid extraction from fish muscles was introduced by Bligh and Dyer [1] using an organic solvent–water mixture. The method was further developed by White et al. [2] to define microbial communities in ecological samples. Frostega˚rd et al. [3, 4] optimized the method to extract microbial PLFAs from soil samples. The protocol presented here is based on the latter method and is one of the most frequently cited methods [5] despite other methods being developed [6, 7]. Critical steps and common problems associated with the method and guidelines for interpretation have been summarized by Watzinger and Frostega˚rd et al. [5, 8]. The protocol contains three major steps: (a) the extraction of lipids from the environmental sample, (b) the separation of lipid classes by solid-phase extraction, and finally (c) the alkaline methylation of the phospholipids for GC analysis.

Marc G. Dumont and Marcela Herna´ndez Garcı´a (eds.), Stable Isotope Probing: Methods and Protocols, Methods in Molecular Biology, vol. 2046, https://doi.org/10.1007/978-1-4939-9721-3_4, © Springer Science+Business Media, LLC, part of Springer Nature 2019

45

46

Andrea Watzinger and Rebecca Hood-Nowotny

Determining the carbon isotope ratio within the fatty acid methyl esters (FAMEs) using an isotope ratio mass spectrometer connected to a gas chromatograph was first published in 1998 by Boschker et al. [9]. Reviews on the analytical and instrumental requirements for successful 13C FAME measurements have been published [8, 10–12]. A step-by-step measuring GC-IRMS procedure will be presented here (d).

2

Materials Solutions are prepared using ultrapure water and chemicals are either GC or HPLC grade. Chloroform stabilized with ethanol must be used. All solutions except the B&D solution and the citric buffer can be stored for several months (see Note 1). A supply of an inert gas (e.g., N2 with a purity of 4.0) at suitable pressure and a block sample concentrator with suitable temperature control is required. Solid-phase extraction is done using a standard vacuum manifold setup. It is recommended that only glassware is used during the entire extraction (see Note 2). Lids for glassware and GC vials must be coated with PTFE. For pipetting, again for best results glass Pasteur pipettes or glass tips are recommended (see Note 3).

2.1

Lipid Extraction

1. Chloroform. 2. 19:0 PLFA–methanol: 15 μM 19:0 phosphatidylcholine (1,2-dinonadecanoyl-sn-glycero-3-phosphatidylcholine). Weigh in on a four-figure balance and note the correct weigh. Store at 4
C in the dark (see Note 4). 3. Citric acid buffer: 0.085 M citric acid, 0.065 M tripotassium citrate, pH 4 (see Note 5). 4. Bligh & Dyer (B&D) solution: Mix chloroform, methanol, and citric buffer in a volumetric ratio of 1:2:0.8 respectively (see Note 6). The Bligh & Dyer solution can be stored for a week at room temperature (see Note 7). 5. Centrifuge tubes (12 mL), 2000  g (tolerance). 6. Evaporation tubes (5 mL).

2.2 Separation of Lipids

1. Chloroform. 2. Acetone. 3. Methanol. 4. 3-mL SPE columns filled with 500-mg unbonded silica, 50 μm diameter, 60 A˚ pore size (Isolute SI). 5. Centrifuge tubes (12 mL), 2000  g.

Stable Isotope Probing of PLFAs

47

6. 13:0 FAME toluene–methanol: mix 30 μM 13:0 FAME (methyltridecanoate) in toluene–methanol 1:1 volumetric solution. Weigh in on a four-figure balance and note the correct weigh. Store at 4
C in the dark. 2.3 Methylation of Phospholipids

1. KOH–methanol: 0.2 M KOH in methanol (see Note 8). 2. Acetic acid: 1 M acetic acid. 3. Hexane–chloroform: 4:1 volumetric solution. 4. Evaporation tubes. 5. Isooctane. 6. GC vials and inserts.

2.4 FAME Measurement

1. FAME Standards: 300, 200, 100, 50 μM 13:0 FAME (methyltridecanoate) and 19:0 FAME (nonadecanoate) in isooctane (see Note 9). 2. GC column: 5% diphenyl 95% dimethylpolysiloxane fused-silica capillary column, length 60 m, 0.25 mm inner diameter, 0.25 μm film thickness (see Note 10). 3. Inlet system: split/splitless injector: deactivate liner with a single tapper or programmable temperature vaporizer (PTV): with a baffled deactivated liner (see Note 11).

3

Methods The extraction is not complete. Experience has shown that any deviation from the procedure will introduce a potential bias in the results, so we suggest following a highly standardized protocol for all samples requiring intercomparison. For health and safety, all procedures must be conducted in a fume hood equipped with an organic solvent filter. 12 samples can easily be extracted and measured overnight, when preparing solutions and conducting steps 1 and 2 on the day before. One blank (sample without soil) per extraction run is typically included. If it is necessary to interrupt the extraction, due to time constraints, it is best done when the lipids are dissolved (e.g., Subheading 3.1, step 17; Subheading 3.2, step 10; Subheading 3.3, step 15), at which point flush the sample tube with an inert gas and store the samples in the dark.

3.1

Lipid Extraction

1. Weigh in 2 g of moist soil sample (on a four-figure balance and note the correct weigh) to the centrifuge tube-A and add 1.7 mL citric buffer (corrected for the soil water content (see Note 12)), 2.1 mL chloroform, 3.2 mL methanol, 1.00 mL 19:0 PLFA–methanol (15 nmol PLFA) (see Note 13).

48

Andrea Watzinger and Rebecca Hood-Nowotny

2. Close the tubes and vortex (at maximum speed) twice for 10 s each (2 rounds) (see Note 14). Store in the dark overnight (see Note 15). 3. Vortex twice, 10 s. 4. Centrifuge the tubes at 2000  g, 10 min (see Note 16). 5. Decant the supernatant into centrifuge tube-B. 6. Add 2 mL of B&D solution to tube-A. 7. Vortex twice, 10 s (see Note 17). 8. Centrifuge, 2000  g, 10 min. 9. Combine the supernatants into tube-B. 10. To tube-B add 1.5 mL chloroform and 1.5 mL water and close. 11. Vortex twice, 10 s. 12. Centrifuge, 2000  g, 5 min (see Note 18). 13. Discard the upper (water–methanol) phase (see Note 19). 14. Transfer 3.6 mL (2 1.8 mL) from the chloroform phase into evaporation tubes-C. 15. Dry the sample at 40
C in heating block concentrator with a gentle flow of an inert gas (N2) (see Note 20). 16. Remove the samples immediately when dry and add 0.5 mL of chloroform. 17. Vortex once, 10 s. 3.2 Lipid Separation (See Note 21)

1. Fix the SPE columns onto the vacuum manifold and condition by rinsing them with 5 mL (2 2.5 mL) chloroform (see Note 22). 2. Pipet 0.5 mL sample into the SPE columns and infiltrate quickly. 3. Rinse out remaining sample from tube-C and the SPE columns using a total of 5 mL of chloroform. 4. Rinse the SPE columns with 20 mL (8 2.5 mL) of acetone. 5. Discard all the drainage solution. 6. Add clean centrifuge tubes-D below the SPE columns. 7. Rinse the SPE columns with 5 mL (2 2.5 mL) methanol, collect till dripping stops. 8. Dry the sample in tubes-D at 40
C with a gentle flow of an inert gas (N2). 9. Remove the sample immediately when dry and add 1.20 mL of 13:0 FAME methanol–toluene (36 nmol FAME). 10. Vortex once, 10 s.

Stable Isotope Probing of PLFAs

3.3 Methylation of Phospholipids

49

1. Add 1.0 mL of KOH methanol, close and shake by hand (2 s). 2. Load onto the sample concentrator incubate at 40
C for 20 min (see Note 23). 3. Add 0.3 mL of acetic acid. 4. Add 2.0 mL of hexane–chloroform (4:1). 5. Add 2.0 mL of water. 6. Close tightly (see Note 24) and vortex twice, 10 s. 7. Centrifuge, 2000  g, 5 min for phase separation. 8. Remove 2 mL of the upper hexane phase into clean evaporation tubes-E. 9. Add another 2.0 mL of hexane-chloroform (4:1) to the remaining lower phase. 10. Close tightly and vortex twice, 10 s. 11. Centrifuge, 2000  g, 5 min. 12. Remove 2 mL of the upper hexane phase and add to the first extraction, tube-E. 13. Dry the sample at 40
C on the sample concentrator with a gentle flow of an inert gas (N2). 14. Add 100 μL of isooctane. 15. Vortex once, 10 s. 16. Transfer the isooctane into GC vials F with inserts. Store in the dark at 20
C (see Note 25).

3.4 FAME Measurement

Prepare the gas chromatograph–isotope ratio mass spectrometer (GC-IRMS) for 13C fatty acid methyl ester (FAME) measurement. Depending on the equipment the exact measures will be different, but the common steps are as follows. 1. Check and/or exchange the liner, GC column, septum, and syringe. The liner repeatedly becomes activated during measurements. Exchange liner after approx. 100 measurements. The septum also needs to be exchanged after approx. 100 measurements. The flow direction of the GC column should be maintained; if the chromatogram gets dirty cut the GC column by 30 cm (after around 1000 measurements). The syringe should be exchanged if any resistance to movement of the plunger is observed. 2. Check the tightness of the GC-IRMS system (see Note 26). 3. Oxidize the combustion reactor (see Note 27). If the oxidation capacity of the combustion reactor becomes insufficient a shift in the isotopic ratios will be observed. This shift might also occur during initial measurements especially if the reactor is new. If the flow resistance of the reactor increases, the

50

Andrea Watzinger and Rebecca Hood-Nowotny

combustion reactor possibly needs to be exchanged. Use the internal standard 13:0 FAME to detect this phenomenon and drift correction. 4. Define the GC protocol: The GC oven temperature is set at 70
C for 2 min in order to allow for the removal of the isooctane, it is then ramped to 160
C at 15
C/min, then to 280
C at 2.5
C/min (FAME peak separation period) and held for 2 min. When using a split/splitless injector the temperature is held at 280
C. If a PTV injector is used, then the temperature will be held at 70
C to allow removal of isooctane and then increased to 280
C to transfer the FAMEs onto the GC column. Subsequently, the GC oven temperature will start with the first temperature ramp. Carrier gas is helium and the column flow is kept at 1 mL/min or a column head pressure of 195 kPa (see Note 28). 5. Target FAME peak heights to fit within the instrument linearity and sensitivity range. For common soil samples, the injection split is kept minimal (1:4) and the injection volume high (4 μL) in order to improve signal height of FAME peaks. Evaporation of solvent (isooctane) to around 50 μL might be necessary to further increase signal intensity. 6. Adjust the CO2 working gas introduced to check IRMS performance, to the average signal height of FAME measurements. Check and note the standard deviation (1σ, n ¼ 10) of the δ13C values of the working gas and the linearity (signal height dependence) of the instrument (see Note 29). 7. Define and document the GC-IRMS precision and linearity (signal height dependence) using a 13:0 and 19:0 FAME standards. The isotopic ratio the standards 13:0 and 19:0 are referenced against international standards by elemental analyzer–isotope ratio mass spectrometer (EA-IRMS) (see Note 30). 8. Introduce a minimum of three working gas peaks before the FAME peaks emerge on the trace. The sample peaks are anchored to the working gas standard (ws) by the following formula. This method is usually implemented in the IRMS software. δtrue sample ¼ δmeasured sample þ δtrue ws þ δmeasured sample  δtrue ws =1000 9. Start the measurements and monitor/record the δ13C of the 13:0 and 19:0 internal standards. Stop and/or repeat the measurements when the precision is not met (see Note 31). 10. Check the chromatogram and manually integrate the FAMEs if auto integration fails because peaks are too close and/or the background selected by the auto integration contains peaks (see Note 32). A sample chromatogram with integration is added in Fig. 1.

Stable Isotope Probing of PLFAs

51

Fig. 1 Typical soil microbial PLFA chromatogram. Peaks are identified and integration is exemplified

11. Report the carbon isotopic composition of FAMEs in the delta (δ) notation relative to the Vienna Pee Dee Belemnite (VPDB) standard. Therefore, both the working gas and the isotopic ratio of the internal standard (std) 13:0 are referenced against international standards by elemental analyzer–isotope ratio mass spectrometer (EA-IRMS). In the GC chromatogram the sample FAME peaks are reference against the 13:0 internal standard using the following formula (see Note 33). δtrue sample ¼ ½ðδmeasured sample þ 1000Þ  ðδtrue std þ 1000Þ= ðδmeasured std þ 1000Þ  1000 12. To derive the isotopic signature of the PLFA, the δ13C values of the FAME need to be corrected for the methyl-group deriving from methanol (MeOH) introduced to the fatty acid (FA) during alkaline methylation. The methanol is referenced against international standards by EA-IRMS. This correction is not needed for 13:0, which was added as FAME during extraction, using following formula: δFA sample ¼ ðmFAME δFAME  mMeOH δMeOH Þ=m FA m is the amount of carbon atoms in the molecule; for example, the methyl-nonadecanoate has 20 carbon atoms, one deriving from the methanol and 19 from the fatty acid.

52

Andrea Watzinger and Rebecca Hood-Nowotny

13. Calculate the amount of FA per g dry soil by referring to the amount of 19:0 added at the first step of PLFA extraction and by using the following formula (see Note 34): FAðμmol=gÞ ¼ area FAME  ð2  PLFA nmolÞ=area 19 : 0 FAME=dry soilðgÞ

4

Notes 1. Solutions must be shaken immediately before being used as mixtures of different densities of solvents might lead to phase separation. 2. If the blank contains interfering peaks, the glassware and reusable PTFE lids which could have been machine washed must be rinsed with chloroform and dried before being used. 3. This is especially important when pipetting the extracted lipids as the lipids will likely attach to plastic tips. When pipetting organic solvents, prepipet the solvents to avoid dripping. 4. The amount of the internal standards 19:0 PLFA and 13:0 FAME added need to be adapted to the samples such as to reach average peak heights on the chromatogram. The amounts given here are suitable for temperate climate agricultural soils. 5. Dissolve 1.78 g of citric acid monohydrate and 2.11 g of tripotassium citrate monohydrate in 100 mL of water. 6. For all volumetric solutions it is sufficiently precise to use a volumetric cylinder to determine the volume. 7. Thoroughly mix before use. If citric crystals are formed and cannot be dissolved by mixing, prepare a new solution. 8. It is difficult to dissolve the exact amount of KOH. To avoid intra experimental variation in methylation efficiency, make sure you use the same KOH–methanol for all of your samples within one experiment. Dissolution of KOH in methanol is an exothermal reaction! Add the KOH slowly. 9. Adapt the FAME standard solution to the linearity range of the GC-IRMS while using the same method as for the sample measurements. 10. Measurement is limited by the peak resolution of the GC system and the combustion unit and the sensitivity of the IRMS. Longer GC columns do not lead to much better separation but increase measurement times considerably. The high film thickness is needed because relative high amounts of compounds are introduced onto the column.

Stable Isotope Probing of PLFAs

53

11. The activation of the liner can be monitored by checking the ratio of unsaturated versus saturated FAMEs. Unsaturated FAMEs decrease when the liner is getting more reactive. 12. Determine the water content in 2 g of soil/substrate and reduce the volume of the citric buffer (1.7 mL) by the soil water volume (e.g., 10 weigh percent water content of 2 g soil ¼ 0.2 mL). 13. Freeze drying is an ideal preparation step. Freeze drying is also useful when extracting PLFA from waters. For drainage waters freeze drying of 100 mL of water is usually sufficient. 14. Check your pH after adding the B&D solution to a specific soil/substrate. Sometimes, especially in freeze-dried waters, the citric buffer might be too weak to achieve a pH of 4. In these cases, add 10–100 μL of 1.5 M HCl to remove carbonates prior to extraction. 15. Check in the morning if the Bligh and Dyer above the soil has not separated between water–methanol and a chloroform phase. A one-phase solution is important to enable proper extraction. If you obtain a phase separation, check the sample water correction. 16. In organic soils and substrates some organic particles might still float on the surface after centrifugation. You can increase speed 2500  g and/or time of centrifugation, but the particles usually do not cause a problem, as later steps remove these particles. If you change speed and/or time of centrifugation keep this altered methodology constant for all experimental samples. 17. After centrifugation the soil sometimes sticks on the bottom of the tube and cannot be released by vortexing. In this case, shake manually until the soil is removed before vortexing. 18. The water and organic phase must be clear (i.e., well separated). Increasing the temperature of the solution might help to separate organic and water phase. However, if the phase separation is insufficient, always check the pH of the water phase and the sample water correction. 19. In order to avoid removing the lower phase, suck the water phase from the side as the meniscus bends downward. 20. The flow of the inert gas speeds up the evaporation and avoids oxidation. The gas flow is optimum if the surface shakes slightly. Except for the methanol step (around 1.5 h) the drying is typically very fast (a few minutes). 21. Conduct the soil phase extraction under very low vacuum to achieve a flow of around 2 mL/min in the SPE columns. It is important to avoid that the SPE columns run dry. Always add the next solution, when the former just hits the silica surface.

54

Andrea Watzinger and Rebecca Hood-Nowotny

The flow between different SPE columns with different samples will be different. A vacuum manifold, where the flow of the single SPE columns can be regulated is helpful. 22. The SPE column can contain water, which will interfere with the lipid separation efficiency. Rinsing the SPE columns with chloroform stabilized with ethanol removes residual water. Prerinsing the SPE columns with methanol before chloroform conditioning can also be useful. The package of the SPE columns should be stored closed/airtight. 23. The methylation is not complete. The duration must therefore be kept constant for all comparable samples. Methylation starts when KOH–methanol is added and stops after addition of the acetic acid solution. 24. Hexane is easily lost if the lid is not tightly closed. 25. Extracted FAMEs can be stored for a few months until measurement. Before measurement carefully shake the GC vial to mix the isooctane solution. After measurement close the vials with a new unpunctured lid. 26. Slowly baking out the GC column at 20
C below maximum GC column temperature and holding the temperature overnight whilst maintaining the He flow and the liner at operation temperature will clean the column and simultaneously softens the ferrules to ease tightening. 27. On a combustion II Interface using a NiO/CuO/Pt reactor we oxidize the reactor for 30 min at 620
C (0.5 bar O2 pressure) and flush the reactor at 1.0 bar He pressure while increasing the temperature stepwise by 100
C (10 min hold) until 940
C. Additionally we introduce a 2 s O2 flush after each run. 28. Constant flow slightly improves separation of high molecular weight FAMEs. 29. A standard deviation of 0.3‰ sometimes occurs, even if the measurement is good.

Stable Isotope Probing of PLFAs

55

32. When the FAMEs are transferred through the GC column, heavy isotopes are eluted earlier than light ones, a phenomenon which is called the isotopic swing. To correctly define the stable isotope ratio of the FAME the full isotopic swing needs to be integrated. With microbial FAMEs, this is sometimes not possible, because the peaks are eluted too close to each other. 33. If the working standards are correctly defined and the GC-IRMS system works well, we usually do not observe an offset between the 13:0 measured by EA-IRMS (δtrue std) and GC-IRMS (δraw std). In this case the δtrue sample ¼ δmeasured sample. 34. A PLFA contains two FAs; therefore, the mol PLFA needs to be multiplied by 2.

