1 Microbial Transformations of Organic Chemicals in Produced Fluid From

Microbial transformations of organic chemicals in produced fluid from hydraulically

fractured natural-gas wells

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy

in the Graduate School of The Ohio State University

Morgan V. Evans

Graduate Program in Environmental Science

The Ohio State University

2019

Dissertation Committee

Professor Paula Mouser, Advisor

Professor Gil Bohrer, Co-Advisor

Professor Matthew Sullivan, Member

Professor Ilham El-Monier, Member

Professor Natalie Hull, Member

Copyrighted by

Morgan Volker Evans

2019

Abstract

Hydraulic fracturing and horizontal drilling technologies have greatly improved the production of oil and natural-gas from previously inaccessible non-permeable rock formations. Fluids comprised of water, chemicals, and proppant (e.g., sand) are injected at high pressures during hydraulic fracturing, and these fluids mix with formation porewaters and return to the surface with the hydrocarbon resource. Despite the addition of biocides during operations and the brine-level salinities of the formation porewaters, microorganisms have been identified in input, flowback (days to weeks after hydraulic fracturing occurs), and produced fluids (months to years after hydraulic fracturing occurs). Microorganisms in the hydraulically fractured system may have deleterious effects on well infrastructure and hydrocarbon recovery efficiency. The reduction of oxidized sulfur compounds (e.g., sulfate, thiosulfate) to sulfide has been associated with both well corrosion and souring of natural-gas, and proliferation of microorganisms during operations may lead to biomass clogging of the newly created fractures in the shale formation culminating in reduced hydrocarbon recovery. Consequently, it is important to elucidate microbial metabolisms in the hydraulically fractured ecosystem.

The numerous nitrogen and carbon sources injected in input fluid mixtures may sustain shale-associated microorganisms, prompting a need to investigate the capacity of

ii microbial life to enzymatically transform organic chemicals commonplace to hydraulic fracturing operations.

In Chapter 2, we investigated the putative microbial metabolisms of two bacterial genera frequently identified in the first few weeks to months after hydraulic fracturing occurs

(Marinobacter and Arcobacter). Using microbial culture-dependent methods (e.g., genomics, salinity range and carbon source growth testing) and microbial culture- independent methods (e.g., metagenomics) coupled to geochemical measurements from four Appalachian Basin natural-gas wells, we determined Marinobacter and Arcobacter likely play significant roles in biogeochemical cycling weeks to months after fracturing.

There is evidence that Marinobacter can utilize a wide variety of nitrogen and carbon compounds including hydrocarbons, whereas Arcobacter can use a reductive TCA cycle coupled to sulfur oxidation.

In Chapter 3, we tested the ability of the dominant shale-associated bacterial genera,

Halanaerobium, to transform frequently used polyglycol surfactants. We used a variety of microbial and analytical chemical methods both in situ during production of a hydraulically fractured Utica-Point Pleasant natural-gas well, and in the laboratory during batch growth of Halanaerobium congolense WG10. Our results revealed that

Halanaerobium can enzymatically transform alkyl polyethoxylates, polypropylene glycols, and monomeric glycols, under anaerobic conditions.

iii

In Chapter 4, we investigated microbial (de)halogenation pathways during hydraulic fracturing of natural-gas wells using a metagenomic approach. We identified genes encoding for halogenation, hydrolytic dehalogenation, and reductive dehalogenation, months after fracturing occurred. The presence of these pathways indicates the potential for microbially-generated organohalides in produced fluids as well as reduction of organohalides in wastewaters.

In Chapter 5, we surveyed the microbial community at six stages of treatment in a class

(II) injection well facility. The microbial community was highly similar to produced fluids from the Marcellus Shale, despite the transport of wastewaters in trucks, exposure to oxygen, and addition of chemicals in the treatment process.

Dedication

This dissertation is dedicated to my husband, Derek.

Acknowledgments

First, I would like to express my gratitude to my advisor, Dr. Paula J. Mouser. She has taught me how to be a quality scientist and has been dedicated to my growth as a researcher and has gone to great lengths to ensure my success. My appreciation and respect for her are unquantifiable. I would like to thank my committee members, Dr. Gil

Bohrer, Dr. Matthew Sullivan, Dr. Ilham El-Monier, and Dr. Natalie Hull, for their valuable time and service on my committee. Thank you to Dr. Kelly Wrighton, Dr. Mike

Wilkins, Rebecca Daly, Mikayla Borton, and Dr. Anne Booker, for their assistance on this dissertation. I would like to express my appreciation to Dr. Andrea Hanson and Dr.

Jenna Luek, whose advice, editing, and moral support helped me endure even the toughest times in my graduate career. Thanks to Jenny Panescu, our lab manager and fellow student, who helped me transition into graduate school and taught me proper microbiology procedures in the lab. Thanks to my fellow students, July Laszakovits,

Billy Fagan, Sharon Scott Grove, Anton Rosi, Nick Nastasi, Dr. Michael Brooker, Kate

Villars, Katie Heyob, Ryan Trexler. Their assistance in my research and thought- provoking scientific discussions helped on even the most difficult of days. I would like to express my appreciation for my family, friends, and husband, who were patient with me during this time, and helped keep me afloat these last 4 years. Funding for this research was provided by the Fay Graduate Fellowship (ESGP), the National Science Foundation vi

(CBET award no. 1342701), and DOE National Energy Technology Laboratory through the Marcellus Shale Energy and Environmental Laboratory (project #DE-FE0024297).

vii

Vita

2015………………………………………... B.S. Chemistry, The Ohio State University

Fay Fellow, Environmental Science

2015-2016………………………………….. Graduate Program, The Ohio State

University

2016-2017………………………………….. Graduate Administrative Assistant, Ohio

Water Resources Center, The Ohio State

University

2017-2019………………………………….. Graduate Research Associate, Department

of Civil, Environmental, and Geodetic

Engineering, The Ohio State University

viii

Publications

Evans, M. V., Panescu, J., Hanson, A. J., Welch, S. A., Sheets, J. M., Nastasi, N., et al.

(2018). Members of Marinobacter and Arcobacter influence system biogeochemistry during early production of hydraulically fractured natural gas wells in the appalachian basin. Frontiers in Microbiology 9, 2646. doi:10.3389/fmicb.2018.02646.

Evans M. V., Getzinger G., Luek J. L., Hanson A. J., McLaughlin M. C., Blotevogel J., et al. In situ transformation of ethoxylate and glycol surfactants by shale-colonizing microorganisms during hydraulic fracturing. (In Review at the ISME Journal).

Evans M. V., Daly R. A., Luek J. L., Wrighton K. C., Mouser P. J.. Microbial

(de)halogenation pathways in hydraulically fractured shale. (In Preparation)

Evans M. V. and Mouser P. J. Microbial communities in a class (II) injection well disposal facility receiving produced fluids from the Marcellus Shale. (In Preparation)

Processes Impacts. doi:10.1039/C8EM00291F.

Heyob, K. M., Blotevogel, J., Brooker, M., Evans, M. V., Lenhart, J. J., Wright, J., et al.

(2017). Natural Attenuation of Nonionic Surfactants Used in Hydraulic Fracturing Fluids:

Degradation Rates, Pathways, and Mechanisms. Environ. Sci. Technol. 51, 13985–13994. doi:10.1021/acs.est.7b01539.

Fields of Study

Major Field: Environmental Science

Table of Contents

Abstract ...... ii Dedication ...... v Acknowledgments...... vi Vita ...... viii Table of Contents ...... xi List of Tables ...... xvi List of Figures ...... xvii Chapter 1. Introduction ...... 1 1.1 Problem Description ...... 1 1.1.1 Microbial biogeochemical cycling in hydraulically fractured natural-gas wells ...... 4 1.1.2 Microbial xenobiotic pathways in hydraulically fractured shale ...... 6 1.1.3 Microbial community responses to produced fluid treatment ...... 9 1.2 Research objectives ...... 10 1.3 Dissertation overview ...... 11 1.3.1 Members of Marinobacter and Arcobacter influence system biogeochemistry during early production of hydraulically fractured natural-gas wells in the Appalachian Basin ...... 12 1.3.2 In situ transformation of ethoxylate and glycol surfactants by shale-colonizing microorganisms during hydraulic fracturing ...... 13 1.3.3 Microbial (de)halogenation pathways in hydraulically fractured shale ...... 14 1.3.4 Microbial communities in a class (II) injection well disposal facility receiving produced fluids from the Marcellus Shale ...... 14 1.4 References ...... 15 Chapter 2. Members of Marinobacter and Arcobacter influence biogeochemistry during early production of hydraulically fractured shale gas wells in the Appalachian Basin .... 25

2.1 Introduction ...... 26 2.2 Materials and Methods ...... 30 2.2.1 Methods Overview ...... 30 2.2.2 Sample Collection ...... 30 2.2.3 Non-purgeable Organic Carbon, Nitrogen, Sulfur, and Chloride Analysis ..... 31 2.2.4 Carbon Dioxide Analysis ...... 32 2.2.5 Microbial Enumeration ...... 33 2.2.6 Bacterial Isolation ...... 34 2.2.7 Salinity Growth Curves...... 34 2.2.8 Utilization of Carbon Substrates ...... 36 2.2.9 Isolation and Sequencing of DNA ...... 37 2.2.10 Phylogenetic analysis ...... 37 2.2.11 Scanning electron microscopy (SEM) sample preparation and imaging ...... 38 2.2.12 Metagenomic sequencing, assembly, and annotation ...... 39 2.2.13 Genomic and Metagenomic Analyses ...... 40 2.2.14 Data Accession...... 40 2.3 Results ...... 41 2.3.1 Marinobacter and Arcobacter persist in unconventional source ecosystem for weeks to months after hydraulic fracturing ...... 41 2.3.2 Shale Marinobacter and Arcobacter are slight to moderate halophiles ...... 44 2.3.3 Marinobacter and Arcobacter utilize a wide variety of carbon sources ...... 46 2.3.4 Marinobacter and Arcobacter influence nitrogen, sulfur, and iron cycles in unconventional systems ...... 49 2.3.5 Marinobacter and Arcobacter possess unique virulence strategies in produced fluid ...... 53 2.3.6 Isolate genomes largely agree with metagenomic data...... 54 2.4 Discussion ...... 55 2.4.1 Marinobacter and Arcobacter dominate early produced fluid communities through unique osmoadaptations and defense mechanisms...... 55 2.4.2 Marinobacter and Arcobacter are carbon opportunists ...... 57 2.4.3 Importance of Marinobacter and Arcobacter on biogeochemical cycling in flowback and produced fluids ...... 59 2.5 Conclusion ...... 62 2.6 References ...... 63 xii

2.7 Figures...... 77 Chapter 3. In situ transformation of ethoxylate and glycol surfactants by shale-colonizing microorganisms during hydraulic fracturing ...... 83 3.1 Introduction ...... 84 3.2 Materials and Methods ...... 87 3.2.1 Produced water sampling and pre-processing ...... 87 3.2.2 Bacterial growth experiments ...... 88 3.2.3 Shotgun proteomic analysis ...... 89 3.2.4 Polyglycol analysis ...... 90 3.2.5 Metagenomic sequencing and analysis ...... 91 3.2.6 Data Accession...... 92 3.3 Results & Discussion ...... 92 3.3.1 Surfactants attenuate in situ during the months following hydraulic fracturing ...... 92 3.3.2 Geochemical mixing model discerns physical from biochemical surfactant trends ...... 93 3.3.3 Metagenomic identification of surfactant degrading genes ...... 95 3.3.4 Surfactant degradation observed in isolate cultures ...... 97 3.4 Conclusion ...... 102 3.5 References ...... 105 3.6 Figures...... 113 Chapter 4. Hydraulically fractured natural-gas well microbial communities contain genomic (de)halogenation potential ...... 117 4.1 Introduction ...... 118 4.2 Methods...... 120 4.2.1 Sample collection and analysis ...... 120 4.2.2 Metagenomic sequencing, assembly, annotation, and binning ...... 120 4.2.3 Target gene phylogenies ...... 121 4.2.4 Data accession ...... 122 4.3 Results and Discussion ...... 122 4.3.1 Organohalides present in fluids produced from HF O&G wells ...... 122 4.3.2 Halogenation and dehalogenation potential in Appalachian Basin natural-gas wells ...... 124

xiii

4.3.3 Phylogeny of haloperoxidase and reductive dehalogenation genes in HF O&G wells ...... 126 4.4 References ...... 129 4.5 Figures...... 136 Chapter 5. Marcellus Shale wastewater microbial communities persist through pre- treatment at a class (II) injection well disposal facility ...... 139 5.1 Introduction ...... 140 5.2 Materials & Methods ...... 142 5.2.1 Site and sampling information ...... 142 5.2.2 Cell counts ...... 143 5.2.3 DNA extraction, sequencing, and genomic analyses ...... 144 5.3 Results ...... 145 5.4 Discussion ...... 147 5.4.1 Class (II) injection well treatment processes alter microbial community members ...... 147 5.4.2 Injection well communities resemble produced fluid communities from hydraulically fractured wells...... 148 5.4.3 Putative metabolic implications for the disposal facility ...... 150 5.5 Conclusion ...... 151 5.6 References ...... 152 5.7 Figures...... 159 Chapter 6. Conclusions and Future Work ...... 163 6.1 Conclusions and Future Studies ...... 163 6.1.1 Examine members of the Marinobacter and Arcobacter genera for gene pathways that may impact nitrogen, carbon, and sulfur cycles during the early months after HF...... 163 6.1.2 Determine if microorganisms in produced fluids transform xenobiotic glycol surfactants commonly used during HF operations...... 165 6.1.3 Elucidate putative microbial (de)halogenation pathways in HF natural-gas wells to determine if organohalides in produced fluids may be biotically produced and/or transformed...... 167 6.1.4 Identify changes in microbial communities and evaluate potential mechanisms for persistence in the treatment processes of a class (II) injection well facility disposing of produced fluids from the Marcellus Shale...... 169 6.2 Concluding Thoughts ...... 171

xiv

Bibliography ...... 173 Appendix A. Supporting Information for Chapter 2 ...... 195 Appendix B. Supporting Information for Chapter 3 ...... 204 Supplemental Methods...... 204 Supplemental Text ...... 208 Supplemental Figures...... 209 Supplemental Tables ...... 218 Supplemental References ...... 224 Appendix C. Supporting Information for Chapter 4 ...... 226 Supplemental Methods...... 227 Supplemental Text ...... 228 Supplemental Tables ...... 235 Supplemental References ...... 242 Appendix D. Supporting Information for Chapter 5 ...... 244

List of Tables

Table 2.1. Identifiers for hydraulically fractured natural-gas wells analyzed and/or summarized in this study...... 82 Table24.1. Reported organohalides in produced fluid from HF oil and natural-gas wells...... 138 Table4A1. Salinity curve results for Marinobacter and Arcobacter ...... 200 Table5A2. Pulled metagenome genes meeting bitscore and identity cutoffs ...... 201 Table6B1. Frequency of use as disclosed in the FracFocus database, adapted from Rogers et al. 2015*Frequency is dependent on structure...... 218 Table7B2. Listed constituents of industrial Revert Flow used for culturing experiments ...... 218 Table8B3. Detection limits, linear range, and R2 values for GC and IC organic acid data. (*LOD/LOQ is given as the lowest calibration curve point) ...... 218 Table9B4. Surfactant data from the Utica-Point Pleasant natural-gas well produced fluids including retention times, rate constants, and LOD/LOQs ...... 219

Table10B5. PPG and AEO species detected in culture samples with R2 values, LOD/LOQs (n=16), and intraday variability ...... 220

Table11B6. Relative standard deviation (n=3) for the sum of PPGs and AEOs at C/C0=0.1 to C/C0=1 ...... 220

Table12B7. Reported Cfinal/C0 values from Figure 3 ...... 221

Table13B8. BLASTp results from Halanaerobium congolense WG10 to metagenomes 221

Table14B9. Relevant geochemical measurements from the Utica-Point Pleasant well fluid through time. NPOC (Non-purgeable organic carbon). Gray cells indicate samples where data was not measured...... 222

Table15B10. Presence (red) or absence (gray) of key genes in all Halanaerobium isolates from the Utica-Point Pleasant natural-gas well ...... 223

Table16C1. Metagenomic information and accession numbers ...... 235

Table17C2. V3-V4 region of the EMIRGE reconstructed 16S rRNA for reported taxa .. 236

Table18C3. Isolate genes used for reference gene database ...... 237

Table19C4. Reference gene database amino acid bit-score and identity (%) cutoffs ...... 238

Table20C5. Gene annotations and blastp taxonomy ...... 239

Table21D1. Constituents in disclosed chemical additives used in the disposal well facility treatment process ...... 244

xvi

List of Figures

Figure 2.1 Flowchart summarizing methods used for isolating and characterizing Arcobacter and Marinobacter strains from produced fluid samples...... 77 Figure 2.2 Geochemical and microbial community data from four Appalachian Basin natural-gas wells...... 78 Figure 2.3. SEM images and phylogenetic placement of Marinobacter and Arcobacter. 79 Figure 2.4. Salinity growth curves and carbon source utilization by Arcobacter and Marinobacter...... 80 Figure 2.5. Conceptual metabolic models for A) Marinobacter and B) Arcobacter isolates incorporating isolate genomic (black arrows) and metagenome data (red arrows)...... 81 Figure63.1 Trends in AEOs and PEGs in a Utica-Point Pleasant natural-gas well ...... 113 Figure73.2 Surfactant enzymatic degradation pathway, metagenomic gene copies relevant to pathway, and Halanaerobium congolense WG10 contig...... 114 Figure83.3. Relative change in initial and final concentrations of polypropylene glycols (PPGs) and C8 alkyl polyethoxylates (AEOs) in surfactant amended cell cultures...... 115 Figure93.4. Proposed surfactant metabolic reconstruction of Halanaerobium congolense WG10 cultured with surfactants compared to glucose controls...... 116

Figure104.1. Organohalide metabolism genes, geochemical, and microbial biomass measurements in four Appalachian Basin natural-gas wells ...... 136

Figure114.2. Phylogenetic placement of dehalogenase and halogenase genes ...... 137 Table3A1. Biolog plate reader carbon source testing results ...... 196

Figure13A2. FracFocus Disclosure for Utica-S1 ...... 197

Figure14A3. FracFocus Disclosure for Utica-S4 ...... 198

Figure15A4. FracFocus Disclosure for Marcellus-MIP3H ...... 199

Figure16B1. Chromatogram of PEGs (0, EOx) and alkyl polyethoxylates (AEOs) (Cx,EOy) in Utica-Point Pleasant natural-gas well produced fluid at 86 days after hydraulic fracturing...... 209

Figure17B2. A) Chromatogram of detected PPGs and AEOs adducts in culture samples at t=0. B) Chromatogram zoom-in of C8 AEO ethoxymers...... 210

Figure18B3. Maximum-likelihood phylogenetic tree of pduC genes from Utica-Point Pleasant natural-gas well metagenomes and isolates. Bootstrap values (0-100%) are denoted by blue circles...... 211

Figure19B4. Metabolite trends in produced fluids...... 214

Figure20B5. Two end member conservative mixing models for strontium, chloride, and lithium-chloride ratios over 120 days assuming input fluid and day 204 as end members...... 215 xvii

Figure21B6. Growth of H. congolense WG10 on glucose alone (left) and glucose + revert flow (right)...... 215

Figure22B7. Growth curves and EG and PEG trends during H. congolense WG10 growth ...... 216

Figure23B8. Average z-score in proteins associated with glycol-metabolism surfactant treated cultures (Revert Flow, RF) and glucose controls (GC)...... 217

Figure25D1. Rarefaction curves for observed OTUs at the lowest sequencing depth (18,054 sequences) ...... 244

xviii

Chapter 1. Introduction

1.1 Problem Description

Horizontal drilling and hydraulic fracturing (HF) technologies have been increasingly utilized to recover hydrocarbon resources from shale formations (Arthur et al., 2009;

Kerr, 2010; Tour et al., 2010). Petroleum resources captured from shale formations have greatly impacted energy production in the United States, as HF wells now make up the majority of newly drilled oil and gas (O&G) wells (U.S. Energy Information

Administration (EIA), 2018). Moreover, half of all crude oil production and two-thirds of all natural-gas production come from shale formations (U.S. Energy Information

Administration (EIA), 2016a, 2016b).

HF involves the high-pressure injection of water, proppant (e.g., sand), and chemicals into the shale rock after lateral drilling occurs along the target formation (Arthur et al.,

2009; Vidic et al., 2013). This generates a network of fractures held open by the injected proppant (Arthur et al., 2009; Kerr, 2010; Tour et al., 2010), allowing oil and/or gas reserves to flow to the surface along with a mixture of injected fluid and formation porewater (Kerr, 2010; Vidic et al., 2013). As the well matures, produced fluid chemical composition diverges from the surface injected fluids and converges towards subsurface

1 formation brines, with characteristically high salinity and total dissolved solids (Barbot et al., 2013; Haluszczak et al., 2013; Kondash and Vengosh, 2015; Vidic et al., 2013).

Microbial communities present in produced fluids shift simultaneously with the fluid chemistry changes (Cluff et al., 2014). Microbial communities in injected fluids typically resemble the microorganisms present in freshwater inputs (Cluff et al., 2014; Daly et al.,

2016). However, as the well matures, oxygen, electron acceptors, and carbon sources are depleted, and the communities become dominated by halotolerant anaerobes and methanogens (Cluff et al., 2014; Daly et al., 2016; Davis et al., 2012; Fichter et al., 2012;

Lipus et al., 2018; Mohan et al., 2013; Struchtemeyer and Elshahed, 2012; Wuchter et al.,

2013). Similar microbial communities have been documented in produced fluids from

O&G wells in geographically-distinct shale formations, including the Marcellus Shale

(West Virginia, Pennsylvania), Utica-Point Pleasant Formation (Ohio), Barnett Shale

(Texas), Bakken Shale (North Dakota), and the Antrim Shale (Michigan) (Booker et al.,

2017; Cluff et al., 2014; Daly et al., 2016; Davis et al., 2012; Fichter et al., 2012; Lipus et al., 2018; Mohan et al., 2013; Struchtemeyer and Elshahed, 2012; Wuchter et al., 2013).

The metabolic capacities of specific microorganisms in produced fluids from HF petroleum wells have increased our knowledge of microbial life and their impact on fluid chemistry and well infrastructure. Halanaerobium is the dominant taxa in several wells from the Utica-Point Pleasant, Marcellus Shale, Bakken Shale, Antrim Shale, and Barnett

Shale O&G wells (Booker et al., 2017; Cluff et al., 2014; Daly et al., 2016; Liang et al.,

2016; Lipus et al., 2018; Mohan et al., 2013; Wuchter et al., 2013). This genus participates in a community metabolism through the fermentation of microbially produced osmoprotectants, which was initially catalyzed by the addition of choline in fracturing fluid inputs (Daly et al., 2016). Halanaerobium also has the capacity to metabolize fracturing fluid gelling agent guar gum and reduce thiosulfate, a redox reaction frequently associated with corrosion of steel (Booker et al., 2017; Choudhary et al., 2015; Liang et al., 2016). Co-fermentation of amino acids (Stickland reaction) has been shown to sustain several dominant microbial taxa (Halanaerobium, Geotoga, Ca.

Uticabacter, Methanohalophilus) in HF natural-gas wells from the Appalachian Basin

(Borton et al., 2018b). Methanohalophilus strains in produced fluid may utilize methanol, which is frequently disclosed in HF fluid inputs (Borton et al., 2018a; Elsner and

Hoelzer, 2016).

However, very little else is known about how produced fluid microorganisms interact with the numerous other chemicals disclosed in HF fluid recipes (Elsner and Hoelzer,

2016). During HF, both naturally-derived and xenobiotic compounds are injected into the subsurface, supplying potential carbon and nitrogen sources for the proliferation of halotolerant bacteria and archaea (Elsner and Hoelzer, 2016; Mouser et al., 2016). Due to the potential for well corrosion, natural-gas souring, and fracture-network clogging by microbial biomass, it is imperative to understand if injected chemicals can be readily transformed by produced fluid microorganisms during HF operations.

The vast majority of produced fluid wastewater in certain states (e.g. Ohio, West

Virginia, Texas) is disposed of by injection into deep subsurface class (II) disposal wells

(U.S. Environmental Protection Agency, 2016; Veil, 2015). Although microbial communities have been well-documented in fluids during and after HF, changes during disposal operations remains unknown. Given that induced seismicity has been observed in class (II) disposal wells receiving HF produced fluids (Kim, 2013; Weingarten et al.,

2015) and that HF-associated microorganisms are capable of corroding infrastructure and forming pressure-inducing clogs (Fichter et al., 2012), understanding the microbial ecology of these wastewaters prior to injection is imperative.

Using a variety of chemical and microbial techniques for both field and laboratory data, we sought to further elucidate microbial metabolisms in produced fluids from HF shale O&G wells. Below, three specific knowledge gaps, which are the primary concerns of this dissertation, are introduced in detail.

1.1.1 Microbial biogeochemical cycling in hydraulically fractured natural-gas wells

In the weeks to months following HF, produced fluids have low-to-moderate levels of salinity and high concentrations of nitrogen and carbon from input fluid chemicals

(Barbot et al., 2013; Chapman et al., 2012; Cluff et al., 2014; Haluszczak et al., 2013;

Harkness et al., 2015; Lester et al., 2015; Warner et al., 2013). In contrast, in the months to years after HF, the composition changes as produced fluids become characterized by anoxic conditions, brine-level salinities, high sulfide levels, and low carbon and nitrogen 4 concentrations (Barbot et al., 2013; Chapman et al., 2012; Cluff et al., 2014; Haluszczak et al., 2013; Harkness et al., 2015; Lester et al., 2015; Warner et al., 2013).

Consequently, microbial communities change over the course of production as the O&G well matures (Cluff et al., 2014; Daly et al., 2016). Several studies have investigated the microbial metabolisms of dominant microorganisms found in later produced fluids (e.g.

Halanaerobium, Methanohalophilus), but little has been reported on the microbial metabolisms of the community members identified during the early period after production (Booker et al., 2017; Borton et al., 2018b; Liang et al., 2016; Lipus et al.,

2017).

The unique well-maturation period in the first few months after HF is likely marked by microbially-mediated redox transformations that are thermodynamically unfavorable in later production times. These metabolic shifts may influence carbon, nitrogen, and sulfur cycles, with unknown impacts on the well infrastructure and, more broadly, on the shale formation itself. Microorganisms in O&G environments have previously been associated with the souring of natural-gas reserves from sulfide production, the formation of biofilms which clog formation fracture networks resulting in decreased hydrocarbon yields, and the corrosion of infrastructure from microbially-mediated redox reactions

(Arensdorf et al., 2009; Choudhary et al., 2015; Fichter et al., 2012).

Several studies have identified Marinobacter and Arcobacter genera in produced fluids several months after production occurred from multiple non-permeable formations including the Barnett, Antrim, Haynesville, Utica-Point Pleasant, and Marcellus (Cluff et al., 2014; Daly et al., 2016; Davis et al., 2012; Fichter et al., 2012; Liang et al., 2016;

Mohan et al., 2013; Mouser et al., 2016; Struchtemeyer and Elshahed, 2012; Wuchter et al., 2013). These taxa have been previously identified in other environments, and their ability to participate in biogeochemical cycling has been well-described. Marinobacter aquaeolei was called a biogeochemical opportunist for its ability to metabolize a wide variety of carbon and nitrogen compounds, as well as utilize both oxygen and nitrate as terminal electron acceptors (Singer et al., 2011). An Arcobacter strain was described as capable of coupling CO2 fixation to sulfur oxidation, which was never before identified in

HF produced fluid microorganisms (Wirsen et al., 2002). While one particular study highlighted the potential for Marinobacter to utilize organic nitrogen and hydrocarbon- containing compounds (Daly et al., 2016), the full metabolic capabilities of Marinobacter and Arcobacter and their capacity for biogeochemical cycling during the initial period following HF have yet to be elucidated.

1.1.2 Microbial xenobiotic pathways in hydraulically fractured shale

A myriad of chemicals is reportedly used during fracturing operations for many purposes, including improving hydrocarbon extraction efficiency and protecting well infrastructure

(Elsner and Hoelzer, 2016). Several of these chemicals are naturally-derived or labile carbon compounds capable of sustaining microbial life in produced fluids. For example, 6 guar gum, a plant-produced polysaccharide, is a gelling agent utilized as a carbon source by a Halanaerobium strain isolated from a HF Bakken Shale oil well (Elsner and

Hoelzer, 2016; Liang et al., 2016). The vitamin choline is the most commonly used clay stabilizer in hydraulically fractured petroleum wells (Elsner and Hoelzer, 2016), and putatively sustains a community metabolism culminating in methanogenesis (Daly et al.,

2016).

However, there have been no investigations into the potential of produced fluid microbial communities to transform xenobiotic compounds injected during fracturing operations.

Xenobiotic glycols are frequently-disclosed fracture fluid components, reported in as many as 19% of HF wells as surfactants, solvents, cross-linkers, and scale inhibitors

(Elsner and Hoelzer, 2016; Rogers et al., 2015). Alkyl polyethoxylates (AEOs) are surfactants containing a glycol-chain connected to an alkyl group and have been identified in fluids produced from HF O&G wells (Nell and Helbling, 2018; Rosenblum et al., 2017; Thurman et al., 2014). A recent report described attenuation of these compounds over one year in a HF O&G well with little to no causal explanation

(Rosenblum et al., 2017). Polyglycols (e.g., AEOs, polyethylene glycols (PEGs)) are readily biodegraded under anaerobic conditions (e.g., sludge, methanogenic consortia)

(Huang et al., 2005; Huber et al., 2000; Kawai, 2002; Straß and Schink, 1986; Wagener and Schink, 1988), and several studies have investigated the fate of these compounds in anoxic microcosms simulating spills of HF fluid in groundwater and sediments (Heyob et al., 2017; Rogers et al., 2018). However, this biotransformation pathway has not been

7 investigated during HF operations, and could explain the aforementioned attenuation trends (Rosenblum et al., 2017). Biotransformation of these compounds may negatively influence hydrocarbon recovery efforts through degradation of injected chemicals leading to reduction in chemical efficacy, in addition to the production of organic acid metabolites capable of corroding infrastructure.

Further, recent work has described the presence of non-indigenous organic compounds that were not disclosed in fracture fluid recipes (Hoelzer et al., 2016). In particular, organohalides (haloaliphatics and haloaromatics) have been detected in produced fluids from numerous shale formations months to years after HF occurred (Akyon et al., 2019;

Hayes, 2009; Hoelzer et al., 2016; Luek et al., 2017, 2018; Luek and Gonsior, 2017;

Maguire-Boyle and Barron, 2014; Thacker et al., 2015). These compounds are suspected to be generated during HF through reactions between surface-sourced oxidizers (e.g. ammonium persulfate, sodium hypochlorite), geogenic-derived halides at g/L concentrations, and organic carbon present in fluids (Hoelzer et al., 2016; Luek et al.,

2017). Microbial organohalide transformation pathways are frequently found in similar anaerobic or saline environments (e.g. chemically-contaminated groundwater, seafloor sediments) and these pathways may be present in produced fluids from HF O&G wells

(Fetzner, 1998; Holliger and Schumacher, 1994; Kawai et al., 2014; Mohn and Tiedje,

1992). Biotic generation of organohalides has been associated with oxidative stress (e.g., peroxidases) and antagonistic relationships (e.g., antibacterial or antifungal agents) (van

Pée and Unversucht, 2003). Hydrolytic and reductive mechanisms of halide removal

8 from haloaliphatics and haloaromatics have been well-studied (Mohn and Tiedje, 1992; van Pée and Unversucht, 2003). Yet, there have been no investigations of microbial

(de)halogenation pathways in fluids produced from HF O&G wells. Halogenation reactions may explain the presence of organohalides in produced fluids after oxidative additive (e.g. ammonium persulfate, sodium hypochlorite) kinetics have slowed, and dehalogenation reactions may assist in the removal of toxic organohalides from wastewaters.

1.1.3 Microbial community responses to produced fluid treatment

Wastewaters from HF O&G wells are difficult to treat due to the high total dissolved solids (TDS), nitrogen, and carbon concentrations (Getzinger et al., 2015; Haluszczak et al., 2013; Vidic et al., 2013). Most wastewaters generated from shale gas development in

Ohio are not treated or reused in other HF O&G wells, but instead are disposed of through class (II) injection well facilities (Veil, 2015). However, induced seismicity has been increasingly observed in disposal formations receiving fluids from HF activity

(Ellsworth, 2013; Kell, 2011; Kim, 2013; Weingarten et al., 2015). Little is known about how wastewaters are treated prior to injection into disposal wells. Microbial life in fluids from HF O&G wells influence fluid chemistry during well production and the subsequent biomass may clog the generated fractures in the hydrocarbon formation (Daly et al.,

2016; Fichter et al., 2012). Consequently, these microorganisms may impact fluid chemistries during wastewater treatment prior to injection, and pore-clogging biomass may negatively impact disposal efficiency. One study identified degraded water and 9 sediment quality in a stream near a class (II) injection well facility using geochemical and microbial community analyses (Akob et al., 2016); however, no studies to date have investigated produced fluid microbial communities prior to injection into class (II) disposal wells. Identifying the key microorganisms and changes in community structure during treatment of HF wastewaters prior to injection will inform industry of best practices to avoid corrosion and bio-clogging of disposal formations.

1.2 Research objectives

This work is focused on identifying the microbial metabolisms in the HF system relevant to biogeochemical cycling and xenobiotic transformation, as well as determining how the microbial community changes when produced fluid is treated for injection well disposal.

Elucidating microbial metabolisms in this system helps us understand how microorganisms behave in extreme environments, and results will help inform industry of best practices to avoid microbially-mediated souring, corrosion, and unintended chemical transformations. There are four research objectives in this dissertation:

1. Examine members of the Marinobacter and Arcobacter genera for gene

pathways that may impact nitrogen, carbon, and sulfur cycles during the early

months after HF.

2. Determine if and how microorganisms in produced fluids are responsible for

transforming xenobiotic glycol surfactants commonly used during HF

operations.

3. Elucidate putative microbial (de)halogenation pathways in HF natural-gas

wells to determine if organohalides in produced fluids may be biotically

produced and/or transformed.

4. Identify changes in microbial communities and evaluate potential mechanisms

for persistence in the treatment processes of a class (II) injection well facility

disposing of produced fluids from the Marcellus Shale.

1.3 Dissertation overview

This dissertation has six chapters from introduction (Chapter 1) to conclusion (Chapter

6). Chapter 2 investigates the putative metabolisms of Marinobacter and Arcobacter using isolate genomics, field metagenomics, and field geochemical measurements.

Chapter 3 examines the ability of the dominant shale microorganism Halanaerobium to transform xenobiotic glycol surfactants using field data (surfactant chemistry, metagenomics) and laboratory-scale experiments where cells were cultured with surfactants (proteomics, surfactant chemistry). Chapter 4 uses a metagenomic approach to determine the potential for biotic halogenation and dehalogenation reactions in four

Appalachian Basin natural-gas wells. Chapter 5 surveys the microbial community structure at six points throughout the treatment process of a class (II) injection well disposal facility (Chapter 5). Chapter 6 summarizes the research in chapters 2-5 and assesses the future directions of the work.

1.3.1 Members of Marinobacter and Arcobacter influence system biogeochemistry during early production of hydraulically fractured natural-gas wells in the

Appalachian Basin

This chapter investigated the putative metabolisms of two bacterial genera commonly identified in produced fluids during the first 6 months after HF occurs. Marinobacter and

Arcobacter isolates were briefly studied for the range of salinities capable of sustaining growth, and for carbon source utilization. The genomes of these isolates were sequenced and mined for genes associated with nitrogen, carbon, and sulfur cycling pathways.

Metagenomic data from produced fluid samples of six HF natural-gas wells were also investigated for relevant biogeochemical cycling genes and compared to genomic data.