Acknowledgments The authors would like to thank all the coworkers and students for critical questions and comments when conducting PLFA extraction and GC-IRMS measurements, which over the years improved the ¨ Wirtschafts- und protocol. The County of Lower Austria, NO Tourismusfonds (WST3-F-5031245/003-2018) is acknowledged for funding. References 1. Bligh EG, Dyer WJ (1959) A rapid method of total lipid extraction and purification. Can J Biochem Physiol 37:911–917 2. White DC, Davis WM, Nickels JD et al (1979) Determination of the sedimentary microbial biomass by extractable lipid phosphate. Oecologia 40:51–62 3. Frostega˚rd A˚, Tunlid A, Ba˚a˚th E (1991) Microbial biomass measured as total lipid phosphate in soils of different organic content. J Microbiol Methods 14:151–163 4. Frostega˚rd A, Tunlid A, Ba˚a˚th E (1993) Phospholipid fatty acid composition, biomass, and activity of microbial communities from two soil types experimentally exposed to different heavy metals. Appl Environ Microbiol 59:3605–3617 5. Frostega˚rd A˚, Tunlid A, Ba˚a˚th E (2011) Use and misuse of PLFA measurements in soils. Soil Biol Biochem 43:1621–1625 6. Zelles L (1999) Fatty acid patterns of phospholipids and lipopolysaccharides in the characterisation of microbial communities in soil: a review. Biol Fertil Soils 29:111–129 7. Fernandes MF, Saxena J, Dick RP (2013) Comparison of whole-cell fatty acid (MIDI)

or phospholipid fatty acid (PLFA) extractants as biomarkers to profile soil microbial communities. Microb Ecol 66:145–157 8. Watzinger A (2015) Microbial phospholipid biomarkers and stable isotope methods help reveal soil functions. Soil Biol Biochem 86:98–107 9. Boschker H, Nold S, Wellsbury P et al (1998) Direct linking of microbial populations to specific biogeochemical processes by 13C-labelling of biomarkers. Nature 392:801–805 10. Brenna JT, Corso TN, Tobias HJ et al (1998) High-precision continuous-flow isotope ratio mass spectrometry. Mass Spectrom Rev 16:227–258 11. Meier-Augenstein W (2002) Stable isotope analysis of fatty acids by gas chromatography–isotope ratio mass spectrometry. Anal Chim Acta 465:63–79 12. Evershed RP, Crossman ZM, Bull ID et al (2006) 13C-labelling of lipids to investigate microbial communities in the environment. Curr Opin Biotechnol 17:72–82

Chapter 5 SIP-Metaproteomics: Linking Microbial Taxonomy, Function, and Activity Martin Taubert Abstract Stable isotope probing combined with metaproteomics enables the detection and characterization of active key species in microbial populations under near-natural conditions, which greatly helps to understand the metabolic functions of complex microbial communities. This is achieved by providing growth substrates labeled with heavy isotopes such as 13C, which will be assimilated into microbial biomass. After subsequent extraction of proteins and proteolytic cleavage into peptides, the heavy isotope enrichment can be detected by high-resolution mass spectrometric analysis, and linked to the functional and taxonomic characterization of these biomarkers. Here we provide protocols for obtaining isotopically labeled proteins and for downstream SIP-metaproteomics analysis. Key words Stable isotope probing, Metaproteomics, Microbial ecology, Metabolic labeling, Mass spectrometry

1

Introduction Metabolic labeling of active microbial populations using growth substrates enriched in heavy isotopes, termed stable isotope probing (SIP), has drastically enhanced our ability for a cultureindependent characterization of function and diversity of complex microbial communities in the environment. SIP has unparalleled potential to select for active key organisms while minimizing an enrichment bias when employing near-natural conditions [1, 2]. Since its first application, SIP has been combined with various molecular biology methods, including the analysis of DNA [3], RNA [4], phospholipid-derived fatty acids [5], proteins [6] and whole cells [7, 8]. The advent of omics technologies has resulted in studies employing SIP-metagenomics [9], SIP-metatranscriptomics [10] and SIP-metaproteomics [11], providing in-depth characterization of the active key players in microbial communities. The flexibility in applications makes SIP a

Marc G. Dumont and Marcela Herna´ndez Garcı´a (eds.), Stable Isotope Probing: Methods and Protocols, Methods in Molecular Biology, vol. 2046, https://doi.org/10.1007/978-1-4939-9721-3_5, © Springer Science+Business Media, LLC, part of Springer Nature 2019

57

58

Martin Taubert

useful tool to tackle a variety of questions in microbial ecology. SIP-metagenomics provides access to genomes of rare, uncultured organisms and allows for the construction of hypotheses about their metabolic capabilities. SIP-metaproteomics can confirm such hypotheses by detecting the expression of metabolic pathways and demonstrating activity based on the heavy isotope incorporation in the involved proteins. Metaproteomics relies on mass spectrometric (MS) analysis of peptides obtained from a microbial community by proteolytic cleavage of extracted proteins. This method is inherently able to realize the two critical steps of a SIP experiment: The detection of heavy isotope incorporation, via the corresponding increase in the masses of peptide molecules, and the functional and taxonomic characterization, by identifying a peptide’s sequence based on its MS/MS fragmentation patterns and comparison to a reference database. Performed for tens of thousands of peptides by contemporary liquid chromatography (LC) coupled MS, this approach can shed light on entire microbial communities, linking taxonomy, function, and activity. The sensitive and accurate detection of heavy isotope incorporation furthermore enables a differentiation of the direct uptake of a labeled substrate and cross-feeding, and allows for the use of alternative elements such as sulfur (33S, 34S, 36 S) or hydrogen (2D) [12–14]. Hence, SIP combined with metaproteomics can trace elemental fluxes in microbial communities to identify the metabolic and biogeochemical roles of active microorganisms in the environment [11, 15–17]. Here we provide a general guideline for protein-based SIP experimentation, which will aid users in experimental design and establishing a workflow for recovery and characterization of labeled peptides in a metaproteomics analysis.

2

Materials In addition to the materials described below, further information on the equipment and reagents required for protein-SIP are described elsewhere in detail [6].

2.1 Reagents and Equipment for Protein Extraction

1. SET buffer: 0.75 M sucrose, 40 mM EDTA, 50 mM Tris base pH 9, autoclave, store at room temperature. 2. 100 mM phenylmethylsulfonyl fluoride (PMSF, a protease inhibitor), dissolved in 2-propanol. 3. 10% m/v SDS. 4. Shaking/heating incubator for 15-mL tubes. 5. Centrifuge for 15-mL tubes. 6. PCI: phenol–chloroform–isoamyl alcohol (25:24:1, pH 8.0).

SIP-Metaproteomics

59

7. 0.1 M ammonium acetate, dissolved in methanol. 8. 80% v/v acetone, ice-cold. 9. 70% v/v ethanol, ice-cold. 2.2 Reagents and Equipment for Gel Electrophoresis

1. Apparatus for sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) (see Note 1). 2. Resolving gel (12.5%): 4.35 mL aqua dest., 2.5 mL 1.5 M Tris-HCl buffer pH 8.8, 100 μL 10% m/v SDS, 3 mL 40% acrylamide–bisacrylamide (37.5:1), 7.5 μL tetramethylethylenediamine (TEMED), 75 μL 10% m/v ammonium persulfate (APS) per gel. 3. Stacking gel: 1.58 mL aqua dest., 0.625 mL 0.5 M Tris-HCl buffer pH 6.8, 25 μL 10% m/v SDS, 0.25 mL 40% acrylamide–bisacrylamide (37.5:1), 4 μL TEMED, 19 μL 10% m/v APS per gel. 4. SDS sample buffer: 60 mM Tris-HCl buffer pH 6.8, 10% glycerol, 2% SDS, 5% β-mercaptoethanol, 0.01% bromophenol blue (see Note 2). 5. Electrode buffer (5): 15 g/L Tris base, 72 g/L glycine, 5 g/L SDS. 6. Protein molecular weight marker (e.g., Pierce™ Unstained Protein MW Marker). 7. Coomassie solution: add 40 mg Coomassie G250 to 482.5 mL aqua dest., stir for 2 h at room temperature, then add 17.5 mL 1 M HCl while stirring, filter solution through paper filter if necessary. 8. Microwave oven.

2.3 Reagents and Equipment for InGel Tryptic Cleavage

1. Acetonitrile. 2. Wash solution: 40% v/v acetonitrile, 20 mM ammonium bicarbonate. 3. 5 mM ammonium bicarbonate. 4. 10 mM ammonium bicarbonate. 5. DTT solution: 10 mM 1,4-dithiothreitol in 10 mM ammonium bicarbonate. 6. IAA solution: 100 mM 2-iodoacetamide in 10 mM ammonium bicarbonate. 7. Trypsin buffer: dissolve 20 μg Trypsin Proteomics Sequencing Grade in 20 μL 1 mM HCl. Add 5 μL of dissolved trypsin to 495 μL of 5 mM ammonium bicarbonate directly before use. 8. Extraction buffer: 50% v/v acetonitrile, 5% v/v formic acid. 9. Vacuum centrifuge. 10. 0.1% v/v formic acid.

60

Martin Taubert

11. ZipTip™ C18 Pipette Tips (Merck). 12. NanoDrop™ 2000 spectrophotometer (Thermo Fisher Scientific). 13. Proteome Discoverer™ (Thermo Fisher Scientific).

3

Methods

3.1 Considerations for Planning a SIP Experiment

Based on the questions to be addressed, different design considerations have to be taken into account for a successful SIP experiment, and no universal guideline can be provided. The following steps highlight points of general importance for the experimental design. 1. Include control incubations with unlabeled (12C) substrate. The determination of 13C incorporation is only possible with an unlabeled sample for comparison, and proteomics relies on the unlabeled samples for data analysis. 2. The substrate concentration supplied in a SIP experiment needs to be high enough to stimulate sufficient activity for detection, but still should be in an environmentally relevant range (see Note 3). 3. The use of a fully labeled substrate (i.e., all carbon is replaced by 13C) is recommended. Also make sure that the native substrate concentration in the sampled material is negligible in comparison to the concentration of the added 13C substrate, as otherwise a dilution of the label will occur (see Note 4). 4. When conducting an incubation experiment with SIP, multiple time points should be used to follow the utilization of the substrate and the carbon flux through the microbial community (see Note 5). 5. During the incubations, monitoring of the metabolic activity is highly recommended. This can include measuring the substrate concentration, the concentration and isotope ratio of CO2 or other metabolic products, oxygen or other electron acceptors/ donors etc. 6. Certain microorganisms might use the 13C labeled compound as an energy source but not as a carbon source. Hence, such organisms will not be labeled directly but might become labeled later on via cross-feeding, for example, on the 13CCO2 produced by oxidation of the labeled compound.

3.2 Protein Extraction from Sediment or Water Samples (See Note 6)

1. For water samples: filter through a 0.2 μm pore size filter. 2. Transfer filter or up to 1 g of sediment to a 15-mL-tube. 3. Add 2.6 mL SET buffer. 4. Add 30 μL PMSF.

SIP-Metaproteomics

61

5. Add 350 μL SDS. 6. Incubate for 2 h at 55  C shaking (see Note 7). 7. Add 2 mL PCI, mix vigorously. 8. Centrifuge at 12,000  g for 10 min (see Note 8). 9. Transfer the upper, aqueous phase to new 15-mL tube. 10. Add 2 mL PCI to aqueous phase, mix vigorously. 11. Centrifuge again at 13,000  g for 10 min. 12. Remove the upper, aqueous phase (see Note 9). 13. Combine both PCI phases (see Note 10). 14. Add 2 mL SET buffer to PCI phases, mix vigorously. 15. Centrifuge again at 12,000  g for 10 min. 16. Remove the upper, aqueous phase. 17. Add 20 mL ammonium acetate in methanol to combined PCI phases. 18. Incubate overnight at

20  C for protein precipitation.

19. Centrifuge at 12,000  g and 4  C for 30 min (see Note 8). 20. Remove the supernatant. 21. Wash the pellet with 1 mL ammonium acetate in methanol (twice), 1 mL 80% acetone (twice), 1 mL 70% ethanol (once). 22. For each washing step, incubate for 20 min at 20  C, then centrifuge at 12,000  g for 10 min and remove the supernatant. 23. Air-dry the pellet for 15 min. 24. Store at 3.3 Protein Purification by SDS-PAGE

80  C.

SDS-PAGE traditionally is used for separation of proteins by size. Here, the main purpose is a purification of the sample from (e.g., larger cellular debris or inorganic particles by the filtering effect of the gel matrix); hence, the full length of the gel is not used. 1. Cast sodium dodecyl sulfate–polyacrylamide gels for electrophoresis, consisting of resolving gel (12.5%) and stacking gel, following standard laboratory procedures (see Note 11). 2. Resuspend extracted protein pellets in 35 μL SDS sample buffer. 3. Incubate samples for 10 min at 95  C shaking. 4. Centrifuge samples at 13,000  g for 10 min and transfer the supernatant to a new tube. 5. Assemble electrophoresis apparatus and fill in 1 electrode buffer, rinse wells of stacking gel by pipetting up and down.

62

Martin Taubert

Fig. 1 Gel image of SDS-PAGE with protein extracts from different samples. The broken line indicates the area of the gel to be cut out for in-gel proteolytic cleavage of the respective sample. The gel lane marked with an asterisk shows indications of overloading (see Note 12). Depicted is only the separating gel; the stacking gel above has been removed before staining with Coomassie G250. M: Protein size ladder

6. Load 25 μL of samples into the wells of the stacking gel (see Note 12). 7. Load protein molecular weight marker. 8. Run the gel for 90 min at 10 mA (per gel). 9. Stop electrophoresis when the samples have entered the resolving gel by about 1–2 cm (see Fig. 1). 10. Remove the gel from electrophoresis apparatus. 11. Add 100 mL of aqua dest. to the gel and microwave until it almost boils. 12. Gently shake for 2 min, discard aqua dest. 13. Repeat the washing steps 11 and 12 two more times. 14. Add 300 mL Coomassie solution to the gel and microwave until it almost boils. 15. Gently shake for 10 min and discard Coomassie solution (see Note 13). 16. Add 100 mL of aqua dest and gently shake for 10 min, discard aqua dest. 17. Store the gel in 100 mL fresh aqua dest. 18. Visualize stained protein bands in gel lane for all samples. 3.4 In-Gel Tryptic Cleavage

In this section, the proteins are proteolytically cleaved into shorter peptides. To prevent the formation of disulfide bonds, prior to this, cysteine residues of the proteins are reduced and alkylated with DTT and IAA. The proteolytic cleavage is followed by extraction of the peptides from the gel, purification and quantification. 1. For each sample, excise the stained part of the gel lane and transfer it to one 0.5 mL tube (see Note 14). 2. Add 200 μL wash solution to each gel piece, shake for 10 min (see Note 15), and remove wash solution.

SIP-Metaproteomics

63

3. Add 200 μL acetonitrile, shake for 5 min, and remove acetonitrile. 4. Dry gel pieces in a vacuum centrifuge for 5 min. 5. Add 30 μL DTT solution and shake for 30 min, then remove DTT solution. 6. Add 30 μL IAA solution and shake for 30 min, then remove IAA solution. 7. Add 200 μL acetonitrile, shake for 5 min, and remove acetonitrile. 8. Add 200 μL 10 mM ammonium bicarbonate solution, shake for 10 min, and remove the ammonium bicarbonate solution. 9. Add 200 μL acetonitrile, shake for 5 min, and remove acetonitrile. 10. Dry gel pieces in a vacuum centrifuge for 5 min. 11. Add 20 μL of trypsin buffer and incubate at 37  C overnight. 12. Add 30 μL 5 mM ammonium bicarbonate solution and shake for 10 min. 13. Remove and collect the 5 mM ammonium bicarbonate solution in separate tube for each sample. 14. Add 30 μL extraction buffer to gel piece and shake for 10 min. 15. Remove extraction buffer and add it to the collected 5 mM ammonium bicarbonate solution. 16. Repeat steps 14 and 15 once more. 17. Evaporate the collected solutions in a vacuum centrifuge (see Note 16). 18. Resuspend in 0.1% formic acid and perform ZipTip™ desalting and enrichment of peptide lysates according to manufacturer’s manual (see Note 17). 19. Evaporate the peptide containing eluate in a vacuum centrifuge. 20. Resuspend in 10–20 μL of 0.1% formic acid. 21. Determine peptide concentrations via the absorbance at 260 nm and 280 nm by measuring 1–2 μL peptide lysate on a NanoDrop spectrophotometer (see Note 18). 22. Adjust the volumes of the peptide lysates so that 1 μg of peptides is injected in each LC-MS/MS run. 3.5 SIP-Proteomics Analysis by HighResolution MS

The peptide lysates are separated by liquid chromatography and mass spectrometry is used to analyze peptide masses and perform fragmentation analysis, which allows for a subsequent identification. The measurement of peptide masses also allows for the detection of the mass increase due to the 13C incorporation.

64

Martin Taubert

Fig. 2 Mass spectra obtained from peptide extracts by LC-MS analysis. The upper panel shows peak patterns of multiple unlabeled (12C) peptides, marked with plus signs. The leftmost peak of each pattern is the monoisotopic peak containing only light isotopes. The right-tailing peaks are a result of the natural abundance of heavy isotopes. The lower panel shows peak patterns of multiple labeled (13C) peptides, marked with asterisks, along with further unlabeled peptides. The labeled peptide patterns show Poisson-distribution-like shapes with left- and right-tailing peaks

1. Perform LC-MS analysis using a high-resolution mass spectrometer (e.g., Orbitrap™), or send samples to a proteomics core facility for analysis (see Note 19). 2. Manually inspect mass spectra of 13C labeled samples for an easy and quick way to determine whether 13C labeled peptides are present and the SIP experiment has worked (see Fig. 2). 3. Perform peptide identification based on a suitable reference database (see Note 20), for example, using Proteome Discoverer™ (Thermo Fisher Scientific) or open source software such as OpenMS [18] or MaxQuant [19]. 4. Perform phylogenetic classification of identified peptides by determining the lowest common ancestor, for example, using the tools available at unipept.ugent.be, prophane.de or the MetaProteome Analyzer [20] (see Note 21). 5. Perform identification of 13C labeled peptides by mapping identifications from LC-MS data of unlabeled samples onto data of 13C samples, for example, using the MetaProSIP package [21] of OpenMS or the implemented MetaProSIP workflow in Galaxy [22] (see Note 22). 6. Quantify 13C incorporation in labeled peptides of taxonomic groups of interest to determine the extent of assimilation of carbon from the labeled carbon source (see Note 23).

SIP-Metaproteomics

4

65

Notes 1. The volumes for resolving and stacking gel depend on the size of the gel tank used and have to be scaled accordingly. The volumes given here are for mini-gels (10  8 cm, 0.75 mm thickness). 2. Instead of mercaptoethanol, 100 mM dithiothreitol (final concentration) can be used. 3. When little experience with SIP experiments is present, the applied substrate concentration should not be below 1 mM of carbon to facilitate analysis. Even if this is way above the environmental concentrations observed, such an experiment can still give insights into substrate utilization and carbon transfer. For more experienced experimenters, the use of lower concentrations combined with regular respiking of substrate can help to stay in an environmentally relevant concentration range and still achieve sufficient labeling for detection. 4. To reduce the content of native compounds in the sample, a starvation period can be included before adding the labeled compound. 5. If no information about the expectable activity of the microbial community of interest is available, consider performing a preliminary experiment to identify the best incubation time points to be used in the SIP experiment. With three time points, the first time point should be sampled at the first detection of microbial activity, the second when activity rates reach a maximum and the third when the substrate becomes depleted. 6. Optimal protocols for protein extraction strongly depend on the sample type used. Phenol extraction works well with a broad range of sample materials and hence is a good starting point if no optimised protocols are available. Soils, especially when rich in humic substances, are the most challenging materials for protein extraction. See [23] for more information regarding protein extraction from soils. 7. If no shaking incubator is available, then incubation without shaking (e.g., in a water bath) interrupted by shaking every 30 min (e.g., on a vortex) will yield comparable results. 8. If available, these centrifugation steps can be performed in a swing-out rotor at 4000  g, which typically results in an improved phase separation/pellet formation than a fixed angle rotor. 9. This protocol allows for parallel extraction of DNA from the aqueous phase [24] for 16S rRNA gene analysis or metagenomics (see also Note 20).