We hypothesized that these microbial taxa may play significant roles in biogeochemical cycling with the potential to negatively impact well infrastructure and hydrocarbon recovery efforts, as well as fuel other microorganisms in produced fluids through regeneration of electron acceptors and labile carbon sources. This chapter has been published in Frontiers in Microbiology with co-authors Jenny Panescu, Andrea Hanson,

Susan Welch, Julia Sheets, Nicholas Nastasi, Rebecca Daly, David Cole, Thomas Darrah,

Michael Wilkins, Kelly Wrighton and Paula Mouser. My contribution to this chapter was performing the genomic and metagenomic analyses, constructing the manuscript figures, and writing the manuscript. These contributions enabled connection of the laboratory data to the field data.

1.3.2 In situ transformation of ethoxylate and glycol surfactants by shale-colonizing microorganisms during hydraulic fracturing

The research described in this chapter identified losses in xenobiotic glycol surfactant abundances through time in a HF natural-gas well using liquid chromatography coupled to mass spectrometry. We hypothesized that surfactant losses were due to microbial enzymatic transformations during production. To determine the cause of these surfactant losses, we used a combination of chemical analyses, microbial culture-independent methods (e.g. metagenomics), and culturing techniques (e.g. shotgun proteomics on cells cultured with surfactants) to identify a previously unrecognized co-metabolic glycol surfactant transformation pathway in Halanaerobium congolense WG10 isolated from this well. This chapter is currently under review at the International Society for Microbial

Ecology journal with co-authors Gordon Getzinger, Jenna Luek, Andrea Hanson, Molly

McLaughlin, Susan Welch, Carrie Nicora, Samuel Purvine, Chengdong Xu, David Cole,

Thomas Darrah, David Hoyt, Thomas Metz, P. Lee Ferguson, Mary Lipton, Michael

Wilkins, and Paula Mouser. My contributions to this chapter included assisting in the experimental design, analyzing the degradation rates of the polyglycol surfactants from the environmental samples, performing genomic and metagenomic analyses, performing laboratory experiments with the bacterial isolate, developing a GC-FID method for organic acids, analyzing polyglycol surfactants in the culture samples, analyzing shotgun proteomic data, designing the figures, and writing the manuscript. These efforts confirmed hypotheses generated from field data by examining microbial polyglycol

13 metabolism in a laboratory setting using an isolate from the hydraulically fractured natural-gas well.

1.3.3 Microbial (de)halogenation pathways in hydraulically fractured shale

This chapter used a metagenomic approach to identify microbial genes associated with organohalide cycling in produced fluids from natural-gas wells. We hypothesized that genes encoding for dehalogenation would be present but that genes associated with halogenation reactions would not be identified in produced fluids. Consequently, this would indicate that the organohalides identified in numerous produced fluid samples by others are formed abiotically but that microorganisms may dehalogenate the generated organohalides. This chapter was written for submission to Environmental Science and

Technology Letters with co-authors Rebecca Daly, Jenna Luek, Susan Welch, Kelly

Wrighton, and Paula Mouser. My contribution to this chapter was performing the metagenomic analyses, constructing the manuscript figures, and writing the manuscript.

These contributions unveiled the microbial potential for organohalide transformation in the HF system.

1.3.4 Microbial communities in a class (II) injection well disposal facility receiving produced fluids from the Marcellus Shale

This chapter employed 16S rRNA gene sequencing (V3-V4 region) to identify microbial communities in six samples from sequential stages of wastewater treatment in a class (II)

14 injection well disposal facility. The disposal facility was receiving fluids from producing wells in the Marcellus Shale. We hypothesized that microbial community structures in the fluid treatment process samples would differ from traditional HF O&G well produced fluid community structures due to the introduction of oxygen into the system and because several steps in the treatment process involved chemical and mechanical changes to the fluids. This chapter was written for submission to FEMS Microbiology Ecology with co- authors Jenna Luek and Paula Mouser. My contribution to this chapter was performing sampling at the injection well facility, extracting the DNA from samples for 16S rRNA analysis, assisting in bioinformatic analyses, designing the figures, and writing the manuscript. These contributions enabled a deeper understanding into the microbial ecology of injection well facilities receiving produced fluids from hydraulically fractured natural-gas wells.

1.4 References

Akob, D. M., Mumford, A. C., Orem, W., Engle, M. A., Klinges, J. G., Kent, D. B., et al. (2016). Wastewater disposal from unconventional oil and gas development degrades stream quality at a West Virginia injection facility. Environ Sci Technol 50, 5517–5525. doi:10.1021/acs.est.6b00428.

Akyon, B., McLaughlin, M., Hernández, F., Blotevogel, J., and Bibby, K. (2019). Characterization and biological removal of organic compounds from hydraulic fracturing produced water. Environmental Science: Processes & Impacts. doi:10.1039/C8EM00354H.

Arensdorf, J. J., Miner, K., Ertmoed, R., Clay, W. K., Stadnicki, P., and Voordouw, G. (2009). Mitigation of Reservoir Souring by Nitrate in a Produced-Water 15

Reinjection System in Alberta. in SPE-121731-MS (SPE: Society of Petroleum Engineers), 8. doi:10.2118/121731-MS.

Arthur, J. D., Bohm, B., and Layne, M. (2009). Hydraulic fracturing considerations for natural gas wells of the Marcellus Shale. Available at: http://archives.datapages.com.proxy.lib.ohio- state.edu/data/gcags_pdf/2009/arthetal.htm [Accessed August 2, 2017].

Barbot, E., Vidic, N. S., Gregory, K. B., and Vidic, R. D. (2013). Spatial and temporal correlation of water quality parameters of produced waters from Devonian-age shale following hydraulic fracturing. Environ. Sci. Technol. 47, 2562–2569. doi:10.1021/es304638h.

Booker A.E., Hoyt D.W., Meulia T., Eder E., Nicora C.D., Purvine S.O., et al. (2018) Deep subsurface pressure stimulates metabolic versatility in shale-colonizing Halanaerobium. In review. Appl Environ Microbiol.

Booker, A. E., Borton, M. A., Daly, R. A., Welch, S. A., Nicora, C. D., Hoyt, D. W., et al. (2017). Sulfide generation by dominant Halanaerobium microorganisms in hydraulically fractured shales. mSphere 2. doi:10.1128/mSphereDirect.00257-17.

Borton, M. A., Daly, R. A., O’Banion, B., Hoyt, D. W., Marcus, D. N., Welch, S., et al. (2018a). Comparative genomics and physiology of the genus Methanohalophilus, a prevalent methanogen in hydraulically fractured shale. Environmental Microbiology 20, 4596–4611. doi:10.1111/1462-2920.14467.

Borton, M. A., Hoyt, D. W., Roux, S., Daly, R. A., Welch, S. A., Nicora, C. D., et al. (2018b). Coupled laboratory and field investigations resolve microbial interactions that underpin persistence in hydraulically fractured shales. PNAS 115, E6585. doi:10.1073/pnas.1800155115.

Chapman, E. C., Capo, R. C., Stewart, B. W., Kirby, C. S., Hammack, R. W., Schroeder, K. T., et al. (2012). Geochemical and strontium isotope characterization of produced waters from Marcellus Shale natural gas extraction. Environ. Sci. Technol. 46, 3545–3553. doi:10.1021/es204005g.

Choudhary, L., Macdonald, D. D., and Alfantazi, A. (2015). Role of thiosulfate in the corrosion of steels: a review. CORROSION 71, 1147–1168. doi:10.5006/1709.

Cluff, M. A., Hartsock, A., MacRae, J. D., Carter, K., and Mouser, P. J. (2014). Temporal changes in microbial ecology and geochemistry in produced water from hydraulically fractured Marcellus Shale gas wells. Environ. Sci. Technol. 48, 6508–6517. doi:10.1021/es501173p.

Daly, R. A., Borton, M. A., Wilkins, M. J., Hoyt, D. W., Kountz, D. J., Wolfe, R. A., et al. (2016). Microbial metabolisms in a 2.5-km-deep ecosystem created by hydraulic fracturing in shales. Nature Microbiology 1, 16146. doi:10.1038/nmicrobiol.2016.146.

Ellsworth, W. L. (2013). Injection-induced earthquakes. Science 341, 1225942. doi:10.1126/science.1225942.

Elsner, M., and Hoelzer, K. (2016). Quantitative survey and structural classification of hydraulic fracturing chemicals reported in unconventional gas production. Environ. Sci. Technol. 50, 3290–3314. doi:10.1021/acs.est.5b02818.

Fetzner, S. (1998). Bacterial dehalogenation. Appl Microbiol Biotechnol 50, 633–657. doi:10.1007/s002530051346.

Fichter, J., Wunch, K., Moore, R., Summer, E., Braman, S., and Holmes, P. (2012). How hot is too hot for bacteria? A technical study assessing bacterial establishment in downhole drilling, fracturing And stimulation operations. in (NACE International). Available at: https://www.onepetro.org/conference-paper/NACE- 2012-1310 [Accessed January 9, 2018].

Getzinger, G. J., O’Connor, M. P., Hoelzer, K., Drollette, B. D., Karatum, O., Deshusses, M. A., et al. (2015). Natural gas residual fluids: sources, endpoints, and organic chemical composition after centralized waste treatment in pennsylvania. Environ. Sci. Technol. 49, 8347–8355. doi:10.1021/acs.est.5b00471.

Haluszczak, L. O., Rose, A. W., and Kump, L. R. (2013). Geochemical evaluation of flowback brine from Marcellus gas wells in Pennsylvania, USA. Applied Geochemistry 28, 55–61. doi:10.1016/j.apgeochem.2012.10.002.

Harkness, J. S., Dwyer, G. S., Warner, N. R., Parker, K. M., Mitch, W. A., and Vengosh, A. (2015). Iodide, bromide, and ammonium in hydraulic fracturing and oil and gas wastewaters: environmental implications. Environ. Sci. Technol. 49, 1955– 1963. doi:10.1021/es504654n.

Hayes, T. (2009). Sampling and analysis of water streams associated with the development of Marcellus Shale gas. Des Plaines, IL: Marcellus Shale Coalition Available at: http://www.water- research.net/naturalgasPA/pdffiles/MSCommission-Report.pdf.

Heyob, K. M., Blotevogel, J., Brooker, M., Evans, M. V., Lenhart, J. J., Wright, J., et al. (2017). Natural attenuation of nonionic surfactants used in hydraulic fracturing fluids: degradation rates, pathways, and mechanisms. Environ. Sci. Technol. 51, 13985–13994. doi:10.1021/acs.est.7b01539.

Hoelzer, K., Sumner, A. J., Karatum, O., Nelson, R. K., Drollette, B. D., O’Connor, M. P., et al. (2016). Indications of transformation products from hydraulic fracturing 18

additives in shale-gas wastewater. Environ. Sci. Technol. 50, 8036–8048. doi:10.1021/acs.est.6b00430.

Holliger, C., and Schumacher, W. (1994). Reductive dehalogenation as a respiratory process. Antonie van Leeuwenhoek 66, 239–246. doi:10.1007/BF00871642.

Huang, Y.-L., Li, Q.-B., Deng, X., Lu, Y.-H., Liao, X.-K., Hong, M.-Y., et al. (2005). Aerobic and anaerobic biodegradation of polyethylene glycols using sludge microbes. Process Biochemistry 40, 207–211. doi:10.1016/j.procbio.2003.12.004.

Huber, M., Meyer, U., and Rys, P. (2000). Biodegradation mechanisms of linear alcohol ethoxylates under anaerobic conditions. Environ. Sci. Technol. 34, 1737–1741. doi:10.1021/es9903680.

Kawai, F. (2002). Microbial degradation of polyethers. Appl Microbiol Biotechnol 58, 30–38. doi:10.1007/s00253-001-0850-2.

Kell, S. (2011). State oil and gas agency groundwater investigations and their role in advancing regulatory reforms, a two-state review: Ohio and Texas. Ground Water Protection Council Available at: http://fracfocus.org/sites/default/files/publications/state_oil__gas_agency_ground water_investigations_optimized.pdf.

Kerr, R. A. (2010). Natural gas From shale bursts onto the scene. Science 328, 1624– 1626. doi:10.1126/science.328.5986.1624.

Kim, W.-Y. (2013). Induced seismicity associated with fluid injection into a deep well in Youngstown, Ohio. J. Geophys. Res. Solid Earth 118, 3506–3518. doi:10.1002/jgrb.50247.

Kondash, A., and Vengosh, A. (2015). Water footprint of hydraulic fracturing. Environ. Sci. Technol. Lett. 2, 276–280. doi:10.1021/acs.estlett.5b00211.

Lester, Y., Ferrer, I., Thurman, E. M., Sitterley, K. A., Korak, J. A., Aiken, G., et al. (2015). Characterization of hydraulic fracturing flowback water in Colorado: implications for water treatment. Sci. Total Environ. 512–513, 637–644. doi:10.1016/j.scitotenv.2015.01.043.

Liang, R., Davidova, I. A., Marks, C. R., Stamps, B. W., Harriman, B. H., Stevenson, B. S., et al. (2016). Metabolic capability of a predominant Halanaerobium sp. in hydraulically fractured gas wells and its implication in pipeline corrosion. Front. Microbiol. 988. doi:10.3389/fmicb.2016.00988.

Lipus, D., Roy, D., Khan, E., Ross, D., Vikram, A., Gulliver, D., et al. (2018). Microbial communities in Bakken region produced water. FEMS Microbiol Lett 365. doi:10.1093/femsle/fny107.

Luek, J. L., and Gonsior, M. (2017). Organic compounds in hydraulic fracturing fluids and wastewaters: A review. Water Res. 123, 536–548. doi:10.1016/j.watres.2017.07.012.

Luek, J. L., Harir, M., Schmitt-Kopplin, P., Mouser, P. J., and Gonsior, M. (2018). Temporal dynamics of halogenated organic compounds in Marcellus Shale flowback. Water Research 136, 200–206. doi:10.1016/j.watres.2018.02.055.

Luek, J. L., Schmitt-Kopplin, P., Mouser, P. J., Petty, W. T., Richardson, S. D., and Gonsior, M. (2017). Halogenated organic compounds identified in hydraulic fracturing wastewaters using ultrahigh resolution mass spectrometry. Environ. Sci. Technol. 51, 5377–5385. doi:10.1021/acs.est.6b06213.

Maguire-Boyle, S. J., and Barron, A. R. (2014). Organic compounds in produced waters from shale gas wells. Environ. Sci.: Processes Impacts 16, 2237–2248. doi:10.1039/C4EM00376D.

Mohan, A. M., Hartsock, A., Bibby, K. J., Hammack, R. W., Vidic, R. D., and Gregory, K. B. (2013). Microbial community changes in hydraulic fracturing fluids and produced water from shale gas extraction. Environ. Sci. Technol. 47, 13141– 13150. doi:10.1021/es402928b.

Mohn, W. W., and Tiedje, J. M. (1992). Microbial reductive dehalogenation. Microbiology and Molecular Biology Reviews 56, 482–507.

Mouser, P. J., Borton, M., Darrah, T. H., Hartsock, A., and Wrighton, K. C. (2016). Hydraulic fracturing offers view of microbial life in the deep terrestrial subsurface. FEMS Microbiology Ecology 92, fiw166. doi:10.1093/femsec/fiw166.

Nell, M., and Helbling, D. E. (2018). Exploring matrix effects and quantifying organic additives in hydraulic fracturing associated fluids using liquid chromatography electrospray ionization mass spectrometry. Environmental Science: Processes & Impacts. doi:10.1039/C8EM00135A.

Rogers, J. D., Burke, T. L., Osborn, S. G., and Ryan, J. N. (2015). A Framework for Identifying Organic Compounds of Concern in Hydraulic Fracturing Fluids Based

on Their Mobility and Persistence in Groundwater. Environ. Sci. Technol. Lett. 2, 158–164. doi:10.1021/acs.estlett.5b00090.

Rogers, J. D., Thurman, E. M., Ferrer, I., Rosenblum, J., Evans, M. V., Mouser, P., et al. (2018). Degradation of polyethylene glycols and polypropylene glycols in microcosms simulating a spill of produced water in shallow groundwater. Environ. Sci.: Processes Impacts. doi:10.1039/C8EM00291F.

Rosenblum, J., Thurman, E. M., Ferrer, I., Aiken, G., and Linden, K. G. (2017). Organic chemical characterization and mass balance of a hydraulically fractured well: from fracturing fluid to produced water over 405 days. Environ. Sci. Technol. 51, 14006–14015. doi:10.1021/acs.est.7b03362.

Singer, E., Webb, E. A., Nelson, W. C., Heidelberg, J. F., Ivanova, N., Pati, A., et al. (2011). Genomic Potential of Marinobacter aquaeolei, a Biogeochemical “Opportunitroph.” Appl. Environ. Microbiol. 77, 2763–2771. doi:10.1128/AEM.01866-10.

Straß, A., and Schink, B. (1986). Fermentation of polyethylene glycol via acetaldehyde in Pelobacter venetianus. Appl Microbiol Biotechnol 25, 37–42. doi:10.1007/BF00252510.

Struchtemeyer, C. G., and Elshahed, M. S. (2012). Bacterial communities associated with hydraulic fracturing fluids in thermogenic natural gas wells in North Central Texas, USA. FEMS Microbiol. Ecol. 81, 13–25. doi:10.1111/j.1574- 6941.2011.01196.x.

Thacker, J. B., Carlton, D. D., Hildenbrand, Z. L., Kadjo, A. F., and Schug, K. A. (2015). Chemical analysis of wastewater from unconventional drilling operations. Water 7, 1568–1579. doi:10.3390/w7041568.

Thurman, E. M., Ferrer, I., Blotevogel, J., and Borch, T. (2014). Analysis of Hydraulic Fracturing Flowback and Produced Waters Using Accurate Mass: Identification of Ethoxylated Surfactants. Anal. Chem. 86, 9653–9661. doi:10.1021/ac502163k.

Tour, J. M., Kittrell, C., and Colvin, V. L. (2010). Green carbon as a bridge to renewable energy. Nature Materials. doi:10.1038/nmat2887.

U.S. Energy Information Administration (EIA) (2016a). Hydraulic fracturing accounts for about half of current U.S. crude oil production - Today in Energy -. Available at: https://www.eia.gov/todayinenergy/detail.php?id=25372 [Accessed October 7, 2017].

U.S. Energy Information Administration (EIA) (2016b). Hydraulically fractured wells provide two-thirds of U.S. natural gas production - Today in Energy -. Available at: https://www.eia.gov/todayinenergy/detail.php?id=26112 [Accessed October 16, 2018].

U.S. Energy Information Administration (EIA) (2018). Hydraulically fractured horizontal wells account for most new oil and natural gas wells - Today in Energy. Available at: https://www.eia.gov/todayinenergy/detail.php?id=34732 [Accessed October 16, 2018].

U.S. Environmental Protection Agency (2016). Hydraulic Fracturing for Oil and Gas: Impacts from the Hydraulic Fracturing Water Cycle on Drinking Water Resources in the United States (Final Report). (No. EPA/600/R-16/236F). Available at: https://cfpub.epa.gov/ncea/hfstudy/recordisplay.cfm?deid=332990 [Accessed January 18, 2019]. van Pée, K.-H., and Unversucht, S. (2003). Biological dehalogenation and halogenation reactions. Chemosphere 52, 299–312. doi:10.1016/S0045-6535(03)00204-2.

Veil, J. (2015). U.S. Produced water volumes and management practices in 2012. Oklahoma, City, OK: Ground Water Protection Council Available at: http://www.gwpc.org/sites/default/files/Produced%20Water%20Report%202014- GWPC_0.pdf.

Vidic, R. D., Brantley, S. L., Vandenbossche, J. M., Yoxtheimer, D., and Abad, J. D. (2013). Impact of shale gas development on regional water quality. Science 340, 1235009. doi:10.1126/science.1235009.

Wagener, S., and Schink, B. (1988). Fermentative degradation of nonionic surfactants and polyethylene glycol by enrichment cultures and by pure cultures of homoacetogenic and propionate-forming bacteria. Appl. Environ. Microbiol. 54, 561–565.

Warner, N. R., Christie, C. A., Jackson, R. B., and Vengosh, A. (2013). Impacts of shale gas wastewater disposal on water quality in Western Pennsylvania. Environ. Sci. Technol. 47, 11849–11857. doi:10.1021/es402165b.

Weingarten, M., Ge, S., Godt, J. W., Bekins, B. A., and Rubinstein, J. L. (2015). High- rate injection is associated with the increase in U.S. mid-continent seismicity. Science 348, 1336. doi:10.1126/science.aab1345.

Wirsen, C. O., Sievert, S. M., Cavanaugh, C. M., Molyneaux, S. J., Ahmad, A., Taylor, L. T., et al. (2002). Characterization of an Autotrophic Sulfide-Oxidizing Marine Arcobacter sp. That Produces Filamentous Sulfur. Appl. Environ. Microbiol. 68, 316–325. doi:10.1128/AEM.68.1.316-325.2002.

Wuchter, C., Banning, E., Mincer, T., Drenzek, N. J., and Coolen, M. J. (2013). Microbial diversity and methanogenic activity of Antrim Shale formation waters from recently fractured wells. Front. Microbiol. 4. doi:10.3389/fmicb.2013.00367.

Chapter 2. Members of Marinobacter and Arcobacter influence biogeochemistry during early production of hydraulically fractured shale gas wells in the Appalachian Basin

Published in Frontiers in Microbiology, 2018

Abstract

Hydraulic fracturing is the prevailing method for enhancing recovery of hydrocarbon resources from unconventional shale formations, yet little is understood regarding the microbial impact on biogeochemical cycling in natural-gas wells. Although the metabolisms of certain fermentative bacteria and methanogenic archaea that dominate in later produced fluids have been well studied, few details have been reported on microorganisms prevalent during the early flowback period, when oxygen and other surface-derived oxyanions and nutrients become depleted. Here, we report the isolation, genomic and phenotypic characterization of Marinobacter and Arcobacter bacterial species from natural-gas wells in the Utica-Point Pleasant and Marcellus Formations coupled to supporting geochemical and metagenomic analyses of produced fluid samples.

These unconventional hydrocarbon system-derived Marinobacter sp. are capable of utilizing a diversity of organic carbon sources including aliphatic and aromatic hydrocarbons, amino acids, and carboxylic acids. Marinobacter and Arcobacter can metabolize organic nitrogen sources and have the capacity for denitrification and dissimilatory nitrate reduction to ammonia (DNRA) respectively; with DNRA and ammonification processes partially explaining high concentrations of ammonia measured in produced fluids. Arcobacter is capable of chemosynthetic sulfur oxidation, which could

25 fuel metabolic processes for other heterotrophic, fermentative, or sulfate-reducing community members. Our analysis revealed mechanisms for growth of these taxa across a broad range of salinities (up to 15% salt), which explains their enrichment during early natural-gas production. These results demonstrate the prevalence of Marinobacter and

Arcobacter during a key maturation phase of hydraulically fractured natural-gas wells, and highlight the significant role these genera play in biogeochemical cycling for this economically important energy system.

2.1 Introduction

Hydraulic fracturing has enhanced the extraction of oil and gas from previously inaccessible unconventional petroleum resources (e.g., shales, tight sands, coalbeds), with hydraulically fractured wells accounting for the production of 67% of the natural gas and

51% of the oil in the U.S. in 2015 (EIA 2016, Kerr 2010, Tour et al. 2010). Hydraulic fracturing of unconventional systems is accomplished through the high-pressure injection of water (~90%), proppant (~9%), and chemical additives (~1%) into subsurface formations (Vidic et al. 2013, Arthur et al. 2009). This process generates and sustains a network of fractures in the intrinsically low-permeability formation, liberating hydrocarbons which return through the wellbore to be collected at the surface (Kerr 2010,

Tour et al. 2010, Arthur et al. 2009). Injected fluids return to the surface over several months as they reach equilibrium with deep formation brines and undergo water-rock interactions (Vengosh et al. 2017). The fluids that return to the surface are referred to as flowback fluids (generally the first few weeks after hydraulic fracturing occurs) and 26 produced fluids (several weeks to years after hydraulic fracturing occurs and throughout the production life cycle). Flowback fluids are characterized by high carbon concentrations

(>100 mg/L), moderate to high salinities (1-8 NaCl%), and increasing concentrations of soluble nitrogen and sulfur species (Haluszczak et al. 2013, Harkness et al. 2015, Warner et al. 2013, Cluff et al. 2014, Barbot et al. 2013, Chapman et al. 2012, Lester et al. 2015).

Produced fluids, in contrast, more closely resemble deep formation brines (>10% NaCl).

Produced fluids are extremely high in total dissolved solids, have lower carbon content (50 mg/L and below), and contain elevated concentrations of reduced sulfur species (Vengosh et al. 2017, Haluszczak et al. 2013, Harkness et al. 2015, Warner et al. 2013, Cluff et al.

2014, Barbot et al. 2013, Chapman et al. 2012, Lester et al. 2015, Booker et al. 2017).

Despite dramatic changes in solution chemistry during the gas extraction process, microbial life has been detected in flowback and produced fluids, and likely impacts fluid chemistry changes, infrastructure stability, and hydrocarbon recovery efficiency (Cluff et al. 2014, Daly et al. 2016, Booker et al. 2017, Liang et al. 2016, Nixon et al. 2017, Lipus et al. 2018, Struchtemeyer and Elshahed 2012, Gaspar et al. 2014).

Microbial growth in oil and gas wells can contribute to reservoir souring, well corrosion, and clogged pore space, decreasing hydrocarbon productivity (Fitcher et al. 2012, Gaspar et al. 2014). Although biocides are commonly added to control microbial growth, evidence of a dynamic bacterial and archaeal community has been observed in fluids collected during and after hydraulic fracturing from multiple unconventional systems in the U.S.

(Cluff et al. 2014, Daly et al. 2016, Struchtemeyer and Elshahed 2012, Mohan et al. 2013,

Fitcher et al. 2012, Lipus et al. 2017, Lipus et al. 2018, Borton et al. 2018). These studies suggest that microorganisms are injected from surface input fluids and persist despite extreme environmental conditions such as high salinities, limited access to terminal electron acceptors (e.g., oxygen, nitrate, sulfate), high pressures, and warmer temperatures than near-surface conditions (Mouser et al. 2016). Despite these ecosystem stressors, a low-diversity microbial community persists and plays a significant role in the biogeochemical cycles of unconventional systems, including low porosity shales and limestones (Fitcher et al. 2012, Cluff et al. 2014, Daly et al. 2016, Booker et al. 2017, Liang et al. 2016, Nixon et al. 2017, Lipus et al. 2018, Borton et al. 2018).

Prior studies from our group revealed that Marcellus Shale-associated taxa

Halanaerobium, Marinobacter, and Frackibacter may participate in carbon and nitrogen cycling through fermentation/transformation of both injected chemicals and microbially- produced osmoprotectants (Daly et al. 2016). Laboratory studies have suggested that

Halanaerobium may also contribute to sulfide-induced well-souring and corrosion months to years after fracturing has occurred (Liang et al. 2016, Lipus et al. 2018, Nixon et al.

2017, Booker et al. 2017). The aforementioned studies have focused on taxa that dominate in later production phases characterized by elevated salinity and depleted organic carbon.

Still, relatively little has been reported on the physiological and biogeochemical roles of dominant taxa within the first few months of production, a time period that may be critical for the establishment of later persisting taxa. During this natural-gas well maturation transition period, organic carbon is abundant (e.g indigenous hydrocarbons and injected

28 carbon), salinity is moderate in comparison to later time points (1-10% versus 10-30% in later produced fluids), and terminal electron acceptors (e.g oxygen, nitrate, sulfate) remain available, allowing for a diversity of microbially-mediated redox reactions and enzymatic transformations not thermodynamically possible during later stages of production (Mouser et al. 2016).

Two genera, Marinobacter and Arcobacter, have been detected in produced fluids from numerous unconventional hydrocarbon-producing formations including the Barnett,

Antrim, Haynesville, Utica, Point Pleasant, and Marcellus (Mouser et al. 2016, Davis et al.

2012, Cluff et al. 2014, Daly et al. 2016, Wutcher et al. 2013, Fitcher et al. 2012, Mohan et al. 2013, Struchtemeyer and Elshahed 2012, Liang et al. 2016). Although one study applied a metagenomics approach to infer Marinobacter sp. may be capable of utilizing organic nitrogen and hydrocarbon compounds (Daly et al. 2016), the metabolic capabilities of these taxa in unconventional systems are poorly defined. Here, we used laboratory experiments to investigate salinity tolerance and carbon substrate utilization of

Marinobacter and Arcobacter strains cultivated from natural-gas well produced fluids. We complement these laboratory investigations with genomic, metagenomic, and geochemical data from hydraulically stimulated natural-gas wells, demonstrating the power of using both cultivation and cultivation-independent approaches for elucidating sources and use of carbon and energy for halotolerant microorganisms from a deep terrestrial ecosystem.

2.2 Materials and Methods

2.2.1 Methods Overview

Fluids were collected from six different hydraulically fractured natural-gas wells for geochemical and genomic analyses summarized in this manuscript: four natural-gas wells producing from the Utica-Pt. Pleasant Formation (Utica-3, Utica-4, Utica-5, Utica-6), and two natural-gas wells producing from the Marcellus Shale (Marcellus-4 and Marcellus-5)

(Table 2.1). Nitrogen and sulfur analyses were performed on all six wells while metagenomic analyses, near-full-length reconstruction of the 16S rRNA gene, NPOC, and CO2 analyses were performed on one well (Marcellus-4). Marinobacter sp. UTICA-

S1B6 was isolated from Utica-3, whereas Arcobacter sp. UTICA-S4D1 was isolated from Utica-6, and Arcobacter sp. MARC-MIP3H16 was isolated from Marcellus-4.

Additionally, seven 16S rRNA gene sequences reported in a previous study (Marcellus-1,

Daly et al. 2016) were used in our phylogenetic analysis.

2.2.2 Sample Collection

Produced fluid samples were collected from six hydraulically fractured natural-gas wells in the northern Appalachian Basin: four from the Utica-Point Pleasant Formation (Utica-

3, Utica-4, Utica-5, Utica-6) and two from the Marcellus Shale (Marcellus-4 and

Marcellus-5). Liquid samples were taken from the gas-water separator. Samples were collected from the start of flowback at various intervals over a year into production (496 days) as previously reported (Luek et al. 2018). Samples for bacterial enumeration were 30 collected in 50 mL polypropylene conical vials containing 5.0 mL of 25% paraformaldehyde and stored at 4°C until enumeration. Samples for isolation and culturing

(25 mL) were collected in sterile serum bottles with no headspace and sealed with butyl rubber septa to maintain anaerobic conditions, then upon arrival at the laboratory purged with N2 gas and stored at ambient temperature in the dark until inoculation (Figure 2.1).

Samples for geochemical analysis were collected in 1 L sterile containers (either HDPE or glass) with no headspace and stored at 4°C until analysis. Sulfide measurements were conducted on unfiltered samples (Hach method 8131, see following section). Samples for total dissolved nitrogen, ammonia, and sulfate were filtered using 0.45 µm PES filters

(EMD Millipore, Burlington, MA), with the exception of the Marcellus-4 and Marcellus-5 wells in which 0.22 µm PES filters were used, within 48 hours of sampling and stored at

4°C until analysis.

2.2.3 Non-purgeable Organic Carbon, Nitrogen, Sulfur, and Chloride Analysis

Non-purgeable organic carbon (NPOC) was measured in the Marcellus-4 samples by a

TOC/TN analyzer equipped with autosampler (TOC-V CSN/TNM-1/ASI-V, Shimadzu,

Kyoto, Japan). Total dissolved nitrogen (TDN) was analyzed in samples from all wells by a Shimadzu TOC/TN analyzer (see above), employing appropriate dilutions as needed for samples. Ammonia (NH3) in the produced fluid samples was analyzed colorimetrically using the modified Berthelot reaction method on a Skalar San++ continuous flow nutrient analyzer. Reagents were prepared following the method supplied by the manufacturer.

Sulfide was measured using a Hach DR 900, method 8131. Sulfate and chloride were 31 analyzed using a Thermo-Scientific Dionex ICS-2100 ion chromatograph (Waltham, MA); samples were diluted by a factor of 100–1000 due to the high salinity.

2.2.4 Carbon Dioxide Analysis

Produced gas samples were collected from the Marcellus-4 natural-gas well over a period of 140 days. Produced gas samples were collected with negligible air contamination using thick-walled 0.95 cm (3/8 inch) outside diameter and 40.6 cm (16 inch) long refrigeration- grade copper tubes. The samples were collected by pumping produced fluid that had been collected using HDPE carboys from the producing natural-gas well through the copper tubes, with tubes tapped to remove air bubbles prior to collection. The copper tubes were then sealed using brass refrigeration clamps with a 0.762 mm (0.030 inch) gap (Harkness et al. 2017, Kang et al. 2016). Gas samples in the copper tubes were prepared for analysis by cold welding ~2.5 cm (~1 inch) splits of the copper tubing using stainless steel clamps.

The copper tube was then attached to an ultra-high vacuum steel line (total pressure= 1-3 x 10-9 torr), which is monitored continuously using a 0-20 torr MKS capacitance monometer (accurate to the nearest thousandths) and an isolated ion gauge, using a 0.64 cm (1/4 inch) VCR connection and expanded to obtain aliquots for various gas geochemical analyses.

Carbon dioxide concentrations (CO2) were measured on a SRS Quadrupole MS and an SRI

8610C Multi-Gas 3+ gas chromatograph (GC) equipped with a flame ionization detector

(FID) and thermal conductivity detector (TCD) at the Ohio State University Noble Gas 32

Laboratory (Darrah et al. 2015, Moore et al. 2018, Heilweil et al. 2016). All reported CO2 concentrations were above the method detection limit. The average external precision was determined by measuring a series of synthetic natural gas standards obtained from Praxair and the DCG Partnership. Standard analytical errors for CO2 was less than ± 1.06% based on daily replicate measurements during analyses.

2.2.5 Microbial Enumeration

Microbial (bacterial and archaeal) enumeration was performed using filtration and epifluorescent microscopy as described previously (Diemer et al. 2013, Brussaard 2004,

Chen et al. 2001, Noble and Fuhrman 1998, Patel et al. 2007, Weinbauer et al. 1998). For each sample, the volume filtered was optimized such that at least two filters were prepared to yield approximately 10 to 200 cells per counting field. Between 1 to 15 mL of fixed sample was filtered through a 25 mm dia. 0.2 µm pore black PCTE filter (Sterlitech, Kent,

WA), stained with 2X SYBR Gold (Life Technologies, Carlsbad, CA) in TE buffer, then mounted on a microscope slide with SlowFade reagent (Life Technologies, Carlsbad, CA).

The slide was viewed with a Labomed Lx500 epifluorescent microscope through a 40X air objective under 480 nm excitation. For each sample twenty randomly selected fields were counted per filter.

2.2.6 Bacterial Isolation

Marinobacter sp. UTICA-S1B6 was recovered from produced fluid collected from a

Utica-Point Pleasant natural-gas well (Utica-3) on the first day of flowback using Difco

Marine Broth 2216 (DM2216) medium supplemented with 40 mM nitrate at 30°C

(Figure 2.1, Tummings et al. 2018). Arcobacter strains UTICA-S4D1 and MARC-

MIP3H16 were isolated from produced fluids collected at 159 days (from the Utica-Point

Pleasant formation, Utica-6) and 93 days (from Marcellus Shale, Marcellus-4) after hydraulic fracturing began, respectively (Figure 2.1). Arcobacter strains were grown at

30°C in the dark on DM2216 at 5% NaCl purged with 80:20 N2/CO2. We tested growth under anaerobic and aerobic conditions; because cells grew better in the presence of oxygen, they were maintained under aerobic conditions for growth experiments (Figure

2.1, Panescu et al. 2018). The purity of isolates was verified by 16S rRNA sequencing

(see Isolation and Sequencing of DNA section).