66

Martin Taubert

10. Avoid a potential interphase as well as pellets of cell debris that might have formed at the bottom of the phenol phase. If necessary, transfer the clean phenol phase to a new tube. 11. Add TEMED and APS directly before casting the corresponding gel, as these will start the polymerization process. Cast the resolving gel up to 5 mm under the well comb. Overlay with 0.5 mL 2-propanol to create a smooth surface between the resolving gel and the stacking gel, but remove 2-propanol and rinse with water thoroughly before casting the stacking gel. The resolving gel should polymerize within 45 min, the stacking gel within 30 min. 12. When extracting proteins from environmental samples, often insufficient material for protein quantification is available. Dissolving the protein pellet in 35 μL SDS sample buffer and loading 25 μL onto the gel leaves sufficient material for another gel in case the loaded protein amount was too high. Overloading will result in strong staining by Coomassie combined with a lack of separated bands and stained proteins extending beyond the gel lanes. 13. The microwave staining procedure is optimized to quickly visualize protein bands within 30 min. Alternatively, the procedure can be done without microwaving by washing for 30 min each and staining overnight. 14. Large gel pieces can be crushed or cut into smaller parts to facilitate in-gel proteolytic cleavage. Gel pieces can also be cut horizontally into 2–4 bands and split up into an according number of 0.5 mL tubes to fractionate each sample into multiple parts. This can increase the total number of protein identifications in the mass spectrometric analysis, but will also increase the preparation and measuring time required. 15. This washing step can be repeated several times if the gel piece still shows strong Coomassie staining. 16. At this step, for each gel piece one tube containing the gel piece without any liquid (which can be discarded) and one tube containing the collected solutions (ammonium bicarbonate and extraction buffer) with the tryptic peptides should exist. The complete evaporation of the collected solutions in the vacuum centrifuge can take several hours. 17. The ZipTips have a loading capacity of 5 μg of peptides on the column (large column bed tips). ZipTips provide high reproducibility, reliability, and recovery of peptides for LC-MS/MS analysis [25]. 18. The quantification of peptide concentration before LC-MS/ MS analysis is highly recommended, as it will increase quality

SIP-Metaproteomics

67

and reproducible of the mass spectra. An absolute quantification can be achieved by calibration with a peptide standard. If this is not available, use a relative quantification for equalizing the amount of peptide lysate to be injected into LC-MS/MS between samples, to avoid sample-to-sample variations. 19. The general procedure for SIP-proteomics is the same as for normal proteomics analysis, and thus should be done according to “in house” proteomics protocols. 20. Proteomics analysis is only able to identify peptides present in the reference database used. If the reference database does not reflect the microbial diversity in the sample, analysis will yield a low number of identified peptides and will produce a biased picture of the microbial community. A large database will also increase the chance of false positive identifications [26]. The ideal reference database is an assembled and binned metagenome obtained from the sample investigated. If this is not available, 16S rRNA gene analysis can be used to determine the community composition and compile a suitable database reflecting this composition from publicly available sequence databases. 21. Peptide identification relies on the presence of unlabeled peptides, so 13C labeled peptides cannot be identified. When comparing the number of identified peptides between 12C and 13C samples, a drop in identifications is the first hint that a specific taxonomic group has been labeled. 22. For a low number of peptides, manual analysis can be performed, as previously described [11]. Alternatively, an additional LC-MS analysis of a mix of labeled and unlabeled sample allows for a direct identification of 13C labeled peptides but prevents a quantitative comparison of 12C and 13C peptide signal. For a computationally highly expensive approach for identification of 13C peptides without relying on 12C peptide signals, see [17]. 23. Peptides from the same organism typically show a uniform 13C incorporation, which allows for the characterization of carbon usage by this organism. For an in-depth review of the interpretation of 13C incorporation patterns, see [27].

Acknowledgments I thank Nico Jehmlich for providing helpful inputs and for reviewing the manuscript.

68

Martin Taubert

References 1. Friedrich MW (2006) Stable-isotope probing of DNA: insights into the function of uncultivated microorganisms from isotopically labeled metagenomes. Curr Opin Biotech 17:59–66 2. Neufeld JD, Dumont MG, Vohra J, Murrell JC (2007) Methodological considerations for the use of stable isotope probing in microbial ecology. Microb Ecol 53:435–442 3. Neufeld JD, Vohra J, Dumont MG, Lueders T, Manefield M, Friedrich MW et al (2007) DNA stable-isotope probing. Nat Protoc 2:860–866 4. Whiteley AS, Thomson B, Lueders T, Manefield M (2007) RNA stable-isotope probing. Nat Protoc 2:838–844 5. Boschker HTS, Nold SC, Wellsbury P, Bos D, de Graaf W, Pel R et al (1998) Direct linking of microbial populations to specific biogeochemical processes by C-13-labelling of biomarkers. Nature 392:801–805 6. Jehmlich N, Schmidt F, Taubert M, Seifert J, Bastida F, von Bergen M et al (2010) Proteinbased stable isotope probing. Nat Protoc 5:1957–1966 7. Huang WE, Stoecker K, Griffiths R, Newbold L, Daims H, Whiteley AS et al (2007) Raman-FISH: combining stableisotope Raman spectroscopy and fluorescence in situ hybridization for the single cell analysis of identity and function. Environ Microbiol 9:1878–1889 8. Musat N, Halm H, Winterholler B, Hoppe P, Peduzzi S, Hillion F et al (2008) A single-cell view on the ecophysiology of anaerobic phototrophic bacteria. Proc Natl Acad Sci U S A 105:17861–17866 9. Dumont MG, Radajewski SM, Miguez CB, McDonald IR, Murrell JC (2006) Identification of a complete methane monooxygenase operon from soil by combining stable isotope probing and metagenomic analysis. Environ Microbiol 8:1240–1250 10. Dumont MG, Pommerenke B, Casper P (2013) Using stable isotope probing to obtain a targeted metatranscriptome of aerobic methanotrophs in lake sediment. Env Microbiol Rep 5:757–764 11. Taubert M, Vogt C, Wubet T, Kleinsteuber S, Tarkka MT, Harms H et al (2012) Protein-SIP enables time-resolved analysis of the carbon flux in a sulfate-reducing, benzene-degrading microbial consortium. ISME J 6:2291–2301 12. Jehmlich N, Kopinke FD, Lenhard S, Vogt C, Herbst FA, Seifert J et al (2012) Sulfur-36S stable isotope labeling of amino acids for quantification (SULAQ). Proteomics 12:37–42

13. Justice NB, Li Z, Wang YF, Spaudling SE, Mosier AC, Hettich RL et al (2014) N-15and H-2 proteomic stable isotope probing links nitrogen flow to archaeal heterotrophic activity. Environ Microbiol 16:3224–3237 14. Taubert M, Sto¨ckel S, Geesink P, Girnus S, Jehmlich N, von Bergen M et al (2018) Tracking active groundwater microbes with D2O labelling to understand their ecosystem function. Environ Microbiol 20:369–384 15. Herbst FA, Bahr A, Duarte M, Pieper DH, Richnow HH, von Bergen M et al (2013) Elucidation of in situ polycyclic aromatic hydrocarbon degradation by functional metaproteomics (protein-SIP). Proteomics 13:2910–2920 16. Lu¨nsmann V, Kappelmeyer U, Benndorf R, Martinez-Lavanchy PM, Taubert A, Adrian L et al (2016) In situ protein-SIP highlights Burkholderiaceae as key players degrading toluene by para ring hydroxylation in a constructed wetland model. Environ Microbiol 18:1176–1186 17. Pan CL, Fischer CR, Hyatt D, Bowen BP, Hettich RL, Banfield JF (2011) Quantitative tracking of isotope flows in proteomes of microbial communities. Mol Cell Proteomics 10:M110.006049 18. Ro¨st HL, Sachsenberg T, Aiche S, Bielow C, Weisser H, Aicheler F et al (2016) OpenMS: a flexible open-source software platform for mass spectrometry data analysis. Nat Methods 13:741–748 19. Tyanova S, Temu T, Cox J (2016) The MaxQuant computational platform for mass spectrometry-based shotgun proteomics. Nat Protoc 11:2301–2319 20. Muth T, Kohrs F, Heyer R, Benndorf D, Rapp E, Reichl U et al (2018) MPA portable: a stand-alone software package for analyzing metaproteome samples on the go. Anal Chem 90:685–689 21. Sachsenberg T, Herbst FA, Taubert M, Kermer R, Jehmlich N, von Bergen M et al (2015) MetaProSIP: automated inference of stable isotope incorporation rates in proteins for functional metaproteomics. J Proteome Res 14:619–627 22. Hoekman B, Breitling R, Suits F, Bischoff R, Horvatovich P (2012) msCompare: a framework for quantitative analysis of label-free LC-MS data for comparative candidate biomarker studies. Mol Cell Proteomics 11: M111.015974

SIP-Metaproteomics 23. Qian C, Hettich RL (2017) Optimized extraction method to remove humic acid interferences from soil samples prior to microbial proteome measurements. J Proteome Res 16:2537–2546 24. Taubert M, Grob C, Howat AM, Burns OJ, Chen Y, Neufeld JD et al (2016) Analysis of active methylotrophic communities: when DNA-SIP meets high-throughput technologies. Methods Mol Biol 1399:235–255 25. Jehmlich N, Golatowski C, Murr A, Salazar G, Dhople VM, Hammer E et al (2014)

69

Comparative evaluation of peptide desalting methods for salivary proteome analysis. Clin Chim Acta 434:16–20 26. Muth T, Kolmeder CA, Saloj€arvi J, Keskitalo S, Varjosalo M, Verdam FJ et al (2015) Navigating through metaproteomics data: a logbook of database searching. Proteomics 15:3439–3453 27. Seifert J, Taubert M, Jehmlich N, Schmidt F, Volker U, Vogt C et al (2012) Protein-based stable isotope probing (protein-SIP) in functional metaproteomics. Mass Spectrom Rev 31:683–697

Chapter 6 Chip-SIP: Stable Isotope Probing Analyzed with rRNA-Targeted Microarrays and NanoSIMS Xavier Mayali, Peter K. Weber, Erin Nuccio, Jory Lietard, Mark Somoza, Steven J. Blazewicz, and Jennifer Pett-Ridge Abstract Chip-SIP is a stable isotope probing (SIP) method for linking microbial identity and function in mixed communities and is capable of analyzing multiple isotopes (13C, 15N, and 18O) simultaneously. This method uses a high-density microarray to separate taxon-specific 16S (or 18S) rRNA genes and a high sensitivity magnetic sector secondary ion mass spectrometer (SIMS) to determine the relative isotope incorporation of the rRNA at each probe location. Using a maskless array synthesizer (MAS), we synthesize multiple unique sequences to target hundreds of taxa at the ribosomal operational taxonomic unit (OTU) level on an array surface, and then analyze it with a NanoSIMS 50, using its high-spatial resolution imaging capability to generate isotope ratios for individual probes. The Chip-SIP method has been used in diverse systems, including surface marine and estuarine water, rhizosphere, and peat soils, to quantify taxon-specific relative incorporation of different substrates in complex microbial communities. Depending on the hypothesis and experimental design, Chip-SIP allows the user to compare the same community incorporating different substrates, different communities incorporating the same substrate(s), or quantify how a community responds to treatment effects, such as temperature or nutrient concentrations. Key words Stable isotope probing, NanoSIMS, 16S rRNA, Microbial ecology, Microarrays, 15 N, 18O

1

13

C,

Introduction One of the goals of microbial ecology is to link the identity of microorganisms with their biogeochemical function in complex communities without cultivation. With stable isotope probing (SIP), this goal is accomplished by incubating an environmental sample with a substrate enriched in a rare stable isotope, followed by analyses that identify the organisms that consumed the substrates and incorporated the stable isotope into their biomass [1]. Multiple SIP techniques take advantage of isotope tracing to link microbial identity and function by targeting biomolecules ranging from DNA to RNA, lipids, proteins, and even intact cells

Marc G. Dumont and Marcela Herna´ndez Garcı´a (eds.), Stable Isotope Probing: Methods and Protocols, Methods in Molecular Biology, vol. 2046, https://doi.org/10.1007/978-1-4939-9721-3_6, © Springer Science+Business Media, LLC, part of Springer Nature 2019

71

72

Xavier Mayali et al.

[1–3]. We briefly describe the most commonly used approach— density gradient SIP—and factors that led to the development of an alternative method called Chip-SIP, where nucleic acids are first sorted on the probe spots of a phylogenetic microarray, and then evaluated for isotopic enrichment via a NanoSIMS imaging mass spectrometer. Like the earlier published “14C isotope array” approach [4], Chip-SIP targets community RNA but is compatible with stable isotope tracing studies. In the most commonly used SIP approach, nucleic acids are extracted after an isotope tracing experiment and enriched fragments of DNA/RNA are separated from unenriched fragments with isopycnic ultracentrifugation, since nucleic acids containing a heavy isotope have a higher buoyant density. After this separation, a number of sequencing approaches can be used to identify the active organisms, targeting 16S rRNA genes [5], metagenomes [6, 7], or metatranscriptomes [8]. Recent improvements to density SIP go a step beyond simple identification and show how SIP results combined with qPCR data can be used to calculate population growth of individual taxa [9, 10]. These SIP approaches are widely used and have contributed to hundreds of published research studies. However, they can generally identify activity with one stable isotope label at a time (e.g., 13C, 15N, or 18O) instead of combinations of multiple stable isotopes (e.g., dual- or triple-isotope experiments, although see [11]), and require relatively high levels of isotope labeling (>8 atom % for 13C for a given taxon; [9]). They also tend to be relatively low throughput and time-intensive because the DNA or RNA gradient must be manually fractionated. Chip-SIP aims to accomplish a similar goal as density gradient SIP but in reverse order by first separating nucleic acids by taxonomic identity and then measuring the isotope incorporation of individual rRNAs (Fig. 1). Taxa are identified based on hybridization of DNA oligonucleotides targeting the small subunit ribosomal RNA (16S for bacteria and archaea, and 18S for eukaryotes), in a manner analogous to the fluorescent in situ hybridization (FISH) approach used for identifying intact cells [12]. Our initial array probe design was modeled on the Phylochip [13], which targets large portions of mixed microbial communities with a 16S phylogenetic microarray, including multiple probes for each taxon. However, for the Phylochip method, PCR is used to amplify community DNA prior to hybridization. For an isotope tracing experiment, this is unacceptable because it would dilute any isotope signal below detection limits. In Chip-SIP, ribosomal RNA is the target (not DNA); since ribosomal RNA is relatively abundant, extracted rRNA is hybridized directly to the array surface without amplification. In order to detect a stable isotope signal from hybridized RNA fragments, we use a secondary ion mass spectrometer (SIMS), which directly ionizes and analyzes the isotopic

Chip-SIP

73

Isotopically RNA enriched

+ 15N

Incubation time varies

13C

Label RNA with fluorescent molecule

No labeling

Hybridization to phylogenetic 16S rDNA microarray

Fluorescence signal (how much RNA?)

NanoSIMS 13C/12C (how enriched is RNA?)

50 µm

13C/12C

enriched

unenriched

fluorescence

Fig. 1 General overview of the Chip-SIP approach

composition of rRNA molecules hybridized to the DNA probes on the array surface. Chip-SIP has been applied in a number of environmental microbiology studies, briefly summarized here. The method was first described in an examination of organic matter niche partitioning in an estuary, and used to illustrate resource partitioning between closely related taxa growing on simple organic substrates such as amino acids, nucleic acids or fatty acids [14]. Chip-SIP was next used to measure the relative incorporation of C versus N from dual labeled amino acids in different taxonomic groups [15], as well as the differentiation of trophic strategies in estuarine taxa adapted to low or high nutrient concentrations [16]. A subsequent study examined the degradation of algal-derived particles over time [17], and another carried out SIP experiments with 14 different

74

Xavier Mayali et al.

substrates to differentiate generalists and specialists in the coastal Pacific Ocean [18]. In 2017, Bryson et al. used Chip-SIP in combination with proteomics-SIP, amplicon and metagenomic sequencing of marine microbial communities, and showed differential assimilation of substrates into protein and ribonucleotide biomass for multiple taxa [19]. In a dual labeling study of soil bacteria, Chip-SIP was used to study bacteria and eukaryotes that consumed plant exudates versus labeled root detritus, and showed that most rhizosphere taxa utilized both root and litter-derived resources [20]. Finally, our group has tested the method in several additional systems, using it to study methane oxidizers in soil from a peat bog, carbon metabolism of microbial communities living in the gut of wood-eating Passalid beetles, and assimilation of different nitrogen sources by forest soil microorganisms (X. Mayali, P.K. Weber, D. Myrold, and J. Pett-Ridge, unpublished data). After a SIP experiment has been conducted and preliminary community composition has been characterized by amplicon sequencing, the first step of the Chip-SIP approach is to design probes for the microarray. As with the PhyloChip, we generate a group (10–20) of nonredundant 25-bp DNA probes that target the most abundant taxa in the community, or specific OTUs of interest, usually on the order of a few hundreds. The overall goal in probe design (Fig. 2) is to minimize cross-hybridization potential yet maintain high specificity. The probe design process creates probes

Fig. 2 Overview of the probe design pipeline

Chip-SIP

75

based on next-generation sequencing (NGS) data and evaluates both probe quality and specificity in silico. Experimental NGS sequences are inserted into the main phylogenetic tree in the ARB software environment [21], and then groups of sequences are selected for probe design. Group clusters within the tree are detected using TreeOTU [22], which uses branch length to select nearby sequences, and then further refines groups using the BLAST percent similarity of those sequences to the original experimental sequence. After group selection is complete, probes are automatically generated for each group using an ARB Probe Design macro, and the resulting probes are checked for specificity using an additional ARB Probe Match macro (macros available at https:// github.com/enuccio/auto_probe_design). Probes are then assessed for quality (e.g., G + C content, homopolymer runs, hairpin formation) and taxonomic specificity to ensure that any outgroup hits are still closely related to the group of interest. The final table provides a list of probes that pass quality thresholds and indicates the consensus taxonomy of the probe to the last common taxonomic level. To reduce the number of nonspecific probes, we often test probe specificity empirically by printing a large array of probes and hybridize with nontarget RNA; probes that bind the nontarget RNA are then discarded. Microarrays for Chip-SIP are made using a photolithographic microarray fabrication method called Maskless Array Synthesis (MAS) [23, 24]. MAS was originally developed for custom synthesis of high-density DNA microarrays for genomics applications (see Note 1). A MAS instrument can be conceptually divided into two components, an optical system and a chemical delivery system. The chemical side consists of an automated oligonucleotide synthesizer, which is used to deliver reagents and solvents to a reaction chamber defined by two functionalized glass surfaces, separated by a thin gasket, where the microarray synthesis takes place [25]. The optical system is similar to that of a photolithographic system, but uses a digital micromirror device (DMD) instead of photomasks to deliver ultraviolet (UV) light in a specific pattern [26]. This pattern, displayed on the DMD, is imaged onto the microarray surface, where the layout and oligonucleotide sequences are determined by selective removal of the photocleavable protecting groups on the phosphoramidites. The setup of the optical system and the synthesis chemistry are illustrated in Fig. 3. A significant advantage of this microarray synthesis approach is that the individual spots are quite small, 14
14 μm in our setup, and are separated by less than one micron (see Note 2). This high density enables efficient isotopic analysis because it requires less time for analysis and tuning during the isotope detection step in the NanoSIMS mass spectrometer. Another advantage of microarray printing through a MAS unit is that there is great flexibility in the choice of surface functional group to allow DNA synthesis. Specifically, we used standard-

76

Xavier Mayali et al.

Fig. 3 (a) Schematic of the optical system of the maskless array synthesizer: A—365 nm source, B—dichroic mirror, C—light homogenizer, D—shutter, E—mirrors, F—DMD, G—Offner relay primary mirror, and H— secondary mirror, I—Reaction chamber, J—Light dump. (b) Array layout determination in MAS showing two simplified dT and dG coupling cycles. (c) The synthesis cycle is similar to that used in solid-phase synthesis except that UV light rather than a strong acid is used to deprotect the 50 hydroxyl. The capping step is omitted for CHIP-SIP arrays. Figure a reproduced from [23] with publisher’s permission

dimension glass microscope slides with a conductive indium tin oxide (ITO) layer, necessary because NanoSIMS analysis requires a conductive surface. In order for the initial phosphoramidite to form a covalent bond to the surface, an epoxy functionalization on top of the ITO is also necessary. In principle, other conductive layers, such as gold or aluminum, and other surface functionalizations, such as hydroxyl or amine, should also be effective [27]. These functional groups are added to the surface using standard silane chemistry. Array synthesis time for Chip-SIP slides is 4.5 h (see Note 3). Generally, probes are not randomly located throughout the microarray surface but are instead grouped into smaller “subarrays” that are specific to a given community. Within the subarrays, probes are arranged by taxon, and the subarray is surrounded (on all four sides) by a row of probes that target the positive control oligonucleotide CPK6. These control probes are later used to measure the background isotope signal for each hybridized sample. Sample hybridization for Chip-SIP follows standard hybridization protocols. In our initial years applying the method, we used the Nimblegen Hybridization buffer but the exact components are not published and it is no longer commercially available, thus we have switched to the Affymetrix hybridization buffer, which can be made from standard laboratory chemicals. Similarly, for efficient hybridization, we originally used Nimblegen mixer slides in combination with the MAUI hybridization system (which uses airflow to mix the sample throughout the 18-h incubation). These

Chip-SIP

77

Fig. 4 Simultaneous NanoSIMS analysis of 13C, 15N, and 18O from triple-labeled RNA from a pure culture of Bacillus cereus hybridized to a microarray targeting 16S rRNA (different probe spots target a different section of the gene). Top: NanoSIMS ratio images; Bottom: extracted ratio data for all probe locations

slides are unfortunately no longer commercially available, and we currently use a SecureSeal hybridization chamber (without airflow). To determine the isotopic composition of individual rRNA probe spots on the microarray surface, we use a nanoSIMS 50 instrument (CAMECA, Gennevilliers, France). Typically, we analyze the probe spots for carbon or nitrogen isotopic composition, but analysis for other isotopic ratios is also possible, for example 18O (Fig. 4). For both carbon and nitrogen isotopic analysis, we use a Cs+ analysis beam and detect CN because the count rate of this species is >10
higher than any other C or N bearing secondary ion species during the relatively short period that the rRNA persists under ion bombardment. To obtain the 13C/12C ratio, we monitor 12C14N at mass 26 and 13C14N at mass 27, and to obtain the 15N/14N ratio, we monitor 12C14N at mass 26 and 12C15N at mass 27. Because of the relatively highcount rate for CN, we find we achieve better results for C by switching detection back and forth between 13C14N and 12C15N at mass 27 than would be obtained by monitoring either C or

78

Xavier Mayali et al.

C2. To measure 18O/16O ratios, we monitor 16O at mass 16 and 18  O at mass 18. A high-sensitivity, high-mass-specificity SIMS instrument like a NanoSIMS is necessary for detection because of the small amount of rRNA in each probe spot and the need to resolve isobaric interferences to obtain precise and accurate isotope measurements. For example, boron is a surface contaminant that yields 11B16O at mass 27; it requires high mass resolving power to resolve 11B16O from the CN ion to obtain accurate measurements (ΔM ¼ 0.0041 amu from 12C15N and ΔM ¼ 0.0022 amu from 13C14N). Other large geometry SIMS instruments such as the CAMECA ims-1280 and the ANU SHRIMP could conceivably make this measurement, but the NanoSIMS series instruments have the advantage of high spatial resolution (which facilitates locating the target probe spots), and a larger sample holder which can accommodate half of a traditional glass microscope slide (25 mm
37 mm). One of the crucial steps for a successful Chip-SIP experiment is finding the appropriate area to analyze in the NanoSIMS. It is not possible to analyze the entire microarray slide surface; due to its size, it would take months of continuous analysis. We design smaller areas of the microarray surface, called “subarrays”, to target the specific community of interest for a given project; these are generally comprised of a few thousand probes, targeting several hundred taxa. This allows for replicate subarrays on the same slide, and the possibility of having different subarrays targeting different communities printed on the same microarray. Subarrays can also be smaller than described above if the sample community is less complex, or for laboratory cultures used for testing. Since the hybridized microarray looks featureless in a brightfield microscope, our approach to find the correct subarray for analysis is to use a fluorescent image taken after hybridization of the RNA and a fluorescently labeled control oligonucleotide that binds to control spots. The control spots can be found in relation to scratch marks made with a diamond pen, which can be identified both in the fluorescent image and the brightfield microscope in the NanoSIMS instrument. Specific NanoSIMS instrument parameters can be found in the methods section below, but generally we aim for ~3–6 pA of current and usually use a L1 value ~1000–1400 to achieve this. Once the top left (or any corner) of the sub-array is found by monitoring the CN- ions, a chain analysis automated run can be started to collect the entire subarray area. Chip-SIP generates a large amount of quantitative NanoSIMS image data that needs to be matched to the original array design. While multiple programs have been developed to extract and quantify NanoSIMS image data, the custom software L’IMAGE from Larry Nittler at the Carnegie Institute in Washington has the capability to

Chip-SIP

79

semiautomatically extract isotope data in an array format, including the ability to account for scanning distortion and stage movement issues. This capability facilitates extracting the large amount of data collected. The final step of matching the isotope enrichment data for each probe spot to the probe’s identity is performed in a spreadsheet program (e.g., Microsoft Excel or similar). To determine the relative isotope incorporation for each taxon, the isotope data are combined with a measurement of how much RNA is hybridized to each probe location (measured by fluorescence intensity). For each taxon, isotopic enrichment of individual probe spots is plotted versus probe fluorescence, and a linear regression slope is calculated with the y intercept constrained to natural isotope abundances (0 per mil). This calculated slope (per mil/fluorescence) is a metric that can be used to compare the relative incorporation of a given substrate by different taxa. Since fluorescent labeling introduces a substantial amount of natural abundance organic matter, we hybridize two separate arrays with the same sample, the first for NanoSIMS analysis, and the second for a fluorescently labeled hybridization (Fig. 1).

2

Materials

2.1 Probe Design Software

1. Python (Packages: Biopython, ViennaRNA, numpy, natsort). 2. ARB (download instructions: http://www.haloarchaea.com/ resources/arb/). 3. BLAST+. 4. TreeOTU (https://github.com/dongyingwu/TreeOTU).

2.2 Microarray Synthesis

1. Expedite 8909 synthesizer (PerSeptive Biosystems).

2.2.1 Synthesis System and Glass Substrates

3. ITO Super Epoxy 2 microscope slides (Arrayit, lot #160119, or Sigma-Aldrich cat # 6369 16 functionalized with Superepoxy2 coating by Arrayit).

2. High-power 365 nm UV-LED (Nichia NVSU333A).

4. Multimeter (Voltcraft). 5. TickoPur RW 77 special purpose cleaner (Sigma-Aldrich). 2.2.2 Phosphoramidites and Synthesis Reagents

1. NPPOC-dA, dC, dG, and dT phosphoramidites (available from Orgentis Chemicals): diluted to 0.03 M in dry acetonitrile. 2. Activator: 0.25 M dicyanoimidazole in ACN (Biosolve). 3. Oxidizer: 0.02 M I2 in THF–H2O–pyridine 90.5:9.05:0.41 v/v/v (Sigma-Aldrich).

80

Xavier Mayali et al.

4. Exposure solvent: 1% imidazole (w/w, Sigma-Aldrich) in DMSO (Biosolve). 5. Acetonitrile, dry (Biosolve). 2.2.3 Microarray Deprotection

1. Ethylenediamine >99.5% (Sigma-Aldrich). 2. Ethanol 99% (Millipore). 3. Deionized H2O.

2.3 RNA Extractions and Hybridization

1. Qiagen RNeasy kit (or alternate RNA extraction kit such as Mobio Powersoil Kit). 2. Ulysis Alexa Fluor 546 nucleic acid labeling kit. 3. Hybridization chamber such as MAUI Hybridization system. 4. SecureSeal™ Hybridization Chambers (7
7 mm; Grace Bio-Labs). 5. Isopropanol, NaCl, 70% ethanol. 6. 2
fragmentation buffer (Affymetrix) or similar for heat activated nucleic acid fragmentation. 7. 12
MES stock (pH 6.5): 70.4 g of MES free acid monohydrate, 193.3 g MES sodium salt per liter; do not use DEPCtreated water (see Note 4). 8. 2
hybridization buffer: 2
MES, 2 M NaCl, 40 mM EDTA, 0.02% Tween 20. 9. Formamide, molecular grade. 10. NSWB (nonstringent wash buffer): 6
SSPE (sodium chloride–sodium phosphate–EDTA), 0.01% Tween 20, pH 7.4. 11. SWB (Stringent Wash Buffer): 100 mM MES, 0.1 M NaCl, 0.01% Tween 20. 12. FWB (Final Wash Buffer): 0.2
SSC (saline–sodium citrate). 13. Microarray High Speed Centrifuge (ArrayIt cat #MHC110V). 14. Axon Instrument Genepix 4000B fluorescent microarray scanner. 15. Nimblescan software. 16. Genepix Pro analysis software.

2.4 SIMS Mass Spectrometry and Data Analysis

1. Preferably a CAMECA NanoSIMS 50 or NanoSIMS 50L, but a CAMECA ims1280 or Australian Scientific Instruments SHRIMP has sufficient sensitivity and specificity. 2. SIMS custom holder that can fit one half of a standard microscope glass slide (ims1280 or SHRIMP would require using smaller slide pieces). 3. L’IMAGE SIMS processing software (http://limagesoftware. net/).

Chip-SIP

3 3.1

81

Methods Probe Design

3.1.1 Create Supporting Files

The full protocol, including downloadable scripts, is available here: https://github.com/enuccio/auto_probe_design. 1. Cluster NGS sequences using clustering algorithm of choice (e.g., UPARSE, dada2) to cluster sequences by OTU or ASV. 2. Create a tree by adding representative OTUs or ASVs to the global ARB Silva phylogeny. Export the tree in newick format. 3. Separately export the entire ARB database in fasta-wide format. 4. Create a BLAST database from the ARB database. BLAST the experimental sequences against the ARB database.

3.1.2 Create Probe Groups and Probe Design

1. Detect group clusters on the tree using TreeOTU (branch length 0.02). 2. Build the PT server for the ARB database containing the experimental sequences. Note the position number of the PT server in the PT server list for the next step. 3. Generate OTU groups by combining TreeOTU and BLAST results, and then generate an ARB macro file (using script: assign_otus.py). This script will only keep members of the TreeOTU that are 97% similar by BLAST to the original sequence. 4. Run the Probe Design macro generated by assign_otus.py to automatically select group clusters and design associated probes (macro: design_probes.amc).

3.1.3 Probe Specificity Check

1. Parse probe files from ARB, then create a macro to run a probe specificity check (script: parse_probe_files.py). 2. Batch test in silico probes against ARB database using macro generated in previous step (macro: check_probes.amc). 3. Parse ARB probe match files and check the quality of the probes (script: parse_probe_match.py). This script parses the ARB probe match files, checks probe quality heuristics (e.g., G + C content, homopolymer runs, hairpin formation), and then groups probes into different sets based on number of mismatches (perfect match; 1 mismatch; 1 mismatch in unique internal 17-mer; 2 mismatches). 4. Generate consensus taxonomy for probes; check probe specificity using taxonomy (script: probe_specificity.py). Output consensus taxonomy that indicates the specificity of the probe (last common taxonomic level). 5. Generate probe specificity statistics by calculating probe specificity for probes using BLAST (script: probe_specificity_statistics.py).

82

Xavier Mayali et al.

Compare the full-length sequence of interest (used to design the probe) against all full-length sequences that could be homologous to the probe (probe hits calculated by ARB probe match). Complete probe specificity check for all sequences. 3.2 Microarray Fabrication

Microarray synthesis on conductive indium tin oxide substrates requires a few adjustments to the general array fabrication protocol, which are described below.

3.2.1 Substrate and Synthesis Cell Preparation

1. Identify the conductive face of the ITO slide by measuring its electrical resistance with a multimeter. The ITO-functionalized face typically gives a resistance of 300–400 Ω corner-to-corner. 2. In a dual array synthesis system with two ITO slides, drill one slide at two locations with a 0.9 mm diamond bit, then proceed with washing with deionized water and rinsing in an ultrasonic bath with the special purpose cleaner (30 min). After rinsing, wash the slide with deionized water then acetonitrile and dry in a microarray centrifuge. During drilling, protect the synthesis area with a self-adhesive chamber (Grace BioLabs SA200), to prevent surface damage. Drilled and nondrilled slides are stored at room temperature in a desiccator. 3. Assemble the synthesis cell by joining the two slides, conductive sides facing each other, using a 50 μm-thick Teflon gasket as a spacer. 4. Before starting the synthesis, measure the radiant intensity R1 of UV light with a calibrated UV intensity meter (Su¨ss MicroTec Model 1000). 5. Record the radiant intensity R2 of UV light when placing an ITO microscope slide directly in front of the UV intensity detector. Calculate the light transmittance R2/R1 through ITO substrates. This value usually oscillates around 0.6.

3.2.2 Microarray Synthesis

The synthesis itself largely follows standard protocols described elsewhere [28]. However, the exposure time parameter, which defines the amount of UV light reaching the synthesis surface, must take into account the loss of UV light when passing through a layer of indium tin oxide. For instance, for a radiant exposure of 6 J/cm2 sufficient to remove >95% of NPPOC protecting groups, and at an output power of 100 mW/cm2 from the UV LED, assuming 0.6 light transmittance, the exposure time is set to 100 s. This corresponds to a 40% increase in exposure time compared to a synthesis on standard microscope slides.

3.2.3 Microarray Deprotection

1. Fill a 55-mL staining glass jar with 1:1 ethylenediamine/ethanol (15 mL each). 2. Place the synthesized slides in different grooves within the glass jar, then firmly close the jar.

Chip-SIP

83

3. Leave the arrays to react, without shaking, for 2 h at room temperature. 4. Take the slides from the jar, rinse thoroughly with deionized water, and dry in a microarray centrifuge. 5. Place the dried microarrays in a 50-mL falcon tube, then store in a desiccator until further use. 3.3 RNA Extraction and Hybridization

1. Extract RNA with the protocol most appropriate for the sample type. For aquatic samples of microbes collected on filters, we typically use the Qiagen RNeasy mini kit according to manufacturer’s instructions for bacterial nucleic acid, using 10 mg/ mL lysozyme in the lysis step and including vortexing for 10 min. 2. Split RNA sample: one fraction is used for fluorescent labeling; the other is hybridized unlabeled for NanoSIMS analysis. 3. Label half the RNA with the Ulysis Alexa Fluor 546 labeling kit (Thermo) for 10 min at 90  C (2 μL of RNA, 10 μL of labeling buffer, 2 μL of Alexa Fluor reagent); cool on ice. 4. Fragment the RNA (fluorescently labeled or not) using 1
fragmentation buffer (Affymetrix) for 10 min at 90  C. Concentrate by isopropanol precipitation (with added salts to make it 0.5 M), pellet, wash in 70% ice cold ethanol, and resuspend to a final concentration of 500 ng/μL in water. 5. To hybridize, add 1-μg RNA sample in 1
hybridization buffer with 30% final concentration formamide (50 μL total volume per sample) to a SecureSeal chamber affixed to the microarray area (see Note 5; 4 samples may be hybridized per array). Add 0.5 μL of the alignment oligomer CPK6, labeled with Cy3 (final 100 μM). Incubate the slide inside a Maui hybridization system (BioMicro® Systems) or other hybridization chamber for 18 h at 42  C. 6. Place the array in a prewarmed (45  C) container of nonstringent wash buffer, remove the SecureSeal chamber, and wash the array for 1 min with agitation at room temperature, alternating between the nonstringent wash buffer, the stringent wash buffer, and the final wash buffer. Spin the array dry with a microarray centrifuge. 7. Image both arrays (fluorescently labeled RNA and nonlabeled RNA) with a Genepix 4000B fluorescence scanner at pmt. ¼ 650 units (adjust the pmt. setting to achieve optimal signal/noise). Open the resulting .tiff images with Genepix Pro software to obtain XY coordinates of the unlabeled samples, this helps to navigate within the NanoSIMS (see Note 6). Open fluorescently labeled array images with the Nimblescan software to extract fluorescent intensity values for each probe spot for subsequent data analysis.

84

Xavier Mayali et al.

3.4 NanoSIMS and Chip-SIP Data Analysis

1. Load the microarray into the holder and insert into NanoSIMS 50 or 50L analysis chamber. 2. Use the NanoSIMS CCD camera to find scratches for navigation (see Note 6). 3. In SIMS mode, in an area distant from the probe spots, tune the instrument with the following parameters: current at FC0 ~3–5 pA (adjust with L1), true mass resolving power (M/Δ M) > 6700 if measuring only 12C15N and >12,300 if measuring 13C14N (adjust with entrance, aperture and energy slits and tuning). Note that the NanoSIMS software overestimates mass resolving power by ~30%. 4. Find the upper left corners of the subarrays to be analyzed, using the XY coordinates from the fluorescent image as a starting point (see Note 6). Since the exact locations cannot be seen with the CCD or the secondary electron detector, they must be found using the 12C14N image in real-time imaging mode. 5. Start a chained analysis to collect the entire area of the subarray, with 50 μm rasters, 256
256 pixels, 3 planes per mass, and 1000 ms per pixel. Secondary ion beam centering and high mass resolving (HMR) automatic centering should be carried out for each location. 6. When analysis is finished, use L’IMAGE software (or similar) to automatically stitch together all the images for each subarray. Regions of interests (ROIs) of the individual probe spots are extracted using the auto-ROI “rubber-sheet” option in L’IMAGE. Isotopic ratios are converted to delta (permil) values using δ ¼ [(Rmeas/Rstandard)  1]
1000, where Rmeas is the measured ratio and Rstandard is the standard ratio (e.g., 0.011237 for 13C/12C, 0.00367 for 15N/14N). Data are then corrected for natural abundance ratios measured in unhybridized locations of the sample (control CPK6 spots). Note that regions of the array where probe spots are not synthesized exhibit greater isotope background signal and should not be used for correction. We determine if a taxon is statistically isotopically enriched based on the slope of the relationship between δ and fluorescence for each OTU. We call the slope of this relationship the ‘hybridization-corrected enrichment’ (HCE) [14]. HCE is a measure of relative enrichment based on the binding affinity of individual probes. OTUs are considered significantly isotopically enriched if the slope minus two calculated standard errors (SE) is greater than zero and if the slope is significant based on a t-score statistic (t ¼ slope/SE) with a p-value of less than 0.05. p-Values are adjusted for multiple tests using the Benjamin–Hochberg procedure for calculating false discovery rates.