2.2.7 Salinity Growth Curves

Two of the three isolates, Marinobacter sp. UTICA-S1B6 and Arcobacter sp. MARC-

MIP3H16, were tested for salinity tolerance (Figure 2.1). Salinity experiments for both strains were conducted aerobically at 30°C in the dark with shaking. In addition, an anaerobic experiment was also carried out for Marinobacter at 37°C in the dark. For

Arcobacter, a stock of Trypticase Soy Broth (Cat. No. 1.05459, EMD Millipore

Corporation, Billerica, MA) was amended with 1% DL Trace Mineral Solution and NaCl

34 at the following concentrations: 0.5%, 2%, 4%, 6%, 8%, 10%, 12% and 14%. For

Marinobacter, bacterial growth medium consisting of 25% DM2216 and 75% Lennox

Broth (Bonis and Gralnick 2015) with 1% NaCl, was prepared from commercially available powdered mixes. To these media, NaCl was added to obtain the following final concentrations: 0.9%, 2.5%, 5%, 7.5%, 10%, 12.5%, 15%, and 20%. Singlet cultures were grown to mid-log phase, 2% was transferred into fresh media, then 2% was transferred again into triplicate cultures from which data were used for the salinity curves. The media was aliquoted into 9 mL increments in borosilicate glass tubes suitable for anaerobic culturing of size 18x150 mm purchased from Bellco Glass (Vineland, NJ) and autoclaved at 121°C for 20 min; the tubes were used for both the aerobic and anaerobic cultures. Aerobic media tubes were covered with silicone foam stoppers while the anaerobic media tubes were closed with butyl rubber stoppers clamped with aluminum seals. Absorbance at a wavelength of 610 nm was measured with a Hach DR

900 Colorimeter (Loveland, Colorado) fitted with a custom adapter.

A standard curve relating optical density to cell numbers was constructed. Cultures were grown to late log phase in the same test media at 2.5% NaCl, fixed (2.5% paraformaldehyde final concentration) and diluted or concentrated to obtain absorbance values of <0.1, ~0.5 and ~1.0 at 610 nm. Two aliquots from each concentration were diluted in 2.5% NaCl to yield 10 to 200 cells per counting field, and counted exactly as described above in the section, Microbial Enumeration. Growth rates were calculated using a first-order model based on cell numbers and time elapsed.

2.2.8 Utilization of Carbon Substrates

Culture plates containing DM2216 with 2% agar (Affymetrix USB, Tewksbury, MA) were streaked with approx. 10 μL of cryogenic glycerol stock containing bacterial isolates and incubated at 30°C for two to three days. Individual 1-2 mm diameter colonies were picked, resuspended in liquid DM2216, re-streaked and re-grown as described above. Second generation colonies were resuspended and inoculated in Biolog PM1 (Hayward, CA) inoculating fluid augmented with 5% NaCl solution, 20 μM pyruvate (carbon source) and

0.12X Marinobacter Iron Medium (MIM) (Bonis and Gralnick 2015, AffymetrixUSB,

Tewksbury, MA) (Figure 2.1). The cultures were grown to late log phase (two-three days) where 2% was transferred into fresh PM1-MIM-pyruvate medium. Once the second generation PM1-MIM-pyruvate culture reached the desired density, the cells were pelleted for 10 minutes at 5,000 x g, washed twice in 5% NaCl, and resuspended in 500 µL PM1-

MIM with no carbon source. The washed culture was quantified on a NanoDrop 2000 spectrophotometer (Thermo Scientific) at 600 nm in an Eppendorf Uvette fitted with an adapter (50 µL of culture used), and diluted to 0.1 Absorbance units with PM1-MIM, of which 880 µL was combined with an additional 12 mL PM1-MIM inoculating fluid and distributed to a new PM1 substrate plate at 110 µL per well. All carbon substrate assessments were performed under aerobic conditions. The plate was sealed with a sterile, optically clear, non-cytotoxic, gas-permeable adhesive film (4titude, Wotton, England).

Optical densities were measured using a plate reader (BioTek Synergy HTX Multi-Mode

Reader, Winooski, VT) at 600 nm every 2 hours for 3 days while incubating at 30 °C. 36

Normalization took place by subtracting the optical densities of a control plate containing media with no cells.

2.2.9 Isolation and Sequencing of DNA

Marinobacter and Arcobacter genomic DNA was extracted using Qiagen DNA extraction kits (Qiagen DNA minikit and Qiagen DNeasy kit, respectively) (Hilden, Germany) as previously described (Tummings et al. 2018, Panescu et al. 2018). The 16S rRNA gene was amplified with universal primer sets for bacteria (27F, 1492R) and archaea (4Fa,

1492R). The 16S rRNA gene was sequenced using the Sanger method on a 3730 DNA

Analyzer (Applied Biosystems) for quality control at the Comprehensive Cancer Center

Genomics Shared Resource at the Ohio State University. After passing quality control, the isolates were subjected to full genome sequencing using an Illumina MiSeq at the Joint

Genome Institute (JGI), Walnut Creek, CA, USA. Genome assemblies were constructed using Spades v. 3.6.2 with annotation performed by the JGI in the Integrated Microbial

Genomes platform v. 4.12.1 (Figure 2.1, Tummings et al. 2018, Panescu et al. 2018,

Huntemann et al. 2016).

2.2.10 Phylogenetic analysis

Maximum-likelihood (ML) phylogenetic trees were constructed in RAxML using the full length 16S rRNA gene in Geneious v. 8 using a MUSCLE alignment with a maximum of

10,000 iterations (v.7.2.8, nucleotide model GTR gamma, rapid bootstrapping and search

37 for best-scoring ML tree algorithm, with 1,000 bootstrap replicates). Nearly full-length 16S rRNA EMIRGE sequences used in the trees were reconstructed from related shale fluid metagenome samples as described previously (Miller et al. 2011, Daly et al. 2016).

2.2.11 Scanning electron microscopy (SEM) sample preparation and imaging

Cells from cultures grown to log phase were fixed by combining 178 µL culture with 22

µL 25% PFA and allowed to equilibrate for 1-2 hours in a microcentrifuge tube. Cells were next pelleted at 5,000 x g for 10 min, supernatant was removed, and cell pellet was washed with 1.0 mL PBS and resuspended in 200 µL PBS. For each isolate, ~100 µL of the resuspended culture was placed on a 0.2 µm Nucleopore polycarbonate filter (Whatman,

GE Healthcare, Chicago, IL) mounted on a 25 mm diameter Swinnex syringe filter holder

(EMD Millipore, Darmstadt, Germany) connected to flexible tubing. The culture was embedded into the filter by applying gentle vacuum (manually with a disposable syringe or by using an electric vacuum pump). The filter was next rinsed with 300-500 µL of PBS at 2% NaCl followed by additional vacuum application. After the culture was embedded onto the filter and all excess liquid was removed, the filter was allowed to dry. Dried samples were subjected to successive 5 minute rinses of increasing concentrations of 35-

100% ethanol. After evaporation, the filter was covered in hexamethyldisilazane, covered with a petri dish lid and allowed to dry overnight. The filters were subsequently mounted with carbon tape on aluminum stubs and imaged with a FEI Quanta FEG 250 scanning electron microscope at the Subsurface Energy Materials Characterization and Analysis

Laboratory (SEMCAL), School of Earth Sciences, The Ohio State University (Figure 2.1). 38

2.2.12 Metagenomic sequencing, assembly, and annotation

Metagenomic sequencing, assembly, and annotation was performed as previously described (Borton et al. 2018). In short, produced fluid samples collected at 7 time points

(drilling mud sample prior to fracturing, days after fracturing 25, 34, 35, 79, 93, 142) from

Marcellus Shale natural-gas well Marcellus-4 (300 to 1000 ml) were concentrated on a

0.22-µm PES filter (Nalgene; Fisher Scientific). Total nucleic acids were extracted using a modified phenol-chloroform extraction method. Briefly, libraries were created and quantified using an Illumina Library creation kit (KAPA Biosystems) with solid-phase reversible 402 immobilization size selection. Libraries were then sequenced on the

Illumina HiSeq 2500 sequencing platform utilizing a TruSeq Rapid paired-end 404 cluster kit. Fastq files were generated with CASSAVA 1.8.2. Illumina sequences from each sample were first trimmed from both the 5′ and 3′ ends with Sickle and then each sample was assembled individually with IDBA-UD (Brown et al. 2015, Wrighton et al. 2012) using default parameters. Scaffolds were annotated as described previously by predicting open reading frames with MetaProdigal (Hyatt et al. 2012). Sequences were compared with

USEARCH (Edgar et al. 2010) to KEGG, UniRef90, and InterProScan (Quevillon et al.

2005) with single and reverse best hit (RBH) matches of >60 bases reported.

2.2.13 Genomic and Metagenomic Analyses

Inquiry into genes contained in isolate genomes was performed using IMG/MER from JGI

(Figure 2.1, Chen et al. 2017). Genes in isolate genomes were pulled from IMG/MER using

EC numbers and KEGG Orthology numbers after genomic annotation following the JGI genome annotation pipeline (Huntemann et al. 2016). Homologous genes in assembled metagenomes were mined by KEGG orthology number, EC number, and name with a cutoff of >200 bit score and >45.9% identity similarity to genus Marinobacter/Arcobacter at day 34 (after fracturing) for Marinobacter and Arcobacter and days 34 and 93 for

Arcobacter (Booker et al. 2016). Arcobacter and Marinobacter normalized relative abundances through time (days 25, 34, 35, 79, 93, 142) were tracked using the 16S rRNA gene reconstructed from the metagenomic data using EMIRGE (Miller et al. 2011, Daly et al. 2016).

2.2.14 Data Accession

Whole genome sequences for Arcobacter sp. MARC-MIP3H16 and UTICA-S4D1 were deposited in NCBI under accession numbers PTIW01000000 and FUYO01000000 and at the JGI IMG/M database under genome IDs 2700989666 and 2703719342, respectively

(Pansecu et al. 2018). Whole genome sequences for Marinobacter sp. UTICA-S1B3,

UTICA-S1B6, UTICA-S1B9 are available in NCBI under accession numbers

PTIV00000000, PTIT00000000, and PTIU00000000, respectively, and at JGI IMG/M database under genome IDs 2700989663, 2700989662, and 2700989665 (Tummings et al.

2018). All metagenomic sequencing data has been deposited under Bioproject

PRJNA308326.

2.3 Results

2.3.1 Marinobacter and Arcobacter persist in unconventional source ecosystem for weeks to months after hydraulic fracturing

Using near-full-length 16S rRNA gene sequences reconstructed from 7 sample metagenomes using EMIRGE, we tracked the relative abundance of two key taxa,

Marinobacter and Arcobacter, during the first 140 days after hydraulic fracturing in a

Marcellus Shale natural-gas well (Marcellus-4). Marinobacter were detected in drill muds (6.3%) used during natural-gas well development, and produced fluid samples collected in the first four weeks after hydraulic fracturing, reaching up to 20% relative abundance on day 34, but decreased thereafter to levels below 0.5% (Figure 2.2a). We compared a representative 16S rRNA gene sequence reconstructed from metagenomes in the Marcellus natural-gas well (Marcellus-4) using EMIRGE (Day 38 EMIRGE

Marinobacter sp, Figure 2.3a) to (i) near-full-length 16S rRNA gene sequences reconstructed from metagenomes in another previously studied Marcellus Shale natural- gas well (Marcellus-1) (Figure 2.3a, Daly et al. 2016), (ii) 16S rRNA gene sequences from genomes available for Marinobacter strains isolated from the Utica-Pt. Pleasant

Formation (Utica-3) (Day 0 after flowback began), and (iii) 16S rRNA gene sequences acquired from NCBI for Marinobacter strains isolated from other marine environments.

Representative Marinobacter sampled from Marcellus-4 were highly similar to the three isolates recovered from the Utica-Pt. Pleasant Formation (97-99% shared nucleotide identity) (UTICA-S1B3; UTICA-S1B6; UTICA-S1B9) and cluster closely to 16S rRNA gene sequences collected from another Marcellus Shale natural-gas well (Marcellus-1,

Figure 2.3a).

Based on visual analysis of SEM micrographs, Marinobacter sp. UTICA-S1B6 is rod- shaped and approximately 2-4 µm long by 0.5 µm thick (Figure 2.3a). The closest related strain to M. UTICA-S1B6, Marinobacter persicus, is a strict aerobe isolated from a hypersaline lake and capable of tolerating salinities up to 20% NaCl (Bagheri et al. 2013).

Other closely related strains M. aquaeolei and M. hydrocarbonoclasticus were discovered in environments with high concentrations of hydrocarbons (oil well, marine petroleum spill, respectively) (Huu et al. 1999, Gauthier et al. 1992)

Arcobacter spp. generally persist in the first weeks to months after hydraulic fracturing and remain abundant for a longer time period than Marinobacter before replacement by other abundant taxa commonly identified in produced fluids including Halanaerobium and Methanohalophilus as natural gas and produced fluid production continues and the natural-gas well matures (Cluff et al. 2014, Mohan et al. 2013, Davis et al. 2012, Wutcher et al. 2013, Liang et al. 2016). Although we did not detect Arcobacter in drill muds, it comprised over 50% of the microbial community in produced fluids from the Marcellus-4 natural-gas well collected on day 34, and subsequently ranged in relative abundance from

1% to 44% until day 142 (Figure 2.2a). In a similar fashion to Marinobacter discussed above, we compared these Arcobacter 16S rRNA sequences reconstructed from the

Marcellus-4 natural-gas well metagenomes using EMIRGE (Days 94, 98 EMIRGE

Arcobacter sp.) to (i) near full-length 16S rRNA sequences reconstructed from metagenomes in another previously studied Marcellus Shale natural-gas well (Marcellus-

1) (Days 0, 7, Figure 2.3b, Daly et al. 2016), (ii) 16S rRNA gene sequences for two

Arcobacter strains isolated from the Utica-Point Pleasant Formation (Utica-6) (Day 159 after hydraulic fracturing occurred) and Marcellus Shale natural-gas well Marcellus-4

(Day 93 after hydraulic fracturing), and (iii) 16S rRNA gene sequences acquired from

NCBI for Arcobacter strains isolated from other marine environments. Representative

Arcobacter sequences recovered from both Marcellus-4 (Day 94 Marcellus-4 and Day 98

Marcellus-4) and Marcellus-1 (T0_EMIRGE36 and T7_EMIRGE21) were very similar to

Arcobacter spp. MARC-MIP3H16 and UTICA-S4D1 isolates. In particular, sequence from an initial flowback sample (Day 0 Marcellus-1) and an uncultured clone were most closely related (99.6%) to laboratory cultured strains (Figure 2.3b).

Arcobacter sp. MARC-MIP3H16 is a curved rod approximately 1-3 µm long by 0.3-0.5

µm thick with a polar flagellum (not shown) (Figure 2.3b). The closest related strains to

A. MARC-MIP3H16, Arcobacter sp. solar lake and Arcobacter marinus, were both isolated from saline aquatic environments (Donachie et al. 2005, Kim et al. 2010).

Arcobacter marinus can grow in 3-5% NaCl under both aerobic and microaerophilic conditions. Another closely related strain, Arcobacter halophilus, isolated from a

43 hypersaline lagoon, is capable of growth across aerobic to anaerobic conditions, provided media contains at least 2% NaCl (Donachie et al. 2005). The close clustering of isolate and sample DNA despite lithogical and geographical differences demonstrates the prevalence of both Marinobacter and Arcobacter in early produced fluid from hydraulically fractured natural-gas wells.

2.3.2 Shale Marinobacter and Arcobacter are slight to moderate halophiles

The highest relative abundance of 16S rRNA gene copy numbers detected for

Marinobacter and Arcobacter was associated with lower salt concentrations in the

Marcellus-4 natural-gas well (approximately 40 g/L Cl- or 4% Cl-, Figure 2.2a). Therefore, to better understand the adaptability of Marinobacter and Arcobacter to increasing salinity levels, we tested the tolerance of both isolates across a broad range of salinities under aerobic (Arcobacter, Marinobacter) and anaerobic (Marinobacter) conditions (Figure

2.4a). Marinobacter sp. UTICA-S1B6 grew in concentrations between 0.9% and 15% Cl-, with optimal growth at 10% Cl- when oxygen was present (Figure 2.4a). In the absence of oxygen, Marinobacter growth rate improved at a lower salinity (5% Cl-). Arcobacter sp.

MARC-MIP3H16 grew in salt concentrations between 2% and 12%, with an optimal growth rate at 4% Cl- (0.6 hr-1), considerably higher than growth at all other salinities

(<0.25 hr-1) (Figure 2.4a). Our laboratory findings were therefore consistent with field salinity levels when Marinobacter and Arcobacter were highest in relative abundance during the first 140 days of production in the Marcellus-4 natural-gas well (4%-8%).

However, Arcobacter growth rate was significantly higher than Marinobacter in this range, 44 which may partially explain its dominance and persistence during this transitional period in Marcellus-4 natural gas and produced fluid well maturation.

To assess possible osmoadaptation mechanisms, we searched the isolate genomes for relevant genes associated with osmoprotectant synthesis and transport, and the movement of ions across the cell membrane (Daly et al. 2016). Genomic evidence of the isolated culture indicated that Marinobacter. sp. UTICA-S1B6 has a glycine betaine transporter for osmolyte uptake as well as the dehydrogenases responsible for conversion of choline into glycine betaine under aerobic conditions for de novo osmoprotectant synthesis. We also discovered several genes responsible for the synthesis of ectoine, L-ectoine synthase (ectC) and ectoine hydroxlase (ectD), and anti-porter genes (H+, Na+, K+, Cl-) responsible for the osmotic regulation within the cell (Figure 2.5, Oren et al. 2008). Arcobacter spp. UTICA-

S4D1 and MARC-MIP3H16 both have the putative ability to synthesize ectoine via ectD, but not ectC. These Arcobacter strains do not have the functional potential for synthesis of osmolytes other than ectoine, but instead contain anti-porters (H+, Na+, K+, Cl-) that may be used for adaptation to changing concentrations of NaCl via the "salt-in" strategy, as well as transporters for glycine betaine and polar amino acids for uptake of extracellular osmolytes (Figure 2.5, Oren et al. 2008, Daly et al. 2016). Neither Marinobacter sp.

UTICA-S1B6 nor Arcobacter spp. UTICA-S4D1 and MARC-MIP3H16 have genes responsible for uptake or synthesis of osmolytes sorbitol/mannitol, or trehalose.

2.3.3 Marinobacter and Arcobacter utilize a wide variety of carbon sources

Marinobacter sp. UTICA-S1B6 can utilize a variety of organic compounds for carbon and energy sources, as evidenced by isolate genomic potential and laboratory physiological assessments. We detected genes for the Entner-Doudoroff pathway for conversion of glucose to pyruvate in M. UTICA-S1B6, however, this pathway was not complete as it lacked all genes responsible for phosphorylation of glucose (EC 3.1.3.9, 2.7.1.1, 2.7.2.2, and 2.7.1.199). Consistent with our genomic inferences, we confirmed the isolate’s inability to grow on glucose under oxic laboratory conditions (see Figure A.1 in Appendix

A). We detected multiple amino acid ABC transporters and di/tricarboxylate transporters in the Marinobacter isolate genome, providing evidence that amino acid uptake is important for this strain. In complementary laboratory experiments, Marinobacter sp.

UTICA-S1B6 grew on two amino acids (L-glutamate, L-proline) and several carboxylic acids (OD>0.1) as the sole carbon source, with citrate, L-malate, and acetate as three of the more optimally utilized compounds (Figure 2.4b). In addition, genes responsible for breakdown and utilization of starch, glycogen, amylose, fructose, and sucrose are present in the isolate genome (Figure 2.5a); however, M. UTICA-S1B6 did not grow on sucrose or fructose as a sole carbon source in the laboratory (see Figure A.1 in Appendix A).

Our genomic investigation also showed that Marinobacter sp. UTICA-S1B6 has the potential to degrade aliphatic and aromatic hydrocarbons. The strain contains the alkane-

1-monooxygenase gene (alkB) responsible for activating alkanes to a primary alcohol using oxygen, as well as the alcohol and aldehyde dehydrogenases that would be necessary for 46 substrate transformation before β-oxidation and processing through the TCA cycle (Figure

2.5a). Although we did not test the strain's growth on aliphatic hydrocarbons, M. UTICA-

S1B6 grew at mid-range optical densities (0.05-0.1) on polysorbate nonionic surfactants

Tween 20, 40, and 80, which contain long-chain hydrocarbons (Figure 2.4b). Genes known to catalyze the anaerobic breakdown of hydrocarbons (benzylsuccinate synthase, alkylsuccinate synthase) through fumarate addition (Agrawal and Gieg 2013) were not identified in the genome, making it unlikely that Marinobacter sp. UTICA-S1B6 degrades hydrocarbons in the absence of oxygen.

Besides aliphatic hydrocarbons, Marinobacter sp. UTICA-S1B6 has the putative ability to degrade benzene and toluene, both of which have been found in produced fluids from the

Marcellus Shale at concentrations in the low part per million range (Hayes et al. 2012). The pathway involving benzene and toluene degradation employs the dmpK-L gene family, found in the M. UTICA-S1B6 genome, which converts benzene to phenol and later to catechol, or toluene to 2-hydroxytoluene and later 3-methylcatechol (Kanehisa and Goto

2000). Catechol and 3-methylcatechol are further metabolized by the xylE-J gene family to acetaldehyde and pyruvate, both of which can be converted to acetyl-CoA and into central metabolism (e.g., the TCA cycle) (Kanehisa and Goto 2000). This strain does not have the capacity to utilize other common aromatic compounds such as xylene, ethylbenzene, or PAHs (e.g., naphthalene). Additionally, we found a haloacid dehalogenase gene (L-DEX) in the M. UTICA-S1B6 genome, encoding for the conversion of (S)- 2-haloacids (e.g., haloacetates, halopropionates, halobutyrates containing fluoro,

47 chloro, bromo, or iodo substiutions) to (R)-2-hydroxyacids (Goldman et al. 1968).

Production of 2-hydroxyacids such as glycolate could be further broken down into glyoxylate, which can be utilized in the TCA cycle (Figure 2.5a). The genes responsible for this pathway, glcDEF and hprA, were detected in the M. UTICA-S1B6 genome as well as the Marcellus-4 metagenomes (Figure 2.5a).

Our genomic investigation revealed Arcobacter spp. MARC-MIP3H16 and UTICA-S4D1 may fix CO2 via the reverse TCA cycle, in a similar fashion to other reported Arcobacter strains (Figure 2.5a) (Wirsen et al. 2002, Hugler et al. 2005, Klatt and Polerecky 2015,

Roalkvam et al. 2015). These two Arcobacter strains contained all nine enzymes involved in the reductive TCA cycle (Hugler et al. 2005). We followed trends in both organic (non- purgeable organic carbon (NPOC)) and inorganic carbon in the form of CO2 gas from the

Marcellus-4 natural-gas well. Both CO2 and NPOC fluctuate over time, but there is a strong association between higher CO2 concentrations and higher relative abundance of

Arcobacter based on the 16S rRNA gene around day 35 and 80 (Figures 2.2a and 2.2b).

Although growth on inorganic carbon was not tested with these Arcobacter isolates, and

CO2 trends in natural-gas wells are likely influenced by a variety of biogeochemical factors, this association suggests higher concentrations of CO2 may enable growth of

Arcobacter through an autotrophic metabolism or chemosynthesis of reduced sulfur species during certain times of production.

In addition to the genomic potential to fix inorganic carbon, Arcobacter sp. MARC-

MIP3H16 utilized 25 different amino acids and carboxylic acids (Figure 2.4b). In the genome, we found amino acid ABC transporters, monosaccharide transporters, dicarboxylate transporters, C4-transporters, and anaerobic C4-transporters confirming the ability to uptake these compounds. Amino acids L-glutamine, L-glutamate, L-aspartate, and L-proline were optimally utilized as sole carbon sources, (OD>0.1). The carboxylic acids L-malate, propionate, pyruvate, L-lactate, and succinate were optimally converted for growth as sole carbon sources while several other carboxylic acids were minimally utilized (optical density changes between 0.05 and 0.1), such as acetate and methyl pyruvate.

2.3.4 Marinobacter and Arcobacter influence nitrogen, sulfur, and iron cycles in unconventional systems

We tracked total dissolved nitrogen (TDN) and ammonium/ammonia in six northern

Appalachian Basin natural-gas wells and observed a consistent trend after hydraulic fracturing. TDN levels increased from under 5 mg/L in initial flowback samples to above

100 mg/L within six months of production in natural-gas wells from both the Utica-Point

Pleasant Formation and Marcellus Shale (Figure 2.2c). The majority of TDN (59% or greater at all time points) was comprised of fixed sources of nitrogen

(ammonia/ammonium). Nitrate was analyzed but was below detection (< 1 ppm N) at all time points in flowback and produced fluid samples. Other possible sources of nitrogen in the system include organic nitrogen compounds, which were not measured. 49

Although nitrate was not detected in ppm concentrations for these natural-gas wells, nitrate reduction is a possible mode of energy generation for both Marinobacter and Arcobacter, especially at levels below our analytical detection limits. Our genomic analysis showed that Marinobacter sp. UTICA-S1B6 is capable of denitrification, possessing the proper transporters and catalytic genes for complete conversion of nitrate to N2 (Figure 2.5a). In addition to utilizing inorganic nitrogen as an electron acceptor, M. UTICA-S1B6 can convert urea to ammonia with urease, which is especially interesting given urea is a component in the hydraulic fracturing fluid formulation for one of the Marcellus natural- gas wells in this study (see Figure A.2 in Appendix A). Arcobacter spp. MARC-MIP3H16 and UTICA-S4D1, on the other hand, are capable of dissimilatory nitrate reduction to ammonia. Like Marinobacter sp. UTICA-S1B6, Arcobacter spp. MARC-MIP3H16 and

UTICA-S4D1 can also oxidize nitroalkanes via the NMO gene, which may serve as an alternative source of nitrate within the cell.

The genome of Marinobacter sp. UTICA-S1B6 contains genes responsible for importing and assimilating sulfur from both inorganic (e.g., sulfate) and organic sulfur forms.

Specifically, M. UTICA-S1B6 can putatively utilize alkanesulfonates via the ssuD gene, producing sulfite and an aldehyde group (Figure 2.5a) (Eichhorn et al. 1999). The ability to convert alkane-derived nutrients to sources of carbon and macronutrients (e.g., sulfate, nitrate) is consistent in the Marcellus-4 metagenomes that cluster closely to our

Marinobacter isolate (Figure 2.5a, see Table A.1 in Appendix A).

One of the more unique aspects of Arcobacter sp. MARC-MIP3H16 relative to other taxa derived from these hydrocarbon bearing systems is its ability to completely oxidize reduced sulfur compounds (inc. sulfide, elemental sulfur, and thiosulfate) to sulfate using the sox pathway (Figure 2.5b). Like other Arcobacter, A. MARC-MIP3H16 may be coupling this electron transfer process to chemosynthesis (Friedrich et al. 2005, Klatt and

Polerecky 2015, Roalkvam et al. 2015). Our geochemical analysis of sulfate and sulfide in produced fluids from natural-gas wells in both the Utica-Point Pleasant Formation and

Marcellus Shale supports the potential for these metabolisms to occur in these systems.

Sulfate was present at part per million concentrations (20-140 mg/L) in source waters used for hydraulic fracturing of these natural-gas wells (Figure 2.2d). In the Marcellus natural- gas wells, within 10 days after hydraulic fracturing, sulfate dropped below 1 mg/L and remained below detection thereafter (Figure 2.2d). In the Utica-Point Pleasant natural-gas wells, sulfate concentrations varied between 11 and 60 mg/L during the first 59 days post- fracturing. After this time, sulfate concentrations decreased to below 6 mg/L while sulfide increased to between 10-22 mg/L until about day 150 after hydraulic fracturing. Sulfide levels plateaued at levels of 20-35 mg/L as the natural-gas well matured (150 to >450 days,

Figure 2.2d). This period of intermediate sulfate and sulfide concentrations was associated with a higher abundance of 16S rRNA gene sequences for Arcobacter (Figure 2.2a) and could be related to the reoxidation of reduced sulfur species during this time. Also present in the Arcobacter genomes was the sqr gene responsible for oxidation of sulfide to polysulfides, the pshABC gene which reduces sulfur and thiosulfate to sulfide, and the cysIJ

51 gene which oxidizes sulfide to sulfite (Kanehisa and Goto 2000). Both isolates are also capable of reducing tetrathionate to thiosulfate (ttr gene) (Kanehisa and Goto 2000). One isolate, Arcobacter sp. MARC-MIP3H16, contains the gene responsible for oxidation of thiosulfate to sulfite (sseA) (Kanehisa and Goto 2000).

We detected 36 cytochromes in the genome of Marinobacter sp. UTICA-S1B6, some of which may be involved in extracellular electron transfer to redox active species (e.g., iron and oxygen). One of these cytochromes was a (per)oxidase also detected in Arcobacter sp.

MARC-MIP3H16. Additionally, Marinobacter contained a thiol (per)oxidase and a catalase (per)oxidase. Closely related Marinobacter subterrani is known to oxidize Fe(II), and Marinobacter aquaeolei oxidizes Fe(II) from metal structures during growth in biofilms (Bonis and Gralnick 2015, Singer et al. 2011). The (per)oxidase cytochrome is one of many cytochromes thought to be responsible for iron oxidation, while a type II secretion system is associated with Marinobacter aquaeolei's ability to form a biofilm that interacts with metal surfaces (Singer et al. 2011). Indeed, the (per)oxidase encoding gene

(pfam 00141) in the genome of Marinobacter sp. UTICA-S1B6 is homologous to the peroxidase encoding gene of Marinobacter subterrani JG233 (82% identity, 1234 bit score) and Marinobacter aquaeolei VT8 (88% identity, bit score 1321). Interestingly,

Marinobacter sp. UTICA-S1B6 also contains the type II secretion system, suggesting a capacity to attach to reduced iron surfaces (e.g., minerals such as clays, silicates, and carbonates common in shale, carbon steel casings) and oxidize reduced forms of iron abundant within the organic-rich subsurface systems responsible for the generation of

52 petroleum (i.e., source rocks such as shale, mudstones, organic-rich limestones, etc.)

(Singer et al. 2011).

2.3.5 Marinobacter and Arcobacter possess unique virulence strategies in produced fluid

During our assessment of these Marinobacter and Arcobacter genomes, we identified genes comprising a type VI secretion system (Figure 2.5a and 2.5b). This system is best described as a phage-like apparatus which can be used to lyse other cells by injecting enzymes designed to degrade cellular components. Although energetically costly (Alteri and Mobley 2016), this virulence strategy may provide a competitive advantage to these taxa under nutrient limited conditions or environmental stress (Alteri and Mobley 2016,

Chakraborty et al. 2011). Said strategy could also contribute to horizontal gene transfer within this system (Borgeaud et al. 2015). To date, no literature has been published identifying this virulence survival mechanism in the environment that occurs following hydraulic fracturing, although there is strong genomic evidence for persistent viral attacks to these taxa (Daly et al. 2016). Indeed, the genomes of M. UTICA-S1B6, A. UTICA-

S4D1, and A. MARC-MIP3H16 all contain genes indicative of a type I CRISPR-Cas system which serve as acquired immunity to predatory bacteriophage (Makarova et al.

2015).

2.3.6 Isolate genomes largely agree with metagenomic data

The key pathways summarized above for the isolate genome of Marinobacter sp. UTICA-

S1B6 were also discovered with high homology in the Marinobacter metagenome sampled from fluids produced on day 34 from Marcellus Shale natural-gas well Marcellus-4 (bit score >200, identity >45.9%). One notable exception was the difference in the pathway for aromatic hydrocarbon degradation. The metagenome contained genes associated with catechol degradation via ortho-cleavage as opposed to the isolate genome which contained genes encoding for the meta-cleavage pathway (Caspi et al. 2014, Kanehisa and Goto

2000). Several studies report that the meta-pathway is used for both non-substituted aromatics in addition to alkyl and halosubstituted aromatics such as toluene and chlorobenzene, whereas the ortho-pathway is generally observed exclusively during growth on non-substituted benzene or phenol (Mars et al. 1997, Veenagayathri and

Vasudevan 2011). The exclusive possession of the ortho-pathway by Marinobacter in the produced fluid metagenome may indicate sole preference for non-substituted or hydroxylated aromatic hydrocarbons in the environment. In addition, the gene for utilizing nitroalkanes through conversion to an aldehyde or ketone plus nitrate was discovered in the metagenomes with high homology to Marinobacter, but was absent in the genomes.

Largely, genes belonging to Arcobacter mined from the metagenome of the Marcellus

Shale natural-gas well Marcellus-4 at days 34 and 93 meeting homology cutoffs agreed with the isolate genomes. However, a few exceptions are noted (see Table A.1 in Appendix

A) including the absence of two sulfur genes in the metagenomes, cysJ (sulfide to sulfite) 54 and sseA (thiosulfate to sulfite), as well as TMO (transforms trimethylamine n-oxide to trimethylamine), which were all present in the Arcobacter isolate genome.

2.4 Discussion

2.4.1 Marinobacter and Arcobacter dominate early produced fluid communities through unique osmoadaptations and defense mechanisms

Marinobacter are one of a dozen microbial taxa commonly found in producing natural-gas wells from hydraulically stimulated formations across the U.S. (Mouser et al. 2016). The presence of this taxa during early production in Marcellus-4 is consistent with trends reported for other geographically and lithologically distinct unconventional hydrocarbon- producing systems (e.g., shales, mudstones, organic-rich limestones), which show

Marinobacter present in early flowback fluid samples (<49 days after flowback began) but largely absent after that time (>49 days) (Cluff et al. 2014, Mohan et al. 2013, Davis et al.

2012, Fitcher et al. 2012, Struchemeyer and Elshahed 2012). Similarly, the temporal trends reported here for Arcobacter abundance (Figure 2.2a) are consistent with those reported by

Wutcher and coauthors (2013), who detected Arcobacter marinus in produced fluids from the Antrim Shale ranging from 35% to 85% of the total ε-proteobacteria abundance during the five months after hydraulic fracturing. Previous studies on fluids produced from the

Marcellus Shale report Arcobacter abundance peaking (upwards of 66% of total reported microbial community based on the 16S rRNA gene) within the first two weeks after production begins and decreasing thereafter (Cluff et al. 2014, Mohan et al. 2013). This

55 consistent detection of Marinobacter during the first few weeks after fracturing and

Arcobacter during the first few months after fracturing suggests these taxa play an important, albeit fleeting, role in the subsurface following natural-gas well completion and hydraulic fracturing.

Based on genomic and metagenomic inference, there are several survival strategies these organisms possess that could enable their persistence in fluids derived from tight shales and other unconventional hydrocarbon-producing systems during this transitional biogeochemical phase. Arguably, the largest bottleneck for the survival of microorganisms in an ecosystem such as those generated following fracturing of shales or other unconventional systems is the high salinity observed in produced fluids from these formations, which is known to increase as hydrocarbon production proceeds. Both

Arcobacter and Marinobacter possess strategies for adapting to high salt, namely via anti- porters using a “salt-in” strategy, where cells import ions into the cytoplasm to avoid osmotic stress under increasing concentrations in the environment. Both taxa can also synthesize the osmolyte ectoine and import extracellular osmolytes. Additionally,

Marinobacter can produce osmoprotectant glycine betaine in response to high salinity, which can subsequently be fermented by other microbial community members to fuel methanogenesis (Daly et al. 2016).

The ability to prey upon other bacterial taxa is a survival strategy that could allow predatorial taxa to have a significant advantage in subsurface systems with strong

56 competition for nutrient or electron acceptor resources. The type VI secretion system detected in both Marinobacter and Arcobacter genomes may allow these taxa to attack neighboring cells with a phage-like appendage, insert effector proteins, and capture cellular materials from the victim including lipids, polysaccharides, proteins, amino acids, and nucleic acids. Besides antagonism, the type VI secretion system may be used for cell signaling, biofilm remodeling, or lysing phage-infected bacteria (Russell et al. 2014).