Chip-SIP

4

85

Notes 1. MAS is not limited to DNA synthesis and other nucleic acids can be used instead [29, 30]; however, for Chip-SIP, we synthesize DNA probes on the array surface, and hybridize these to RNA targets from each sample. 2. The density of the spots on the Chip-SIP microarray can be increased by using DMDs with smaller mirrors or with reduction optics. 3. Two mirror image arrays are synthesized in this time. The synthesis time can be greatly reduced by using phosphoramidites with the Bz-NPPOC or SPh-NPPOC photolabile groups [31]. 4. MES stock should not be autoclaved but instead filtersterilized, stored at 4  C away from light, and discarded if yellow. 5. Trim the SecureSeal to 2
2 to fit over four subarrays. The SecureSeal chamber is adhered to the slide at room temperature for 2 h and 95  C for 5 s. 6. We design our subarrays for each microbial community in a square orientation and surround it with a frame of fiducials (probe spots that target a control oligonucleotide) that are printed with a nonprinted spot in every other location, resulting in a “checkerboard” pattern (Fig. 5). We use these fiducials to find the corners of the array when the slide is in the a

b 20000 17142 14285 11428 8571 5714 2857 0

Fig. 5 (a) Fluorescence micrograph of fluorescently labeled RNA hybridized to a subarray; (b) corresponding NanoSIMS 13C permil image of the same sample (without fluorescent labeling) hybridized to a separate microarray. The NanoSIMS image is a stitched composite of multiple 50
50 μm images. Notice that the square area of probes is surrounded by a frame of control probes, used for (1) ensuring that the hybridization is successful, (2) normalizing for background isotope enrichment, and (3) finding the array corners in both the fluorescence image in relation to the scratches and with the NanoSIMS using the 12C14N image

86

Xavier Mayali et al.

NanoSIMS. To facilitate locating these spots, fiducial scratches are made on the arrays with a diamond pen prior to fluorescent imaging. The fluorescent image shows both the scratches as well as the control oligonucleotide CPK6, which targets the outside frame of each subarray. This makes it possible to find the appropriate starting corners of the subarrays when starting the NanoSIMS analyses.

Acknowledgments The work leading to the results presented in the chapter was performed under the auspices of the US Department of Energy at the Lawrence Livermore National Laboratory under Contract DEAC52-07NA27344. Funding provided by the Genomic Sciences Program from DOE-OBER through the Biofuels Science Focus Area Grant SCW1039 and the microbial carbon cycle program grant # SCW1590. LLNL’s Laboratory Directed Research and Development Program (07-ERD-053) and the Austrian Science Fund (FWF P23797 P27275 and P30596) is acknowledged. References 1. Murrell JC, Whiteley AS (2011) Stable isotope probing and related technologies. ASM Press, Washington, DC 2. Jehmlich N, Schmidt F, Taubert M, Seifert J, Bastida F, von Bergen M et al (2010) Protein stable isotope probing (protein-SIP). Nat Protoc 5:1957–1966 3. Behrens S, Losekann T, Pett-Ridge J, Weber PK, Ng W, Stevenson BS et al (2008) Linking microbial phylogeny to metabolic activity at the single-cell level by using enhanced element labeling-catalyzed reporter deposition Fluorescence In Situ Hybridization (EL-FISH) and NanoSIMS. Appl Environ Microbiol 74:3143 4. Adamczyk J, Hesselsoe M, Iversen N, Horn M, Lehner A, Nielsen PH et al (2003) The isotope array: a new tool that employs substratemediated labeling of rRNA for determination of microbial community structure and function. Appl Environ Microbiol 69:6875–6887 5. Radajewski S, Ineson P, Parekh NR, Murrell JC (2000) Stable-isotope probing as a tool in microbial ecology. Nature 403:646–649 6. Dumont MG, Radajewski SM, Miguez CB, McDonald IR, Murrell JC (2006) Identification of a complete methane monooxygenase operon from soil by combining stable isotope probing and metagenomic analysis. Environ Microbiol 8:1240–1250

7. Kalyuzhnaya MG, Lapidus A, Ivanova N, Copeland AC, McHardy AC, Szeto E et al (2008) High-resolution metagenomics targets specific functional types in complex microbial communities. Nat Biotechnol 26:1029–1034 8. Dumont MG, Pommerenke B, Casper P (2013) Using stable isotope probing to obtain a targeted metatranscriptome of aerobic methanotrophs in lake sediment. Environ Microbiol Reports 5:757–764 9. Hungate BA, Mau RL, Schwartz E, Caporaso JG, Dijkstra P et al (2015) Quantitative microbial ecology through stable isotope probing. Appl Environ Microbiol 81:7570–7581 10. Koch BJ, McHugh TA, Hayer M, Schwartz E, Blazewicz SJ, Dijkstra P et al (2018) Estimating taxon-specific population dynamics in intact microbial communities. Ecosphere. https://doi.org/10.1002/ecs22090 11. Mau RL, Liu CM, Aziz M, Schwartz E, Dijkstra P, Marks JC et al (2014) Linking soil bacterial biodiversity and soil carbon stability. ISME J 9:1477 12. Amann RI, Krumholz L, Stahl DA (1990) Fluorescent-oligonucleotide probing of whole cells for determinative phylogenetic and environmental studies in microbiology. J Bacteriol 172:762–770

Chip-SIP 13. Brodie EL, DeSantis TZ, Joyner DC, Baek SM, Larsen JT, Andersen GL et al (2006) Application of a high-density oligonucleotide microarray approach to study bacterial population dynamics during uranium reduction and reoxidation. Appl Environ Microbiol 72:6288–6298 14. Mayali X, Weber PK, Brodie EL, Mabery S, Hoeprich P, Pett-Ridge J (2012) Highthroughput isotopic analysis of RNA microarrays to quantify microbial resource use. ISME J 6:1210–1221 15. Mayali X, Weber PK, Pett-Ridge J (2013) Taxon-specific C:N relative use efficiency for amino acids in an estuarine community. FEMS Microbiol Ecol 83:402–412 16. Mayali X, Weber PK, Mabery S, Pett-Ridge J (2014) Phylogenetic patterns in the microbial response to resource availability: amino acid incorporation in San Francisco Bay. PLoS One 9:e95842 17. Mayali X, Stewart B, Mabery S, Weber PK (2016) Temporal succession in carbon incorporation from macromolecules by particleattached bacteria in marine microcosms. Environ Microbiol Rep 8:68–75 18. Mayali X, Weber PK (2018) Quantitative substrate-specific incorporation reveals niche differentiation in a coastal microbial community. FEMS Microbiol Ecol 94:fiy047 19. Bryson S, Li Z, Chavez F, Weber PK, PettRidge J, Hettich RL, Pan C et al (2017) Phylogenetically conserved resource partitioning in the coastal microbial loop. ISME J 11:2781 20. Pett-Ridge J, Firestone MK (2017) Using stable isotopes to explore root-microbe-mineral interactions in soil. Rhizosphere. https://doi. org/10.1016/jrhisph201704016 21. Ludwig W, Strunk O, Westram R, Richter L, Meier H, Yadhukumar et al (2004) ARB: a software environment for sequence data. Nucleic Acids Res 32:1363–1371 22. Wu D, Doroud L, Eisen JA (2013) TreeOTU: operational taxonomic unit classification based on phylogenetic trees. arXiv. arXiv:13086333 23. Agbavwe C, Kim C, Hong D, Heinrich K, Wang T, Somoza MM (2011) Efficiency error

87

and yield in light-directed maskless synthesis of DNA microarrays. J Nanobiotechnol:57. https://doi.org/10.1186/1477-3155-9-57 24. Sack M, Ho¨lz K, Holik A-K, Kretschy N, Somoza V, Stengele K-P et al (2016) Express photolithographic DNA microarray synthesis with optimized chemistry and high-efficiency photolabile groups. J Nanobiotechnol 14. https://doi.org/10.1186/s12951-016-01660 25. Sack M, Kretschy N, Rohm B, Somoza V, Somoza MM (2013) Simultaneous lightdirected synthesis of mirror-image microarrays in a photochemical reaction cell with flare suppression. Anal Chem 85:8513–8517 26. Ho¨lz K, Lietard J, Somoza MM (2017) HighPower 365 nm UV LED mercury arc lamp replacement for photochemistry and chemical photolithography. ACS Sustain Chem Eng 5:828–834 27. Franssen-van Hal NLW, van der Putte P, Hellmuth K, Matysiak S, Kretschy N, Somoza MM (2013) Optimized light-directed synthesis of aptamer microarrays. Anal Chem 85:5950–5957 28. Singh-Gasson S, Green RD, Yue Y, Nelson C, Blattner F, Sussman MR et al (1999) Maskless fabrication of light-directed oligonucleotide microarrays using a digital micromirror array. Nat Biotechnol 17:974–978 29. Lietard J, Abou Assi H, Go´mez-Pinto I, Gonza´lez C, Somoza MM, Damha MJ (2017) Mapping the affinity landscape of thrombinbinding aptamers on 20 F-ANA/DNA chimeric G-quadruplex microarrays. Nucleic Acids Res 45:1619–1632 30. Lietard J, Ameur D, Damha M, Somoza MM (2018) High-density RNA microarrays synthesized in situ by photolithography. Angew Chem Int Ed. https://doi.org/10.1002/ anie201806895 31. Kretschy N, Holik A-K, Somoza V, Stengele K-P, Somoza MM (2015) Next-generation o-nitrobenzyl photolabile groups for lightdirected chemistry and microarray synthesis. Angew Chem Int Ed 54:8555–8559

Chapter 7 Quantification of Methanogenic Pathways Using Stable Carbon Isotopic Signatures Quan Yuan Abstract In many anaerobic environments methane (CH4) is produced by methanogens, with either H2/CO2 or acetate (i.e., the methyl group) as precursors, through what are referred to as hydrogenotrophic and acetoclastic methanogenic pathways respectively. Their relative contribution to total CH4 production can be quantified by determining the stable carbon isotopic fractionation factors for both pathways as well as the isotopic signatures of CO2, CH4, and the methyl group in acetate of the sample. The procedures for measuring carbon isotopic fractionation factors of both methanogenic pathways and isotopic composition of these compounds by isotope ratio mass spectrometry are described in this chapter. The results are very helpful in evaluating the activity of the methanogens involved in each methanogenic pathway as well as those of other biological pathways with different fractionation factors. Key words Stable carbon isotope, Isotopic fractionation factor, Methane, Methanogenic pathway, GC-C-IRMS

1

Introduction While some information on microbial functioning can be obtained by using stable isotope probing techniques [1, 2] or combining genomic and metaproteomic approaches [3, 4], the in situ functions of the microbial communities usually can only be analyzed by incubation and measurement of the temporal change of biomarkers including DNA, RNA, and protein. However, analysis of stable isotope signatures in soil samples might overcome this problem, since the isotopic signatures partially reflect the microbial functioning [5]. Just under 99% of all carbon on earth consists of the stable isotope 12C and approximately 1.11% of the stable isotope 13 C. The 13C isotopic signature of a particular carbon compound is given by its ratio R ¼ 13C/12C and is usually denoted relative to a standard (st) as δ13C ¼ 103 (R/Rst  1) [6]. The reactions in a

Marc G. Dumont and Marcela Herna´ndez Garcı´a (eds.), Stable Isotope Probing: Methods and Protocols, Methods in Molecular Biology, vol. 2046, https://doi.org/10.1007/978-1-4939-9721-3_7, © Springer Science+Business Media, LLC, part of Springer Nature 2019

89

90

Quan Yuan

biochemical pathway, especially those involving the cleavage of carbon bonds, often discriminate against the heavy 13C isotope (kinetic isotope effect), because the reaction rate constants are larger for substrates with 12C than 13C [7]. As a result, the δ13C of the product is always lower than that of the substrate. The fractionation factors (α) have been applied to quantify how much a given biochemical reaction (or pathway) discriminates against the substrate molecules containing the 13C. For a reaction A ! B the fractionation factor is defined as αA/B ¼ (δA + 1000)/(δB + 1000) [8], sometimes also expressed as an isotopic enrichment factor ε  103  (1  α). If two biochemical pathways display sufficiently different fractionation factors, reflected in the difference of δ13C between substrate and product, these pathways can be differentiated by stable carbon isotope signatures [5, 9]. Indeed, fractionation factors are sufficiently different for some key biochemical pathways in anaerobic biodegradation of organic substrate. Therefore, it is possible to quantify the relative contribution of hydrogenotrophic and acetoclastic methanogenic pathways to CH4 production, and of chemolithotrophic (acetyl-CoA synthase) and heterotrophic (fermentation) pathways to acetate formation. In addition, stable carbon isotope analysis may allow partitioning the contribution of different organic substrates to end products of degradation, for example, the relative contribution of root exudation versus soil organic matter to CH4 production in rice field soil [10, 11], provided the different substrates have substantial difference in δ13C values (e.g., a mixture of C3 and C4 plants), and the carbon conversion pathways have negligible fractionation factors or the fractionation factors could be solved [10]. Here we present the methods of using stable carbon isotopic signatures for elucidating the microbial functional pathways of methane production.

2

Materials 1. Soil or sediment samples. 2. 26-mL borosilicate glass pressure tubes with crimp top. 3. Butyl rubber stoppers, aluminum crimps, and a crimping tool. 4. N2 gas. 5. CH3F (methyl fluoride) (see Note 1). 6. 1 mmol/L H2SO4. 7. 0.42 mol/L sodium peroxodisulfate. 8. 1.35 mol/L phosphoric acid. 9. NaOH. 10. Gas-tight pressure lock syringe.

Isotope Fractionation

91

11. 0.2-μm polytetrafluoroethylene (PTFE) membrane filters. 12. Gas chromatograph (GC) with flame ionization detector and methanizer (Ni catalyst at 350  C). 13. High-performance liquid chromatograph (HPLC) with ion exclusion column, refractive index and UV detectors. 14. Isotope ratio mass spectrometer (IRMS). 15. Finnigan Standard GC Combustion Interface III, Finnigan LC IsoLink. 16. Pyrolytic oven.

3

Methods

3.1 Incubation Experiments

Set up all batches of anoxic rice field soil in multiple replicates, of which triplicates will be opened at different time points during the incubation, and analyzed as described below. 1. Prepare anoxic microcosms by adding 5 g soil + 5 mL of deionized water into 26-mL pressure tubes. Close the tubes with butyl rubber stoppers, and exchange the gas phase with N2. 2. Add CH3F to the headspace of half the treatments to a final concentration of 2%. Leave the remaining tubes without added CH3F. 3. At regular intervals take gas samples from the headspace of the tubes and analyze for CH4 and CO2 as well as δ13C value of CH4 and CO2, as described below.

3.2 Analyses of Gas and Liquid Samples

1. After vigorously shaking the bottles by hand, take gas samples (200 μL) with a gas-tight pressure lock syringe, and analyze immediately using gas chromatography (GC). CH4, CH3F, and CO2 are analyzed using GC with a flame-ionization detector. CO2 is detected after conversion to CH4 with a methanizer. 2. Take liquid samples with a sterile syringe, membrane-filtered (0.2 μm) and store frozen (20  C) until analysis. Acetate is measured using high-performance liquid chromatography (HPLC) with refractive index and UV detectors.

3.3 Stable Carbon Isotope Analysis of CH4 and CO2 with Gas Chromatograph Combustion Isotope Ratio Mass Spectrometry (GC-C-IRMS)

1. The CH4 and CO2 in the gas samples are first separated by GC; after conversion of CH4 to CO2 in the Finnigan Standard GC Combustion Interface III, the 13C/12C is determined by the IRMS instrument.

92

Quan Yuan

3.4 Carbon Isotopic Measurements of Acetate Using an HPLC-LC-IRMS System

1. Load the acetate on an ion exclusion column with 1 mmol/L of H2SO4 at 0.3 mL/min as eluent, and coupled to a Finnigan LC IsoLink for oxidation of the separated compounds to CO2 at 99.9  C with 0.42 mol/L sodium peroxodisulfate and 1.35 mol/L phosphoric acid [12]. 2. Detect the isotope ratios on an IRMS; the analysis results in determination of δ13C of total acetate.

3.5 Measuring δ13C of the Methyl Group of Acetate (δac-methyl) by Off-Line Pyrolysis

1. Purify the acetate in the liquid sample with HPLC by collecting the acetate fraction from each run. 2. Add the purified sample to a strong NaOH solution (final molar ratio of acetate to NaOH of 1:200), and dry in a Pyrex tube under vacuum. 3. Pyrolyze the dried reactants under vacuum at 400  C, so as to convert the carboxyl carbon to CO2 and the methyl carbon to CH4 [13]. 4. Take the gas samples and analyze the δ13C of the produced CH4 by GC-C-IRMS (see Subheading 3.3). This is identical to the δ13C of the methyl carbon.

3.6

Calculations

3.6.1 Determination of the δCH4

Calculate the isotopic signature for newly formed CH4 (δn) from the isotopic signatures at two time points t ¼ 1(δ1) and t ¼ 2(δ2) with the following mass balance equation:
δ2 ¼ f n δn þ 1  f n δ1

ð1Þ

with fn the fraction of the newly formed C compound relative to the total at t ¼ 2. 3.6.2 Calculation of Fractionation Factor for Conversion of CO2 to CH4 (αmc)

The apparent fractionation factor for conversion of CO2 to CH4 is given by αapp ¼ ðδCO2 þ 1000Þ=ðδCH4 þ 1000Þ

ð2Þ

the term “apparent“ is used, since the isotope signature of CH4 might be determined by acetoclastic plus hydrogenotrophic methanogenesis, while for the calculation only the isotope signature of the methanogenic substrate CO2 is used. While in the presence of CH3F, the αapp will be taken as αmc since the acetoclastic methanogenesis is inhibited (see Note 2). 3.6.3 Calculation of Fractionation Factor for Conversion of Acetate to CH4 (αma)

The isotopic effect εac-methyl/CH4 associated with acetoclastic methanogenesis is calculated according to the Mariotti equation [14]: δr ¼ δri þ ε½ln ð1  f Þ

ð3Þ

Isotope Fractionation

93

where δri is the isotopic composition of the reactant (ac-methyl) at the beginning, which in our case is the maximum accumulation of acetate in the incubation; δr is the isotope composition of the residual reactant, when f was determined; and f is the fractional yield of the products based on the consumption of acetate (0 < f < 1). Linear regression of δr against ln(1  f ) gives ε as the slope. The enrichment factor could be converted to the fractionation factor according to ε  103  (1  α). When the acetate concentration approaches threshold values (200 μM), no fractionation occurs during conversion of fermentatively produced acetate to CH4 (αma ¼ 0). In that case, δ13C of CH4 derived from acetate equals to δac-fermentation, which is the δ13C value of acetate methyl produced by fermentation and is equal to δac-methyl in the presence of CH3F. 3.6.4 Determination of Relative Contribution of Hydrogenotrophic and Acetoclastic Methanogenic Pathways

Determine the relative contribution of H2/CO2-derived CH4 to total CH4 with the following mass balance equation [5]:
δCH4 ¼ f H2 δmc þ 1  f H2 δma

ð4Þ

f H2 ¼ ðδCH4  δma Þ=ðδmc  δma Þ

ð5Þ

solved for f H2

where f H2 is the fraction of CH4 formed from H2/CO2, δCH4 the δ13C of total produced CH4, and δma and δmc are the isotope ratios of CH4 derived from acetate and H2/CO2, respectively. The relative contribution of acetoclastic methanogenic pathway equals to 1  f H2 (see Note 3).