Interestingly, Marinobacter hydrocarbonoclasticus grown on hexadecane overexpressed type VI secretion system proteins relative to controls. The authors concluded a role for the type VI secretion system in alkane assimilation (Vaysse et al. 2009). As both Arcobacter and Marinobacter possess this system, its presence could confer their advantage in fluids derived from tight shales and other unconventional hydrocarbon-producing systems through biofilm formation, virulence, or hydrocarbon assimilation. Furthermore, the presence of CRISPR-associated encoding genes in the genomes of all three isolates may confer additional immunity in this system in the presence of lysogenic phage (Daly et al.

2016).

2.4.2 Marinobacter and Arcobacter are carbon opportunists

Our investigation into the metabolic potential of Marinobacter sp. UTICA-S1B6 revealed the capacity for complex carbon oxidation, including aliphatic compounds, aromatic compounds, and haloacids. Our phylogenetic analysis based on the 16S rRNA gene showed

M. UTICA-S1B6 is closely related to Marinobacter hydrocarbonoclasticus (96.6%), a bacterium isolated from seawater collected near an oil refinery in the Mediterranean Sea 57

(Mounier et al. 2014). M. hydrocarbonoclasticus is able to degrade a wide variety of aliphatic and aromatic hydrocarbons through oleolytic biofilm formation. Marinobacter sp. UTICA-S1B6 could access and transform hydrocarbons in a similar manner based on our genomic evaluation by coupling these reactions to oxygen and/or nitrate reduction, enabling its persistence during the first few weeks of flowback before oxygenated resources are depleted. Furthermore, Marinobacter may persist as labile carbon substrates are depleted due to its ability to utilize a wide range of organic carbon sources much like other opportunistic Marinobacter species (Singer et al. 2011).

An autotrophic metabolism potentially confers advantage for Arcobacter species among a primarily heterotrophic and fermentative microbial community. As the natural-gas well matures and the sources of injected labile carbon dwindle, Arcobacter may alternate between heterotrophic and autotrophic metabolisms. Further experimentation is required to confirm its ability to fix CO2, but geochemical and genomic evidence suggest this enzymatic pathway may be important for the persistence of Arcobacter during the first few months following hydraulic fracturing and hydrocarbon production within a given natural-gas well. Much like opportunistic Marinobacter, Arcobacter species in unconventional systems may also utilize diverse carbon substrates (organic and inorganic), electron acceptors (oxygen, nitrate), and electron donors (reduced sulfur and iron species) that are generally present during early flowback.

2.4.3 Importance of Marinobacter and Arcobacter on biogeochemical cycling in flowback and produced fluids

Nitrate is added in certain oil and gas systems to thermodynamically control growth of sulfate reducers and fermenters, thereby reducing biocorrosion damage to infrastructure through sulfide and/or acid generation (An et al. 2017, Voorduow 2008, Arensdorf et al.

2009, Voorduow et al. 2009). As a result, there is need to understand the utilization of both organic and inorganic nitrogen species by microorganisms in subsurface environments such as shale or other unconventional hydrocarbon-bearing systems. In contrast to other subsurface terrestrial systems like deep groundwater environments that have limited nitrogen sources (Bomberg and Ahonen 2017, Kutvonen et al. 2015), excess sources of fixed nitrogen are available in fluids produced from shale or other unconventional hydrocarbon resources.

Our results show that both TDN and ammonia increase significantly as production proceeds in six hydraulically fractured natural-gas wells in the northern Appalachian Basin. The individual constituents of TDN may originate from a variety of anthropogenic and indigenous sources. Since nitrate was below detection in produced fluids, it is not a considerable source of TDN here. In the case of the Utica-3, Utica-4, Utica-5, and Utica-6 natural-gas wells, produced fluids from other hydraulically fractured natural-gas wells were used as a source of slickwater for hydraulic fracturing (i.e., recycled produced fluid), which may partially explain high initial TDN values. In terms of other nitrogen sources, organic nitrogen is frequently added to fracture fluid formulations for several different 59 purposes, including clay stabilization, scale inhibition, friction reduction, gel formation, and as surfactants and solvents (Elsner and Hoelzer 2016). For example, urea was one chemical component of the hydraulic fracturing fluid in the Marcellus Shale natural-gas well Marcellus-4, and its enzymatic conversion to ammonia and carbon dioxide may contribute to TDN and ammonia trends in these wells. Ammonia is also a potential byproduct of (poly)acrylamide and related polymer degradation (biotic or abiotic).

Polyacrylamides are contained in many slickwater fracturing fluid recipes (Elsner &

Hoelzer 2016), and are listed on the FracFocus report for the three Marcellus and Utica-

Point Pleasant natural-gas wells where these three strains were isolated (Marinobacter sp.

UTICA-S1B6, Arcobacter sp. S4D1, and Arcobacter sp. MARC-MIP3H16) (see Figures

A.2, A.3, A.4 in Appendix A). Interestingly, Marinobacter contains amidase-encoding genes, suggesting that this taxa could degrade amide-based polymers in this system, such as polyacrylamides. Furthermore, both Marinobacter and Arcobacter can produce nitrate from nitroalkanes, a pre-cursor biochemical process that could enable denitrification or dissimilatory nitrate reduction despite low (<1 ppm) nitrate concentrations in produced fluids.

Nitrogen fixation is energetically costly, and the presence of bioavailable nitrogen sources

(e.g., nitrate, urea, amines, ammonia) makes it unlikely nitrogen fixation occurs in these unconventional hydrocarbon-producing systems (Houlton et al. 2008). Fixed sources of nitrogen must therefore be biologically produced or sourced within the rock (e.g., shale).

Arcobacter may produce ammonia during early production through dissimilatory nitrate

60 reduction. A methanogen that thrives during later production, Methanohalophilus, may also contribute to ammonia production during methanogenesis fueled by methylamines

(Borton et al. 2018). Altogether, it is plausible that Marinobacter and Arcobacter in early production of hydraulically fractured natural-gas wells contribute to marked increases in total nitrogen concentrations and ammonia in produced fluids by reducing nitrate and/or degrading xenobiotic organic nitrogen sources injected by well operators.

The primary cause of natural-gas well infrastructure fouling is attributed to the acidic and reactive properties of sulfides, making the sulfur cycle a vitally important process to understand in this system (Booker et al. 2017, Pirzadeh et al. 2014). Previous studies indicate thiosulfate reduction by Halanaerobium spps. as the primary mechanism for sulfide formation in later produced fluids (Liang et al. 2016, Booker et al. 2017).

Contributing to this biogeochemical cycle, Arcobacter spp. have the capacity to refuel sulfide production, at least temporarily, through oxidation of reduced sulfur species (e.g., sulfides, sulfur) to thiosulfate and sulfate. This process would initially reduce sulfide levels until oxygen and/or nitrate are depleted from the system and provide electron acceptors

(e.g., thiosulfate, sulfate) for other dominant taxa. Although Arcobacter spps. studied here have the ability to oxidize sulfide to sulfate, we propose thiosulfate to be the primary endpoint of Arcobacter sulfur metabolism due to (1) the lack of sulfate present in later produced fluids, and (2) the presence of known thiosulfate reducers in these wells (e.g.

Halanaerobium).

2.5 Conclusion

Bacterial taxa common to early production stages of hydraulically fractured natural-gas wells play significant roles in carbon, nitrogen, and sulfur cycling in flowback and produced fluids. Marinobacter can oxidize hydrocarbons as well as capture carbon, nitrate, and sulfur species from alkane-derived hydrocarbons, which are most likely present from fracture fluid additives and/or brines produced from unconventional hydrocarbon systems during the early stages of natural-gas production. Arcobacter uniquely participates in carbon and sulfur cycling through coupling chemosynthesis to oxidation of reduced sulfur species, which may fuel heterotrophic, fermentative, and/or thiosulfate-reducing microorganisms. These taxa may further contribute to high ammonia concentrations in these natural-gas wells through urea conversion or nitrate reduction. Our genomic and experimental investigations of Marinobacter and Arcobacter physiology reveals the impact these slight to moderate halophilic taxa have on biogeochemical cycles during early production stages in hydraulic fractured natural-gas wells as oxygen and carbon resources are diminishing and salinities increase.

Author contributions:

P. Mouser and J. Panescu conceived and designed the experiments. J. Panescu and M.

Evans performed genomic analyses. M. Evans performed metagenomic analyses, assisted with geochemical measurements, and wrote the manuscript. J. Panescu performed the laboratory-based experiments. J. Panescu, S. Welch, and J. Sheets acquired the SEM images. N. Nastasi compiled geochemical data and contributed a figure. D. Cole, M. 62

Wilkins, K. Wrighton and P. Mouser designed field experiment while A. Hanson, R. Daly,

S. Welch, and J. Sheets carried out field sampling and performed geochemical measurements. K. Wrighton and R. Daly extracted DNA and processed metagenomes. T.

Darrah performed CO2 analysis.

Acknowledgements: This research was supported by funding from the National Sciences

Foundation Dimensions of Biodiversity (award no. 1342701) and DOE National Energy

Technology Laboratory through the Marcellus Shale Energy and Environmental

Laboratory (project #DE-FE0024297) under a subcontract from West Virginia University.

We are especially grateful for support from Northeast Natural Energy for site access and sample support. Sequencing was performed under an award to K.C.W., M.J.W. and P.J.M.

(no. 1931), and conducted by the U.S. Department of Energy Joint Genome Institute, a

DOE Office of Science User Facility, supported by the Office of Science of the U.S.

Department of Energy under contract no. DE-AC02-05CH11231.

2.6 References

Agrawal, A., and Gieg, L. M. (2013). In situ detection of anaerobic alkane metabolites in subsurface environments. Front Microbiol 4. doi:10.3389/fmicb.2013.00140.

Alteri, C. J., and Mobley, H. L. T. (2016). The versatile type VI secretion system. Microbiol Spectr 4. doi:10.1128/microbiolspec.VMBF-0026-2015.

An, B. A., Shen, Y., and Voordouw, G. (2017). Control of sulfide production in high salinity Bakken shale oil reservoirs by halophilic bacteria reducing nitrate to nitrite. Front. Microbiol. 8. doi:10.3389/fmicb.2017.01164.

Arensdorf, J. J., Miner, K., Ertmoed, R., Clay, W. K., Stadnicki, P., and Voordouw, G. (2009). Mitigation of reservoir souring by nitrate in a produced-water reinjection system in Alberta. Society of Petroleum Engineers. doi:10.2118/121731-MS.

Bagheri, M., Amoozegar, M. A., Didari, M., Makhdoumi-Kakhki, A., Schumann, P., Spröer, C., et al. (2013). Marinobacter persicus sp. nov., a moderately halophilic bacterium from a saline lake in Iran. Antonie Van Leeuwenhoek 104, 47–54. doi:10.1007/s10482-013-9923-3.

Bomberg, M., and Ahonen, L. (2017). Editorial: Geomicrobes: life in terrestrial deep subsurface. Front Microbiol 8. doi:10.3389/fmicb.2017.00103.

Bonis, B. M., and Gralnick, J. A. (2015). Marinobacter subterrani, a genetically tractable neutrophilic Fe(II)-oxidizing strain isolated from the Soudan Iron Mine. Front. Microbiol. 6. doi:10.3389/fmicb.2015.00719.

Borgeaud, S., Metzger, L. C., Scrignari, T., and Blokesch, M. (2015). The type VI secretion system of Vibrio cholerae fosters horizontal gene transfer. Science 347, 63–67. doi:10.1126/science.1260064.

Brown, C. T., Hug, L. A., Thomas, B. C., Sharon, I., Castelle, C. J., Singh, A., et al. (2015). Unusual biology across a group comprising more than 15% of domain Bacteria. Nature 523, 208. doi:10.1038/nature14486

Brussaard, C. P. D. (2004). Optimization of procedures for counting viruses by flow cytometry. Appl. Environ. Microbiol. 70, 1506–1513. doi:10.1128/AEM.70.3.1506- 1513.2004.

Caspi, R., Altman, T., Billington, R., Dreher, K., Foerster, H., Fulcher, C. A., et al. (2014). The MetaCyc database of metabolic pathways and enzymes and the BioCyc collection of pathway/genome databases. Nucleic Acids Res 42, D459–D471. doi:10.1093/nar/gkt1103.

Chakraborty, S., Sivaraman, J., Leung, K. Y., and Mok, Y.-K. (2011). Two-component PhoB-PhoR regulatory system and ferric uptake regulator sense phosphate and iron to control virulence genes in Type III and VI secretion systems of Edwardsiella tarda. J Biol Chem 286, 39417–39430. doi:10.1074/jbc.M111.295188.

Chapman, E. C., Capo, R. C., Stewart, B. W., Kirby, C. S., Hammack, R. W., Schroeder, K. T., et al. (2012). Geochemical and strontium isotope characterization of produced waters

65 from Marcellus Shale natural gas extraction. Environ. Sci. Technol. 46, 3545–3553. doi:10.1021/es204005g.

Chen, F., Lu, J., Binder, B. J., Liu, Y., and Hodson, R. E. (2001). Application of digital image analysis and flow cytometry to enumerate marine viruses stained with SYBR Gold. Appl. Environ. Microbiol. 67, 539–545. doi:10.1128/AEM.67.2.539-545.2001.

Chen, I.-M. A., Markowitz, V. M., Chu, K., Palaniappan, K., Szeto, E., Pillay, M., et al. (2017). IMG/M: integrated genome and metagenome comparative data analysis system. Nucleic Acids Res 45, D507–D516. doi:10.1093/nar/gkw929.

Cluff, M. A., Hartsock, A., MacRae, J. D., Carter, K., and Mouser, P. J. (2014). Temporal changes in microbial ecology and geochemistry in produced water from hydraulically fractured Marcellus shale gas wells. Environ. Sci. Technol. 48, 6508–6517. doi:10.1021/es501173p.

Darrah, T. H., Jackson, R. B., Vengosh, A., Warner, N. R., Whyte, C. J., Walsh, T. B., Kondash, A.J., Poreda, R.J. (2015). The evolution of Devonian hydrocarbon gases in shallow aquifers of the northern Appalachian Basin: Insights from integrating noble gas and hydrocarbon geochemistry. Geochimica et Cosmochimica Acta 170, 321–355. doi:10.1016/j.gca.2015.09.006.

Davis, J. P., Struchtemeyer, C. G., and Elshahed, M. S. (2012). Bacterial communities associated with production facilities of two newly drilled thermogenic natural gas wells

66 in the Barnett shale (Texas, USA). Microb Ecol 64, 942–954. doi:10.1007/s00248-012- 0073-3.

Diemer, G., E Kyle, J., and Stedman, K. (2013). Counting viruses using polycarbonate Track EtchTM membrane filters as an alternative to AnodiscTM membrane filters. Available at: http://web.pdx.edu/~kstedman/PCTE_virus_counting_protocol.pdf.

Donachie, S. P. (2005). Arcobacter halophilus sp. nov., the first obligate halophile in the genus Arcobacter. International Journal of Systematic and Evolutionary Microbiology 55, 1271–1277. doi:10.1099/ijs.0.63581-0.

Edgar, R. C. (2010). Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26, 2460–2461. doi:10.1093/bioinformatics/btq461

Eichhorn, E., van der Ploeg, J. R., and Leisinger, T. (1999). Characterization of a two- component alkanesulfonate monooxygenase from Escherichia coli. J. Biol. Chem. 274, 26639–26646. doi:10.1074/jbc.274.38.26639

Fichter, J., Wunch, K., Moore, R., Summer, E., Braman, S., and Holmes, P. (2012). How hot is too hot for bacteria? A technical study assessing bacterial establishment in downhole drilling, fracturing and fracturing operations. in (NACE International). Available at: https://www.onepetro.org/conference-paper/NACE-2012-1310

Friedrich, C. G., Bardischewsky, F., Rother, D., Quentmeier, A., and Fischer, J. (2005). Prokaryotic sulfur oxidation. Curr. Opin. Microbiol. 8, 253–259. doi:10.1016/j.mib.2005.04.005.

Gaspar, J., Mathieu, J., Yang, Y., Tomson, R., Leyris, J. D., Gregory, K. B., (2014). Microbial dynamics and control in shale gas production. Environ. Sci. Technol. Lett. 1, 465–473. doi:10.1021/ez5003242.

Gauthier, M. J., Lafay, B., Christen, R., Fernandez, L., Acquaviva, M., Bonin, P., et al. (1992). Marinobacter hydrocarbonoclasticus gen. nov., sp. nov., a new, extremely halotolerant, hydrocarbon-degrading marine bacterium. Int. J. Syst. Bacteriol. 42, 568– 576. doi:10.1099/00207713-42-4-568.

Goldman, P., Milne, G. W. A., and Keister, D. B. (1968). Carbon-halogen bond cleavage III. Studies on bacterial halidohydrolases. J. Biol. Chem. 243, 428–434. Available at: http://www.jbc.org/content/243/2/428.full.pdf

Harkness, J.S., Darrah, T.H., Warner, N.R., Whyte, C.J., Moore, M.T., Millot, R., Kloppmann, W., Jackson, R.B., Vengosh, A. (2017). The geochemistry of naturally occurring methane and saline groundwater in an area of unconventional shale gas

68 development. Geochimica et Cosmochimica Acta 208, 302-334. doi:10.1016/j.gca.2017.03.039

Hayes, T., Severin, B. F., Engineer, P. S. P., and Okemos, M. I. (2012). Barnett and Appalachian Shale water management and reuse technologies. Research Partnership to Secure Energy for America, Report no. 8122-05. Available at: https://www.netl.doe.gov/File%20Library/Research/Oil- Gas/Natural%20Gas/shale%20gas/08122-05-final-report.pdf

Heilweil, V. M., Solomon, D. K., Darrah, T. H., Gilmore, T. E., and Genereux, D. P. (2016). Gas-tracer experiment for evaluating the fate of methane in a coastal plain stream: degassing versus in-stream oxidation. Environ. Sci. Technol. 50, 10504–10511. doi:10.1021/acs.est.6b02224.

Houlton, B. Z., Wang, Y.-P., Vitousek, P. M., and Field, C. B. (2008). A unifying framework for dinitrogen fixation in the terrestrial biosphere. Nature 454, 327. doi: 10.1038/nature07028

Hügler, M., Wirsen, C. O., Fuchs, G., Taylor, C. D., and Sievert, S. M. (2005). Evidence for autotrophic CO2 fixation via the reductive tricarboxylic acid cycle by members of the ɛ subdivision of Proteobacteria. J Bacteriol 187, 3020–3027. doi:10.1128/JB.187.9.3020- 3027.2005.

Huntemann, M., Ivanova, N. N., Mavromatis, K., Tripp, H. J., Paez-Espino, D., Tennessen, K., et al. (2016). The standard operating procedure of the DOE-JGI Metagenome Annotation Pipeline (MAP v.4). Standards in Genomic Sciences 11, 17. doi:10.1186/s40793-016-0138-x.

Huu, N. B., Denner, E. B. M., Ha, D. T. C., Wanner, G., and Stan-Lotter, H. (1999). Marinobacter aquaeolei sp. nov., a halophilic bacterium isolated from a Vietnamese oil- producing well. International Journal of Systematic and Evolutionary Microbiology 49, 367–375. doi:10.1099/00207713-49-2-367.

Hyatt, D., LoCascio, P. F., Hauser, L. J., and Uberbacher, E. C. (2012). Gene and translation initiation site prediction in metagenomic sequences. Bioinformatics 28, 2223– 2230. doi: 10.1093/bioinformatics/bts429

Hydraulic fracturing accounts for about half of current U.S. crude oil production - Today in Energy - U.S. Energy Information Administration (EIA) Available at: https://www.eia.gov/todayinenergy/detail.php?id=25372

Hydraulically fractured wells provide two-thirds of U.S. natural gas production - Today in Energy - U.S. Energy Information Administration (EIA) Available at: https://www.eia.gov/todayinenergy/detail.php?id=26112

Kang, M., Christian, S., Celia, M.A., Mauzerall, D.L., Bill, M., Miller, A.R., Chen, Y., Conrad, M.E., Darrah, T.H., Jackson, R.B. (2016) Identification and characterization of high methane-emitting abandoned oil and gas wells. PNAS, 13636-13641. doi:10.1073/pnas.1605913113.

Kanehisa, M., and Goto, S. (2000). KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Res 28, 27–30. doi:10.1093/nar/28.1.27.

Kim, H. M., Hwang, C. Y., and Cho, B. C. (2010). Arcobacter marinus sp. nov. International Journal of Systematic and Evolutionary Microbiology 60, 531–536. doi:10.1099/ijs.0.007740-0.

Klatt, J. M., and Polerecky, L. (2015). Assessment of the stoichiometry and efficiency of

CO2 fixation coupled to reduced sulfur oxidation. Front Microbiol 6, 484. doi:10.3389/fmicb.2015.00484.

Kutvonen, H., Rajala, P., Carpén, L., and Bomberg, M. (2015). Nitrate and ammonia as nitrogen sources for deep subsurface microorganisms. Front. Microbiol. 6. doi:10.3389/fmicb.2015.01079.

Lipus, D., Roy, D., Khan, E., Ross, D., Vikram, A., Gulliver, D., et al. (2018). Microbial communities in Bakken region produced water. FEMS Microbiol Lett 365. doi:10.1093/femsle/fny107.

Makarova, K. S., Wolf, Y. I., Alkhnbashi, O. S., Costa, F., Shah, S. A., Saunders, S. J., et al. (2015). An updated evolutionary classification of CRISPR-Cas systems. Nat. Rev. Microbiol. 13, 722–736. doi:10.1038/nrmicro3569.

Mars, A. E., Kasberg, T., Kaschabek, S. R., Agteren, M. H. van, Janssen, D. B., and Reineke, W. (1997). Microbial degradation of chloroaromatics: use of the meta-cleavage pathway for mineralization of chlorobenzene. J. Bacteriol. 179, 4530–4537. doi:10.1128/jb.179.14.4530-4537.1997.

Moore, M. T., Vinson, D. S., Whyte, C. J., Eymold, W. K., Walsh, T. B., and Darrah, T. H. (2018). Differentiating between biogenic and thermogenic sources of natural gas in coalbed methane reservoirs from the Illinois Basin using noble gas and hydrocarbon geochemistry. Geological Society, London, Special Publications 468, 151. doi:10.1144/SP468.8.

Mounier, J., Camus, A., Mitteau, I., Vaysse, P.-J., Goulas, P., Grimaud, R., et al. (2014). The marine bacterium Marinobacter hydrocarbonoclasticus SP17 degrades a wide range of lipids and hydrocarbons through the formation of oleolytic biofilms with distinct gene expression profiles. FEMS Microbiol Ecol 90, 816–831. doi:10.1111/1574-6941.12439.

Nixon, S. L., Walker, L., Streets, M. D. T., Eden, B., Boothman, C., Taylor, K. G., et al. (2017). Guar gum stimulates biogenic sulfide production at elevated pressures: implications for shale gas extraction. Front. Microbiol. 8. doi:10.3389/fmicb.2017.00679.

Noble, R., and Fuhrman, J. (1998). Use of SYBR Green I for rapid epifluorescence counts of marine viruses and bacteria. Aquatic Microbial Ecology 14, 113–118. doi:10.3354/ame014113.

Oren, A. (2008). Microbial life at high salt concentrations: phylogenetic and metabolic diversity. Saline Syst. 4, 2. doi:10.1186/1746-1448-4-2.

Patel, A., Noble, R. T., Steele, J. A., Schwalbach, M. S., Hewson, I., and Fuhrman, J. A. (2007). Virus and prokaryote enumeration from planktonic aquatic environments by epifluorescence microscopy with SYBR Green I. Nature protocols 2, 269. doi:10.1038/nprot.2007.6

Pirzadeh, P., Lesage, K. L., and Marriott, R. A. (2014). Hydraulic fracturing additives and the delayed onset of hydrogen sulfide in shale gas. Energy Fuels 28, 4993–5001. doi:10.1021/ef501059k.

Quevillon, E., Silventoinen, V., Pillai, S., Harte, N., Mulder, N., Apweiler, R., et al. (2005). InterProScan: protein domains identifier. Nucleic acids research 33, W116– W120. doi:10.1093/nar/gki442

Roalkvam, I., Drønen, K., Stokke, R., Daae, F. L., Dahle, H., and Steen, I. H. (2015). Physiological and genomic characterization of Arcobacter anaerophilus IR-1 reveals new metabolic features in Epsilonproteobacteria. Front. Microbiol. 6. doi:10.3389/fmicb.2015.00987.

Russell, A. B., Peterson, S. B., and Mougous, J. D. (2014). Type VI secretion system effectors: poisons with a purpose. Nature Reviews Microbiology 12, 137–148. doi:10.1038/nrmicro3185.

Singer, E., Webb, E. A., Nelson, W. C., Heidelberg, J. F., Ivanova, N., Pati, A., et al. (2011). Genomic potential of Marinobacter aquaeolei, a biogeochemical “opportunitroph.” Appl. Environ. Microbiol. 77, 2763–2771. doi:10.1128/AEM.01866- 10.

Tummings, S., Panescu, J., Daly, R. A., Wrighton, K. C., and Mouser, P. J. (2018). Draft genome sequences of Marinobacter strains recovered from Utica Shale-produced fluids. Genome Announc. 6, e00155-18. doi:10.1128/genomeA.00155-18.

Vaysse, P.-J., Prat, L., Mangenot, S., Cruveiller, S., Goulas, P., and Grimaud, R. (2009). Proteomic analysis of Marinobacter hydrocarbonoclasticus SP17 biofilm formation at the alkane-water interface reveals novel proteins and cellular processes involved in

74 hexadecane assimilation. Research in Microbiology 160, 829–837. doi:10.1016/j.resmic.2009.09.010.

Veenagayathri, K., and Vasudevan, N. (2011). Ortho and meta cleavage dioxygenases detected during the degradation of phenolic compounds by a moderately halophilic bacterial consortium. Int Res J Microbiol 2, 406–414. Available online at: http://www.interesjournals.org/IRJM

Vengosh, A., Kondash, A., Harkness, J., Lauer, N., Warner, N., and Darrah, T. H. (2017). The geochemistry of hydraulic fracturing fluids. Procedia Earth and Planetary Science 17, 21–24. doi:10.1016/j.proeps.2016.12.011.

Voordouw, G. (2008). “Impact of Nitrate on the Sulfur Cycle in Oil Fields” in Microbial Sulfur Metabolism Springer, 296–302. doi:10.1007/978-3-540-72682-1_23.

Warner, N. R., Christie, C. A., Jackson, R. B., and Vengosh, A. (2013). Impacts of shale gas wastewater disposal on water quality in western Pennsylvania. Environ. Sci. Technol. 47, 11849–11857. doi:10.1021/es402165b.

Weinbauer, M. G., Beckmann, C., and Höfle, M. G. (1998). Utility of green fluorescent nucleic acid dyes and aluminum oxide membrane filters for rapid epifluorescence enumeration of soil and sediment bacteria. Applied and Environmental Microbiology 64, 5000–5003. Available at: http://aem.asm.org/content/64/12/5000.long

Wirsen, C. O., Sievert, S. M., Cavanaugh, C. M., Molyneaux, S. J., Ahmad, A., Taylor, L. T., et al. (2002). Characterization of an autotrophic sulfide-oxidizing marine Arcobacter sp. that produces filamentous sulfur. Appl. Environ. Microbiol. 68, 316–325. doi:10.1128/AEM.68.1.316-325.2002

2.7 Figures

Figure 2.1 Flowchart summarizing methods used for isolating and characterizing Arcobacter and Marinobacter strains from produced fluid samples.

Figure 2.2 Geochemical and microbial community data from four Appalachian Basin natural-gas wells. (A) Relative abundance of the 16S rRNA gene corresponding to Arcobacter and Marinobacter reconstructed from metagenomes using EMIRGE combined with chloride trends in the same Marcellus Shale natural-gas well (Marcellus-4). (B) CO2 and NPOC (non-purgeable organic carbon) trends in the Marcellus Shale natural-gas well (Marcellus-4). Temporal trends in nitrogen (C) and sulfur (D) in produced fluids from Utica-Point Pleasant formation and Marcellus Shale natural-gas wells. Plotted values are averages of 4 Utica-Pt. Pleasant natural-gas wells (Utica-3, Utica-4, Utica-5, Utica-6) and 2 Marcellus natural-gas wells (Marcellus-4 and Marcellus-5); error bars are standard deviations between measurements for Utica wells, range for Marcellus wells. Averages were calculated using the same day after flowback began and converted to days after fracturing. Offset graphs to the left of main graphs indicate measurements in drill muds, + injected fluid, or source waters. NH3/NH4 indicates total ammonia/ammonium which are both measured in this method.

Figure 2.3. SEM images and phylogenetic placement of Marinobacter and Arcobacter.

(A) Marinobacter sp. UTICA-S1B6 and (B) Arcobacter sp. MARC-MIP3H16 and phylogenetic placement of (A) Marinobacter sp. UTICA-S1B6 and (B) Arcobacter spps. MARC-MIP3H16 and UTICA-S4D1. Orange text denotes near-full-length 16S rRNA gene sequences reconstructed from Marcellus Shale natural-gas well Marcellus-4 metagenomes using EMIRGE; blue text denotes near-full-length 16S rRNA gene sequences reconstructed from metagenomes using EMIRGE in a previous study, Marcellus-1; red text denotes full-length 16S rRNA gene sequences from Marcellus or Utica Pt. Pleasant natural- gas well isolates. Branches marked with blue dots indicate bootstrap support greater than or equal to 80%.

Figure 2.4. Salinity growth curves and carbon source utilization by Arcobacter and Marinobacter. A) Superimposed salinity growth rate curves for Arcobacter sp. MARC-MIP3H16 and Marinobacter sp. UTICA-S1B6. Error bars denote standard deviation from triplicate measurements. B) Carbon sources utilized by Marinobacter sp. UTICA-S1B6 and Arcobacter sp. MARC-MIP3H16. Negative controls registered absorbance values of 0.023 and 0.014, respectively. Any absorbance readings under 0.05 were considered negative and assigned a white color in this figure. 80

Figure 2.5. Conceptual metabolic models for A) Marinobacter and B) Arcobacter isolates incorporating isolate genomic (black arrows) and metagenome data (red arrows).

Table 2.1. Identifiers for hydraulically fractured natural-gas wells analyzed and/or summarized in this study.

Sample Geochemical Microbial Well Bacteria Figures Formation days after analyses analyses Identifier isolated present fracturing performed performed

2- Utica-Pt. Marinobacter N-NH3, TN, S , Utica-3 38-460 2- n/a Figure 2 Pleasant sp. UTICA-S1B6 SO4 2- Utica-Pt. N-NH3, TN, S , Utica-4 38-460 n/a 2- n/a Figure 2 Pleasant SO4 2- Utica-Pt. N-NH3, TN, S , Utica-5 38-460 n/a 2- n/a Figure 2 Pleasant SO4 2- Utica-Pt. Arcobacter sp. N-NH3, TN, S , Utica-6 38-460 2- n/a Figure 2 Pleasant UTICA-S4D1 SO4 Marcellus-1 Marcellus 4-328 n/a n/a 16S EMIRGE Figure 3 Arcobacter sp. NH , TN, S2-, 3 Metagenomics Marcellus-4 Marcellus 24-485 MARC- SO 2-, Cl-, NPOC, Figure 2, 3, 5 4 , 16S EMIRGE MIP3H16 CO2 2- NH3, TN, S , Marcellus-5 Marcellus 35-496 n/a 2- n/a Figure 2 SO4

Chapter 3. In situ transformation of ethoxylate and glycol surfactants by shale-colonizing

microorganisms during hydraulic fracturing

In review at the ISME journal

Abstract:

In the last decade, extensive application of hydraulic fracturing technologies to unconventional low-permeability hydrocarbon-rich formations has significantly increased natural gas production in the United States and abroad. The injection of surface-sourced fluids to generate fractures in the deep subsurface introduces microbial cells and substrates to low-permeability rock. A subset of injected organic additives has been investigated for their ability to support biological growth in shale microbial community members; however, to date, little is known on how complex xenobiotic organic compounds undergo biotransformations in this deep rock ecosystem. Here, high-resolution chemical, metagenomic, and proteomic analyses reveal that widely-used surfactants are degraded by the shale-associated taxa Halanaerobium, both in situ and under laboratory conditions. These halotolerant bacteria exhibit surfactant substrate specificities, preferring polymeric propoxylated glycols (PPGs) and longer alkyl polyethoxylates (AEOs) over polyethylene glycols (PEGs) and shorter AEOs. Enzymatic transformation occurs through repeated terminal-end polyglycol chain shortening during co-metabolic growth through the methylglyoxal bypass. This work provides the first evidence that shale microorganisms can transform xenobiotic surfactants in fracture fluid formulations, likely affecting the efficiency of

83 hydrocarbon recovery, and demonstrating an important association between injected substrates and microbial growth in an engineered subsurface ecosystem.

3.1 Introduction

The integration of horizontal drilling and hydraulic fracturing (HF) technologies for the recovery of hydrocarbons from unconventional low permability (“tight”) formations has increased the production of natural gas in the United States and abroad (Arthur et al., 2009). In 2017, the United

States became a net exporter of natural gas for the first time since 1957, producing on average more than 70 billion cubic feet per day (U.S. Energy Information Administration (EIA), 2018).

During HF, fluids are injected at high pressures to induce new and reopen existing fractures in extremely low-permeability (nano-Darcy conductivities) rock formations, increasing production of oil and natural gas (O&G) resources (Arthur et al., 2009; Vidic et al., 2013). Injected fluids are comprised primarily of water, proppant (e.g. fine sand), and organic chemical additives, with formulations optimized by the industry to increase natural gas yields while concurrently protecting well infrastructure (Vidic et al., 2013). Once the petroleum well has been completed and natural gas production commences, injected fluids continue to mix with highly saline formation pore waters, generating produced fluid brines that return to the surface over the course of several months to years (Barbot et al., 2013; Cluff et al., 2014; Haluszczak et al., 2013; Vidic et al., 2013).

The interconnected fractures generated by HF create a new microbial ecosystem through an infusion of surface-derived water, nitrogen, and carbon sources to the shale (Cluff et al., 2014;

Hanson et al., in prep; Mouser et al., 2016). Microbial community dynamics in samples from HF natural-gas wells have been described for several different U.S. black shale formations (Mouser et

84 al., 2016). Within months, the microbial community converges from diverse freshwater-associated taxa to low diversity, anaerobic, halotolerant bacteria and archaea (Mouser et al., 2016; Cluff et al., 2014; Hanson et al., in prep; Daly et al., 2016; Booker et al., 2017; Lipus et al., 2018; Mohan et al., 2013; Struchtemeyer and Elshahed, 2012; An et al., 2017; Akob et al., 2015; Evans et al.,

2018). This microbial community shift occurs as HF-injected carbon and electron acceptors are depleted while salinity becomes enriched from formational brine (Akob et al., 2015; Barbot et al.,

2013; Booker et al., 2017; Chapman et al., 2012; Cluff et al., 2014; Daly et al., 2016; Haluszczak et al., 2013; Struchtemeyer and Elshahed, 2012; Vengosh et al., 2017; Warner et al., 2013). Recent studies have begun to describe the unique metabolisms and adaptations enabling microbial persistence in this engineered subsurface (Booker et al., 2017; Borton et al., 2018; Daly et al.,

2016; Evans et al., 2018; Liang et al., 2016; Lipus et al., 2017), with some influenced by the lability of chemical additives used by the O&G industry to protect the petroleum well and enhance its production (Borton et al., 2018; Daly et al., 2016; Elsner and Hoelzer, 2016; Evans et al., 2018;

Liang et al., 2016).

Organic chemical components of HF fluids play an important, yet under-characterized role in sustaining microbial communities in shales. For example, urea is a naturally-derived additive that reduces friction in fracturing fluids (Ahrenst et al., 2015), and may be consumed as a carbon and nitrogen source through conversion to CO2 and NH3 by taxa present in the first few months in HF shale gas wells (e.g. Arcobacter, Marinobacter) (Evans et al., 2018). Additionally, the vitamin choline (added as choline chloride) is the most frequently disclosed clay stabilizer used by the industry, and may fuel a community metabolism culminating in biogenic production of osmoprotectants and ultimately methanogenesis (Daly et al., 2016; Elsner and Hoelzer, 2016).

Furthermore, the gelling agent guar gum, a polysaccharide derived from plants, is readily degraded by Halanaerobium, which couples guar gum fermentation to sulfide production months to years after HF occurs (Liang et al., 2016). The above referenced studies focus on the biotransformation of labile or naturally-derived compounds, underscoring the need to better understand metabolisms influencing xenobiotic compound degradation during and after HF operations (Elsner and Hoelzer,

2016).

Xenobiotic polyglycols are common and essential components of HF fluids in natural-gas wells

(up to 49.7%) as crosslinkers, scale inhibitors, solvents, and surfactants (see Table B1 in Appendix

B) (Elsner and Hoelzer, 2016; Rogers et al., 2015). Specifically, alkyl polyethoxylates (AEOs) containing a branched or linear alkyl hydrocarbon connected to repeating ethoxylate chains have been identified in produced fluid samples from HF petroleum wells (Rosenblum et al., 2017;

Thurman et al., 2014; Nell and Helbling, 2018). In one study examining fluids produced from the

Niobrara Formation, AEO relative concentrations decreased as the O&G well matured, but the mechanism(s) responsible for AEO changes were not thoroughly explored (Rosenblum et al.,

2017). One possible explanation for surfactant losses in HF shale wells is biological attenuation.

Anaerobic degradation of glycols is accomplished via stepwise cleavage of ethoxylate units catalyzed by the propanediol dehydratase gene cluster, generating alcohols, carboxylic acids, and ketone metabolites (Heyob et al., 2017; Huber et al., 2000; Wagener and Schink, 1988). This biotransformation pathway has not been investigated in situ during HF operations, but was described in groundwater microcosms simulating spills of fracturing and produced fluids, highlighting environmental impacts that might arise from an accidental release of these constituents (Heyob et al., 2017; Rogers et al., 2018). Moreover, this surfactant degradation

86 mechanism may have a negative effect on hydrocarbon recovery due to unintended changes in produced fluid chemistry, as the transformation pathway can (1) reduce concentrations of injected chemicals resulting in a loss of chemical efficacy, in addition to (2) produce corrosive organic acids that damage well infrastructure (Fichter et al., 2012). Biological degradation of surfactants may help explain the aforementioned surfactant trends; nonetheless, the environmental and industrial consequences of this pathway are significant.

To determine whether xenobiotic polymers are attenuated by microorganisms in HF systems, we analyzed changes in unsubstituted and alkylated polyglycols in fluid samples collected up to 204 days after production began in a Utica-Point Pleasant Formation natural-gas well in Ohio, U.S.A.

Temporal changes in surfactant chemistry were associated with putative biotransformation genes from produced fluid metagenomes. To confirm the capacity of relevant taxa to enzymatically transform these xenobiotic compounds, we applied genome sequencing, proteomics, and metabolite analysis in laboratory batch experiments, uncovering a previously uncharacterized, co- metabolic pathway for surfactant chain shortening by the halotolerant bacterial strain

Halanaerobium congolense WG10. Our results show that key microbial strains are capable of transforming xenobiotic organic additives in HF natural-gas wells.

3.2 Materials and Methods

3.2.1 Produced water sampling and pre-processing

Fluid samples were recovered from a natural-gas well drilled in the Utica-Point Pleasant

Formation (2.6 km depth) in eastern Ohio between July 2014 and February 2015 as reported previously (Booker et al., 2017). Due to a ~3 month “shut-in”, the first fluid sampling occurred

86 days after HF. Samples were collected from the gas-water separator in 1 L sterile Nalgene

HDPE bottles (Thermo-Fisher Scientific, Waltham, MA) with no headspace. The flow rate in the separator ranged from 190 000 to 380 000 L day-1, resulting in a maximum 8 hour residence time before any sampling event (Booker et al., 2017). Biomass from produced fluid samples (300 to 1

000 mL) was concentrated onto 0.22 µm PES filters (Nalgene, Thermo-Fisher Scientific,

Waltham, MA) within 24 hours of sampling and stored at -80 °C until further use (Booker et al.,

2017). Filtrate was collected and preserved for ion (4 °C), elemental (4 °C, nitric acid), dissolved carbon analysis (4 °C, HCl), and surfactants (-20 °C) as described below and in the Supplemental methods located in Appendix B.

3.2.2 Bacterial growth experiments

Surfactant biodegradation experiments were performed using H. congolense WG10 in SWDM media (see the Supplemental Methods in Appendix B) containing 10 mM D-Glucose and amended with a commercial surfactant additive “Revert Flow”, which contains C6-C12 alcohol ethoxylates, isopropyl alcohol, orange terpenes, and a polyethylene glycol-polypropylene glycol (PEG-PPG) co-block polymer (see Table B2 in Appendix B). Revert Flow was added to culture samples at concentrations of approx. 150 mg/L total organic carbon (TOC), with AEOs and PEG-PPGs each between 15-45% of total Revert Flow mass according to the manufacturer (see Table B2 in

Appendix B). Biotic samples (containing Revert Flow and 10 mM D-glucose) and glucose controls

(10 mM D-glucose) were cultured in triplicate using a 10% inoculum. Killed controls, conducted in duplicate, were initially grown to mid-log phase with 10 mM D-glucose, heat killed using an autoclave, then transferred to fresh media (10%). Abiotic controls (in duplicate) contained a 10% transfer of sterile SWDM. All transfers were performed under anoxic conditions (80% N2, 20%

CO2) and were incubated at 37°C in the dark. Growth was monitored using optical density readings

(OD, 610 nm) on a Hach DR900 Colorimeter (Loveland, Colorado). Samples were collected aseptically immediately after inoculation/transfer (t0), at mid-log phase (tm) when OD reached ~0.3

(25.5 hours for glucose controls, 43 hours for surfactant amended cultures), stationary phase (ts)

(51 hours for both treatments), and death phase (td) (66 hours for both treatments). Dissolved constituents were filtered (0.22 μM PES, EMD Millipore, Burlington, MA) for GC and LC analyses (see Table B3 and Supplemental Methods in Appendix B), while unfiltered samples were used for ATP analysis (see Supplemental Methods in Appendix B).

3.2.3 Shotgun proteomic analysis

Biomass for shotgun proteomic analysis was collected from biological triplicates at mid-log phase and pelleted by centrifugation at 10,000 x g for 10 minutes at 4 °C. Pellets were immediately flash frozen and stored at -80 °C until shipment to Pacific Northwest National Laboratory for analysis.

Protein extraction, digestion into peptides, and analysis on a 2D ACQUITY ultra high- performance liquid chromatography (UPLC) M-class system (Waters, Milford, MA) was performed as previously described (Booker et al., 2017; Wiśniewski et al., 2009). Measured peptides were searched against predicted peptides derived from the H. congolense WG10 genome.

Resulting peptide identifications were filtered via MS-GF+ using a Q value of ≤0.01 (Kim and

Pevzner, 2014) and a reverse decoy database search using the reported filters generated an FDR of

0.9%. For comparative analyses of replicates, protein spectral counts were normalized by the normalized spectral abundance frequency (NSAF) method (Paoletti et al., 2006) and Z-score values were calculated from the mean protein abundance across all conditions as previously reported (Booker et al., 2017). Values between glucose controls and surfactant grown cells were

89 considered significantly different if the difference in the protein z-score was ≥1.65 (90% confidence).

3.2.4 Polyglycol analysis

AEOs (C4, nC6, C6, C8) and PEGs from field samples were analyzed using a high-performance liquid chromatograph coupled to a high-resolution mass spectrometer (LTQ-Orbitrap Velos with positive polarity electrospray ionization (ESI)) (Thermo-Fisher Scientific, Waltham, MA) at Duke

University. Chromatographic separation of branched and linear AEOs (e.g. nC6 vs. C6) was achieved (see Figure B1 in Appendix B). Sample preparation, instrument methodology, and identification methods have been previously reported (Getzinger et al., 2015). Limits of detection and quantification were determined for each measured species (see Table B4 in Appendix B).

C8 AEOs and PPGs in laboratory samples were analyzed using UPLC with quadrupole time-of- flight mass spectrometry (qTOF-MS) at the Colorado State University Central Instrument Facility

(see Figure B2 in Appendix B) (Heyob et al., 2017). The separation and identification methods have been previously reported (Hanson et al., 2019) with the following exceptions: C8 alkyl polyethoxylates (C8 AEOn) and polypropylene glycols (PPGn) were manually quantified on the

+ most abundant adduct [M+NH4] , except for PPGs with propoxymers n=4-8, which were quantified on the more abundant adduct [M+Na]+ (see Table B5 in Appendix B). A calibration

2 curve (R =0.9879) was constructed using a C10EO8 standard (Sigma Aldrich) for quality assurance and to establish a linear range for similar species for accurate semi-quantitation based on relative abundances (see Table B5 and B6 in Appendix B). For both field and culture sample analysis, external calibrants for individual species or mixtures with known species concentrations were

90 unavailable, therefore a semi-quantitative approach was employed which normalized all AEO,

PEG, and PPG species to the starting concentrations of each individual polyglycol, as has been reported previously (see Table B7 in Appendix B) (Heyob et al., 2017; Hanson et al., 2019;

McLaughlin et al., 2016; Rogers et al., 2018; Rosenblum et al., 2017).

3.2.5 Metagenomic sequencing and analysis

DNA was extracted from produced fluid filters and sequenced as previously reported (Booker et al., 2017). Metagenomes in IMG/M were assembled by the Joint Genome Institute using

MEGAHIT v. 1.0.3 and annotated using the IMG Annotation Pipeline v.4.10.0 (Huntemann et al.,

2016). Gene inquiry for isolate genomes and metagenomes were performed using IMG/M from

JGI (Chen et al., 2017). Genes in H. congolense WG10 were identified using annotated EC numbers and by comparison to characterized genes in Acetobacterium woodii (see Supplemental

Methods in Appendix B). Genes in assembled metagenomes were identified through comparison to H. congolense WG10 using BLASTp (within IMG/M) with quality cutoffs (≥90% identity and bitscore ≥200) (see Tables B8 and Figure B3 in Appendix B) (Altschul et al., 1990; Chen et al.,

2017).

3.2.6 Data Accession

Metagenomic sequence data from this study is available under NCBI BioProject ID PRJNA433267 and the Joint Genome Institute (JGI) Integrated Microbial Genomes and Microbiomes (IMG/M) database. Genomic sequence data is available under the JGI-IMG/M database under project number Ga0073285. Proteomic data is available in supplemental information (see Appendix C).

3.3 Results & Discussion

3.3.1 Surfactants attenuate in situ during the months following hydraulic fracturing

Five different polyglycols were detected in the produced fluid of this Utica-Point Pleasant natural- gas well, with relative abundances varying through time based on surfactant class and chain length.

Branched C6 AEOs, containing two to nine ethoxylate units (EO2 to EO9) decreased by 97% on average between the first day of flowback (day 86) and later production at day 204 (Figure 3.1a).

During earlier flowback (days 86-92 after HF), branched C6 AEOs with short ethoxylate chain lengths (EO3, EO4) doubled in abundance concurrent with approximately 50% loss for longer chains by day 94 (EO5-EO9) (Figure 3.1b). This trend is consistent with enzymatic chain shortening of linear AEOs under anaerobic conditions (Huber et al., 2000; Wagener and Schink,

1988). Pseudo-first order rate constants based on temporal changes in surfactant relative abundances were used to estimate half-lives for individual ethoxylate lengths and ranged from 10 and 26 days for branched C6 AEOs (Figure 3.1c). In contrast to the trends described for branched

C6 AEOs, linear structures with the same alkyl length (nC6 AEO) decreased only 35% between days 86 and 204, resulting in half-lives on the order of 52 to 106 days (Figure 3.1a, 3.1c). Shorter alkyl lengths (C4 AEOs) were removed more slowly (Figure 3.1d) resulting in significantly longer half-lives than other AEOs (115 to 990 days) (Figure 3.1c). On the other hand, C8 AEOs decreased

92 the quickest, with 99.7% reduction by day 204, and the shortest half-life of all detected AEO species (10-20 days) (Figure 3.1a, 3.1c). Here, branched C6 and C8 AEOs attenuated at a considerably faster rate than linear C6 or C4 AEO structures, indicating a clear preference for the longer, branched alkyl chains over shorter or linear structures in this system. To further support this point, we also tracked polyethylene glycols (PEGs) through time, which contain an ethoxylate backbone without an alkyl chain. PEGs decreased only 32% by day 204, with estimated half-lives between 65 and 266 days (Figure 3.1c, 3.1d). Similarly, monomeric ethylene glycol only decreased

17% between days 86 and 204, while monomeric propylene glycol decreased 55% during this time

(see Figure B4 in Appendix B). Earlier studies have revealed similar temporal declines in specific

AEOs compared to PEGs, yet little attention has been given to examining potential causes other than sorption or mixing effects (Rosenblum et al., 2017).

3.3.2 Geochemical mixing model discerns physical from biochemical surfactant trends

To elucidate whether observed field trends in AEOs and PEGs were a result of geochemical mixing, we applied an end-member mixing model to inorganic elements and ions measured in the input and produced fluid samples using input fluid and day 204 as end-members. Mg, Ca, Sr, and

Br were strongly related to Cl in formation waters (Spearman’s >0.94), suggesting a conservative mixing model was appropriate to discern physical (i.e. dilution) effects in this system (see Table

B11 in Appendix B) (Warner et al., 2013). Ten days after flowback began (day 96), Li/Cl molar ratios approached a plateau near 2x10-3 (see Figure B5 in Appendix B), indicating fluid mixing effects within the formation had diminished. In addition, pH and conductivity measurements were consistent for the first 12 days after flowback began (see Table B9 in Appendix B), suggesting solubility variations from redox changes were minimal during this period of time. Throughout the

93 temporal flowback series, Cl and Sr increased by 1.5 and 2.8 fold, respectively (see Figure B5 in

Appendix B). By comparison, the abundance of branched C6 and C8 AEOs decreased monotonically through time by 2 orders of magnitude, while PEGs, C4 and nC6 AEOs remained steady or increased during the flowback period (Figure 3.1). If trends in surfactant abundance were due solely to dilution, we would expect to see the same magnitude of change in both inorganic ions and organic constituents in the absence of appreciable pH/redox driven changes in anticipated adsorption, which was not the case. These combined trends indicate dilution and/or mixing could not completely explain surfactant temporal trends, and that another attenuation mechanism occurred.

We considered whether decreases in branched C6 and C8 AEOs could be due to sorption effects from interactions with shale (Rosenblum et al., 2017), but ruled this out as the primary loss mechanism based on two possible explanations. Firstly, as observed here, AEOs with shorter ethoxylate chains increased in abundance before decreasing (e.g. branched C6 EO3, EO4 between day 86 and 89, Figure 3.1b), while other chain lengths for the same alkyl group monotonically decreased (e.g. branched C6 EO2, and EO5-EO9). A system dominated by sorption should show similar trends for ethoxylate units of the same alkyl structure (e.g. decreases, not increases).

Secondly, if we consider partitioning of surfactants from an aqueous phase (fracture fluid) into an organic phase (e.g. kerogen within source rock), we would expect the more hydrophobic surfactant

(higher alkyl chain lengths, linear structures) to have a greater attraction for other organic phases than the less hydrophobic surfactant (lower alkyl chain lengths, branched structures). Sorption may explain the rapid losses in C8 AEOs relative to C4 AEOs or PEGs; however, the opposite trend is observed in the linear versus branched isomers of the C6 AEOs. Specifically, the slightly less

94 hydrophobic branched C6 AEOs decreased to a greater extent than the more hydrophobic linear nC6 AEOs at all observed time points (Figure 3.1a, see Table B4 in Appendix B). It is therefore unlikely the surfactant trends described here are dictated solely by differences in hydrophobicity or sorption to an organic phase. Sorption to cell biomass is also unlikely to dictate surfactant trends as viable cells rarely exceeded 106 cells/ mL in produced fluid (Daly et al., 2016). Altogether these data suggest a third mechanism, such as biotransformation, may govern surfactant fate in this system.

3.3.3 Metagenomic identification of surfactant degrading genes

Enzymatic chain shortening of AEOs, PEGs, and ethylene glycol can occur under anaerobic conditions through the propanediol dehydratase gene family (pduCDE). The biochemical reaction generates acetaldehyde, which can be dismutated to ethanol and acetate by an aldehyde dehydrogenase (pduP) (Huber et al., 2000; Trifunović et al., 2016) (Figure 3.2a). Although there is currently no characterized enzyme for the biotransformation of PPGs under anaerobic conditions, propylene glycol, the monomer of PPGs, can be degraded by a diol dehydratase under both aerobic and anaerobic conditions, generating propionaldehyde, n-propanol, propionate, and acetone as products (Bobik et al., 1999; Booker, 2018; Heyob et al., 2017; Schuchmann et al.,

2015). We therefore mined metagenomes from Utica-Point Pleasant produced fluid samples for genes encoding enzymes known to degrade polyglycols to determine whether the aforementioned microbial biotransformation pathway existed in this system.

We detected the surfactant biotransformation gene, pduC, in produced fluid metagenomes throughout the lifetime of the Utica-Point Pleasant natural-gas well (see Figure B3 in Appendix

B) (Altschul et al., 1990). Observed pduC genes were highly similar (top hit, bit score >100, identity >90%) to dominant Halanaerobium, Thermoanaerobacter, and Clostridiales (see Figure

B3 in Appendix B) previously identified as key microbial community members in this natural-gas well (Booker et al., 2017). Despite the association of pduC to these three different bacterial genera,

86.5% of all pduC genes mapped with high homology (bit-score >70, identity % >70) to

Halanaerobium, and all pduC genes detected after day 112 were matched to Halanaerobium (see

Figure B3 in Appendix B). The halotolerant taxa Halanaerobium commonly dominates the microbial community in later produced fluids from HF natural-gas wells, making it an important model organism for understanding biogeochemical changes in the highly saline, HF shale ecosystem (Booker et al., 2017; Cluff et al., 2014; Daly et al., 2016; Liang et al., 2016; Lipus et al., 2017; Mohan et al., 2013). Since Halanaerobium was persistent in this natural-gas well and its reconstructed genome contained putative surfactant degrading genes, we focused our search to genes within metagenomic data having high homology to this genus.

We compared metagenomic data to a genome sequenced from a Halanaerobium strain isolated from the Utica-Point Pleasant natural-gas well on day 140 (Booker et al., 2017). The isolate, H. congolense WG10, contained several important surfactant degrading genes with homology (bit score >200, identity > 30%) to a known glycol-degrading bacterium, Acetobacterium woodii

(Figure 3.2c) (see Table B8 in Appendix B) (Trifunović et al., 2016). The normalized metagenomic gene counts for the propanediol dehydratase gene family (pduCDE), aldehyde dehydrogenase

(pduP), and iron-containing alcohol dehydrogenase (Fe-ADH) paralleled the increasing abundance of Halanaerobium as the natural-gas well matured, except on day 204 where gene absence is believed to be related to viral predation of Halanaerobium (Figure 3.2b) (Daly et al., 2016, 2018).

The decreasing abundance of branched C6 and C8 AEOs in the natural-gas well corresponds with higher pduCDE gene abundances (Figure 3.1a, Figure 3.2b), suggesting Halanaerobium may be primarily responsible for surfactant chain shortening in this system. In addition to detecting the putative enzymes in the most dominant bacterial taxa, we also identified four possible metabolites that could result from polyglycol chain shortening in produced fluid using LC-MS and the MetFish approach (Xu et al., 2019), including aldehydes, carboxylic acids, alcohols, and a ketone (see

Figure B4 in Appendix B), further supporting the possibility that this biotransformation pathway is occurring in situ. Metabolite data generated by LC-MS analysis was further supported by NMR analysis (see Figure B4 in Appendix B).

3.3.4 Surfactant degradation observed in isolate cultures

Given the likely presence of a surfactant degrading metabolism in Halanaerobium, we tested the ability of the isolate H. congolense WG10 to metabolize a commercially-available surfactant mixture (Revert Flow, see Table B2 in Appendix B) containing PPGs and C6-C12 AEOs by tracking changes in surfactant chemistry, protein abundances, and metabolite production during batch growth (see Figure B6 in Appendix B). As this strain did not grow on the commercially available surfactant mixture, PEGs, ethylene glycol, or propylene glycol as a sole carbon source, surfactant treatments were amended with D-glucose and compared with a glucose control (see Figure B7 in

Appendix B).

Total PPGs decreased in surfactant treatments by 19% with no PPG losses observed in killed or abiotic controls (Figure 3.3). PPGs containing higher propoxylate (PO) chain lengths (PO6 through

PO10) displayed more significant decreases in relative concentration (26-47%) as compared with

97 shorter propoxylate lengths (PO4 and PO5) (8% and 12%, respectively), a trend which may be attributed to the shortening of longer PPG chains (Figure 3.3). Metabolite data showed significantly higher propionate concentrations in surfactant treatments (0.54 mM) compared to glucose controls during the death phase (Figure 3.4), supporting the enzymatic production of carboxylic acids from PPG chain shortening. To our knowledge, H. congolense WG10 represents the first report of a microbial isolate transforming PPG under anaerobic conditions.

Biotransformation of AEOs also occurred during growth of H. congolense WG10 on surfactants.

Total AEOs decreased in relative concentration by 74%, with losses in several ethoxylate chain lengths (EO4 through EO7) significantly lower than killed and abiotic controls (Figure 3.3).

Acetaldehyde/propionaldehyde, the initial product(s) of AEO/PPG chain shortening, had a higher concentration in the surfactant treatment (>0.6 mM) at stationary phase compared to glucose controls, which never surpassed 0.2 mM at any time point (Figure 3.4). Acetate concentrations were significantly higher during mid-log phase in surfactant treatments; however, acetate continued to increase in glucose controls during the death phase whereas concentrations plateaued in surfactant treatments (Figure 3.4). No significant differences in concentration of ethanol/propanol were detected between the two treatments (Figure 3.4). It is important to note that we measured considerable losses in the relative concentration of AEOs in abiotic and killed controls (35% and 64%, respectively, Figure 3.3), which we attributed to interaction with growth media reagents (abiotic) and/or sorption to dead biomass (killed) estimated at 109 cells/mL. It is also possible high salt concentrations caused matrix suppression during analysis of AEO species using mass spectrometry (Nell and Helbling, 2018); however, media conditions were held constant, therefore these effects are similar across treatments.

Using shotgun proteomics, we sought to determine which biogeochemical pathway was responsible for surfactant transformations in H. congolense WG10. Protein relative abundance for each of the three genes in the propanediol dehydratase gene cluster pduCDE (Gene 6-8 in Figure

3.4/Ga0072835_105104-105106) was significantly higher in surfactant grown cultures as compared with glucose controls (Figure 3.4, see Figure B8 in Appendix B). The activity of pduCDE generates one mole of an aldehyde per ethoxylate (AEO) or propoxylate (PPG) cleaved.

As aldehydes are toxic to bacterial cells, they are immediately captured by a bacterial microcompartment (BMC) before conversion to propionyl/acetyl CoA using an aldehyde dehydrogenase (pduP, Gene 9/105116) (Bobik et al., 1999). Here we saw a significant increase in the relative abundance of proteins associated with BMCs (105102, 105103, 105110) as well as overall increases in pduP for surfactant grown cells (Gene 9/105116) (see Figure B8 in Appendix

B). In addition, proteins responsible for converting propionyl/acetyl-phosphate to their corresponding carboxylic acids (propionate/acetate) were identified at significantly higher relative abundances in surfactant grown cultures compared to glucose controls (Gene 11/10188) (Figure

3.4, see Figure B8 in Appendix B). Furthermore, at least eleven proteins associated with Vitamin

B12 synthesis, a key cofactor during pduCDE activity, were observed at significantly higher levels for cells grown on surfactants (see Figure B8 in Appendix B) (Bobik et al., 1999).

In addition to increased abundance of proteins associated with AEO and PPG chain shortening, we also observed clear differences in methylglyoxal synthase (Gene 2/11740), methylglyoxal reductase (Gene 4/10770), and (S)-lactaldehyde reductase (Gene 5/10563) protein abundance for surfactant grown cultures (Figure 3.4, see Figure B8 in Appendix B). These three enzymes convert

99 dihydroxyacetone phosphate derived from glucose to methylglyoxal, next reducing it to lactaldehyde, and finally producing propylene glycol. Propylene glycol is then taken into the propanediol dehydratase gene cluster (pduCDE) for conversion to carboxylic acids and alcohols as described previously. When grown in the presence of glycol-containing compounds, we infer that H. congolense WG10 utilizes the methylglyoxal bypass during glycolysis, thereby “turning up” the pathway for propanediol dehydratase. Consequently, surfactant grown cells likely shuttle more glucose through the methylglyoxal bypass relative to glucose controls. This would culminate in decreased glycolytic activity and increased production of aldehydes

(acetaldehyde/propionaldehyde) and carboxylic acids (acetate/propionate) from the generation of propylene glycol and from shortening of AEO and PPG chains (Figure 3.4).

Research by others suggests the methylglyoxal bypass is important for bacteria grown under stress or nutrient limited conditions, including excess organic carbon (Booth et al., 2003; Cameron and

Cooney, 1986; Chandrangsu et al., 2014; Tran-Din and Gottschalk, 1985). Although carbon was not limited in our experiments, we observed significantly higher abundances of the universal stress protein (uspA, 10238) in surfactant amended cultures (see Appendix C). Moreover, in a recent study on a closely-related Halanaerobium strain isolated from the same natural-gas well (H. congolense WG8), the methylglyoxal bypass was initiated under growth at high pressures with glucose as the sole carbon source, possibly to dispose of excess reducing equivalents (Booker,

2018). Here, we propose co-metabolism of glycols occurs either through cellular stress and/or high concentrations of glycols within the media, activating both the methylglyoxal bypass and the propanediol dehydratase pathways.

100

Our combined field and laboratory data provide evidence for several important findings regarding bacteria living in highly saline anaerobic systems. Firstly, the halotolerant bacterial strain isolated from natural-gas well produced fluids, H. congolense WG10, can co-metabolize two commonly used surfactant classes, namely, AEOs and PPGs, when grown on a labile substrate. Importantly, the pduCDE gene cluster in H. congolense WG10 acts on monomeric ethylene glycol, PPGs, and

AEOs, but not PEGs (see Figure B7 in Appendix B), indicating a predilection for specific polymers. Our field data suggest that enzymatic preference extends to alkyl length (C8 and C6 over

C4 AEOs) and structure (branching over linear AEOs, and AEOs over PEGs). This finding is surprising in light of previous work which described branched AEOs as significantly more recalcitrant to biodegradation than linear AEOs (Mösche, 2004). Here, the presence of glucose is key to surfactant biotransformation through precursor activation of the propanediol dehydratase gene cluster after sending fructose-1,6-biphosphate through the methylglyoxal bypass. The propanediol dehydratase gene family (pduCDE) as well as the aldehyde and alcohol dehydrogenases (pduP, Fe-ADH) were detected in metagenomes assembled from fluids produced from the Utica-Point Pleasant natural-gas well (Figure 3.2b). However, not all Halanaerobium strains contain these genes (see Table B10 in Appendix B), suggesting that this pathway may be unique to certain strains occupying specific niches in the HF shale ecosystem.

During HF operations, both labile carbon and recalcitrant xenobiotic carbon compounds are injected into the subsurface along with microbial cells. Guar gum and other polysaccharides are frequently disclosed in fracture fluid formulations, and are readily metabolized by microorganisms in produced fluids (Daly et al., 2016; Elsner and Hoelzer, 2016; Liang et al., 2016). In particular,

Halanaerobium is capable of cleaving the mannose and galactose monomers from guar gum,

101 yielding galactose which can be converted to glucose-1-phosphate for use in glycolysis (Liang et al., 2016). Guar gum transformation may therefore initiate the methylglyoxal bypass in a similar way to glucose, triggering the propanediol dehydratase pathway and resulting in chain shortening of surfactants in produced fluid. The production of carboxylic acids and alcohols associated with this co-metabolic pathway in Halanaerobium may provide bottom-up support to other shale- associated microorganisms (Borton et al., 2018). Moreover, the biological depletion of injected surfactants in parallel with the acidification of produced fluids from accumulated metabolites may lead to sub-optimal hydrocarbon extraction, advanced petroleum well corrosion, and an overall lower rate of return for the petroleum industry (Fichter et al., 2012).

3.4 Conclusion

Specific classes of AEOs were rapidly attenuated during the first few months after HF of a natural- gas well. Although this trend has been observed in other HF petroleum wells, the mechanism(s) behind surfactant losses were not well understood. In this Utica-Point Pleasant natural-gas well,

Halanaerobium dominated the microbial community structure within months, despite the decreasing concentrations of labile carbon and increasing salinity. An isolate from this natural-gas well (H. congolense WG10) transformed AEOs and PPGs in the presence of glucose, utilizing the methylglyoxal bypass during glycolysis to initiate chain-shortening enzymes. The co-metabolic pathway described here may be initiated by other polymeric compounds that enter glycolysis (e.g. guar gum), suggesting this pathway may inadvertently occur until polymers are completely transformed. In contrast to the surfactant types attenuated in field and laboratory findings (AEOs,

PPGs, monomer glycols), PEGs were not degraded in the Utica-Point Pleasant natural-gas well produced fluids nor by H. congolense WG10, suggesting that PEGs may be useful tracers of injected hydraulic fracturing fluid in this system. Our findings highlight the versatility of the 102 halotolerant bacterial taxa, Halanaerobium, to metabolize xenobiotic organic additives injected during HF of natural-gas wells.

Author contributions:

P. Mouser and M. Evans designed the study and conceived of the experiments with insight from

M. Wilkins and K. Wrighton. P. Mouser designed the mixing model. M. Evans performed laboratory experiments with the bacterial isolate, analyzed culture organic acids and alcohols, analyzed culture polyglycols, compiled all data, and wrote the manuscript. G. Getzinger analyzed field samples for polyglycols with insight from P. L. Ferguson. J. Luek performed supplemental organic acid analysis. A. Hanson advised on culture polyglycol analysis and organic acid analysis.

T. Darrah advised on analysis methods for organic acids. M. McLaughlin and J. Blotevogel assisted with methods development and advised with culture PEG and polyglycol analyses. D.

Cole and S. Welch contributed geochemical data. C. Nicora, S. Purvine, M. Lipton performed shotgun proteomics. D. Hoyt, T. Metz, and C. Xu analyzed metabolites in produced fluids. M.

Wilkins provided the bacterial isolate and advised on the manuscript.

103

Acknowledgements:

This research was supported by funding from the National Science Foundation Dimensions of Biodiversity (award no. 1830742). We express thanks to Mikayla Borton, Kelly

Wrighton, and Rebecca Daly for guidance on genomic analyses, and Desiree Plata and

Nathaniel Warner for helpful conversations on chemical mixing analysis. We are especially grateful for cooperation from our industry partner for site access and sample support.

Genomic and metagenomic sequencing for this research was performed by the Department of Energy’s Joint Genome Institute (JGI) via a large-scale sequencing award (no. 1931).

NMR metabolite support was provided by the Environmental Molecular Sciences

Laboratory (EMSL) via a JGI-EMSL Collaborative Science Initiative (award no. 48483).

LC-MS metabolite analysis using MetFish was performed at the Pacific Northwest

National Laboratory (PNNL) through support from the Department of Energy’s Genomic

Science Program via the Metabolic and Spatial Interactions in Communities Scientific

Focus Area. The JGI, EMSL, and PNNL are sponsored by the Office of Biological and

Environmental Research and operated under contracts DE-AC02-05CH11231 (JGI) and

DE-AC05-76L01830 (EMSL and PNNL).

104

3.5 References

Ahrenst, A., Lungwitz, B., Fredd, C. N., Abad, C., Gurmen, N., Chen, Y., et al. (2015). Friction reduction fluids. Available at: http://www.freepatentsonline.com/9034802.html [Accessed January 4, 2019].

Akob, D. M., Cozzarelli, I. M., Dunlap, D. S., Rowan, E. L., and Lorah, M. M. (2015). Organic and inorganic composition and microbiology of produced waters from Pennsylvania shale gas wells. Applied Geochemistry 60, 116–125. doi:10.1016/j.apgeochem.2015.04.011.

Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990). Basic local alignment search tool. J. Mol. Biol. 215, 403–410. doi:10.1016/S0022- 2836(05)80360-2.

Arthur, J. D., Bohm, B., and Layne, M. (2009). Hydraulic fracturing considerations for natural gas wells of the Marcellus Shale. Available at: http://archives.datapages.com/data/gcags_pdf/2009/arthetal.htm [Accessed August 2, 2017].

Bobik, T. A., Havemann, G. D., Busch, R. J., Williams, D. S., and Aldrich, H. C. (1999). The propanediol utilization (pdu) operon of Salmonella enterica serovar Typhimurium LT2 includes genes necessary for formation of polyhedral

105

organelles involved in coenzyme B(12)-dependent 1, 2-propanediol degradation. J. Bacteriol. 181, 5967–5975.

Booth, I. R., Ferguson, G. P., Miller, S., Li, C., Gunasekera, B., and Kinghorn, S. (2003). Bacterial production of methylglyoxal: a survival strategy or death by misadventure? Biochemical Society Transactions 31, 1406–1408. doi:10.1042/bst0311406.

Borton, M. A., Hoyt, D. W., Roux, S., Daly, R. A., Welch, S. A., Nicora, C. D., et al. (2018). Coupled laboratory and field investigations resolve microbial interactions that underpin persistence in hydraulically fractured shales. PNAS 115, E6585. doi:10.1073/pnas.1800155115.

Cameron, D. C., and Cooney, C. L. (1986). A novel germentation: The production of R(– )–1,2–propanediol and acetol by Clostridium thermosaccharolyticum. Bio/Technology 4, 651.

Chandrangsu, P., Dusi, R., Hamilton, C. J., and Helmann, J. D. (2014). Methylglyoxal resistance in Bacillus subtilis: Contributions of bacillithiol-dependent and independent pathways. Mol Microbiol 91, 706–715. doi:10.1111/mmi.12489.

Chapman, E. C., Capo, R. C., Stewart, B. W., Kirby, C. S., Hammack, R. W., Schroeder, K. T., et al. (2012). Geochemical and strontium isotope characterization of

106

produced waters from Marcellus Shale natural gas extraction. Environ. Sci. Technol. 46, 3545–3553. doi:10.1021/es204005g.

Daly, R. A., Roux, S., Borton, M. A., Morgan, D. M., Johnston, M. D., Booker, A. E., et al. (2018). Viruses control dominant bacteria colonizing the terrestrial deep biosphere after hydraulic fracturing. Nature Microbiology. doi:10.1038/s41564- 018-0312-6.

Evans, M. V., Panescu, J., Hanson, A. J., Welch, S. A., Sheets, J. M., Nastasi, N., et al. (2018). Members of Marinobacter and Arcobacter influence system biogeochemistry during early production of hydraulically fractured natural gas wells in the Appalachian Basin. Front. Microbiol. 9, 2646. doi:10.3389/fmicb.2018.02646.

107

Hanson, A. J., Luek, J. L., Tummings, S. S., McLaughlin, M. C., Blotevogel, J., and Mouser, P. J. (2019). High total dissolved solids in shale gas wastewater inhibit biodegradation of alkyl and nonylphenol ethoxylate surfactants. Science of The Total Environment 668, 1094–1103. doi:10.1016/j.scitotenv.2019.03.041.