4

Notes 1. CH3F is a specific inhibitor of acetoclastic methanogens, which does not affect hydrogenotrophic methanogens [15]. 2. Fractionation factors have to be determined under well-defined conditions, which are usually only met by assaying defined microbial cultures or biochemical reactions in which the desired pathway operates. 3. Use of carbon isotopic signatures in CH4 emitted from a production site (e.g., a wetland) requires even more complex models, since isotopic discrimination in addition occurs during transport and oxidation of the produced CH4.

94

Quan Yuan

Acknowledgments This work was funded by the National Natural Science Foundation of China (41573083) and State Key Laboratory of Environmental Geochemistry (SKLEG2015403), China. References 1. Blagodatskaya E, Kuzyakov Y (2013) Active microorganisms in soil: critical review of estimation criteria and approaches. Soil Biol Biochem 67:192–211 2. Uhlik O, Leewis MC, Strejcek M, Musilova L, Mackova M, Leigh MB, Macek T (2013) Stable isotope probing in the metagenomics era: a bridge towards improved bioremediation. Biotechnol Adv 31:154–165 3. Ram RJ, VerBerkmoes NC, Thelen MP, Tyson GW, Baker BJ, Blake RC, Shah M, Hettich RL, Banfield JF (2005) Community proteomics of a natural microbial biofilm. Science 308:1915–1920 4. Simon C, Daniel R (2011) Metagenomic analyses: past and future trends. Appl Environ Microb 77:1153–1161 5. Conrad R (2005) Quantification of methanogenic pathways using stable carbon isotopic signatures: a review and a proposal. Org Geochem 36:739–752 6. Conrad R, Claus P, Casper P (2009) Characterization of stable isotope fractionation during methane production in the sediment of a eutrophic lake, Lake Dagow, Germany. Limnol Oceanogr 54:457–471 7. Blaser M, Conrad R (2016) Stable carbon isotope fractionation as tracer of carbon cycling in anoxic soil ecosystems. Curr Opin Biotech 41:122–129 8. Hayes JM (1993) Factors controlling C-13 contents of sedimentary organic-compounds—principles and evidence. Mar Geol 113:111–125

9. Penning H, Conrad R (2007) Quantification of carbon flow from stable isotope fractionation in rice field soils with different organic matter content. Org Geochem 38:2058–2069 10. Yuan Q, Pump J, Conrad R (2012) Partitioning of CH4 and CO2 production originating from rice straw, soil and root organic carbon in rice microcosms. PLoS One 7:e49073 11. Yuan Q, Pump J, Conrad R (2014) Straw application in paddy soil enhances methane production also from other carbon sources. Biogeosciences 11:237–246 12. Krummen M, Hilkert AW, Juchelka D, Duhr A, Schluter HJ, Pesch R (2004) A new concept for isotope ratio monitoring liquid chromatography/mass spectrometry. Rapid Commun Mass Spectrom 18:2260–2266 13. Penning H, Tyler SC, Conrad R (2006) Determination of isotope fractionation factors and quantification of carbon flow by stable carbon isotope signatures in a methanogenic rice root model system. Geobiology 4:109–121 14. Mariotti A, Germon JC, Hubert P, Kaiser P, Letolle R, Tardieux A, Tardieux P (1981) Experimental-determination of nitrogen kinetic isotope fractionation—some principles—illustration for the denitrification and nitrification processes. Plant Soil 62:413–430 15. Janssen PH, Frenzel P (1997) Inhibition of methanogenesis by methyl fluoride: studies of pure and defined mixed cultures of anaerobic bacteria and archaea. Appl Environ Microbiol 63:4552–4557

Chapter 8 Stable Isotope-Labeled Single-Cell Raman Spectroscopy Revealing Function and Activity of Environmental Microbes Li Cui, Kai Yang, and Yong-Guan Zhu Abstract Microorganisms play a key role in driving the global element (C, N, H, P, and S) cycling. However, the function and activity of environmental microbes remain largely elusive because the vast majority of them are yet uncultured. Recent achievements in single cell stable isotope-labeled Raman spectroscopy enable direct investigation of function and activity of individual microbes in complex environmental communities. Here, this protocol describes a workflow to investigate environmental microbes in soil and water by combining 15 N, 2D, and 13C stable isotope labeling with different single-cell Raman techniques, including normal Raman, resonance Raman (RR), and surface-enhanced Raman spectroscopy (SERS). Their applications in investigating functional bacteria driving the N and C cycles, and metabolically active cells are described. Key words Single-cell Raman spectroscopy, Surface-enhanced Raman spectroscopy (SERS), 15N, C, 2D stable isotope probing, Function, Activity, N2-fixing bacteria, Metabolically active bacteria, Microbial communities 13

1

Introduction Microorganisms play a key role in driving the biogeochemical cycling of elements on our planet. However, the vast majority of them have not yet been cultivated in the laboratory, hindering the deciphering of “talented” microbial resources [1]. Single-cell Raman spectroscopy combined with stable isotope probing (Raman-SIP) provides an efficient and promising way to uncover the largely unexplored microbes in important ecosystems such as soil, sea, and river water [2–12]. Raman spectroscopy is a nondestructive method capable of providing intrinsic molecular profiles of individual microorganism based on vibrational frequencies of intracellular biomolecules, such as proteins, nucleic acids, lipids, and carbohydrates [2, 3, 13, 14]. More importantly, when combined with SIP such as 15N, 13C, and 2D, assimilation of isotopelabeled substrates by microbes can generate characteristic Raman

Marc G. Dumont and Marcela Herna´ndez Garcı´a (eds.), Stable Isotope Probing: Methods and Protocols, Methods in Molecular Biology, vol. 2046, https://doi.org/10.1007/978-1-4939-9721-3_8, © Springer Science+Business Media, LLC, part of Springer Nature 2019

95

96

Li Cui et al.

Table 1 Application of Raman microscopy to study microbial populations in environmental samples

Raman techniques

Unlabeled cells SIP (cm
1)

Fully labeled cells (cm
1)

Assignment

Applications and references

Single-cell resonance Raman

15

N 1129

1114

Nitrogen fixation [9] Cytochrome c, Stretching mode of C–N

Surfaceenhanced Raman

15

N 730–735

717

Ring breathing mode of adeninecontaining molecules

Single-cell Raman

13

C 1004

967

Symmetric ring Naphthalene degradation [4, 5], breathing mode phenol degradation [26], of phenylalanine phenylalanine uptake [27]

Single-cell Raman

2

D

Nitrogen assimilation [8]

2800–3100 2040–2300 Stretching mode of Metabolically active bacteria C–H in lipid and [15, 31], recovery of metabolic protein activity of viable but nonculturable cells [12], antimicrobial effect [11, 28, 30], photosynthetic activity [29]

shifts due to replacement of light atoms with heavier stable isotopes in chemical bonds of de novo synthesized biomolecules, resulting in decreases of vibrational energy [3]. The shifted marker bands can be used not only to discern functional or active bacteria but also to quantify their activity in situ, as it has been demonstrated in naphthalene-degrading bacteria [4, 5], CO2-fixing photosynthetic microorganisms [6, 7], N2-fixing bacteria [8, 9], physiologically active bacteria [12, 15], and metabolically active bacteria [11], among others (Table 1). In addition, single-cell level detection provides an important means to bypass the necessity of pure culture and directly investigate individual bacteria including largely unexplored uncultured cells in their natural habitat [4, 5, 9, 16–18]. There are several types of Raman spectroscopic techniques. In addition to normal (spontaneous) Raman that is relatively weak, for some bacteria with specific pigments such as cytochrome c (cyt c) and carotenoids, resonance Raman (RR) signals can be excited when the incident laser frequency matches the electronic transition frequency of molecules, leading to around 3 orders of magnitude of enhanced Raman signals [6, 9]. Surface-enhanced Raman spectroscopy (SERS) is another way to provide as high as 6–14 orders of magnitude of enhanced signal for molecules adsorbed on the surface of metallic nanostructures (e.g., silver or gold nanoparticles)

Raman-SIP Deciphers Function and Activity of Microbes

97

[8, 13, 14, 19–23]. Moreover, SERS can selectively enhance N-related molecules of bacteria, enabling it a sensitive way to study N-cycle related bacteria [3, 8]. In this protocol, we will describe several single-cell Raman-SIP techniques combining 15N, 2D, 13C SIP with different Raman techniques. Their applications in studying functional and active bacteria in environmental communities and detailed procedures to achieve them will be presented, including single-cell resonance Raman (RR) combined with 15N2 to study N2-fixing bacteria in soil; SERS combined with 15NH4Cl, Na15NO3, and 15N2 to study assimilation of different N sources; single-cell Raman combined with D2O to study metabolically active bacteria; and single-cell Raman combined with 13C to study carbon sourcedegrading bacteria.

2 2.1

Materials Raman Setup

1. Confocal micro-Raman system (LabRAM Aramis, HORIBA Jobin-Yvon Ltd.) equipped with a multichannel air-cooled CCD utilizing a 1024  256 pixels with thermoelectric cooling down to
70  C for negligible dark current [8, 9]. 2. 100 dry objective (numerical aperture (NA) ¼ 0.9, Olympus) and a 532 nm Nd:YAG laser for sample observation, Raman signal excitation, and acquisition using single-cell normal and resonance Raman measurements [9] (see Note 1). 3. 50 objective (NA ¼ 0.55, Olympus) with a working distance of 8 mm and a 633-nm He-Ne laser for SERS measurements. Proper laser power attenuation should be applied (see Note 1). 4. 300 g/mm grating (for acquiring spectra ranging from 500 to 3200 cm
1 with a spectral resolution of around 4 cm
1). 5. 600 g/mm grating (for acquiring spectra ranging from 400 to 1800 cm
1 with a spectral resolution of around 1 cm
1). 6. Computer-controlled XY motorized stages (X ¼ 75 mm, Y ¼ 50 mm) with a step size of 1.5 μm (minimal step size ¼ 0.1 μm) for Raman mapping, automated stage movement. The stage is controlled by a positioning joystick, computer interface card, drive electronics, and software Labspec 5 (Horiba Jobin-Yvon Ltd.).

2.2 Stable IsotopeLabeled Substrates

1.

15

N2 (99 atom % 15N, purity >98.5%).

2.

15

NH4Cl (98 atom % 15N).

3. Na15NO3 (99 atom % 15N, purity >98.5%). 4. D2O (99.8 atom % D). 5.

13

C-Glucose (99 atom % 13C, purity 99%).

98

Li Cui et al.

2.3 Bacterial Strains and Culture Media

1. Escherichia coli. 2. Nitrogen-fixing strains of Azotobacter sp. (AS1.222) and Azotobacter chroococcum (ACCC10096). 3. Luria–Bertani (LB) broth: 10 g/L tryptone, 5 g/L yeast extract, and 5 g/L NaCl. 4. Minimal medium (MM): 2.5 g/L Na2HPO4, 2.5 g/L KH2PO4, 1.0 g/L NH4Cl, 0.1 g/L MgSO4·7H2O, 10 μL/L saturated CaCl2 solution, 10 μL/L saturated FeSO4 solution per liter. After autoclaving, 1 mL of filter-sterilized Bauchop and Elsden solution is then supplemented [24]. 5. Bauchop and Elsden solution: 10.75 g/L MgSO4, 4.5 g/L FeSO4·7H2O, 2.0 g/L CaCO3, 1.44 g/L ZnSO4·7H2O, 1.12 g/L MnSO4·4H2O, 0.25 g/L CuSO4·5H2O, 0.28 g/L CoSO4·7H2O, 0.06 g/L H3BO3, 51.3 mL/L concentrated HCl. 6. 10 mM glucose. 7. 10 mM sodium succinate. 8. 10 mM sodium citrate. 9. Nitrogen-free Azotobacter medium: 20 g/L mannitol, 0.2 g/L KH2PO4, 0.8 g/L K2HPO4, 0.2 g/L MgSO4·7H2O, 0.1 g/L CaSO4·2H2O, 0.00025 g/L Na2MoO4·2H2O, and 0.0015 g/L FeCl3. 10. Oxygen (O2) gas. 11. 12-mL crimp-top vials, butyl rubber stoppers, aluminum caps, and a crimping device. 12. Aluminum (Al) foil.

2.4 Extraction of Bacteria from Soil and Water Communities

1. 2-mm mesh sieve. 2. Phosphate-buffered saline (PBS): 8 g/L NaCl, 0.2 g/L KCl, 1.44 g/L Na2HPO4, 0.24 g/L KH2PO4/L, autoclaved. 3. 0.5% (v/v) Tween 20.

2.4.1 Fresh Soil Samples

4. 1.42 g/mL Nycodenz solution: dissolve 8 g of Nycodenz ( 98%) in 10 mL sterile water.

2.4.2 Water Samples

1. Membrane filters (10, 3, and 0.22 μm).

2.5 Synthesis of Silver and Gold Nanoparticles

1. 1% HAuCl4·4H2O. 2. Ultrapure water (18.2 MΩ cm). 3. 1% (wt/vol) trisodium citrate. 4. AgNO3.

Raman-SIP Deciphers Function and Activity of Microbes

3

99

Methods Figure 1 shows a schematic illustrating the workflow of Raman-SIP approach in investigating functional and active microbes in environmental samples, including incubation of isotope-labeled substrate with pure bacterial culture and environmental samples (see Note 2), extraction of bacteria from soil and water, Raman/SERS and Raman mapping spectral acquisition, and spectral output. 1. Add 6 mL of N-free Azotobacter medium into a 12-mL crimptop vial. 2. Seal the vials, then replace the air in the headspace with a gas mixture consisting of 15N2 and O2 (volume ratio of N2 to O2 is 4:1) of different volumes to achieve 15N2 of different percentages. Because 15N content is 99% in commercial 15N2 and 0.36% in natural abundance, the final 15N abundance relative to

Fig. 1 Schematic illustrating workflow of single cell Raman-SIP approach, including incubation of water or soil sample with SIP substrates, bacterial extraction, Raman/SERS measurement (including Raman mapping), and spectral output. Abbreviations: SCRS single cell Raman spectroscopy, SCRRS single cell resonance Raman spectroscopy, SERS surface-enhanced Raman spectroscopy.

100

Li Cui et al.

3.1 Single-Cell Resonance Raman (RR) Combined with 15N2 to Study N2-Fixing Diazotrophs in Soil 3.1.1 Incubation of N2Fixing Bacteria with 15N2 as the Sole N Source

N2 in air in the headspace is 99.36%, 49.68%, 25.02%, 10.22%, and 0.36%, respectively. 3. Inoculate 10 μL of stationary-phase Azotobacter sp. (AS1.222) or A. chroococcum (ACCC10096) into the vials by injection and culture at 28  C and 180 rpm for 48 h. 4. Harvest and wash bacteria by centrifugation at 2,500  g for 3 min. 5. Adjust bacterial suspension to a proper concentration (ca. 106 cells/mL) by diluting with ultrapure water. 6. Spot 1–3 μL of bacteria onto an aluminum (Al) foil substrate [25] and air dry for further Raman measurement.

3.1.2 Incubation of Soil Microcosm with 15N2 and Extraction of Bacteria from Soil

1. Sieve soil through a 2-mm sieve. 2. Transfer 2 g of sieved soil to a 12-mL crimp-top vial. 3. Seal the vials, purge with O2 for 10 min to remove air and then replace O2 with an appropriate volume (ca. 8 mL) of 15N2 to achieve 80% 15N2 and 20% oxygen in the headspace (see Note 3). 4. Inject 500 μL of 0.5 M glucose solution into the soil microcosms and incubate at room temperature (ca. 25  C) under low light conditions for 12 days. 5. After incubation, homogenize the soil in 10 mL of PBS supplemented with 0.5% (v/v) Tween 20. 6. Detach soil particle-associated cells by vigorously vortexing for 30 min at room temperature. 7. Carefully introduce 10 mL of the soil slurries into an Eppendorf tube containing 10 mL of Nycodenz stock solution (see Note 4). 8. Centrifuge the tubes at 14,000  g for 30 min at 4  C. 9. Transfer the top Nycodenz phase containing the bacteria, that is, the cell layer, (Fig. 2) to a new centrifugation tube containing 5 mL of PBS and vortex for 30 s. 10. Centrifuge the tubes at 2,500  g for 3 min and wash with ultrapure water twice. 11. Adjust the bacterial suspension to ca. 106 cells/mL by diluting with ultrapure water. 12. Spot 1–3 μL of bacteria onto an Al foil substrate and air dry for further Raman measurement.

Raman-SIP Deciphers Function and Activity of Microbes

101

Fig. 2 Procedure for extracting bacteria from soil by Nycodenz density gradient separation method. (a) Soil slurry is carefully transferred on top of the Nycodenz solution which is denser than water. (b) After centrifugation, the heaviest soil particles in the water move to the bottom of the tube, while bacteria move to the interface layer between water supernatant and Nycodenz phases and can thus be extracted by pipetting

3.2 SERS Combined with 15NH4Cl, Na15NO3 and 15N2 to Study Assimilation of Different N Sources

1. Dilute 1.21 mL 1% HAuCl4·4H2O aqueous solution into 100 mL ultrapure water (18.2 MΩ cm) in a round-bottomed flask.

3.2.1 Synthesis of Silver and Gold Nanoparticles

3. Immediately pipette 0.6 mL of 1% (wt/vol) trisodium citrate solution into the reaction flask and keep boiling for about 1 h.

2. Heat the as-prepared HAuCl4 solution to boil under vigorous stirring with a magnetic stir bar.

4. Stop heating and cool down to room temperature. Synthesis of Au Nanoparticles (Around 120 nm)

Synthesis of Ag Nanoparticles (Average Size of 80 nm)

5. Wash the as-synthesized Au NPs once by ultrapure water through centrifugation at 800  g for 5 min. 6. Discard the supernatant and obtain concentrated Au NPs of around 5000 mg/L. 1. Dissolve 72 mg of AgNO3 in 400-mL ultrapure water (18.2 MΩ cm) in a round-bottomed flask. 2. Heat the as-prepared AgNO3 solution to boil under vigorous stirring with a magnetic stir bar. 3. Immediately pipette 8 mL of 1% (wt/vol) trisodium citrate solution into the reaction flask and keep boiling for around 1 h. 4. Stop heating and cool down to room temperature. 5. Wash the as-synthesized Ag NPs once by ultrapure water through centrifugation at 3,500  g for 5 min. 6. Discard the supernatant and dilute the Ag NPs to around 10,000 mg/L.