Hanson A.J., Lipp J.S., Daly R.A., Sheets J.M., Welch S.A., Sharma S., et al. Microbial membrane lipids show de novo ecosystem transformation in hydraulically fractured Marcellus Shale. In preparation.

Heyob, K. M., Blotevogel, J., Brooker, M., Evans, M. V., Lenhart, J. J., Wright, J., et al. (2017). Natural attenuation of nonionic surfactants used in hydraulic fracturing fluids: Degradation rates, pathways, and mechanisms. Environ. Sci. Technol. 51, 13985–13994. doi:10.1021/acs.est.7b01539.

Huber, M., Meyer, U., and Rys, P. (2000). Biodegradation mechanisms of linear alcohol ethoxylates under anaerobic conditions. Environ. Sci. Technol. 34, 1737–1741. doi:10.1021/es9903680.

108

Kim, S., and Pevzner, P. A. (2014). MS-GF makes progress towards a universal database search tool for proteomics. Nat Commun 5: 5277.

Lipus, D., Roy, D., Khan, E., Ross, D., Vikram, A., Gulliver, D., et al. (2018). Microbial communities in Bakken region produced water. FEMS Microbiol Lett 365. doi:10.1093/femsle/fny107.

McLaughlin, M. C., Borch, T., and Blotevogel, J. (2016). Spills of hydraulic fracturing chemicals on agricultural topsoil: Biodegradation, sorption, and co-contaminant interactions. Environ. Sci. Technol. 50, 6071–6078. doi:10.1021/acs.est.6b00240.

109

Mösche, M. (2004). Anaerobic degradability of alcohol ethoxylates and related non-ionic surfactants. Biodegradation 15, 327–336.

Paoletti, A. C., Parmely, T. J., Tomomori-Sato, C., Sato, S., Zhu, D., Conaway, R. C., et al. (2006). Quantitative proteomic analysis of distinct mammalian Mediator complexes using normalized spectral abundance factors. PNAS 103, 18928. doi:10.1073/pnas.0606379103.

Rogers, J. D., Burke, T. L., Osborn, S. G., and Ryan, J. N. (2015). A framework for identifying organic compounds of concern in hydraulic fracturing fluids based on their mobility and persistence in groundwater. Environ. Sci. Technol. Lett. 2, 158– 164. doi:10.1021/acs.estlett.5b00090.

110

Schuchmann, K., Schmidt, S., Lopez, A. M., Kaberline, C., Kuhns, M., Lorenzen, W., et al. (2015). Nonacetogenic growth of the acetogen Acetobacterium woodii on 1,2- propanediol. Journal of Bacteriology 197, 382–391. doi:10.1128/JB.02383-14.

Thurman, E. M., Ferrer, I., Blotevogel, J., and Borch, T. (2014). Analysis of hydraulic fracturing flowback and produced waters using accurate mass: identification of ethoxylated surfactants. Anal. Chem. 86, 9653–9661. doi:10.1021/ac502163k.

Tran-Din, K., and Gottschalk, G. (1985). Formation of d(-)-1,2-propanediol and d(-)- lactate from glucose by Clostridium sphenoides under phosphate limitation. Arch. Microbiol. 142, 87–92. doi:10.1007/BF00409243.

Trifunović, D., Schuchmann, K., and Müller, V. (2016). Ethylene glycol metabolism in the acetogen Acetobacterium woodii. J. Bacteriol. 198, 1058–1065. doi:10.1128/JB.00942-15.

U.S. Energy Information Administration (EIA) (2018). The United States exported more natural gas than it imported in 2017 - Today in Energy. Available at: https://www.eia.gov/todayinenergy/detail.php?id=35392 [Accessed November 13, 2018].

111

Wiśniewski, J. R., Zougman, A., Nagaraj, N., and Mann, M. (2009). Universal sample preparation method for proteome analysis. Nature Methods 6, 359.

Xu, C., Couvillion, S. P., Sontag, R. L., Isern, N. G., Maezato, Y., Lindemann, S. R., et al. (2019). MetFish: A metabolomics platform for studying microbial communities in chemically extreme environments. bioRxiv, 518647. doi:10.1101/518647.

112

3.6 Figures

Figure63.1 Trends in AEOs and PEGs in a Utica-Point Pleasant natural-gas well

A) Produced fluid temporal change in total AEO species for linear nC6, branched C6, and C8 AEOs where C0 is the sum of all ethoxylates of a particular alkyl species at Day 86. Error bars correspond to the standard deviation of the average change experienced by the sum of all ethoxylates for a given alkyl chain. B) Branched C6 AEO temporal trends relative to Day 86, with each bar representing a different C6 ethoxylate. C) First-order rate constants (bars, left y-axis) and half-lives (circles, right y-axis) for PEGs, C4, nC6, C6, and C8 AEOs. Colors correspond to legend given in A,D. D) Produced fluid temporal change in total AEO for C4 AEOs and PEGs where C0 is the sum of all ethoxylates of a particular alkyl species at Day 86. Error is the same as described in A.

113

Figure73.2 Surfactant enzymatic degradation pathway, metagenomic gene copies relevant to pathway, and Halanaerobium congolense WG10 contig.

A) Scheme for enzymatic biotransformation of a branched C6 alcohol ethoxylate. B) Relative abundance of Halanaerobium through time compared to the normalized gene copies ((gene count/assembled metagenome size)x108) present in produced fluid metagenomes using blastp (>90% identity and bit-score >250) compared to the genes present in the genome of H. congolense WG10. C) H. congolense WG10 genomic contig containing pdu and related genes with homologous genes to Acetobacterium woodii denoted by green circles (bitscore > 200, identity > 35%)

114

Figure83.3. Relative change in initial and final concentrations of polypropylene glycols (PPGs) and C8 alkyl polyethoxylates (AEOs) in surfactant amended cell cultures.

A) The sum of all PPG and C8 AEO polyglycol species. B) Relative change in each individual ethoxylate/propoxylate. Error bars correspond to standard deviations between triplicate measurements, and statistical significance for biotic cells (p<0.1) is denoted by (*) compared to abiotic controls only and (**) compared to both killed and abiotic controls.

115

Figure93.4. Proposed surfactant metabolic reconstruction of Halanaerobium congolense WG10 cultured with surfactants compared to glucose controls. Arrow color represents average z-score in proteins from surfactant cultures. Each arrow denotes one biochemical step, whereas each number corresponds to individual proteins (see Figure B6 in Appendix B). Statistical significance (p<0.1) for differences in protein relative abundance in surfactant grown cultures are denoted by a black outline around the corresponding arrow. Dashed arrows indicate the protein was solely found in surfactant grown cultures and not in glucose controls. Metabolite concentrations are reported here at t0, tm (mid-log), ts (stationary phase), and td (death phase), and statistical significance (p<0.1) is denoted by (*) next to the measurement.

116

Chapter 4. Hydraulically fractured natural-gas well microbial communities contain

genomic (de)halogenation potential

Abstract

Organohalides are routinely detected in fluid produced from hydraulically fractured oil and natural-gas wells, yet the origin and fate of these compounds remains largely unknown despite their probable toxicity. Since few organohalides are disclosed as fracturing fluid additives, one suspected formation mechanism is the reaction of geogenic halides oxidized by injected additives with natural or xenobiotic organic carbon constituents. However, the potential role of microorganisms in organohalide production and transformation is currently underappreciated.

Here we uncover the microorganisms and enzyme systems that contribute to organohalide transformations during hydraulic fracturing. Our survey of 25 metagenomes sampled from four hydraulically fractured Appalachian Basin natural-gas wells identified genes encoding for both halogenation and dehalogenation mechanisms in produced fluids, weeks and months after hydraulic fracturing occurred. Putative genes were most closely related to genes in members of the

Pseudomonas genus, the Halomonadaceae bacterial family, and the Burkholderiales bacterial order. These results suggest microbial organohalide formation occurs through enzymatically- produced hypohalous acids that oxidize geogenic or xenobiotic organic matter. Microbial reductive dehalogenation is inferred to operate on chlorinated or brominated alkanes and

117 aromatics. These results indicate that microorganisms could play a significant role in organohalide transformations in hydraulically fractured oil and gas systems.

4.1 Introduction

Organohalides associated with industrial activity are frequently recalcitrant, toxic pollutants

(Krasner et al., 2006; Rook, 1977). These constituents are produced both intentionally by the chemical industry and unintentionally through engineered processes such as drinking water disinfection (Krasner et al., 2006; Rook, 1977). Due to their risk to human health, organohalides comprise over half of the compounds on the United States Environmental Protection Agency priority pollutant list (Lu et al., 2017; US EPA, 2015). Recently, organohalides have been detected in the wastewaters of hydraulically fractured (HF) oil and natural-gas (O&G) wells (Hoelzer et al.,

2016; Luek et al., 2017; Parker et al., 2014). Organohalides are not regularly disclosed as HF fluid additives, rather, the current paradigm indicates in situ generation through abiotic reactions between fracture fluid oxidants (e.g., sodium hypochlorite, ammonium persulfate) and geogenic halides (inc., chloride, bromide, iodide) that dissolve into subsurface brines and further react with natural or xenobiotic constituents (Elsner and Hoelzer, 2016; Luek et al., 2017; Stringfellow et al.,

2014). Both halogenated aliphatics (e.g., dichloromethane, iodo- and bromoalkanes, bromo- and chloroacetones) and aromatic structures (e.g., benzyl chloride, 1,4,-dichlorobenzene) have been detected in fluids produced from HF O&G wells, even though these compounds are not reported in the input fluids (Akyon et al., 2019; Elsner and Hoelzer, 2016; Hayes, 2009; Luek et al., 2017,

2018; Maguire-Boyle and Barron, 2014; Thacker et al., 2015). While abiotic reactions may generate organohalides during HF operations, microorganisms may also participate in organohalide transformations, as has been observed in related environments (Hoelzer et al., 2016;

Luek et al., 2017). 118

Microorganisms are capable of both organohalide production and transformation (e.g., halide cleavage) (van Pée and Unversucht, 2003). Microbially-produced organohalides play an important role in combating cellular oxidative stress (e.g., peroxidases) and offer a competitive advantage to certain microorganisms in antagonistic relationships (e.g., antibacterial or antifungal agents)

(Bengtson et al., 2013; Gribble, 2015; Weigold et al., 2016), but the biochemical mechanisms responsible for halogenation remain understudied. On the other hand, many pathways for microbial dehalogenation have been characterized in soils, seafloor sediments, and groundwater environments (Hug et al., 2013; Jugder et al., 2016; Kawai et al., 2014; Mohn and Tiedje, 1992;

Weigold et al., 2016). Reductive dehalogenation couples halide cleavage to hydrogen oxidation under strictly anaerobic conditions (Holliger and Schumacher, 1994; Mohn and Tiedje, 1992), while hydrolytic or oxidative dehalogenation can occur across a range of redox conditions

(Fetzner, 1998). Recent studies have described the metabolic potential of halotolerant bacteria in

HF O&G wells, including their role in biogeochemical cycles, amino acid fermentation, and guar gum degradation during natural-gas operations (Daly et al., 2016; Evans et al., 2018; Liang et al.,

2016). Similar metagenomic techniques can be applied to assess the role microorganisms play in

(de)halogenation reactions not previously assessed in HF O&G systems.

To assess the abundance of microbial (de)halogenation pathways in HF systems, we screened metagenomes reconstructed from samples collected 7 to 328 days after HF from four Appalachian

Basin natural-gas wells. (De)halogenation genes were identified based on similarity to genes from known (de)halogenating bacterial isolates in databases (e.g. NCBI, IMG) and reported in the context of produced fluid geochemical measurements and organohalides previously detected in

119 fluids produced from HF O&G wells. These results indicate that microorganisms likely influence organohalide cycles in HF O&G systems.

4.2 Methods

4.2.1 Sample collection and analysis

Fluids were collected at the well head or gas-fluid separators from four natural-gas producing wells in the Appalachian Basin: one in Pennsylvania (Marcellus-1), two in West Virginia (Marcellus-3,

Marcellus-4), and one in Ohio (Utica-3). Input or produced fluid was collected in 1 L HDPE bottles

(Nalgene) with no headspace. Samples were immediately filtered (0.22 µm PES, Millipore, EMD,

Burlington, VT) to collect biomass for DNA extraction and to preserve for geochemical analyses as previously described (see Supplemental Methods in Appendix C) (Borton et al., 2018; Daly et al., 2016; Evans et al., 2018; Luek et al., 2018). For microbial cell counts, unfiltered produced fluids were immediately fixed in paraformaldehyde (5% final concentration) and stored at 4°C for no more than one week prior to cell counting as previously described (see Supplemental Methods in Appendix C) (Evans et al., 2018). Method details for geochemical measurements and cell counts are reported in Appendix D.

4.2.2 Metagenomic sequencing, assembly, annotation, and binning

Approximately 300-1,000 mL of each fluid sample was filtered using 0.22 µM PES filters

(Millipore, EMD, Burlington, VT). Total nucleic acids were extracted from the filter using a

PowerSoil DNA isolation kit (MoBio, Carlsbad, CA) for Marcellus-1, and a modified phenol chloroform extraction technique was used for Utica-1, Marcellus-3, and Marcellus-4 (Borton et al., 2018). Illumina HighSeq 2000/2500 library preparation, DNA preparation, sequencing, and

120 fastq file generation were performed as described previously (Daly et al., 2016). Illumina sequences were trimmed from the 5’ and 3’ ends using Sickle (https://github.com/najoshi/sickle) and assembled using IDBA-UD (Brown et al., 2015; Wrighton et al., 2012) with default parameters

(Daly et al., 2016). Metagenome data statistics are available (see Table D1 in Appendix D). Near- full-length 16S rRNA genes were reconstructed from metagenomic data using EMIRGE as previously described (see Table D2 in Appendix D) (Daly et al., 2016; Miller et al., 2011).

Annotation of scaffolds was performed as previously described (Daly et al., 2016).

4.2.3 Target gene phylogenies

Microbial (de)halogenation pathways were investigated across all samples by first building reference datasets for each target gene. A previously characterized functional gene (e.g., halogenation or dehalogenation gene) retrieved using NCBI (see Table D3 in Appendix D) was used to collect similar sequences from IMG isolates using blastp (e-score < 1e-5) (Altschul et al.,

1990; Chen et al., 2017). Fasta files containing the functional gene and blastp results were next aligned using MUSCLE v. 3.8 in Geneious 8.1.9 (https://www.geneious.com). Alignments were automatically edited to remove alignment blocks of low quality using Gblocks (Castresana, 2002), models of evolution were selected for each gene via ProtTest (Posada et al., 2011), and Maximum

Likelihood (RAxML) phylogenetic trees constructed with 100 bootstrap replicates (Stamatakis,

2014) (https://github.com/lmsolden/protpipeliner). For each of the key functional genes, we created a reference database, and used this to identify homologs (based on BLAST (Altschul et al.,

1990), minimum percent identity (25-50%) and bit scores (100-200) for each gene reported in

Table D4 in Appendix D). Next, we used Bowtie2 to non-competitively map the assembled metagenomic reads to each gene reference database with zero mismatches (Langmead and

121

Salzberg, 2012). Genes were included in a sample if they had 5X read covered over a minimum of

90% of the gene. Phylogenetic trees (in Figure 4.2) were generated as reported above with the following exceptions: Maximum-likelihood (ML) trees were constructed using RAxML (v.7.2.8) with a maximum of 10 iterations, using the nucleotide model GTR gamma, rapid bootstrapping

(100 replicates) and search for best-scoring ML tree algorithm (Stamatakis, 2014). The genes were assigned to organisms either based on recovery in a resolved genome bin (e.g., Pseudomonas and

Halomonadaceae in Marcellus-1), or in the case of low quality genome bins, genes were annotated by the strongest taxonomic signal for the contig (e.g., Pseudomonas and Burkholderiales in Utica-

3, Paramaledivibacter in Marcellus-4).

4.2.4 Data accession

Metagenomic accession numbers are available in the Supporting Information (see Table D1 in

Appendix D).

4.3 Results and Discussion

4.3.1 Organohalides present in fluids produced from HF O&G wells

Organohalides have been identified in flowback and produced fluids from five different unconventional formations including the Fayetteville, Marcellus, Utica, Bakken, and Eagle Ford, with maturation dates ranging from days to years (Table 4.1) (Abualfaraj et al., 2018; Akyon et al., 2019; Hayes, 2009; Hoelzer et al., 2016; Luek et al., 2017, 2018; Maguire-Boyle and Barron,

2014). The majority of analyses were performed within the first few weeks to months of initial flowback using gas chromatography-mass spectrometry (GC-MS) based methods (Luek and

Gonsior, 2017). Temporal trends in organohalide concentrations are unclear due to limited

122 sampling of fluid time series and little overlap in individual compounds analyzed across studies and basins (Luek et al., 2018). Nevertheless, haloalkanes have been detected most frequently

(Maguire-Boyle and Barron, 2014; Orem et al., 2014; Strong et al., 2014), and although this is unsurprising considering the high abundance of aliphatics in black shales (Petsch et al., 2000), existing analytical methods for these compounds would also bias towards their detection. Within the haloalkanes, halomethanes were the most frequently detected, with dichloromethane, the only known haloalkane fracture fluid additive, the most commonly identified structure. Haloalkenes and halobenzenes were also detected in multiple studies, with reports of haloalcohols, halophenols, haloacids, haloketones, and haloacetones. Organochlorides were more frequently detected than organobromides and organoiodides in targeted GC-MS based studies, but organoiodides were most frequently detected using a non-targeted analysis (Luek et al., 2018). Although the majority of analyses to date have identified haloaliphatics and haloaromatics, non-target analyses indicate a wider array of halogenated aromatic compounds are likely present in produced fluids (Hoelzer et al., 2016; Luek et al., 2017, 2018). The extent of organohalide fracture fluid additives may also be broader than originally thought; chloromethyl alkanoates may have been applied as delayed release acids (Hoelzer et al., 2016), while a recent survey of fracture fluid additives in California uncovered > 20 haloaromatics as likely applied as chemical tracers (Stringfellow and Camarillo,

2019).

123

4.3.2 Halogenation and dehalogenation potential in Appalachian Basin natural-gas wells

Genes encoding for halogenation and reductive dehalogenation were identified in all wells surveyed: three Marcellus Shale natural-gas wells and one Utica-Point Pleasant natural-gas well

(Figure 4.1a). Produced fluids from all Marcellus Shale and Utica Point-Pleasant natural-gas wells surveyed here contain microbial biomass on the order of 103 to 107 cells/mL, confirming the presence of microorganisms up to one year after HF (Figure 4.1b) (Daly et al., 2016). In particular, biomass remained between 105 and 106 cells/mL during the first 50 days after flowback when these

(de)halogenation pathways were detected. Non-heme chloroperoxidase and tryptophan halogenase, genes responsible for the halogenation of organic compounds, were present in metagenomes sampled 30-150 days after HF in Marcellus-4, Marcellus-5, and Utica-3. Non-heme chloroperoxidases are a class of serine hydrolases catalyzing the non-specific halogenation of organic compounds in the presence of H2O2 (van Pée and Unversucht, 2003). The enzyme first generates peracetic acid capable of oxidizing dissolved halide ions (Cl-, Br-, I-) to form hypohalous acids (e.g. X-OH), which then act as the halogenating agent (Hofmann et al., 1998). While this mechanism is unlikely to generate high concentrations of organohalides in produced fluids, the highly oxidizing hypohalous acid may reduce inorganic and organic compounds in the system, as well as halogenate organic compounds (Hurst, 2014). High chloride (>50,000 mg/L) and bromide

(>500 mg/L) concentrations in all four HF Appalachian Basin natural-gas wells that are susceptible to oxidation (Figure 4.1b), as well as high organic carbon concentrations (40-400 mg/L) persisting months to years after HF occurs (Figure 4.1b) represent a significant likelihood that reactions between organic carbon and in situ generated hypohalous acids may occur long after the kinetics of injected oxidative additives have slowed. Tryptophan halogenases catalyze the chlorination or bromination of organic compounds and are implicated in the production of the antitumor antibiotic

124 rebeccamycin, as well as the antifungal pyrrolnitrin (van Pée and Unversucht, 2003). Although biocide resistance has been studied in the HF system (Kahrilas et al., 2016; Vikram et al., 2014), biologically-generated antibiotics during HF operations have not been thoroughly investigated.

The presence of tryptophan halogenase genes indicates microbially-derived biocides could offer select bacteria a competitive advantage under antagonistic conditions in the HF ecosystem or in produced fluid holding tanks.

Five different classes of reductive dehalogenase genes and three classes of hydrolytic dehalogenation genes were identified in these Appalachian Basin natural-gas wells (Figure 4.1a).

One ortho-chlorophenol reductive dehalogenase was found in Marcellus-1 on day 328 (Figure

4.1a). Several aliphatic and aromatic reductive dehalogenase genes were also identified between

30-150 days after HF in Marcellus-4, Marcellus-5, and Utica-3. A haloacid dehalogenase, known to remove halogens from (S)-2-haloacids (e.g. halosubstituted C2-C4 carboxylic acids), was identified in two Utica-3 samples (days 38, 54). This gene was described in a reconstructed genome for Marinobacter isolated from this Utica natural-gas well (Evans et al., 2018).

Haloacetate dehalogenase, responsible for the dehalogenation of haloacetic acids, was detected in

Marcellus-4 on (day 34) and Utica-3 (day 54). We also observed haloalkane dehalogenases in at least one sample from all four wells (Figure 4.1a). The types of genes observed here are capable of transforming the organohalides most commonly identified in HF O&G wells (e.g., haloalkanes, haloacids, halobenzenes).

125

4.3.3 Phylogeny of haloperoxidase and reductive dehalogenation genes in HF O&G wells

Phylogenetic analyses of non-heme chloroperoxidase genes and reductive dehalogenases genes confirmed metabolic potential for (de)halogenation reactions in the natural-gas wells based on high similarity (bitscore ≥ 100-200, % identity ≥ 25-45, see Table D4 in Appendix D) to genes from characterized isolates. Non-heme chloroperoxidase genes were present in all four Appalachian

Basin natural-gas wells and clustered closely with characterized genes from Pseudomonas,

Kitatospira, Burkholderia, and Streptomyces (Figure 4.2a). Marcellus-1 (day 0) and Utica-3 (day

38) genes formed a distinct cluster from other characterized isolates, while genes from Marcellus-

4 and Marcellus-5 had best alignment to Pseudomonas (Figure 4.2a). Pseudomonas genome bin relative abundance was greater than 50% in Marcellus-1 (day 0) (Daly et al., 2016), and comprised

19% of the measured 16S rRNA gene diversity in the Utica-3 sample on day 38 (Figure 4.2b).

Consequently, Pseudomonas is likely to play an important role in the production of reactive hypohalous acids, leading to the generation of organohalides in produced fluids during the first few weeks after HF.

Five reductive dehalogenase genes from Marcellus-4 (days 79, 205) clustered with genes from well-known, strictly anaerobic organohalide-respiring bacteria Desulfitobacterium dichloroeliminans, Sulfurospirillum multivorans, and two Dehalococcoides species (Figure 4.2a).

Largely, these genes are known to be responsible for halide removal from haloaliphatic substrates

(e.g., halomethanes, haloethanes, haloethenes, halocyclohexanes) (Mohn and Tiedje, 1992). Two of these genes had best alignments to Paramaledivibacter (bit scores 550, 73% identity for both genes), which was present in low abundance (1%) in the metagenome on day 79 as determined by

16S rRNA gene relative abundance (Figure 4.2b). Ten reductive dehalogenase genes from Utica-

126

3 (days 38, 54) and one from Marcellus-1 (day 328) clustered with the lesser-studied organohalide respirers Comamonas sp., Nitratireductor pacificus, and Sphingobium chlorophenolicum (Figure

4.2a), facultative anaerobic organisms known for their ability to respire aromatic organohalides

(chloro- and bromobenzenes, tetrachlorohydroquinone, and others) under anoxic conditions (Chen et al., 2013; Dai et al., 2003; Jugder et al., 2016). One reductive dehalogenase gene (Utica-3, day

38) was similar to genes found in Pseudomonas (Figure 4.2a), which comprises 19% of the microbial community on that day (Figure 4.2b). The other nine genes from Utica-3 had best blast hits to genera belonging to the order Burkholderiales, and sequences from this order were present in input fluids (day 0) and comprised 36% and 24% of the microbial community on days 38 and

54, respectively (Figure 4.2b). The reductive dehalogenase gene from Marcellus-1 (day 328) was present in a reconstructed Halomonadaceae genome from an earlier study (Figure 4.2b), with a relative abundance of 43.5% on that day (Daly et al., 2016). These findings suggests microbial reductive dehalogenation has the potential to persist in produced fluids a year after HF, under a range of engineering operations in Appalachian Basin shale formations.

Putative genes encoding for the halogenation of organic compounds were identified in all four natural-gas wells studied here in the weeks following HF, indicating that a portion of organohalides in produced fluids may be microbially-generated. In contrast, hydrolytic and reductive dehalogenase genes in fluids months after HF may reduce the organohalide concentration in produced wastewaters. Many of the organohalides reported in produced fluids from HF O&G wells are susceptible to transformation by the dehalogenation genes described here. Organohalide removal from haloaliphatics (e.g. haloalkanes, haloacids) can be performed by hydrolytic/oxidative dehalogenases as well as reductive dehalogenases present here, whereas

127 haloaromatics (e.g. halophenols) can be dehalogenated by reductive dehalogenases. Moreover, substrate specificity for reductive dehalogenation was identified, and may vary depending on the types of microorganisms present, fluid additives, or formation biogeochemistry. Recent atmospheric analyses have detected an increase in halo-volatile organic carbons (halo-VOCs) in regions with active HF activities, likely from volatilization of these compounds from produced fluid storage in vented or open storage units (Li et al., 2017; Rich and Orimoloye, 2016). These halo-VOCs are of environmental concern, as they may be transported long distances in the atmosphere, depositing in nearby watersheds where they may bioaccumulate (Xia et al., 2019).

Ultimately, further research is needed to quantify organohalide metabolic activity in HF O&G systems, and to discern the proportion of organohalides produced biotically versus abiotically.

Author contributions:

P. Mouser and M. Evans designed the study. M. Evans wrote the manuscript. R. Daly and K.

Wrighton provided the metagenomes and metagenomic pipeline, and advised on metagenomic analyses. J. Luek collected data for Table 1 and edited the manuscript.

Acknowledgements:

This research was supported by the National Science Foundation (CBET award 1823069) and the

Department of Energy (DOE) National Energy Technology Laboratory through the Marcellus

Shale Energy and Environmental Laboratory (project #DE-FE0024297) under a subcontract from

West Virginia University. We are grateful to Jenny Panescu, David Cole, Susan Welch, and

Mikayla Borton for their scientific insight, and offer thanks to Northeast Natural Energy and another industry partner for site acces and samples. Genomic and metagenomic sequencing was

128 performed by DOE Joint Genome Institute (JGI) via a large-scale sequencing award (no. 1931).

The JGI is sponsored by the Office of Biological and Environmental Research and operated under contract DE-AC02-05CH11231.

4.4 References

Abualfaraj, N., Gurian, P. L., and Olson, M. S. (2018). Assessing residential exposure risk from spills of flowback water from Marcellus Shale hydraulic fracturing activity. Int J Environ Res Public Health 15. doi:10.3390/ijerph15040727.

Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990). Basic local alignment search tool. J. Mol. Biol. 215, 403–410. doi:10.1016/S0022-2836(05)80360-2.

Bengtson, P., Bastviken, D., and Öberg, G. (2013). Possible roles of reactive chlorine II: assessing biotic chlorination as a way for organisms to handle oxygen stress. Environmental Microbiology 15, 991–1000. doi:10.1111/j.1462-2920.2012.02807.x.

Brown, C. T., Hug, L. A., Thomas, B. C., Sharon, I., Castelle, C. J., Singh, A., et al. (2015). Unusual biology across a group comprising more than 15% of domain Bacteria. Nature 523, 208.

129

Castresana, J. (2002). GBLOCKS: selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Version 0.91 b. Copyrighted by J. Castresana, EMBL.

Chen, K., Huang, L., Xu, C., Liu, X., He, J., Zinder, S. H., et al. (2013). Molecular characterization of the enzymes involved in the degradation of a brominated aromatic herbicide. Molecular Microbiology 89, 1121–1139. doi:10.1111/mmi.12332.

Dai, M., Rogers, J. B., Warner, J. R., and Copley, S. D. (2003). A previously unrecognized step in pentachlorophenol degradation in Sphingobium chlorophenolicum is catalyzed by tetrachlorobenzoquinone reductase (pcpd). Journal of Bacteriology 185, 302–310. doi:10.1128/JB.185.1.302-310.2003.

Evans, M. V., Panescu, J., Hanson, A. J., Welch, S. A., Sheets, J. M., Nastasi, N., et al. (2018). Members of Marinobacter and Arcobacter influence system biogeochemistry during early production of hydraulically fractured natural gas wells in the appalachian basin. Front. Microbiol. 9, 2646. doi:10.3389/fmicb.2018.02646.

Fetzner, S. (1998). Bacterial dehalogenation. Appl Microbiol Biotechnol 50, 633–657. doi:10.1007/s002530051346.

130

Gribble, G. W. (2015). A recent survey of naturally occurring organohalogen compounds. Environ. Chem. 12, 396–405. doi:10.1071/EN15002.

Hayes, T. (2009). Sampling and analysis of water streams associated with the development of Marcellus shale gas. Des Plaines, IL: Marcellus Shale Coalition Available at: http://www.water-research.net/naturalgasPA/pdffiles/MSCommission-Report.pdf.

Hoelzer, K., Sumner, A. J., Karatum, O., Nelson, R. K., Drollette, B. D., O’Connor, M. P., et al. (2016). Indications of transformation products from hydraulic fracturing additives in shale- gas wastewater. Environ. Sci. Technol. 50, 8036–8048. doi:10.1021/acs.est.6b00430.

Holliger, C., and Schumacher, W. (1994). Reductive dehalogenation as a respiratory process. Antonie van Leeuwenhoek 66, 239–246. doi:10.1007/BF00871642.

Hug, L. A., Maphosa, F., Leys, D., Löffler, F. E., Smidt, H., Edwards, E. A., et al. (2013). Overview of organohalide-respiring bacteria and a proposal for a classification system for reductive dehalogenases. Philosophical Transactions of the Royal Society B: Biological Sciences 368, 20120322. doi:10.1098/rstb.2012.0322.

Hurst, J. K. (2014). Hypohalous acid. doi:10.1036/1097-8542.334300.

Jugder, B.-E., Ertan, H., Bohl, S., Lee, M., Marquis, C. P., and Manefield, M. (2016). Organohalide respiring bacteria and reductive dehalogenases: key tools in organohalide bioremediation. Front. Microbiol. 7. doi:10.3389/fmicb.2016.00249.

131

Kahrilas, G. A., Blotevogel, J., Corrin, E. R., and Borch, T. (2016). Downhole transformation of the hydraulic fracturing fluid biocide glutaraldehyde: implications for flowback and produced water quality. Environ. Sci. Technol. 50, 11414–11423. doi:10.1021/acs.est.6b02881.

Krasner, S. W., Weinberg, H. S., Richardson, S. D., Pastor, S. J., Chinn, R., Sclimenti, M. J., et al. (2006). Occurrence of a new generation of disinfection byproducts. Environ. Sci. Technol. 40, 7175–7185. doi:10.1021/es060353j.

Langmead, B., and Salzberg, S. L. (2012). Fast gapped-read alignment with Bowtie 2. Nat Meth 9, 357–359.

Li, S.-M., Leithead, A., Moussa, S. G., Liggio, J., Moran, M. D., Wang, D., et al. (2017). Differences between measured and reported volatile organic compound emissions from oil sands facilities in Alberta, Canada. PNAS 114, E3756–E3765. doi:10.1073/pnas.1617862114.

Liang, R., Davidova, I. A., Marks, C. R., Stamps, B. W., Harriman, B. H., Stevenson, B. S., et al. (2016). Metabolic capability of a predominant halanaerobium sp. in hydraulically fractured gas wells and its implication in pipeline corrosion. Front. Microbiol. 988. doi:10.3389/fmicb.2016.00988.

Lu, Q., Yu, L., Liang, Z., Yan, Q., He, Z., Luan, T., et al. (2017). Dehalococcoides as a potential biomarker evidence for uncharacterized organohalides in environmental samples. Front. Microbiol. 8. doi:10.3389/fmicb.2017.01677.

132

Luek, J. L., and Gonsior, M. (2017). Organic compounds in hydraulic fracturing fluids and wastewaters: A review. Water Res. 123, 536–548. doi:10.1016/j.watres.2017.07.012.

Maguire-Boyle, S. J., and Barron, A. R. (2014). Organic compounds in produced waters from shale gas wells. Environ. Sci.: Processes Impacts 16, 2237–2248. doi:10.1039/C4EM00376D.

Mohn, W. W., and Tiedje, J. M. (1992). Microbial reductive dehalogenation. Microbiol. Rev. 56, 482.

Orem, W., Tatu, C., Varonka, M., Lerch, H., Bates, A., Engle, M., et al. (2014). Organic substances in produced and formation water from unconventional natural gas extraction in coal and shale. International Journal of Coal Geology 126, 20–31. doi:10.1016/j.coal.2014.01.003.

Parker, K. M., Zeng, T., Harkness, J., Vengosh, A., and Mitch, W. A. (2014). Enhanced formation of disinfection byproducts in shale gas wastewater-impacted drinking water supplies. Environ. Sci. Technol. 48, 11161–11169. doi:10.1021/es5028184.

133

Petsch, S. T., Berner, R. A., and Eglinton, T. I. (2000). A field study of the chemical weathering of ancient sedimentary organic matter. Organic Geochemistry 31, 475–487. doi:10.1016/S0146-6380(00)00014-0.

Posada, D., Darriba, D., Taboada, G. L., and Doallo, R. (2011). ProtTest 3: fast selection of best- fit models of protein evolution. Bioinformatics 27, 1164–1165. doi:10.1093/bioinformatics/btr088.

Rich, A. L., and Orimoloye, H. T. (2016). Elevated atmospheric levels of benzene and benzene- related compounds from unconventional shale extraction and processing: human health concern for residential communities. Environ Health Insights 10, 75–82. doi:10.4137/EHI.S33314.

Rook, J. J. (1977). Chlorination reactions of fulvic acids in natural waters. Environ. Sci. Technol. 11, 478–482. doi:10.1021/es60128a014.

Stamatakis, A. (2014). RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30, 1312–1313. doi:10.1093/bioinformatics/btu033.

Stringfellow, W.T., and Camarillo, M. K. (2019). Flowback verses first-flush: new information on the geochemistry of produced water from mandatory reporting. Environmental Science: Processes & Impacts. doi:10.1039/C8EM00351C.

Strong L. C., Gould T., Kasinkas L., Sadowsky M. J., Aksan A., and Wackett L. P. (2014). Biodegradation in waters from hydraulic fracturing: chemistry, microbiology, and

134

engineering. Journal of Environmental Engineering 140, B4013001. doi:10.1061/(ASCE)EE.1943-7870.0000792.

US EPA (2015). Toxic and priority pollutants under the clean water act. Available at: https://www.epa.gov/eg/toxic-and-priority-pollutants-under-clean-water-act [Accessed January 28, 2019]. van Pée, K.-H., and Unversucht, S. (2003). Biological dehalogenation and halogenation reactions. Chemosphere 52, 299–312. doi:10.1016/S0045-6535(03)00204-2.

Vikram, A., Lipus, D., and Bibby, K. (2014). Produced water exposure alters bacterial response to biocides. Environ. Sci. Technol. 48, 13001–13009. doi:10.1021/es5036915.

Weigold, P., El-Hadidi, M., Ruecker, A., Huson, D. H., Scholten, T., Jochmann, M., et al. (2016). A metagenomic-based survey of microbial (de)halogenation potential in a German forest soil. Scientific Reports 6, 28958. doi:10.1038/srep28958.

Xia, Z., Idowu, I., Marvin, C., Thomas, P. J., Johnson, W., Francisco, O., et al. (2019). Identification of halogenated polycyclic aromatic hydrocarbons in biological samples from Alberta Oil-Sands Region. Chemosphere 215, 206–213. doi:10.1016/j.chemosphere.2018.10.050.

135

4.5 Figures

Figure104.1. Organohalide metabolism genes, geochemical, and microbial biomass measurements in four Appalachian Basin natural-gas wells A) Presence (red) or absence (black) of organohalide metabolism genes in four Appalachian Basin HF natural-gas wells. Numbers in the top row correspond to date of sampling after HF occurred. B) Geochemical and microbial biomass measurements in the natural-gas wells.

136

Figure114.2. Phylogenetic placement of dehalogenase and halogenase genes A) Dehalogenase and halogenase genes from characterized isolates and metagenomes. Metagenomic genes are color-coded by well, and taxonomic assignment in reconstructed 16S rRNA gene data is denoted by a colored circle. B) 16S rRNA gene or reconstructed genome bin relative abundances of suspected organohalide metabolizing microorganisms plotted by taxonomy (circle color) and well (background color).

137

Table24.1. Reported organohalides in produced fluid from HF oil and natural-gas wells.

138

Chapter 5. Marcellus Shale wastewater microbial communities persist through pre-

treatment at a class (II) injection well disposal facility

Intended for submission to FEMS Microbial Ecology

Abstract

A significant volume of produced fluid from hydraulically fractured oil and gas wells is disposed of through injection into class (II) disposal wells. These disposal facilities generally treat fluids prior to injection to remove constituents capable of clogging, corroding, or fouling the target formation or the well infrastructure. Shale-associated microoorganisms have previously demonstrated a capacity to corrode well infrastructure and produce biofilms under subsurface pressures, necessitating investigation into how the unit processes of pre-treatment alter wastewater microbial communities. Here we investigate the microbial communities present during pre-treatment prior to disposal of Marcellus Shale wastewaters into a class (II) injection well. Wastewater biomass concentrations varied with treatment, as did the microbial diversity and community structure. Although produced fluids were transported by trucks, exposed to oxygen, and subject to chemical treatment, microbial communities closely resembled those commonly observed in produced fluids from a hydraulically fractured Marcellus Shale natural- gas well. Taxa commonly associated with biofilm-formation, iron reduction, and sulfur cycling were identified (e.g., Halanaerobium, Arcobacter, Shewanella). Due to the complications associated with the continuance of shale-associated microbial taxa in class (II) injection wells, further understanding of the metabolisms reinforcing the persistence of these microorganisms is important to prevent undesirable corrosion and formation clogging.

139

5.1 Introduction

Horizontal drilling and hydraulic fracturing technologies have been widely used over the last decade to recover energy reserves from previously inaccessible hydrocarbon-bearing formations, greatly increasing the production of natural gas worldwide (Arthur et al., 2009; U.S. Energy

Information Administration (EIA), 2016). In mid-2018, two unconventional non-permeable systems in the Appalachian Basin, the Marcellus Shale and Utica Formation, collectively produced 29% of the dry natural gas in the U.S. (U.S. Energy Information Administration (EIA),

2018). The high density of hydraulic fracturing operations in the Appalachian Basin states of

Pennsylvania, West Virginia, and Ohio has promulgated new regulations for the management of produced fluids, as the wastewaters would have hazardous environmental and human health impacts if improperly treated or accidentally released (Kargbo et al., 2010; Vidic et al., 2013).

Produced fluids from black shale wells contain TDS concentrations reaching 200 g/L or higher, including high levels of salinity, iron, sulfur, radionuclides, and heavy metals (Barbot et al.,

2013; Chapman et al., 2012). Further, microorganisms surviving high-pressure injection, biocide treatment, and rapidly changing biochemical conditions are present in produced fluids (Akob et al., 2015; Cluff et al., 2014; Daly et al., 2016; Fichter et al., 2012; Lipus et al., 2018; Mohan et al., 2013; Mouser et al., 2016; Wuchter et al., 2013). Treatment of produced fluids is therefore expensive and difficult, and the efficacy of traditional industrial wastewater treatments in managing these complex mixtures has been questioned (Getzinger et al., 2015). Thus, other methods such as re-use in other hydraulically fractured oil and gas wells, or injection into class

(II) disposal wells are regularly employed instead. For example, Kansas, Oklahoma, and Texas have 3,000-8,000 active oil and gas disposal wells and inject over 100 million gallons of 140 wastewater per day per state (Veil, 2015). While the number of wells and volume of wastewater disposed in the Appalachian Basin states (e.g., Ohio, West Virginia, Pennsylvania) is relatively small in comparison (9-190 wells, 0.045-1.6 million gallons per day) (U.S. Environmental

Protection Agency, 2016); injection well disposal is the most overwhelmingly used management practice in Ohio and West Virginia (91% and 53% of managed produced fluids) (Veil, 2015).

However, recent research has shed light on the risks and hazards associated with injection well disposal, namely induced seismicity, which is thought to be the result of high injection rate from large volumes of wastewater (Kell, 2011; Kim, 2013; Weingarten et al., 2015). In addition, one study identified negative environmental impacts of class (II) injection wells in a nearby stream

(Akob et al., 2016).

Brine wastewaters from black shale wells often require processing prior to injection well disposal to remove residual hydrocarbons, suspended solids, and microbial cells to prevent corrosion or fouling (Arthur et al., 2005). Hydrocarbons remaining in oil and gas wastewaters present an economic benefit for well operators if separated and recovered, with the added benefit of decreased fluid viscosity prior to high-pressure injection disposal. Insoluble particulates (e.g., iron) may further complicate disposal efficacy and are capable of participating in redox reactions upon injection into an anoxic subsurface environment, leading to corrosion of steel casings.

Further, microbial life present in oil and gas wastewaters have a demonstrated capacity for undesirable well infrastructure corrosion and biofilm formation leading to clogging, mechanisms that may play an important yet understudied role in class (II) injection wells (Arensdorf et al.,

2009; Fichter et al., 2012).

141

We collected (6) samples from unit process of brine wastewater treatment at a class (II) injection well disposal facility in Trumbull County, Ohio, USA, to determine the effect microbial life may have on these disposal wells and the role of pre-treatment processes on microbial communities.

The microbial community structure was analyzed in wastewaters prior to water conditioning intended to separate hydrocarbons from aqueous phase, after water conditioning in a Hudson tank used to extract the hydrocarbon layer from the aqueous phase, and after particulate filtration with a 30 µm filter prior to injection. We then compared these microbial communities to input, flowback, and produced fluid microbial communities from a previously studied Marcellus Shale natural-gas well. Our results revealed the viability of shale-associated microbial communities and demonstrated the impact of specific unit processes on microbial community structures prior to injection in class (II) disposal wells.

5.2 Materials & Methods

5.2.1 Site and sampling information

The class (II) injection well was located in Trumbull County, OH, USA and was drilled to a depth of approx. 2400 m in the Mount Simon sandstone formation. The disposal facility reportedly received produced fluids from Marcellus Shale wells. Prior to injection, produced fluids were pre-treated, beginning with the addition of a corrosion inhibitor and scale inhibitor

(see Table E1 in Appendix E). The operators indicated fluids remained in holding tanks during this pre-treatment stage for several days. After this, fluids were pumped into an enhanced oil recovery phase separator, the WILSA ElectroWaveConditioner®, to separate soluble hydrocarbons from fluids by altering the interfacial tension of fluids (Figure 5.1,

142 http://www.wilsa.com). Fluids were then pumped into a separator tank where the hydrocarbon top layer was collected and the remaining fluid flowed into a pipe until it reached a 30 µm filtration apparatus for removing particulate iron, after which fluid was then pumped into the subsurface (Figure 5.1). At the discretion of the operators, our sampling was limited to microbial analyses, the residence time of fluids during pre-treatment was not explicitly provided, and the disclosed chemical additives may not be the only chemicals used by the operators.

Fluid samples were collected in January 2016 at 6 different points in the treatment process

(Figure 5.1), in sterile 1 L HDPE Nalgene bottles. Sample (45 mL) was aliquoted into 50 mL sterile polypropylene centrifuge tubes (VWR, Radnor, PA) containing 5 mL of 25% paraformaldehyde for cell counts and kept on ice for transport back to the laboratory, then stored at 4 °C. The remaining collected fluid was immediately filtered using 0.22 µm polyethersulfone

Sterivex filters (MilliporeSigma, Burlington, MA) to obtain DNA using 250 – 1000 mL of fluid, which were placed in sterile bags, kept on ice for transport back to the laboratory, and stored at

-20 °C until DNA extractions were performed.

5.2.2 Cell counts

Cell counts were performed within 1 week of sampling according to a previously established method (Evans et al., 2018). Samples were diluted such that each filter was prepared to yield approximately 3–120 cells per counting field. Sample (2 mL) was filtered through a 25 mm dia.

0.2 μm pore PCTE filter (Sterlitech, Kent, WA), stained with 2X SYBR Gold (Life

Technologies, Carlsbad, CA) in TE buffer, then mounted on a microscope slide with SlowFade reagent (Life Technologies, Carlsbad, CA). The slide was viewed with a Labomed Lx500

143 epifluorescent microscope through a 40X air objective under 480 nm excitation. The final cell count value was calculated using an average of 15-16 counting field slides per sample with error derived from standard deviations.

5.2.3 DNA extraction, sequencing, and genomic analyses

DNA was extracted from samples using a modified phenol-chloroform method described previously (Daly et al., 2016) and was quality checked using a Qubit 3.0 Fluorometer (Thermo

Fisher Scientific, Waltham, MA). Library preparation and amplification of the V4 region of the

16S rRNA gene was performed at the Argonne National Lab according to a previously established protocol (16S Illumina Amplicon Protocol : Earth Microbiome Project) using bacterial and archaeal primers 515F-806R. The 16S rRNA iTag sequence data was obtained by using an Illumina MiSeq at the Argonne National Lab. Sequences received from Argonne

National Lab were processed using QIIME (v. 2) (Bolyen et al., 2018, 2) and the Ohio

Supercomputing Center. All sequences were initially demultiplexed, quality trimmed, and denoised using DADA2 to construct an amplicon sequence variant (ASV) table (Callahan et al.,

2016, 2). Taxonomy was assigned at 97% using the Silva132 99% database with vsearch

(Rognes et al., 2016).

Diversity statistics and non-metric multidimensional scaling (NMDS) analysis were performed in

R (v. 3.5.5) using the Vegan package. Comparison of the injection well treatment facility microbial communities was performed against input, flowback, and produced fluid sample microbial communities from our earlier study (Cluff et al., 2014). All OTUs/ASVs were

144 converted to relative abundances within each sample to minimize variations in sequence depth between the two studies. The NMDS analysis was performed using Bray-Curtis Dissimilarity indices with the MetaMDS function, ggplot2 was used to visualize the NMDS plot, and statistically significant differences between groupings calculated with ADONIS and ANOSIM.

Dendogram analysis was conducted with the Agnes function in Vegan. Indicator species analysis was conducted on Bray-Curtis Dissimilarity indices using the package IndSpecies.

5.3 Results

The 6 produced fluid samples revealed notable changes in microbial community diversity along the fluid pre-treatment process. Higher diversity was observed in the earliest and latest stages of the fluid treatment process, whereas the samples from the middle stages of pre-treatment were marked by lower diversity (Table 5.1). The earliest sample, Pre-Wilsa, had a Shannon’s diversity index of 5.01, in contrast to the later sample, Hudson Surface, which had a Shannon’s diversity index of 2.35 (Table 5.1). The sample at the end of the pre-treatment, the Filter Effluent, had the highest diversity of all samples, with a Shannon’s diversity index of 8.41 (Table 5.1). Cell counts employing a nucleic acid stain revealed 9.04*104 -1.34*106 cells mL-1 in all fluid samples (Table

5.1).

Twenty distinct amplicon sequence variants (ASVs) (relative abundance > 0.01) comprised the majority of the microbial communities in this class (II) injection well facility (Figure 5.2).

Halanaerobium was the prevailing ASV in the injection well wastewaters (30-63% relative abundance, Figure 5.2). Further, Marinobacter was dominant in this system (7-20% relative abundance), reaching the highest relative abundance in the bottom of the Hudson Tank (Figure

5.2). Notable taxa (relative abundance >5%) in the early to middle stages of treatment also

145 included Acinetobacter, Halomonas, Flexistipes, Halopeptonella, and Pseudoalteromonas

(Figure 5.2). Several ASVs which were previously in low abundance or below detection were observed in the 30 µm Filter sample and/or Filter Effluent. Pseudomonas, Arcobacter,

Methanohalophilus, Idiomarina, Alcanivorax, Thalassolituus, uncultured Rhodobacteraceae, and

Methylobacter were detected at a higher relative abundance in these later samples (>2%) compared to all earlier samples (<2%). In particular, Methylobacter was detected solely in the 30

µm Filter and Filter Effluent samples, and reached 7.4% of the overall microbial community in the Filter Effluent.

We compared the injection well sample microbial communities with input, flowback, and produced fluid samples from a previously studied Marcellus Shale natural-gas well (Cluff et al.,

2014). The injection well samples were most similar to produced fluid samples from the

Marcellus Shale well (Figure 5.3). In particular, the Hudson Surface community was highly similar to the community in the Marcellus Shale Well-2 sample 82 days after fracturing (Figure

5.3). The injection well samples were dissimilar to input and early flowback samples (Figure

5.3). This finding was confirmed by ADONIS and ANOSIM significance tests, which indicated the injection well and produced fluid samples were statistically different from the input and flowback fluid samples (ADONIS significance test, R2=0.35784, p < 0.001; ANOSIM significant test, R2=0.8157, p<0.001). Seven ASVs were identified as indicator species (p<0.01); four

ASVs in the injection well and produced fluid samples (Halanaerobium, Flexistipes,

Methanohalophilus, Geotoga), and three in the input and early flowback fluid samples (Vibrio,

Halolactibacillus, unclassified Gammaproteobacteria).

146

5.4 Discussion

5.4.1 Class (II) injection well treatment processes alter microbial community members

The various unit processes of the pre-treatment process in the injection well facility impacted the diversity and microbial community structure. The initial separation of soluble hydrocarbons from wastewaters by the WILSA® water conditioner decreased microbial diversity, as evidenced by the Hudson Bottom and Hudson Surface samples (Table 5.1). Halanaerobium relative abundance increased substantially after water conditioning in the hydrocarbon-containing Hudson Surface sample (36% in Pre-Wilsa, 33% in Hudson Bottom, 63% in Hudson Surface), whereas most other bacterial genera decreased in relative abundance. Halanaerobium is the only gram-positive bacteria present in this system, and may be enriched in the hydrocarbon-layer (Hudson Surface) due to the effects of the WILSA® water conditioner. Slight increases in relative abundance were observed in the Hudson Surface sample in Pseudoalteromonas, Pseudomonas,

Methanohalophilus, Shewanella, Thalassospira, and Thalassolituus; however, these changes were very minor compared to the change in Halanaerobium and may be related to the ability of these organisms to metabolize hydrocarbons. The Hudson tank appeared to have concentrated cells in the de-solubilized hydrocarbon layer, as the process of separating aqueous (Hudson

Effluent) from non-aqueous phases (Hudson Surface) displayed an order of magnitude decrease in cell counts (Table 5.1). Further, the utilization of the 30 µm filter to remove particulates likely enriched certain bacterial genera and increased diversity (Table 5.1, Figure 5.2). Methylobacter was enriched in the wastewaters following filtration prior to injection, as evidenced by the absence of this genus in all samples prior to iron filtration, followed by its presence in the 30 µm

Filter and subsequent enrichment in the Filter Effluent (Figure 5.2). Nevertheless, the major bacterial genera present in this disposal facility remained at all time points, and cell counts did

147 not decrease below 104 cells/mL, indicating that pre-treatment did not reduce or remove microorganisms from wastewaters.

5.4.2 Injection well communities resemble produced fluid communities from hydraulically fractured wells

NMDS analysis revealed similarities between the injected well samples and produced fluids (49-

328 days after hydraulic fracturing) with Halanaerobium, Flexistipes, Methanohalophilus, and

Geotoga as indicator species in these samples. Halanaerobium and Methanohalophilus frequently dominate in hydraulically fractured oil and gas wells as produced fluid chemistry diverges from surface conditions and begins to resemble highly saline formation pore waters

(Daly et al., 2016). These two microbial community members and Geotoga (along with Ca.

Uticabacter, not identified in the injection well community) have been identified as participating in an amino acid fermentation network (Borton et al., 2018b). Flexistipes has been previously identified in Utica-Pt. Pleasant and Marcellus Shale natural-gas wells but its metabolic potential has not been investigated in the hydraulically fractured shale ecosystem (Booker et al., 2017;

Cluff et al., 2014). However, a genome bin of Flexistipes reconstructed from metagenomic data of an Halfdan oil field well revealed metabolic capacities for nitrate reduction, nitrogen fixation, iron reduction, glycolysis, glucogeonesis, Entner-Doudoroff, pentose phosphate, TCA cycle, hydrocarbon degradation, peptide degradation, and fatty acid degradation (Vigneron et al., 2017).

Similar metabolic capacities may be present in Flexistipes observed in this class (II) injection well facility.

148

The other major microbial community members observed in this class (II) injection well disposal facility have been widely identified in produced fluids from hydraulically fractured oil and gas wells. Pseudomonas, Arcobacter, Marinobacter, Halomonas, Pseudoalteromonas,

Acinetobacter, and Idiomarina, were found in the Marcellus Shale natural-gas well produced fluids from the first few weeks to months after hydraulic fracturing, a unique period in which oxygen, electron acceptors, and surface-injected organic chemicals are abundant (Cluff et al.,

2014). Although not found in Appalachian Basin shale produced fluids, Chromohalobacter has been documented in fluids from the Barnett Shale (An et al., 2017; Davis et al., 2012).

Thalassospira was identified in enrichments from Marcellus Shale flowback fluids (Eastham,

2012). Thalassolituus has not been documented in hydraulically fractured oil and gas wells, but a characterized strain from marine harbor seawater/sediment was reportedly halophilic and obligately utilized hydrocarbons (aliphatic C7-C20) as carbon sources. Alcanivorax has not been associated with hydraulic fracturing activities, but a strain of this genus has been well-studied for its ability to degrade oil hydrocarbons (Kasai et al., 2002; Schneiker et al., 2006).

Microorganisms such as Halanaerobium, Methanohalophilus, Marinobacter, Arcobacter,

Pseudomonas, and Flexistipes appear to persist throughout the produced fluid transport and treatment process, which includes but is not limited to: transportation by truck, exposure to surface conditions, and disposal facility holding tank chemical treatment. While previous studies have investigated survival strategies, community metabolisms, and virulence mechanisms of microbes in produced fluids from the Appalachian Shale (Borton et al., 2018b; Daly et al., 2016), studies are limited to wastewater microbial communities present during and right after fracturing.

149

One study used transcriptomics to investigate Marcellus Shale produced fluid storage impoundments and truck-transported produced fluids, determining several stress response mechanisms such as biofilm formation and oxidative stress, which may also be relevant in this class (II) injection well facility (Vikram et al., 2016). However, the microbial communities in the aforementioned study less closely resembled shale well produced fluid microbial communities

(Vikram et al., 2016). Namely, 16S rRNA gene data revealed the dominant orders were

Alphaproteobacteria and Epsilonproteobacteria, whereas the dominant order observed in this class (II) injection well facility was Gammaproteobacteria followed by Halanaerobiia,

Alphaproteobacteria, Deferribacteres, and Campylobacteria. Therefore, the apparent persistence of select microbes following hydraulic fracturing into wastewater disposal operations indicates noteworthy adaptability, prompting the need for further study.

5.4.3 Putative metabolic implications for the disposal facility

The microbial taxa detected in the wastewaters of this class (II) injection well are capable of having deleterious impacts on well infrastructure and injection efficacy through corrosion, biofilm formation, and formation clogging. Halanaerobium, the dominant genus identified in the wastewater samples, has previously been demonstrated to grow under the high pressures present in the subsurface through biofilm formation (Booker et al., 2018). Consequently, these bacteria may colonize the injection well target formation leading to pore clogging and pressure build-ups.

The continued addition of produced fluids containing cells and remnant carbon and energy sources may further encourage proliferation of biofilm-producing bacteria. Moreover, microbially-mediated redox reactions (e.g., sulfate and thiosulfate reduction, iron reduction) can corrode infrastructure vital to the longevity of the disposal well. Halanaerobium is capable of thiosulfate reduction (Booker et al., 2017), and Arcobacter present here may refuel thiosulfate

150 reduction via sulfur oxidation (Evans et al., 2018). Shewanella strains have been well-studied for their unique iron-reducing metabolisms and may be capable of iron reduction prior to particulate filtration (Lies et al., 2005; Nevin and Lovley, 2002). Due to the metabolic potential for damage to the well and injection formation by the identified microbial taxa, investigation of active microbial metabolisms in wastewaters prior to injection is still needed.

Produced fluids were augmented with two chemical additives intended to reduce corrosion and scale in the infrastructure of the disposal facility (see Table E1 in Appendix E). However, the chemical constituents disclosed here are unlikely to be effective as biocides unless large volumes are added to achieve high concentrations. Indeed, microbial biomass was present at 104-106 cells/mL in all samples during the cold winter month of January, a number likely to increase in the warmer summer months more suited for microbial proliferation. Halanaerobium and

Methanohalophilus are found in produced fluids years after hydraulic fracturing occurs in natural-gas wells, long after the injection of carbon and nitrogen-containing input fluids, suggesting these taxa may subsist in a produced fluid-shale pore network ecosystem (Booker et al., 2018; Booker et al., 2017; Borton et al., 2018b, 2018a; Daly et al., 2016). Likewise, these taxa and/or others may colonize the Mount Simon sandstone formation targeted in this class (II) injection well. Since the microorganisms detected in this injection well facility may have deleterious impacts on injection well efficacy, we recommend the use of a targeted biocide to avoid formation pore colonization leading to clogging and pressure increases.

5.5 Conclusion

Microbial life frequently detected in produced fluids from Appalachian Basin natural-gas wells were also discovered in a class (II) injection well disposal facility receiving Marcellus Shale 151 wastewaters. Despite transport, chemical treatment, and hydrocarbon separation, commonly reported microbes from natural gas wells including Halanaerobium, Flexistipes, and

Methanohalophilus persisted in the disposal facility. Specific genera of bacteria detected here have been previously associated with biofilm formation and redox reactions capable of negatively impacting the injection disposal process through formation pore clogging and corrosion of well casings. These results indicate the need for further investigation into how microbial metabolisms alter class (II) injection well disposal efficacy.

Author contributions:

P. Mouser and M. Evans designed the study and collected samples. M. Evans analyzed data and wrote the manuscript. J. Luek assisted with QIIME2 analysis.

Acknowledgements

We are grateful for sample access from the injection well facility. We express thanks to Mikayla

Borton, Rebecca Daly, and Kelly Wrighton for assistance with DNA extraction.

5.6 References

16S Illumina Amplicon Protocol : Earth Microbiome Project Available at: http://press.igsb.anl.gov/earthmicrobiome/protocols-and-standards/16s/ [Accessed January 20, 2019].

152

Arensdorf, J. J., Miner, K., Ertmoed, R., Clay, W. K., Stadnicki, P., and Voordouw, G. (2009). Mitigation of reservoir souring by nitrate in a produced-water reinjection system in Alberta. in (Society of Petroleum Engineers). doi:10.2118/121731-MS.

Arthur, D., Bohm, B., Coughlin, B. J., and Layne, M. (2008). Hydraulic fracturing considerations for natural gas wells of the Marcellus Shale. All Consulting, LLC, Cincinnati, OH.

Arthur, J. D., Langhus, B. G., and Patel, C. (2005). Technical summary of oil & gas produced water treatment technologies. All Consulting, LLC, Tulsa, OK.

Bolyen, E., Rideout, J. R., Dillon, M. R., Bokulich, N. A., Abnet, C., Al-Ghalith, G. A., et al. (2018). QIIME 2: Reproducible, interactive, scalable, and extensible microbiome data science. PeerJ Preprints 6, e27295v2. doi:10.7287/peerj.preprints.27295v2.

153

Callahan, B. J., McMurdie, P. J., Rosen, M. J., Han, A. W., Johnson, A. J. A., and Holmes, S. P. (2016). DADA2: High-resolution sample inference from Illumina amplicon data. Nature Methods 13, 581.

Cluff, M. A., Hartsock, A., MacRae, J. D., Carter, K., and Mouser, P. J. (2014). Temporal changes in microbial ecology and geochemistry in produced water from hydraulically fractured Marcellus shale gas wells. Environ. Sci. Technol. 48, 6508–6517. doi:10.1021/es501173p.

Eastham, J. L. (2012). Enrichment, characterization and identification of microbial communities associated with unconventional shale gas production water. Available at: http://search.proquest.com/docview/1038957062/abstract/43FBC8EE55394CC8PQ/1 [Accessed March 2, 2019].

154

Evans, M. V., Panescu, J., Hanson, A. J., Welch, S. A., Sheets, J. M., Nastasi, N., et al. (2018). Members of Marinobacter and Arcobacter influence system biogeochemistry during early production of hydraulically fractured natural gas wells in the appalachian basin. Frontiers in Microbiology 9, 2646. doi:10.3389/fmicb.2018.02646.

Getzinger, G. J., O’Connor, M. P., Hoelzer, K., Drollette, B. D., Karatum, O., Deshusses, M. A., et al. (2015). Natural gas residual fluids: sources, endpoints, and organic chemical composition after centralized waste treatment in Pennsylvania. Environ. Sci. Technol. 49, 8347–8355. doi:10.1021/acs.est.5b00471.

Kargbo, D. M., Wilhelm, R. G., and Campbell, D. J. (2010). Natural gas plays in the Marcellus Shale: challenges and potential opportunities. Environ. Sci. Technol. 44, 5679–5684. doi:10.1021/es903811p.

Kasai, Y., Kishira, H., Sasaki, T., Syutsubo, K., Watanabe, K., and Harayama, S. (2002). Predominant growth of Alcanivorax strains in oil-contaminated and nutrient-supplemented sea water. Environ. Microbiol. 4, 141–147. doi:10.1046/j.1462-2920.2002.00275.x.

155

Kim, W.-Y. (2013). Induced seismicity associated with fluid injection into a deep well in Youngstown, Ohio. Journal of Geophysical Research: Solid Earth 118, 3506–3518. doi:10.1002/jgrb.50247.

Lies, D. P., Hernandez, M. E., Kappler, A., Mielke, R. E., Gralnick, J. A., and Newman, D. K. (2005). Shewanella oneidensis MR-1 uses overlapping pathways for iron reduction at a distance and by direct contact under conditions relevant for biofilms. Appl. Environ. Microbiol. 71, 4414. doi:10.1128/AEM.71.8.4414-4426.2005.

Lipus, D., Roy, D., Khan, E., Ross, D., Vikram, A., Gulliver, D., et al. (2018). Microbial communities in Bakken region produced water. FEMS Microbiol Lett 365. doi:10.1093/femsle/fny107.

Nevin, K. P., and Lovley, D. R. (2002). Mechanisms for Fe(III) oxide reduction in sedimentary environments. Geomicrobiology Journal 19, 141–159. doi:10.1080/01490450252864253.

Rognes, T., Flouri, T., Nichols, B., Quince, C., and Mahé, F. (2016). VSEARCH: a versatile open source tool for metagenomics. PeerJ 4, e2584. doi:10.7717/peerj.2584.

Schneiker, S., dos Santos, V. A. M., Bartels, D., Bekel, T., Brecht, M., Buhrmester, J., et al. (2006). Genome sequence of the ubiquitous hydrocarbon-degrading marine bacterium Alcanivorax borkumensis. Nature Biotechnology 24, 997–1004. doi:10.1038/nbt1232.

156

U.S. Energy Information Administration (EIA) (2016). Shale gas production drives world natural gas production growth. Available at: https://www.eia.gov/todayinenergy/detail.php?id=27512 [Accessed March 9, 2019].

U.S. Energy Information Administration (EIA) (2018). Appalachia, Permian, Haynesville drive U.S. natural gas production growth. Available at: https://www.eia.gov/todayinenergy/detail.php?id=36934 [Accessed January 20, 2019].

U.S. Environmental Protection Agency (EPA) (2009). Technical development document for the effluent limitations guidelines and standards for the oil and gas extraction point source category. Available at: https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3378370 [Accessed January 18, 2019].

Vidic, R. D., Brantley, S. L., Vandenbossche, J. M., Yoxtheimer, D., and Abad, J. D. (2013). Impact of Shale Gas Development on Regional Water Quality. Science 340, 1235009. doi:10.1126/science.1235009.

Vigneron, A., Alsop, E. B., Lomans, B. P., Kyrpides, N. C., Head, I. M., and Tsesmetzis, N. (2017). Succession in the petroleum reservoir microbiome through an oil field production lifecycle. The ISME Journal 11, 2141–2154. doi:10.1038/ismej.2017.78.

Vikram, A., Lipus, D., and Bibby, K. (2016). Metatranscriptome analysis of active microbial communities in produced water samples from the Marcellus Shale. Microb Ecol 72, 571– 581. doi:10.1007/s00248-016-0811-z.

157

Weingarten, M., Ge, S., Godt, J. W., Bekins, B. A., and Rubinstein, J. L. (2015). High-rate injection is associated with the increase in U.S. mid-continent seismicity. Science 348, 1336. doi:10.1126/science.aab1345.

158

5.7 Figures

Figure 5.1. Photos and schematic of fluid processing steps prior to injection at the class (II) injection well disposal facility

159

163

Figure 5.2. Relative abundance of amplicon sequence variance (ASVs) at the genus level (rel. abund. ≥ 0.01). Each plot, left to right, corresponds to different relative abundance levels, >0.10, 0.05-0.10, and 0.01-0.05.

160

Figure 5.3. Non-metric multidimensional scaling (NMDS) of Bray-Curtis dissimilarity distances for the injection well samples and two hydraulically fractured natural-gas wells from the Marcellus Shale (input, flowback, and produced fluid). Marcellus Shale sample labels correspond to well number (1 or 2) and days after flowback began (D0-D328). Ellipses correspond to the 95% confidence interval (as rendered by ggplot) for two statistically significant clusters. Dendrogram tree indicates relationships between samples based on Bray-Curtis dissimilarity. Indicator species for each sample cluster (k=2) are reported in the plot (p<0.01).

161

Table 5.1. Sequencing results and cell counts in 6 samples throughout the class (II) injection well treatment process Shannon's

Sample Sample Name No. Sequences Cells mL-1 Diversity

1 Pre-Wilsa 30,003 (1.99 ± 0.14)*105 5.01

2 Hudson Bottom 21,468 (4.09 ± 0.20)*105 5.16

3 Hudson Surface 22,162 (1.34 ± 0.05)*106 2.35

4 Hudson Effluent 15,756 (9.04 ± 0.50)*104 4.42

5 30 µm Filter 18,434 n.m. 5.84

6 Filter Effluent 30,689 (2.63 ± 0.13)*105 8.41

162

Chapter 6. Conclusions and Future Work

The results of the objectives for the research listed in Chapter 1 are addressed here, along with the potential future research topics that could further enhance the findings for each chapter.

6.1 Conclusions and Future Studies

6.1.1 Examine members of the Marinobacter and Arcobacter genera for gene pathways that may impact nitrogen, carbon, and sulfur cycles during the early months after HF.

Chapter 2 investigated the putative metabolisms of Marinobacter and Arcobacter strains isolated from Utica-Point Pleasant and Marcellus Shale natural-gas wells using a combination of isolate genomics, field metagenomics and geochemistry, and phenotypic characterization of isolates. Since previous studies observed these taxa in produced fluids during the initial months following HF and suggested nitrogen, carbon, and sulfur cycling capabilities, we specifically sought to identify genes associated with these pathways.

Phenotypic analyses identified salinity growth ranges and carbon sources utilized by

Marinobacter sp. UTICA-S1B6 and Arcobacter spp. UTICA-S4D1 and MARC-

MIP3H16. Phylogenetic analysis of the near-full-length 16S rRNA gene demonstrated these strains are highly similar to other Marinobacter and Arcobacter strains observed in

HF natural-gas wells. Importantly, we identified several pathways for the utilization of

163 nitrogen, carbon, and sulfur compounds in both taxa. The Marinobacter sp. UTICA-

S1B6 genome has genes encoding for denitrification, conversion of urea to ammonia and carbon dioxide, and utilization of aliphatic and aromatic hydrocarbons. The Arcobacter spp. UTICA-S4D1 and MARC-MIP3H16 genomes have genes encoding for dissimilatory nitrate reduction to ammonia (DNRA), sulfur oxidation (sox system) and a reductive TCA cycle. Urea conversion to ammonia by Marinobacter and DNRA by

Arcobacter may explain high concentrations of ammonia observed in the produced fluids from the wells observed here. Moreover, the sox system in Arcobacter may result in oxidation of the reduced sulfur compounds which were in high concentrations in these wells. Generation of oxidized forms of sulfur (e.g. sulfate, thiosulfate) may encourage proliferation of thiosulfate-reducing Halanaerobium that eventually dominates in later produced fluids and may cause souring of natural-gas and corrosion of the O&G well.

Our results demonstrated the important roles these taxa likely play during early production of HF natural-gas wells.

Chapter 2 could be expanded upon in several ways. Other than the salinity and carbon source analyses, described metabolisms were putative and based on metagenomic data.

Therefore, further laboratory testing on Marinobacter and Arcobacter isolates could confirm or deny the proposed metabolic potential. Data could be substantiated by culturing Marinobacter with different hydrocarbon compounds and analyzing protein relative abundance during growth coupled to changes in hydrocarbon concentrations as measured by gas chromatograpy-mass spectrometry (GC-MS). Analyzing the ability of

164

Arcobacter to utilize the reductive TCA cycle could be accomplished by culturing the isolate with CO2 gas as the sole carbon source and quantifying CO2 and total organic carbon (TOC) during the growth cycle. In lieu of culturing experiments, field-scale metatranscriptomic and metagenomic analysis on temporal samples from a newly- fractured well could be performed to determine the activity of Marinobacter and

Arcobacter during early production of HF O&G wells.

6.1.2 Determine if microorganisms in produced fluids transform xenobiotic glycol surfactants commonly used during HF operations.

Chapter 3 sought to identify the mechanism(s) behind the glycol surfactant losses observed in a Utica-Point Pleasant natural-gas well using culture-independent methods

(field surfactant chemistry, geochemistry, and metagenomics) coupled to culture- dependent methods (isolation, proteomics and surfactant chemistry during isolate growth). A geochemical end-member mixing model ruled out the possibility of surfactant losses by dilution and/or mixing, therefore we turned our attention to biotransformation mechanisms. Metagenomic data revealed chain-shortening genes present in produced fluids over 200 days after HF largely belonging to the dominant microbial taxa,

Halanaerobium. We isolated Halanaerobium and cultured cells with a commercially- available surfactant mixture containing alkyl polyethoxylates (AEOs) and polypropylene glycols (PPGs). AEOs and PPGs were both transformed by Halanaerobium in the presence of glucose, and shotgun proteomics revealed evidence of a previously 165 uncharacterized co-metabolic surfactant chain-shortening pathway via the methylglyoxal bypass during glycolysis. Besides glucose, this co-metabolism may occur with carbon sources like guar gum, which is frequently used as a gelling agent in HF fluid formulations, and can also be shuttled through glycolysis to activate the methylglyoxal bypass. Moreover, we discovered that while Halanaerobium could metabolize AEOs,

PPGs, and monomeric glycols (ethylene and propylene glycols), polyethylene glycols

(PEGs) were not transformed either in the field or during batch growth of

Halanaerobium, indicating the existence of surfactant substrate specificity. This work identified the important role microorganisms in produced fluids have on injected xenobiotic organic chemicals.