102

Li Cui et al.

3.2.2 Incubation of Pure Bacteria with Different 15NLabeled Substrates as Sole N Source

1. Supplement MM medium with 15NH4Cl (0.1 g/L), Na15NO3 (0.1 g/L) and 15N2 (98.5% atom) as the sole N source respectively. Supplement 10 mM glucose, sodium succinate, and sodium citrate as the carbon source. To obtain various levels of 15N-labeled cells (take NH4Cl as an instance), mix 15N- and 14 N-NH4Cl-containing media at different volume ratios of 100:0, 75:25, 50:50, 25:75, 10:90, 5:95, and 0:100 to get final 15N content in the growth media of 98.0%, 73.6%, 49.2%, 24.8%, 10.1%, and 0.4% respectively (see Note 5). Inoculate bacteria into the as-prepared medium and incubate at 37 or 28  C, 140 rpm. 2. Harvest bacteria in the stationary growth phase and wash them twice with ultrapure water via centrifugation at 2,500  g for 3 min. 3. Discard the supernatant, mix cell pellets with 10 μL of concentrated Ag and/or Au thoroughly (see Note 6). 4. Spot 1–3 μL of mixture onto a glass slide and air dry prior to SERS measurement (see Note 7).

3.2.3 Incubation of Wetland Water with Different 15N-Labeled Substrates

1. Collect surface wetland water and filter water through a 10-μm and 3-μm microfiltration membrane to remove large particles. 2. Amend the obtained water with 10-mM glucose, sodium succinate, and sodium citrate as carbon source and 15NH4Cl, Na15NO3 and 15N2 as the nitrogen source, respectively. 3. Incubate the sample at room temperature for 24 h at 28  C and 180 rpm. 4. Wash bacteria in water samples twice with ultrapure water and harvest them by centrifugation at 2,500  g for 3 min. 5. Mix cell pellets with 10 μL of concentrated Ag and/or Au nanoparticles thoroughly. 6. Spot 1–3 μL of mixture onto a glass slide and air dry prior to SERS measurement.

3.3 Single-Cell Raman Combined with D2O to Study Metabolically Active Bacteria 3.3.1 Incubation of Pure Bacteria with D2OAmended Culture Medium 3.3.2 Incubation of River Water with D2O

1. Incubate E. coli in LB media amended with 50% heavy water (vol/vol) (see Notes 2 and 8) at 37  C, 140 rpm for 24 h. 2. Harvest and wash bacteria twice by centrifugation at 2,500  g for 3 min. 3. Spot 1–3 μL of bacteria onto an Al foil substrate and air-dry for further Raman measurement.

1. Collect river surface water and filter through a 10-μm and 3-μm microfiltration membrane to remove large particles.

Raman-SIP Deciphers Function and Activity of Microbes

103

2. Mix 2.4 mL of filtered river water with 1.6 mL of D2O to get final D2O concentration of 40% v/v. One vial containing 2.4 mL of river water and 1.6 mL of sterilized ultrapure H2O is set as the negative control without D labeling. 3. Incubate the vials at room temperature and 500 rpm for 24 h. 4. Harvest the bacteria and wash twice with ultrapure water by centrifugation at 2,500  g for 3 min. 5. Spot 1–3 μL of bacteria onto an Al foil substrate and air-dry for further Raman measurement. 3.4 Single-Cell Raman Combined with 13C to Study Carbon SourceDegrading Bacteria 3.4.1 Incubation of Bacteria with 13CGlucose

13

C-glucose is used here to describe the procedure, which can also be applied to other 13C-labeled carbon sources like naphthalene.

1. Inoculate bacteria into 5 mL of MM media amended with filter-sterilized 10-mM 13C-glucose as the sole carbon source. 2. Harvest bacteria after overnight incubation at 37  C, 180 rpm, and then wash bacteria with ultrapure water by centrifugation at 2,500  g for 3 min. 3. Adjust bacterial suspension to 106 cells/mL by diluting with ultrapure water. 4. Spot 1–3 μL of bacteria onto an aluminum (Al) foil substrate and air-dry for further Raman measurement.

3.5 Raman Spectra Acquisition and Data Analysis

1. To acquire single-cell Raman spectroscopy, a 532 nm Nd:YAG laser with a 300 g/mm grating is employed to acquire spectra in the range of 500–3200 cm
1 with an acquisition time of 10 s. 2. To acquire single-cell resonance Raman spectroscopy, a 532 nm Nd:YAG laser with a 600 g/mm grating is employed to acquire spectra in the range of 500–2000 cm
1 with an acquisition time of 2 s. 3. For SERS measurement, a 633 nm He-Ne laser is employed with a 600 g/mm grating. The spectra are obtained in the range of 500–2000 cm
1 with an acquisition time of 4 s. 4. For Raman mapping, the step size is set at 1.5 μm in a rectangular mapping area with acquisition time of 2 s on each point. 5. Raman spectra are processed by cosmic ray removal, baseline correction and normalization using the Labspec 5 software (HORIBA Jobin Yvon Ltd) before further analysis. Principal component analysis (PCA) and other required spectral preprocessing are performed using the IRootLab toolbox (https:// code.google.com/p/irootlab/) running on MATLAB 2012a [9, 13].

104

Li Cui et al.

Fig. 3 Single-cell Raman/SERS spectra of unlabeled (green) and fully isotopelabeled bacteria (red). (a) Raman spectra of Azotobacter sp. labeled by 15N2. (b) SERS spectra of E. coli labeled by 15NH4Cl. (c, d) Raman spectra of E. coli labeled with D2O and 13C-glucose

6. Typical single-cell Raman/SERS spectra of 15N, 2D and 13C isotope-labeled cell are shown in Fig. 3. The bands with characteristic Raman shift induced by stable isotope labeling are listed in Table 1 (see Note 9). Based on these characteristic shifts, such as bands at 1114 cm
1 (C-15N), 717 cm
1 (15Nadenine), 2040–2300 cm
1 (C-D), 967 cm
1 (13C-phenylalanine), functional and active bacteria in microbial communities can be distinguished. Based on the intensity ratios of labeled and unlabeled bands or shifting degree, such as I1114/ (I1114 + I1129), 15N-induced shift/730 cm
1, I(2040–2300)/ (I(2040–2300) + I(2800–3100)), I967/(I967 + I1000), activity of bacteria in fixing N2, assimilating N and C sources, and resistance of bacteria against antibiotics can be revealed.

4

Notes 1. To minimize the possible laser damage to the sample, Duoscan, which can provide a large sampling area of 30  30 μm2 and thus low laser density, is utilized. Duoscan is achieved via the combination of two mirrors scanning the laser beam rapidly across the chosen area.

Raman-SIP Deciphers Function and Activity of Microbes

105

2. Detection limit of Raman-SIP is around 10% stable isotope abundance in bacteria. Differentiation of labeled cells from those unlabeled based on isotope-induced Raman shift can be achieved only for bacteria with >10% isotope abundance. Ideally use of 90 atom % 15N and 13C of stable isotope substrate for labeling is recommended. For heavy water, because of its growth inhibition on bacteria, 50% D2O (vol/vol) is recommended for labeling. 3. To eliminate 14N2 from the culture vessel, a degasification step is performed before the incubation. In order to remove 14N2 in N-free Azotobacter medium and soil microcosms, the headspaces are purged by O2 before incubation. 4. To ensure a clear phase separation and facilitate extraction of interface layer containing extracted soil bacteria, enough volume of Nycodenz solution is added to keep interface layer above soil particle pellet. It is also recommended to cut off the tip of the pipette tip to avoid clogging by the soil slurry. 5. Considering that 15N content is 98% in commercial 15NNH4Cl and 0.4% in its natural abundance, the final 15N content in the growth media was 98.0%, 73.6%, 49.2%, 24.8%, 10.1%, and 0.4%, respectively. 6. All of the concentrated NPs are freshly prepared before each measurement. 7. To prepare single-cell SERS sample, 10 μL of twofold diluted overnight cultured Escherichia coli is washed twice by ultrapure water and then mixed with 10 μL of concentrated Ag NPs. Spot 1–3 μL of mixture onto an Al foil substrate and air-dry prior to SERS measurement. 8. Because deuterium in the D2O may exchange with naturally abundant hydrogen during autoclaving, filter sterilization of D2O via a 0.22-μm membrane is strongly recommended. 9. Isotope-induced Raman band shifts depend on the substrate type and metabolic pathway of cells and thus only occur in the associated molecules within the cell.

Acknowledgments This work was supported by Natural Science Foundation of China (21777154 and 91851101), National Key Research and Development Program of China (2017YFD0200201), Strategic Priority Research Program of Chinese Academy of Sciences (XDB15020302, XDB15020402), and the K.C. Wong Education Foundation.

106

Li Cui et al.

References 1. Amann RI, Ludwig W, Schleifer KH (1995) Phylogenetic identification and in situ detection of individual microbial cells without cultivation. Microbiol Rev 59(1):143–169 2. Huang WE, Griffiths RI, Thompson IP, Bailey MJ, Whiteley AS (2004) Raman microscopic analysis of single microbial cells. Anal Chem 76(15):4452–4458. https://doi.org/10. 1021/ac049753k 3. Wang Y, Huang WE, Cui L, Wagner M (2016) Single cell stable isotope probing in microbiology using Raman microspectroscopy. Curr Opin Biotech 41:34–42. https://doi.org/10. 1016/j.copbio.2016.04.018 4. Huang WE, Stoecker K, Griffiths R, Newbold L, Daims H, Whiteley AS, Wagner M (2007) Raman-FISH: combining stableisotope Raman spectroscopy and fluorescence in situ hybridization for the single cell analysis of identity and function. Environ Microbiol 9 (8):1878–1889. https://doi.org/10.1111/j. 1462-2920.2007.01352.x 5. Huang WE, Ferguson A, Singer AC, Lawson K, Thompson IP, Kalin RM, Larkin MJ, Bailey MJ, Whiteley AS (2009) Resolving genetic functions within microbial populations: in situ analyses using rRNA and mRNA stable isotope probing coupled with single-cell Raman-fluorescence in situ hybridization. Appl Environ Microbiol 75(1):234–241. https://doi.org/10.1128/Aem.01861-08 6. Li MQ, Canniffe DP, Jackson PJ, Davison PA, FitzGerald S, Dickman MJ, Burgess JG, Hunter CN, Huang WE (2012) Rapid resonance Raman microspectroscopy to probe carbon dioxide fixation by single cells in microbial communities. ISME J 6(4):875–885. https:// doi.org/10.1038/ismej.2011.150 7. Jing X, Gou H, Gong Y, Su X, Xu L, Ji Y, Song Y, Thompson IP, Xu J, Huang WE (2018) Raman-activated cell sorting and metagenomic sequencing revealing carbon-fixing bacteria in the ocean. Environ Microbiol 20 (6):2241–2255. https://doi.org/10.1111/ 1462-2920.14268 8. Cui L, Yang K, Zhou GW, Huang WE, Zhu YG (2017) Surface-enhanced Raman spectroscopy combined with stable isotope probing to monitor nitrogen assimilation at both bulk and single-cell level. Anal Chem 89 (11):5794–5801. https://doi.org/10.1021/ acs.analchem.6b04913 9. Cui L, Yang K, Li HZ, Zhang H, Su JQ, Paraskevaidi M, Martin FL, Ren B, Zhu YG (2018) Functional single-cell approach to probing nitrogen-fixing bacteria in soil

communities by resonance Raman spectroscopy with 15N2 labeling. Anal Chem 90 (8):5082–5089. https://doi.org/10.1021/ acs.analchem.7b05080 10. Eichorst SA, Strasser F, Woyke T, Schintlmeister A, Wagner M, Woebken D (2015) Advancements in the application of NanoSIMS and Raman microspectroscopy to investigate the activity of microbial cells in soils. FEMS Microbiol Ecol 91(10). https://doi. org/10.1093/femsec/fiv106 11. Song YZ, Cui L, Lopez JAS, Xu JB, Zhu YG, Thompson IP, Huang WE (2017) Ramandeuterium isotope probing for in-situ identification of antimicrobial resistant bacteria in Thames River. Sci Rep 7. https://doi.org/10. 1038/S41598-017-16898-X 12. Zhang S, Guo L, Yang K, Zhang Y, Ye C, Chen S, Yu X, Huang WE, Cui L (2018) Induction of Escherichia coli into a VBNC state by continuous-flow UVC and subsequent changes in metabolic activity at the single-cell level. Front Microbiol 9(2243). https://doi. org/10.3389/fmicb.2018.02243 13. Cui L, Zhang YJ, Huang WE, Zhang BF, Martin FL, Li JY, Zhang KS, Zhu YG (2016) Surface-enhanced Raman spectroscopy for identification of heavy metal arsenic(V)mediated enhancing effect on antibiotic resistance. Anal Chem 88(6):3164–3170. https:// doi.org/10.1021/acs.analchem.5b04490 14. Cui L, Chen P, Chen S, Yuan Z, Yu C, Ren B, Zhang K (2013) In situ study of the antibacterial activity and mechanism of action of silver nanoparticles by surface-enhanced Raman spectroscopy. Anal Chem 85(11):5436–5443. https://doi.org/10.1021/ac400245j 15. Berry D, Mader E, Lee TK, Woebken D, Wang Y, Zhu D, Palatinszky M, Schintlmeister A, Schmid MC, Hanson BT, Shterzer N, Mizrahi I, Rauch I, Decker T, Bocklitz T, Popp J, Gibson CM, Fowler PW, Huang WE, Wagner M (2015) Tracking heavy water (D2O) incorporation for identifying and sorting active microbial cells. Proc Natl Acad Sci U S A 112(2):E194–E203. https://doi. org/10.1073/pnas.1420406112 16. Orphan VJ, House CH, Hinrichs K-U, McKeegan KD, DeLong EF (2001) Methaneconsuming archaea revealed by directly coupled isotopic and phylogenetic analysis. Science 293(5529):484–487. https://doi.org/10. 1126/science.1061338 17. Musat N, Halm H, Winterholler B, Hoppe P, Peduzzi S, Hillion F, Horreard F, Amann R, Jorgensen BB, Kuypers MMM (2008) A

Raman-SIP Deciphers Function and Activity of Microbes single-cell view on the ecophysiology of anaerobic phototrophic bacteria. Proc Natl Acad Sci U S A 105(46):17861–17866. https://doi. org/10.1073/pnas.0809329105 18. Wagner M (2009) Single-cell ecophysiology of microbes as revealed by Raman microspectroscopy or secondary ion mass spectrometry imaging. Annu Rev Microbiol 63:411–429. https://doi.org/10.1146/annurev.micro. 091208.073233 19. Zhang B, Cui L, Zhang K (2016) Dosage- and time-dependent antibacterial effect of zinc oxide nanoparticles determined by a highly uniform SERS negating undesired spectral variation. Anal Bioanal Chem 408 (14):3853–3865. https://doi.org/10.1007/ s00216-016-9478-z 20. Cui L, Chen P, Zhang B, Zhang D, Li J, Martin FL, Zhang K (2015) Interrogating chemical variation via layer-by-layer SERS during biofouling and cleaning of nanofiltration membranes with further investigations into cleaning efficiency. Water Res 87:282–291. https://doi.org/10.1016/j.watres.2015.09. 037 21. Chen PY, Cui L, Zhang KS (2015) Surfaceenhanced Raman spectroscopy monitoring the development of dual-species biofouling on membrane surfaces. J Membr Sci 473:36–44. https://doi.org/10.1016/j.memsci.2014.09. 007 22. Ru ECL, Etchegoin PG (2012) Singlemolecule surface-enhanced Raman spectroscopy. Annu Rev Phys Chem 63(1):65–87. https://doi.org/10.1146/annurevphyschem-032511-143757 23. Kubryk P, Kolschbach JS, Marozava S, Lueders T, Meckenstock RU, Niessner R, Ivleva NP (2015) Exploring the potential of stable isotope (resonance) Raman microspectroscopy and surface-enhanced raman scattering for the analysis of microorganisms at single cell level. Anal Chem 87(13):6622–6630. https://doi.org/10.1021/acs.analchem. 5b00673 24. Bauchop T, Elsden SR (1960) The growth of micro-organisms in relation to their energy supply. Microbiology 23(3):457–469. https://doi.org/10.1099/00221287-23-3457

107

25. Cui L, Butler HJ, Martin-Hirsch PL, Martin FL (2016) Aluminium foil as a potential substrate for ATR-FTIR, transflection FTIR or Raman spectrochemical analysis of biological specimens. Anal Methods 8(3):481–487. https://doi.org/10.1039/C5AY02638E 26. Zhang DY, Berry JP, Zhu D, Wang Y, Chen Y, Jiang B, Huang S, Langford H, Li GH, Davison PA, Xu J, Aries E, Huang WE (2015) Magnetic nanoparticle-mediated isolation of functional bacteria in a complex microbial community. ISME J 9(3):603–614. https:// doi.org/10.1038/ismej.2014.161 27. Haider S, Wagner M, Schmid MC, Sixt BS, Christian JG, H€acker G, Pichler P, Mechtler K, Mu¨ller A, Baranyi C, Toenshoff ER, Montanaro J, Horn M (2010) Raman microspectroscopy reveals long-term extracellular activity of chlamydiae. Mol Microbiol 77 (3):687–700. https://doi.org/10.1111/j. 1365-2958.2010.07241.x 28. Tao Y, Wang Y, Huang S, Zhu P, Huang WE, Ling J, Xu J (2017) Metabolic-activity-based assessment of antimicrobial effects by D2Olabeled single-cell Raman microspectroscopy. Anal Chem 89(7):4108–4115. https://doi. org/10.1021/acs.analchem.6b05051 29. Yusuke Y, Yuta S, Takuro I, Yoshiko M, Keisuke G, Yasuyuki O, Yu H (2017) Monitoring photosynthetic activity in microalgal cells by Raman spectroscopy with deuterium oxide as a tracking probe. Chembiochem 18 (20):2063–2068. https://doi.org/10.1002/ cbic.201700314 30. Yang K, Li HZ, Zhu X, Su JQ, Ren B, Zhu YG, Cui L (2019) Rapid Antibiotic Susceptibility Testing of Pathogenic Bacteria Using HeavyWater-Labeled Single-Cell Raman Spectroscopy in Clinical Samples. Anal Chem 91 (9):6296–6303. https://doi.org/10.1021/ acs.analchem.9b01064 31. Li HZ, Bi QF, Yang K, Zheng BX, Pu Q, Cui L (2019) D2O-Isotope-Labeling Approach to Probing Phosphate-Solubilizing Bacteria in Complex Soil Communities by Single-Cell Raman Spectroscopy. Anal Chem 91(3):2239–2246. https://dx.doi.org/10. 1021/acs.analchem.8b04820

Chapter 9 Data Analysis for DNA Stable Isotope Probing Experiments Using Multiple Window High-Resolution SIP Samuel E. Barnett, Nicholas D. Youngblut, and Daniel H. Buckley Abstract DNA stable isotope probing (DNA-SIP) allows for the identification of microbes that assimilate isotopically labeled substrates into DNA. Here we describe the analysis of sequencing data using the multiple window high-resolution DNA-SIP method (MW-HR-SIP). MW-HR-SIP has improved accuracy over other methods and is easily implemented on the statistical platform R. We also discuss key experimental parameters to consider when designing DNA-SIP experiments and how these parameters affect accuracy of analysis. Key words DNA-SIP, Stable isotope probing, R, High throughput sequencing

1 1.1

Introduction Background

DNA stable isotope probing (DNA-SIP) is a powerful tool for studying microbial metabolism within complex communities. DNA-SIP has been applied in a wide range of environments (e.g., soils [1–4], marine [5], gut [6, 7]), and in diverse applications (e.g., carbon cycling [1], nitrogen fixation [3, 8], methanotrophy [4, 5, 9], biodegradation [2, 10, 11]). The general approach is to (1) add an isotopically labeled substrate to a microbial community, (2) extract nucleic acids after some period of time, and then (3) to fractionate the nucleic acids on the basis of their buoyant density (BD) in a CsCl or other density gradient (as previously described [1, 3]). Nucleic acids that have incorporated isotopic label will increase in BD in proportion to the degree of isotope incorporation. High throughput sequencing is then used to characterize DNA sequence composition across a range of buoyant densities. However, many factors besides isotopic composition can influence the BD of nucleic acids [12, 13], and hence the accurate identification of isotopically labeled DNA requires careful comparison of labeled treatment samples to appropriate unlabeled controls.

Marc G. Dumont and Marcela Herna´ndez Garcı´a (eds.), Stable Isotope Probing: Methods and Protocols, Methods in Molecular Biology, vol. 2046, https://doi.org/10.1007/978-1-4939-9721-3_9, © Springer Science+Business Media, LLC, part of Springer Nature 2019

109

110

Samuel E. Barnett et al.

A number of techniques have been described for the analysis of DNA-SIP data [1, 12, 14]. Most identify labeled DNA by detecting an increase in amplicon abundance in “heavy” fractions of a treatment gradient compared to fractions of similar buoyant densities in a control gradient. By “heavy” fractions, we are simply referring to fractions with high buoyant densities. The treatment gradient is defined as containing DNA from organisms exposed to isotopically labeled substrates while the control gradient contains DNA from organisms exposed to identical but unlabeled substrates (see Subheading 1.2). DNA labeled with a heavy isotope will equilibrate in heavier fractions in the treatment gradient due to their higher BD. It is important to note that the selected BDs of the heavy fractions are dependent on the expected BD shift of labeled DNA, which differs depending on which stable isotope is used (i.e., 13C, 15 N, 18O). Advances in high-throughput sequencing and laboratory automation have increased the power of DNA-SIP experiments, necessitating the development of more powerful data analysis tools. We recently developed a toolset, the R package HTSSIP, for analyzing high throughput sequencing data from DNA-SIP experiments [15]. HTSSIP provides a variety of functions that facilitate the analysis of DNA-SIP amplicon data including several different analytical methods, namely quantitative SIP (qSIP) [14], highresolution DNA-SIP (HR-SIP) [1], and an improved method of multiple window HR-SIP (MW-HR-SIP) [12]. Code and functions for these methods are all implemented in HTSSIP, which is freely available at the CRAN archives (https://cran.r-project.org/ web/packages/HTSSIP/index.html). Description of the R package can be found in Youngblut et al. [15]. In addition to functions, the package contains vignettes with step-by-step walkthroughs and example datasets for multiple DNA-SIP analysis methods. Here we present the basic protocol for running MW-HR-SIP as well as considerations for planning and conducting DNA-SIP experiments. HR-SIP and MW-HR-SIP utilize the differential gene expression analysis from the R package DESeq2 [16] to identify operational taxonomic units (OTUs) or other amplicon based products that are enriched in high BD windows of the treatment compared to the control gradient. These high BD windows are made up of multiple heavy fractions. MW-HR-SIP uses multiple overlapping heavy windows, unlike HR-SIP which uses a single window. We have found that using multiple windows increases the sensitivity with negligible loss of specificity [12]. Regardless of isotope label, GC content of DNA partially dictates the BD of each DNA fragment. High GC fragments have higher buoyant densities than low GC fragments, leading to a “smearing” of diverse DNA across a gradient. The overlapping windows in MW-HR-SIP allow for detection of isotopic enrichment of OTUs with a wide range of

DNA-SIP Analysis with MW-HR-SIP

111

GC content. qSIP on the other hand uses quantitative PCR in addition to sequencing to generate an estimate of the mean BD of a given OTU in both the treatment and control gradients. This method then uses the change in BD between the treatment and control gradients to calculate isotopic enrichment of the DNA [14]. Unlike both HR-SIP and MW-HR-SIP, qSIP has the benefit of generating estimates of isotopic enrichment for a given OTU, which may be useful for calculations of uptake efficiency or competition. However, based on method comparisons using simulated datasets and using published analysis parameters, we have found that MW-HR-SIP tends to have higher specificity than qSIP, while maintaining similar or higher sensitivity at high isotope enrichment levels [12]. Choice of analysis method depends on the goals and logistics of each project and should be an important consideration when designing and planning a DNA-SIP experiment. In this chapter we will explain the setup and protocol for performing data analysis using MW-HR-SIP as well as identify experimental design considerations that may impact results. As with any DNA-SIP study, accuracy of MW-HR-SIP depends on the experimental design. Accuracy can be split into two equally important aspects, sensitivity and specificity. Sensitivity measures the proportion of true positives that are correctly identified. In the context of MW-HR-SIP, sensitivity measures the proportion of OTUs that incorporate isotope into their DNA and are correctly identified as enriched in the heavy BD windows of the treatment relative to the corresponding windows in the control. Specificity measures the proportion of true negatives that are correctly identified as such. In MW-HR-SIP, specificity measures the proportion of OTUs that do not incorporate isotope and are correctly identified as such. To maximize both sensitivity and specificity, the following aspects should be considered when planning any DNA-SIP analysis (Fig. 1). We highly recommend conducting a pilot experiment prior to a full DNA-SIP study with the aim of optimizing experimental parameters. A pilot study also gives researchers new to DNA-SIP an introduction to the procedures and methods and can help in the development of wet-lab methods, biologically relevant incubation time, and a realistic experimental scales and timelines for studies. 1.2 Isotopically Labeled Substrate

Sensitivity of MW-HR-SIP is dependent on the degree of BD shift between the labeled and unlabeled DNA. A larger shift in an OTU’s BD generally results in higher sensitivity, as this would increase the enrichment of its DNA in heavy windows of the treatment gradient compared to the control. Tests based on simulated data have shown MW-HR-SIP to have the highest sensitivity, with respect to other tested analysis methods, when incorporators are enriched with at least 50% 13C [12].

112

Samuel E. Barnett et al.

Too Low •Labeled substrate diluted by native substrate. •Reduced sensitivity of MW-HR-SIP.

Parameter

Amount of isotopically labeled substrate

Too High

•Excessive cross feeding. •Increased cost.

Time from substrate addition and harvest

•Excessive cross feeding. •Potential loss of Initial substrate incorporators.

•Too few samples for DESeq2. •Low power for statistical tests. •Reduced sensitivity of MW-HR-SIP.

Number of fractions

•Little DNA for further sequencing or errors. •Large logistical burden. •High sequencing cost.

•Not sequencing low abundant organisms. •Reduced sensitivity of MW-HR-SIP

Sequencing depth

•Slow growing organisms may incorporate little substrate.

•Increased cost

Fig. 1 Important experimental parameters to consider when designing a DNA-SIP experiment and what issues may arise with each. Most of these parameters can be tested and optimized during a pilot experiment

For most DNA-SIP experiments, isotopically labeled substrates should be (nearly) completely enriched (>95%). If using specific compounds as substrates, rather than whole plant litter or biomass, they should be pure to ensure no nonspecific source of isotope. Commercially available complex compounds derived from biomass such as hemicellulose or lignin often have impurities. Ensure purity of complex compounds prior to use. Native sources of the substrate should also be minimized. If a natural medium such as soil or wastewater is used there is likely a native source of many target substrates. The dilution effect of the native substrates can be mitigated by adding significantly more of the isotopically labeled substrate or preincubating the sample to reduce native concentrations. Calculating the native concentration of substrate in the experimental medium may help determine the appropriate amount of isotopically labeled substrate to add. Pilot studies may help to determine the rate of substrate uptake in the community as a whole, informing the concentrations to be used in the full study. Cross feeding is a concern in any complex community. Cross feeding will result in OTUs labeled from incorporating carbon from biological byproducts and biomass rather than the initial substrate. Cross feeding can be mitigated by reducing the time from substrate addition to sample harvesting. This timing will be substrate dependent, for instance labile substrates such as glucose will usually be utilized by microbes rapidly, while whole leaf litter may require more time for significant incorporation.

DNA-SIP Analysis with MW-HR-SIP

113

1.3 Unlabeled Controls

Variation in community composition between control and treatment samples results in decreases in both sensitivity and specificity [12]. To reduce beta-diversity between treatments and controls, the control sample should be treated identically to the treatment sample with the only difference being isotopic label of the substrate. For example, an experiment studying the utilization of cellulose by soil bacteria after one or two weeks would include at least four different soil microcosms. Two treatment soil microcosms would be supplemented with 13C-cellulose, while two control soil microcosms would be supplemented with 12C-cellulose. One treatment and one control microcosm would be harvested destructively at each time point. In the analysis stage, the week one treatment will be compared to the week one control while the week two treatment will be compared to the week two control. Since both substrate and time are matched between the treatment and control microcosms, there should be little difference in community composition between treatment and control microcosms when surveying the unfractionated DNA. Unfractionated DNA is the DNA extracted from samples but not processed through a BD gradient before sequencing. For any HR-SIP study we recommend sequencing the unfractionated DNA from all samples. Unfractionated DNA is necessary to measure beta-diversity between treatment and control samples. Ideally, since the same substrates are added to all samples with the only difference being isotope label, there should be no systematic difference in community composition between the treatment and controls. Using simulated datasets we have found that a Bray–Curtis dissimilarity less than 0.2 is ideal [12]. Running this analysis on paired treatment and control samples from a pilot study will help to identify any factors that may be driving differences in community composition prior to running the full study. Study designs using multiple substrates can utilize the same control for multiple treatments, as long as every treatment and control is amended with all target substrates. For example, a study examining the utilization of both glucose and cellulose after one or two weeks would only require six different soil microcosms, each amended with both cellulose and glucose (Fig. 2).

1.4

Fixed angle rotors are the best option for ultracentrifugation of DNA-SIP gradients. During ultracentrifugation, the BD gradient is established perpendicular to the axis of rotation. During deceleration, however, gravity causes reorientation of the gradient into a vertical orientation. It has been proposed that, due to interactions between the gradient solution and the tube wall, a thin layer called the diffusive boundary layer (DBL) introduces nonequilibrium DNA into gradient fractions following gradient reorientation (see supplementary Fig. S2 in ref. 12). This may lead to a smearing of DNA across the gradient, reducing accuracy, with the level of smearing relative to the degree of change in orientation. Swinging

Rotor Geometry

Samuel E. Barnett et al.

Control

12C-Glucose

1 Weeks

+ 12C-Cellulose

1 !

114

2 Weeks

4 !

Treatments

13

C-Glucose + 12C-Cellulose

1 Weeks 2 Weeks

12C-Glucose

1 Weeks

+ 13C-Cellulose

2 substrates per microcosm

2

2 Weeks 6 microcosms

2 sampling timepoints

5

3

Gradient Comparisons Treatment Control 2 1 3 1 5 4 6 4 4 total comparisons

6

6 CsCl gradients

Fig. 2 Experimental setup using glucose and cellulose as substrates with two harvesting time points. The table of contrasts indicates which pairs of gradients will be compared using MW-HR-SIP

bucket rotors should produce minimal DBL as the entire tube reorients along with the gradient during deceleration. However, swinging bucket rotors require a substantially longer time to equilibrium than fixed angle rotors and are hence impractical in SIP experiments. Vertical rotors should produce the worst DBL smearing effects, while fixed angled rotors should have a smaller degree of smearing. 1.5 Fractionation and Sequencing

When deciding on the number and size of gradient fractions to collect and sequence, both analytical and logistical consequences must be considered. There is a tradeoff between collecting more and smaller fractions or fewer and larger fractions from BD gradients. Analytically, collecting and sequencing many small fractions is ideal as this results in higher sensitivity. Logistically however, this reduces the amount of DNA available for sequencing, increases sequencing cost, and requires more logistics in handling many individual samples. On the other hand, collecting fewer but larger fractions will increase the amount of DNA available, possibly allowing for multiple sequencing methods such as 16S rRNA, fungal ITS, or gene specific amplicon sequencing as well as metagenome sequencing. This approach will, however, decrease sensitivity, leading to more false negatives [12]. Deciding on how many fractions to collect involves balancing these tradeoffs as well as the resources, time constraints, and funding available for the project. Regardless, we recommend having no fewer than three fractions per BD window as this is the recommended sample size for basic DESeq2 analyses [16]. A pilot study will help to determine how many fractions may be feasible for a study. Depth of sequencing also affects sensitivity. Based on simulated datasets, we found that sequencing fractions more deeply increases

DNA-SIP Analysis with MW-HR-SIP

115

the sensitivity of MW-HR-SIP analyses. Sensitivity improvement is most significant for OTUs with low relative abundance [12]. Sequencing depth is especially important to consider if barcoding sequencing libraries, as more pooled samples generally reduces depth, and when deciding on sequencing technology to use. 1.6

2 2.1

Replicates

While including more replicates is ideal in most studies, the expense, logistics, and time required to include them in most DNA-SIP experiments often makes them infeasible. MW-HR-SIP relies on multiple fractions within BD windows for statistical power and therefore does not require replicates. As technologies and techniques improve, laboratory automation and rapid high throughput sequencing may increase the use of replicates in DNA-SIP studies in the future.

Materials Data Format

Data input for analysis with HTSSIP should be in the form of a phyloseq object [17]. This is a commonly used format for processed sequencing data in R and can easily be generated from QIIME or other sequence analysis pipelines. An example dataset can be found at https://github.com/buckleylab/HR-SIP_example. A phyloseq object is composed of four main datasets that are linked by a unique sample ID and unique OTU name: l

sample_data: the metadata for all samples. For MW-HR-SIP analysis, each sample will be a separate fraction from each BD gradient. Within this metadata it is essential to have the distinguishing variables that will be used to build contrasts for comparing treatments with their corresponding controls. Variables from the example in the introduction could be “Labeled_substrate” and “Week”. One variable within sample_data must be labeled “Buoyant_density” and contain the BD of each fraction. sample_data format can be checked with HTSSIP function physeq_format(). Table 1 contains the first 10 entries of the example metadata table.

l

otu_table:

l

tax_table:

a table composed of read counts for each OTU in each sample. As with any DESeq2 analysis, the OTU table must contain the raw read counts prior to normalization [16]. a table containing the assigned taxonomy of

each OTU. l

phy_tree: a phylogenetic tree of all OTUs. This component is not essential for HR-SIP, unless UniFrac or other tree-based method is used for beta-diversity measurements and other downstream analyses.

116

Samuel E. Barnett et al.

Table 1 First ten entries in sample metadata for example data used in this work Sample_ID

Sample_type

Substrate

Fraction

Buoyant_density

13C-Ami.Frac5

SIP

13C-Ami

5

1.771206

12C-Con.Frac5

SIP

12C-Con

5

1.772298

13C-Ami.Frac6

SIP

13C-Ami

6

1.766835

12C-Con.Frac6

SIP

12C-Con

6

1.766835

13C-Ami.Frac7

SIP

13C-Ami

7

1.762464

12C-Con.Frac7

SIP

12C-Con

7

1.762464

13C-Ami.Frac8

SIP

13C-Ami

8

1.757

12C-Con.Frac8

SIP

12C-Con

8

1.757

13C-Ami.Frac9

SIP

13C-Ami

9

1.752629

12C-Con.Frac9

SIP

12C-Con

9

1.752629

This list is ordered by fraction and substrate. It is important to note that the BDs do not always match between the treatment and control for a given fraction number. This is due to the stochasticity in the fractionation method. The heavy BD windows were designed to account for this discrepancy

The phyloseq object containing all project data can be saved as an R object using base command saveRDS(). To test if a phyloseq object is in the appropriate format for use with HTSSIP, run the following command. tryCatch(physeq_format(example.physeq), error = function(e) e)

In the protocols below, the example data (SIP_phyloseq.rds) consists of processed MiSeq sequencing data from an experiment adding 13C-labeled amino acid mix to soil. The phyloseq object for the MW-HR-SIP example includes all collected CsCl gradient fractions from both the 13C-amino acid treatment microcosm as well as the 12C-amino acid control microcosm. Most real-world studies will have more than one treatment and control. This situation can be handled by converting the master phyloseq containing all data into a list of phyloseq objects, each containing a single contrast between a treatment and its corresponding control. HRSIP can handle cases where the same control is used for multiple treatments. For more information on this please examine the HTSSIP vignettes or the tutorial at https://github.com/buckleylab/HR-SIP_exam ple. The phyloseq object used for the beta-diversity example (unfractionated_phyloseq.rds) contains just the sequenced unfractionated samples from both the treatment and control microcosms. This was DNA directly sequenced following extraction from the microcosm soil without being fractionated in a CsCl gradient.

DNA-SIP Analysis with MW-HR-SIP

2.2 The HTSSIP Package

117

MW-HR-SIP analysis utilizes the statistical software R [18], which can be installed following instructions found on the r-project website, www.r-project.org. MW-HR-SIP along with other DNA-SIP analysis methods has been assembled into the R package HTSSIP [15]. HTSSIP can be installed from the CRAN archive (https:// cran.r-project.org/) through R using the command install. packages(“HTSSIP”). This package is also available through bioconda at https://anaconda.org/bioconda/r-htssip. The latest development version of HTSSIP can be found at https://github. com/buckleylab/HTSSIP and installed using the devtools command install_github(“buckleylab/HTSSIP”) [19]. HTSSIP version 1.4.0 requires the following packages available on CRAN: R (
3.4.0), igraph (
1.1.2), ape (
4.1), magrittr (
1.5), stringr (
1.2.0), plyr (
1.8.4), dplyr (
0.7.4), tidyr (
0.7.2), ggplot2 (
2.2.1), vegan (
2.4.0), coenocliner (
0.2.2), lazyeval (
0.2.0). HTSSIP version 1.4.0 also requires the following packages available on Bioconductor (https://bioconductor.org/): DESeq2 (
1.16.1), phyloseq (
1.20.0) Bioconductor packages must be installed prior to installation of HTSSIP using the following commands: source(’http://bioconductor.org/biocLite.R’) biocLite(’phyloseq’) biocLite(’DESeq2’)

2.3 Parallel Processing (Optional)

3

DNA-SIP analysis can be computationally intensive and may take hours to run on large datasets. To make analysis faster for those with multiple core systems, the MW-HR-SIP functionality in HTSSIP is capable of parallel processing. Parallel processing requires the additional package doParallel [20], which is available on CRAN. Parallel processing can be activated setting option parallel¼TRUE in the main command HRSIP (see Note 1).

Methods

3.1 Beta Diversity Between Treatment and Control Microcosms

Prior to running MW-HR-SIP, we recommended checking the dissimilarity in the unfractionated community composition between each treatment sample and its corresponding control. Here, this beta-diversity measurement will be done by calculating the Bray–Curtis dissimilarity index. Ideal dissimilarity between a treatment and its control is less than 0.2 [12]. 1. Import the required R package. library(phyloseq)

118

Samuel E. Barnett et al.

2. Import the phyloseq object containing the unfractionated sequencing data Unfrac.physeq