One aspect discussed in this chapter, but not tested, was the coupling of surfactant metabolism with a carbon source present during HF operations such as guar gum. A laboratory experiment culturing Halanaerobium on surfactants with guar gum would provide further evidence of this co-metabolism during HF operations. A challenge related to this work is simulating subsurface conditions during HF. The use of high-pressure reaction chambers with shale rock in future work is advisable to better simulate environmental conditions and connect this metabolism back to field data. Shale particulates may promote abiotic sorption of surfactants in addition to biotic transformation, whereas the high-pressure conditions may accelerate the methylglyoxal bypass during glycolysis while producing significantly less biomass, as has been described previously (Booker, 2018). Stable isotope analysis on field and laboratory

166 surfactant data may further substantiate this work. Since microorganisms preferentially transform lighter isotopes over heavier isotopes, examining the carbon isotopes of AEOs and glycols during both isolate batch growth and in field analyses could further confirm biotically-mediated surfactant transformation. Several lower abundance taxa

(Thermoanaerobacter, Clostridiales) were present in the few days after HF occurred that contained the propanediol dehydratase chain-shortening gene pduC. Isolation and experimental culturing of these taxa with surfactants may shed light on the ability of microorganisms other than Halanaerobium to metabolize glycol surfactants.

6.1.3 Elucidate putative microbial (de)halogenation pathways in HF natural-gas wells to determine if organohalides in produced fluids may be biotically produced and/or transformed.

Chapter 4 employed a metagenomic approach to identify (de)halogenation genes and assess the potential for microbial organohalide cycling in four HF Appalachian Basin natural-gas wells. While several abiotic mechanisms are suspected to generate organohalides in HF produced fluids, there have been no investigations into the potential contributions of microorganisms to the produce or transform these compounds. Genes of one class of halogenases, the non-heme chloroperoxidases, were identified in all four natural-gas wells up to one month after HF occurred and were highly similar to genes found in Pseudomonas, which was a major microbial community member at those time points. Non-heme chloroperoxidase genes non-specifically halogenate organic matter 167 after the enzymatic generation of hypohalous acids (e.g. HOCl, HOBr). Additionally, hydrolytic and reductive dehalogenase genes were identified in all four natural-gas wells up to eleven months after HF. Reductive dehalogenase genes observed in one Utica-Point

Pleasant natural-gas well were homologous to halophenol reductive dehalogenase genes previously observed in facultative microorganisms. These genes were highly similar to strains in the Burkholderiales order, which was present at 4% of the microbial community during the period of gene detection. Results indicate microorganisms likely play a role in organohalide generation and transformation, potentially in non-specifically halogenating organic matter during the month after HF, as well as in halide cleavage from organohalides generated by abiotic and biotic mechanisms months after HF.

This chapter was limited to metagenomic analysis, which can only confirm metabolic potential but not active function. Adding metatranscriptomic analysis on field samples, which can provide a picture of the actively transcribed genes, would help confirm if

(de)halogenation genes are active during HF operations. Another way to investigate active microbial dehalogenation in produced fluids is to enrich for halogenating or dehalogenating microbial taxa in fluids from a newly-fractured O&G well. The dehalogenation enrichments could contain an added organohalide compound (e.g., 2- chlorophenol) supplemented with basal media, and experimental conditions varied in different enrichment sets (e.g., salinity, oxygen concentration, electron donors).

(Meta)genome sequencing and (meta)transcriptome analysis of these enrichments would be performed to identify responsible taxa and actively expressed genes in enrichments.

168

Results would be coupled to GC-MS analysis to track organohalide concentrations from inoculation to the end of the experiment. These experiments would more directly link microbial metagenomic potential to in situ function. They could also help elucidate the portion of organohalides transformed biotically versus abiotically in O&G wastewaters and conditions necessary for these reactions to occur.

6.1.4 Identify changes in microbial communities and evaluate potential mechanisms for persistence in the treatment processes of a class (II) injection well facility disposing of produced fluids from the Marcellus Shale.

Chapter five surveyed the microbial community composition as determined by 16S rRNA gene sequencing at six sampling points through the treatment process of a class (II) injection well disposal facility. Although there are numerous studies on produced fluid microbial communities, and one study describing the decreased water quality of streams near these disposal facilities, there are no reports on the microbial communities present in produced fluids in injection well facilities. This research showed that despite exposure to oxygen, bulk fluid physical changes (e.g., surface and interfacial tension changes), and several chemical additives used in the fluid treatment process, microbial communities closely resembled those coming from HF O&G wells. We also identified changes in specific taxa after several unit processes, namely following the Wilsa® which separated soluble hydrocarbons from the aqueous portion of the fluid through alterations in surface tension and viscosity, as well as during the colloidal and particulate filtration step which 169 is intended to remove insoluble iron. We compared the injection well microbial communities to input, flowback, and produced fluid microbial communities from a previously studied Marcellus Shale natural-gas well, finding that the injection well communities were highly similar to produced fluid samples from the hydraulically fractured well. These results are the first describing microbial communities in class (II) injection well disposal facilities and identifying key changes in microbial community members during common unit treatment process. In addition, this work revealed the capacity for microorganisms to contribute to clogging of the disposal target formation through biofilm formation.

The investigation in chapter 5 was restricted to 16S rRNA gene analysis of the microbial community and did not include any geochemical or organic chemical analyses per the request of the injection well facility operators. In the future, sampling at several sites rather than just one to compare site variations, and performing metagenomic analysis coupled to geochemical (e.g., salinity, iron, redox conditions) and organic chemical (e.g., bulk inorganic and organic carbon) analyses would greatly enrich the study. Using non- targeted analytical chemistry techniques (e.g., gas and liquid chromatography coupled to mass spectrometry) would shed light on the impacts of unit processes on fluid chemistry throughout pretreatment, as well as help identify potential carbon sources for microorganisms. Further, bio-clogging could be investigated in detail by extracting fluid prior to injection at a sampling site and simulating subsurface conditions over the course of several months in an anoxic high-pressure reactor with a porous rock sample

170 resembling target formation rock types (e.g., sandstone). Microbial biomass could be quantified by flow cytometry, ATP could be measured as an indicator of microbial activity, and any present biofilms would be visualized by scanning electron microscopy.

Carbon sources and/or various types of microbial media could be added if initial results indicate no microbial growth. An experiment like this may provide evidence of bio- clogging by produced fluid microorganisms in disposal formations.

6.2 Concluding Thoughts

The research described in this dissertation elucidated several microbial-chemical interactions in the hydraulically fractured shale ecosystem. Genomes of commonly identified microbial taxa (e.g., Marinobacter, Arcobacter, Halanaerobium) were analyzed for metabolic potential relevant to deep subsurface conditions during operations. Interactions between shale-associated taxa and chemicals commonly injected during hydraulic fracturing operations were revealed using culture-dependent methods

(e.g., testing growth of Halanaerobium on a commercial surfactant mixture using shotgun proteomics and genomics) and culture-independent methods (e.g., identifying key genes in shale metagenomes). This work suggests that halotolerant microorganisms play an important role in chemical cycling in saline produced fluids during hydraulic fracturing operations.

The work presented here was enabled by partnerships with industry, allowing site and sample access. These chapters include the work of numerous collaborators, including

171 microbiologists, chemists, engineers, and geologists, who made this interdisciplinary work possible and greatly enriched these studies.

172

Bibliography

16S Illumina Amplicon Protocol : Earth Microbiome Project Available at: http://press.igsb.anl.gov/earthmicrobiome/protocols-and-standards/16s/ [Accessed January 20, 2019]. Abualfaraj, N., Gurian, P. L., and Olson, M. S. (2018). Assessing residential exposure risk from spills of flowback water from Marcellus Shale hydraulic fracturing activity. Int J Environ Res Public Health 15. doi:10.3390/ijerph15040727. Agrawal, A., and Gieg, L. M. (2013). In situ detection of anaerobic alkane metabolites in subsurface environments. Front. Microbiol. 4. doi:10.3389/fmicb.2013.00140. Ahrenst, A., Lungwitz, B., Fredd, C. N., Abad, C., Gurmen, N., Chen, Y., et al. (2015). Friction reduction fluids. Available at: http://www.freepatentsonline.com/9034802.html [Accessed January 4, 2019]. Akob, D. M., Cozzarelli, I. M., Dunlap, D. S., Rowan, E. L., and Lorah, M. M. (2015). Organic and inorganic composition and microbiology of produced waters from Pennsylvania shale gas wells. Applied Geochemistry 60, 116–125. doi:10.1016/j.apgeochem.2015.04.011. Akob, D. M., Mumford, A. C., Orem, W., Engle, M. A., Klinges, J. G., Kent, D. B., et al. (2016). Wastewater disposal from unconventional oil and gas development degrades stream quality at a West Virginia injection facility. Environ Sci Technol 50, 5517–5525. doi:10.1021/acs.est.6b00428. Akyon, B., McLaughlin, M., Hernández, F., Blotevogel, J., and Bibby, K. (2019). Characterization and biological removal of organic compounds from hydraulic fracturing produced water. Environmental Science: Processes & Impacts. doi:10.1039/C8EM00354H. Alteri, C. J., and Mobley, H. L. T. (2016). The versatile type VI secretion system. Microbiol Spectr 4. doi:10.1128/microbiolspec.VMBF-0026-2015. Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990). Basic local alignment search tool. J. Mol. Biol. 215, 403–410. doi:10.1016/S0022-2836(05)80360-2. 173

An, B. A., Shen, Y., and Voordouw, G. (2017). Control of sulfide production in high salinity Bakken Shale oil reservoirs by halophilic bacteria reducing nitrate to nitrite. Front. Microbiol. 8. doi:10.3389/fmicb.2017.01164. Arensdorf, J. J., Miner, K., Ertmoed, R., Clay, W. K., Stadnicki, P., and Voordouw, G. (2009). Mitigation of reservoir souring by nitrate in a produced-water reinjection system in Alberta. in (Society of Petroleum Engineers). doi:10.2118/121731-MS. Arthur, D., Bohm, B., Coughlin, B. J., and Layne, M. (2008). Hydraulic fracturing considerations for natural gas wells of the Marcellus Shale. All Consulting, LLC, Cincinnati, OH. Arthur, J. D., Langhus, B. G., and Patel, C. (2005). Technical summary of oil & gas produced water treatment technologies. All Consulting, LLC, Tulsa, OK. Assessment, U. E. N. C. for E. (2009). Technical development document for the effluent limitations guidelines and standards for the oil and gas extraction point source category. Available at: https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3378370 [Accessed January 18, 2019]. Bagheri, M., Amoozegar, M. A., Didari, M., Makhdoumi-Kakhki, A., Schumann, P., Spröer, C., et al. (2013). Marinobacter persicus sp. nov., a moderately halophilic bacterium from a saline lake in Iran. Antonie Van Leeuwenhoek 104, 47–54. doi:10.1007/s10482-013- 9923-3. Barbot, E., Vidic, N. S., Gregory, K. B., and Vidic, R. D. (2013). Spatial and temporal correlation of water quality parameters of produced waters from Devonian-age shale following hydraulic fracturing. Environ. Sci. Technol. 47, 2562–2569. doi:10.1021/es304638h. Bengtson, P., Bastviken, D., and Öberg, G. (2013). Possible roles of reactive chlorine II: assessing biotic chlorination as a way for organisms to handle oxygen stress. Environmental Microbiology 15, 991–1000. doi:10.1111/j.1462-2920.2012.02807.x. Bobik, T. A., Havemann, G. D., Busch, R. J., Williams, D. S., and Aldrich, H. C. (1999). The propanediol utilization (pdu) operon of Salmonella enterica serovar Typhimurium LT2

174

includes genes necessary for formation of polyhedral organelles involved in coenzyme B(12)-dependent 1, 2-propanediol degradation. J. Bacteriol. 181, 5967–5975. Bolyen, E., Rideout, J. R., Dillon, M. R., Bokulich, N. A., Abnet, C., Al-Ghalith, G. A., et al. (2018). QIIME 2: Reproducible, interactive, scalable, and extensible microbiome data science. PeerJ Preprints 6, e27295v2. doi:10.7287/peerj.preprints.27295v2. Bomberg, M., and Ahonen, L. (2017). Editorial: Geomicrobes: life in terrestrial deep subsurface. Front. Microbiol. 8. doi:10.3389/fmicb.2017.00103. Bonis, B. M., and Gralnick, J. A. (2015). Marinobacter subterrani, a genetically tractable neutrophilic Fe(II)-oxidizing strain isolated from the Soudan Iron Mine. Front. Microbiol. 6. doi:10.3389/fmicb.2015.00719. Booker A.E., Hoyt D.W., Meulia T., Eder E., Nicora C.D., Purvine S.O., et al. (2018) Deep subsurface pressure stimulates metabolic versatility in shale-colonizing Halanaerobium. In review. Appl Environ Microbiol. Booker, A. E., Borton, M. A., Daly, R. A., Welch, S. A., Nicora, C. D., Hoyt, D. W., et al. (2017). Sulfide generation by dominant Halanaerobium microorganisms in hydraulically fractured shales. mSphere 2. doi:10.1128/mSphereDirect.00257-17. Booth, I. R., Ferguson, G. P., Miller, S., Li, C., Gunasekera, B., and Kinghorn, S. (2003). Bacterial production of methylglyoxal: a survival strategy or death by misadventure? Biochemical Society Transactions 31, 1406–1408. doi:10.1042/bst0311406. Borgeaud, S., Metzger, L. C., Scrignari, T., and Blokesch, M. (2015). The type VI secretion system of Vibrio cholerae fosters horizontal gene transfer. Science 347, 63–67. doi:10.1126/science.1260064. Borton, M. A., Daly, R. A., O’Banion, B., Hoyt, D. W., Marcus, D. N., Welch, S., et al. (2018a). Comparative genomics and physiology of the genus Methanohalophilus, a prevalent methanogen in hydraulically fractured shale. Environmental Microbiology 20, 4596–4611. doi:10.1111/1462-2920.14467. Borton, M. A., Hoyt, D. W., Roux, S., Daly, R. A., Welch, S. A., Nicora, C. D., et al. (2018b). Coupled laboratory and field investigations resolve microbial interactions that underpin

175

persistence in hydraulically fractured shales. PNAS, 201800155. doi:10.1073/pnas.1800155115. Brown, C. T., Hug, L. A., Thomas, B. C., Sharon, I., Castelle, C. J., Singh, A., et al. (2015). Unusual biology across a group comprising more than 15% of domain Bacteria. Nature 523, 208. Brussaard, C. P. D. (2004). Optimization of procedures for counting viruses by flow cytometry. Appl. Environ. Microbiol. 70, 1506–1513. doi:10.1128/AEM.70.3.1506- 1513.2004. Callahan, B. J., McMurdie, P. J., Rosen, M. J., Han, A. W., Johnson, A. J. A., and Holmes, S. P. (2016). DADA2: High-resolution sample inference from Illumina amplicon data. Nature Methods 13, 581. Cameron, D. C., and Cooney, C. L. (1986). A novel fermentation: the production of R(–)–1,2– propanediol and acetol by Clostridium thermosaccharolyticum. Bio/Technology 4, 651. Caspi, R., Altman, T., Billington, R., Dreher, K., Foerster, H., Fulcher, C. A., et al. (2014). The MetaCyc database of metabolic pathways and enzymes and the BioCyc collection of Pathway/Genome Databases. Nucleic Acids Research 42, D459–D471. doi:10.1093/nar/gkt1103. Castresana, J. (2002). GBLOCKS: selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Version 0.91 b. Copyrighted by J. Castresana, EMBL. Chakraborty, S., Sivaraman, J., Leung, K. Y., and Mok, Y.-K. (2011). Two-component PhoB- PhoR regulatory system and ferric uptake regulator sense phosphate and iron to control virulence genes in type III and VI secretion systems of Edwardsiella tarda. J Biol Chem 286, 39417–39430. doi:10.1074/jbc.M111.295188. Chandrangsu, P., Dusi, R., Hamilton, C. J., and Helmann, J. D. (2014). Methylglyoxal resistance in Bacillus subtilis: Contributions of bacillithiol-dependent and independent pathways. Mol Microbiol 91, 706–715. doi:10.1111/mmi.12489. Chapman, E. C., Capo, R. C., Stewart, B. W., Kirby, C. S., Hammack, R. W., Schroeder, K. T., et al. (2012). Geochemical and strontium isotope characterization of produced waters

176

from Marcellus Shale natural gas extraction. Environ. Sci. Technol. 46, 3545–3553. doi:10.1021/es204005g. Chen, I.-M. A., Markowitz, V. M., Chu, K., Palaniappan, K., Szeto, E., Pillay, M., et al. (2017). IMG/M: integrated genome and metagenome comparative data analysis system. Nucleic Acids Res 45, D507–D516. doi:10.1093/nar/gkw929. Chen, K., Huang, L., Xu, C., Liu, X., He, J., Zinder, S. H., et al. (2013). Molecular characterization of the enzymes involved in the degradation of a brominated aromatic herbicide. Molecular Microbiology 89, 1121–1139. doi:10.1111/mmi.12332. Choudhary, L., Macdonald, D. D., and Alfantazi, A. (2015). Role of thiosulfate in the corrosion of steels: a review. Corrosion. 71, 1147–1168. doi:10.5006/1709. Cluff, M. A., Hartsock, A., MacRae, J. D., Carter, K., and Mouser, P. J. (2014). Temporal changes in microbial ecology and geochemistry in produced water from hydraulically fractured Marcellus Shale gas wells. Environ. Sci. Technol. 48, 6508–6517. doi:10.1021/es501173p. Dai, M., Rogers, J. B., Warner, J. R., and Copley, S. D. (2003). A previously unrecognized step in pentachlorophenol degradation in Sphingobium chlorophenolicum is catalyzed by tetrachlorobenzoquinone reductase (pcpd). Journal of Bacteriology 185, 302–310. doi:10.1128/JB.185.1.302-310.2003. Daly, R. A., Borton, M. A., Wilkins, M. J., Hoyt, D. W., Kountz, D. J., Wolfe, R. A., et al. (2016). Microbial metabolisms in a 2.5-km-deep ecosystem created by hydraulic fracturing in shales. Nature Microbiology 1, nmicrobiol2016146. doi:10.1038/nmicrobiol.2016.146. Daly, R. A., Roux, S., Borton, M. A., Morgan, D. M., Johnston, M. D., Booker, A. E., et al. (2018). Viruses control dominant bacteria colonizing the terrestrial deep biosphere after hydraulic fracturing. Nature Microbiology. doi:10.1038/s41564-018-0312-6. Darrah, T. H., Jackson, R. B., Vengosh, A., Warner, N. R., Whyte, C. J., Walsh, T. B., et al. (2015). The evolution of Devonian hydrocarbon gases in shallow aquifers of the northern Appalachian Basin: Insights from integrating noble gas and hydrocarbon geochemistry. Geochimica et Cosmochimica Acta 170, 321–355. doi:10.1016/j.gca.2015.09.006.

177

Davis, J. P., Struchtemeyer, C. G., and Elshahed, M. S. (2012). Bacterial communities associated with production facilities of two newly drilled thermogenic natural gas wells in the Barnett Shale (Texas, USA). Microb Ecol 64, 942–954. doi:10.1007/s00248-012- 0073-3. Diemer, G. S., Kyle, J. E., and Stedman, K. M. Counting viruses using polycarbonate Track EtchTM membrane filters as an alternative to AnodiscTM membrane filters. Donachie, S. P. (2005). Arcobacter halophilus sp. nov., the first obligate halophile in the genus Arcobacter. International Journal of Systematic and Evolutionary Microbiology 55, 1271–1277. doi:10.1099/ijs.0.63581-0. Eastham, J. L. (2012). Enrichment, characterization and identification of microbial communities associated with unconventional shale gas production water. Available at: http://search.proquest.com/docview/1038957062/abstract/43FBC8EE55394CC8PQ/1 [Accessed March 2, 2019]. Edgar, R. C. (2010). Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26, 2460–2461. Eichhorn, E., van der Ploeg, J. R., and Leisinger, T. (1999). Characterization of a two- component alkanesulfonate monooxygenase from Escherichia coli. J. Biol. Chem. 274, 26639–26646. Ellsworth, W. L. (2013). Injection-induced earthquakes. Science 341, 1225942–1225942. doi:10.1126/science.1225942. Elsner, M., and Hoelzer, K. (2016). Quantitative survey and structural classification of hydraulic fracturing chemicals reported in unconventional gas production. Environ. Sci. Technol. 50, 3290–3314. doi:10.1021/acs.est.5b02818. Evans, M. V., Panescu, J., Hanson, A. J., Welch, S. A., Sheets, J. M., Nastasi, N., et al. (2018). Members of Marinobacter and Arcobacter influence system biogeochemistry during early production of hydraulically fractured natural gas wells in the Appalachian Basin. Front. Microbiol. 9, 2646. doi:10.3389/fmicb.2018.02646. Fetzner, S. (1998). Bacterial dehalogenation. Appl Microbiol Biotechnol 50, 633–657. doi:10.1007/s002530051346.

178

Fichter, J., Wunch, K., Moore, R., Summer, E., Braman, S., and Holmes, P. (2012). How hot is too hot for bacteria? A technical study assessing bacterial establishment in downhole drilling, fracturing and stimulation operations. in NACE-2012-1310. Friedrich, C. G., Bardischewsky, F., Rother, D., Quentmeier, A., and Fischer, J. (2005). Prokaryotic sulfur oxidation. Curr. Opin. Microbiol. 8, 253–259. doi:10.1016/j.mib.2005.04.005. Gaspar, J., Mathieu, J., Yang, Y., Tomson, R., Leyris, J. D., Gregory, K. B., et al. (2014). Microbial dynamics and control in shale gas production. Environ. Sci. Technol. Lett. 1, 465–473. doi:10.1021/ez5003242. Gauthier, M. J., Lafay, B., Christen, R., Fernandez, L., Acquaviva, M., Bonin, P., et al. (1992). Marinobacter hydrocarbonoclasticus gen. nov., sp. nov., a new, extremely halotolerant, hydrocarbon-degrading marine bacterium. Int. J. Syst. Bacteriol. 42, 568–576. doi:10.1099/00207713-42-4-568. Getzinger, G. J., O’Connor, M. P., Hoelzer, K., Drollette, B. D., Karatum, O., Deshusses, M. A., et al. (2015). Natural gas residual fluids: sources, endpoints, and organic chemical composition after centralized waste treatment in Pennsylvania. Environ. Sci. Technol. 49, 8347–8355. doi:10.1021/acs.est.5b00471. Goldman, P., Milne, G. W. A., and Keister, D. B. (1968). Carbon-halogen bond cleavage III. Studies on bacterial halidohydrolases. Journal of biological chemistry 243, 428–434. Gribble, G. W. (2015). A recent survey of naturally occurring organohalogen compounds. Environ. Chem. 12, 396–405. doi:10.1071/EN15002. Haluszczak, L. O., Rose, A. W., and Kump, L. R. (2013). Geochemical evaluation of flowback brine from Marcellus gas wells in Pennsylvania, USA. Applied Geochemistry 28, 55–61. doi:10.1016/j.apgeochem.2012.10.002. Hanson A.J., Lipp J.S., Daly R.A., Sheets J.M., Welch S.A., Sharma S., et al. Microbial membrane lipids show de novo ecosystem transformation in hydraulically fractured Marcellus Shale. In preparation. Hanson, A. J., Luek, J. L., Tummings, S. S., McLaughlin, M. C., Blotevogel, J., and Mouser, P. J. (2019). High total dissolved solids in shale gas wastewater inhibit biodegradation of

179

alkyl and nonylphenol ethoxylate surfactants. Science of The Total Environment 668, 1094–1103. doi:10.1016/j.scitotenv.2019.03.041. Harkness, J. S., Darrah, T. H., Warner, N. R., Whyte, C. J., Moore, M. T., Millot, R., et al. (2017). The geochemistry of naturally occurring methane and saline groundwater in an area of unconventional shale gas development. Geochimica et Cosmochimica Acta 208, 302–334. doi:10.1016/j.gca.2017.03.039. Harkness, J. S., Dwyer, G. S., Warner, N. R., Parker, K. M., Mitch, W. A., and Vengosh, A. (2015). Iodide, bromide, and ammonium in hydraulic fracturing and oil and gas wastewaters: environmental implications. Environ. Sci. Technol. 49, 1955–1963. doi:10.1021/es504654n. Hayes, T. (2009). Sampling and analysis of water streams associated with the development of Marcellus Shale gas. Des Plaines, IL: Marcellus Shale Coalition Available at: http://www.water-research.net/naturalgasPA/pdffiles/MSCommission-Report.pdf. Hayes, T., Severin, B. F., Engineer, P. S. P., and Okemos, M. I. (2012). Barnett and Appalachian Shale water management and reuse technologies. Contract 8122, 05. Heilweil, V. M., Solomon, D. K., Darrah, T. H., Gilmore, T. E., and Genereux, D. P. (2016). Gas-tracer experiment for evaluating the fate of methane in a coastal plain stream: degassing versus in-stream oxidation. Environ. Sci. Technol. 50, 10504–10511. doi:10.1021/acs.est.6b02224. Heyob, K. M., Blotevogel, J., Brooker, M., Evans, M. V., Lenhart, J. J., Wright, J., et al. (2017). Natural attenuation of nonionic surfactants used in hydraulic fracturing fluids: degradation rates, pathways, and mechanisms. Environ. Sci. Technol. 51, 13985–13994. doi:10.1021/acs.est.7b01539. Hoelzer, K., Sumner, A. J., Karatum, O., Nelson, R. K., Drollette, B. D., O’Connor, M. P., et al. (2016). Indications of transformation products from hydraulic fracturing additives in shale-gas wastewater. Environ. Sci. Technol. 50, 8036–8048. doi:10.1021/acs.est.6b00430.

180

Hofmann, B., Tölzer, S., Pelletier, I., Altenbuchner, J., van Pée, K. H., and Hecht, H. J. (1998). Structural investigation of the cofactor-free chloroperoxidases. Journal of Molecular Biology 279, 889–900. doi:10.1006/jmbi.1998.1802. Holliger, C., and Schumacher, W. (1994). Reductive dehalogenation as a respiratory process. Antonie van Leeuwenhoek 66, 239–246. doi:10.1007/BF00871642. Houlton, B. Z., Wang, Y.-P., Vitousek, P. M., and Field, C. B. (2008). A unifying framework for dinitrogen fixation in the terrestrial biosphere. Nature 454, 327. Huang, Y.-L., Li, Q.-B., Deng, X., Lu, Y.-H., Liao, X.-K., Hong, M.-Y., et al. (2005). Aerobic and anaerobic biodegradation of polyethylene glycols using sludge microbes. Process Biochemistry 40, 207–211. doi:10.1016/j.procbio.2003.12.004. Huber, M., Meyer, U., and Rys, P. (2000). Biodegradation mechanisms of linear alcohol ethoxylates under anaerobic conditions. Environ. Sci. Technol. 34, 1737–1741. doi:10.1021/es9903680. Hug, L. A., Maphosa, F., Leys, D., Löffler, F. E., Smidt, H., Edwards, E. A., et al. (2013). Overview of organohalide-respiring bacteria and a proposal for a classification system for reductive dehalogenases. Philosophical Transactions of the Royal Society B: Biological Sciences 368, 20120322. doi:10.1098/rstb.2012.0322. Hügler, M., Wirsen, C. O., Fuchs, G., Taylor, C. D., and Sievert, S. M. (2005). Evidence for

autotrophic CO2 fixation via the reductive tricarboxylic acid cycle by members of the ɛ subdivision of Proteobacteria. J Bacteriol 187, 3020–3027. doi:10.1128/JB.187.9.3020- 3027.2005. Huntemann, M., Ivanova, N. N., Mavromatis, K., Tripp, H. J., Paez-Espino, D., Palaniappan, K., et al. (2015). The standard operating procedure of the DOE-JGI Microbial Genome Annotation Pipeline (MGAP v.4). Standards in Genomic Sciences 10, 86. doi:10.1186/s40793-015-0077-y. Hurst, J. K. (2014). Hypohalous acid. doi:10.1036/1097-8542.334300. Huu, N. B., Denner, E. B. M., Ha, D. T. C., Wanner, G., and Stan-Lotter, H. (1999). Marinobacter aquaeolei sp. nov., a halophilic bacterium isolated from a Vietnamese oil-

181

producing well. International Journal of Systematic and Evolutionary Microbiology 49, 367–375. doi:10.1099/00207713-49-2-367. Hyatt, D., LoCascio, P. F., Hauser, L. J., and Uberbacher, E. C. (2012). Gene and translation initiation site prediction in metagenomic sequences. Bioinformatics 28, 2223–2230. Jesenská, A., Pavlová, M., Strouhal, M., Chaloupková, R., Těšínská, I., Monincová, M., et al. (2005). Cloning, biochemical properties, and distribution of mycobacterial haloalkane dehalogenases. Appl. Environ. Microbiol. 71, 6736–6745. doi:10.1128/AEM.71.11.6736- 6745.2005. Jugder, B.-E., Ertan, H., Bohl, S., Lee, M., Marquis, C. P., and Manefield, M. (2016). Organohalide respiring bacteria and reductive dehalogenases: key tools in organohalide bioremediation. Front. Microbiol. 7. doi:10.3389/fmicb.2016.00249. Kahrilas, G. A., Blotevogel, J., Corrin, E. R., and Borch, T. (2016). Downhole transformation of the hydraulic fracturing fluid biocide glutaraldehyde: implications for flowback and produced water quality. Environ. Sci. Technol. 50, 11414–11423. doi:10.1021/acs.est.6b02881. Kanehisa, M., and Goto, S. (2000). KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 28, 27–30. Kang, M., Christian, S., Celia, M. A., Mauzerall, D. L., Bill, M., Miller, A. R., et al. (2016). Identification and characterization of high methane-emitting abandoned oil and gas wells. Proc Natl Acad Sci USA 113, 13636. doi:10.1073/pnas.1605913113. Kargbo, D. M., Wilhelm, R. G., and Campbell, D. J. (2010). Natural gas plays in the Marcellus Shale: challenges and potential Opportunities. Environ. Sci. Technol. 44, 5679–5684. doi:10.1021/es903811p. Kasai, Y., Kishira, H., Sasaki, T., Syutsubo, K., Watanabe, K., and Harayama, S. (2002). Predominant growth of Alcanivorax strains in oil-contaminated and nutrient- supplemented sea water. Environ. Microbiol. 4, 141–147. doi:10.1046/j.1462- 2920.2002.00275.x. Kawai, F. (2002). Microbial degradation of polyethers. Appl Microbiol Biotechnol 58, 30–38. doi:10.1007/s00253-001-0850-2.

182

Kawai, M., Futagami, T., Toyoda, A., Takaki, Y., Nishi, S., Hori, S., et al. (2014). High frequency of phylogenetically diverse reductive dehalogenase-homologous genes in deep subseafloor sedimentary metagenomes. Front. Microbiol. 5. doi:10.3389/fmicb.2014.00080. Kell, S. (2011). State oil and gas agency groundwater investigations and their role in advancing regulatory reforms, a two-state review: Ohio and Texas. Ground Water Protection Council Available at: http://fracfocus.org/sites/default/files/publications/state_oil__gas_agency_groundwater_i nvestigations_optimized.pdf. Kerr, R. A. (2010). Natural gas from shale bursts onto the scene. Science 328, 1624–1626. doi:10.1126/science.328.5986.1624. Kim, H. M., Hwang, C. Y., and Cho, B. C. (2010). Arcobacter marinus sp. nov. International Journal of Systematic and Evolutionary Microbiology 60, 531–536. doi:10.1099/ijs.0.007740-0. Kim, S., and Pevzner, P. A. (2014). MS-GF makes progress towards a universal database search tool for proteomics. Nat Commun 5: 5277. Kim, W.-Y. (2013). Induced seismicity associated with fluid injection into a deep well in Youngstown, Ohio. Journal of Geophysical Research: Solid Earth 118, 3506–3518. doi:10.1002/jgrb.50247. Klatt, J. M., and Polerecky, L. (2015). Assessment of the stoichiometry and efficiency of CO2 fixation coupled to reduced sulfur oxidation. Front. Microbiol. 6, 484. doi:10.3389/fmicb.2015.00484. Kondash, A., and Vengosh, A. (2015). Water footprint of hydraulic fracturing. Environ. Sci. Technol. Lett. 2, 276–280. doi:10.1021/acs.estlett.5b00211. Krasner, S. W., Weinberg, H. S., Richardson, S. D., Pastor, S. J., Chinn, R., Sclimenti, M. J., et al. (2006). Occurrence of a new generation of disinfection byproducts. Environ. Sci. Technol. 40, 7175–7185. doi:10.1021/es060353j. Kurihara, T., Yamauchi, T., Ichiyama, S., Takahata, H., and Esaki, N. (2003). Purification, characterization, and gene cloning of a novel fluoroacetate dehalogenase from

183

Burkholderia sp. FA1. Journal of Molecular Catalysis B: Enzymatic 23, 347–355. doi:10.1016/S1381-1177(03)00098-5. Kutvonen, H., Rajala, P., Carpén, L., and Bomberg, M. (2015). Nitrate and ammonia as nitrogen sources for deep subsurface microorganisms. Front. Microbiol. 6. doi:10.3389/fmicb.2015.01079. Langmead, B., and Salzberg, S. L. (2012). Fast gapped-read alignment with Bowtie 2. Nat Meth 9, 357–359. Lester, Y., Ferrer, I., Thurman, E. M., Sitterley, K. A., Korak, J. A., Aiken, G., et al. (2015). Characterization of hydraulic fracturing flowback water in Colorado: implications for water treatment. Sci. Total Environ. 512–513, 637–644. doi:10.1016/j.scitotenv.2015.01.043. Letunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for the display and annotation of phylogenetic and other trees. Nucleic Acids Research 44, W242–W245. doi:10.1093/nar/gkw290. Li, S.-M., Leithead, A., Moussa, S. G., Liggio, J., Moran, M. D., Wang, D., et al. (2017). Differences between measured and reported volatile organic compound emissions from oil sands facilities in Alberta, Canada. Proc Natl Acad Sci U S A 114, E3756–E3765. doi:10.1073/pnas.1617862114. Liang, R., Davidova, I. A., Marks, C. R., Stamps, B. W., Harriman, B. H., Stevenson, B. S., et al. (2016). Metabolic capability of a predominant Halanaerobium sp. in hydraulically fractured gas wells and its implication in pipeline corrosion. Front. Microbiol. 988. doi:10.3389/fmicb.2016.00988. Lies, D. P., Hernandez, M. E., Kappler, A., Mielke, R. E., Gralnick, J. A., and Newman, D. K. (2005). Shewanella oneidensis MR-1 uses overlapping pathways for iron reduction at a distance and by direct contact under conditions relevant for biofilms. Appl. Environ. Microbiol. 71, 4414. doi:10.1128/AEM.71.8.4414-4426.2005. Lipus, D., Roy, D., Khan, E., Ross, D., Vikram, A., Gulliver, D., et al. (2018). Microbial communities in Bakken region produced water. FEMS Microbiol Lett 365. doi:10.1093/femsle/fny107.

184

Lipus, D., Vikram, A., Ross, D., Bain, D., Gulliver, D., Hammack, R., et al. (2017). Predominance and metabolic potential of Halanaerobium in produced water from hydraulically fractured Marcellus Shale wells. Appl. Environ. Microbiol., AEM.02659- 16. doi:10.1128/AEM.02659-16. Lu, Q., Yu, L., Liang, Z., Yan, Q., He, Z., Luan, T., et al. (2017). Dehalococcoides as a potential biomarker evidence for uncharacterized organohalides in environmental samples. Front. Microbiol. 8. doi:10.3389/fmicb.2017.01677. Luek, J. L., and Gonsior, M. (2017). Organic compounds in hydraulic fracturing fluids and wastewaters: A review. Water Res. 123, 536–548. doi:10.1016/j.watres.2017.07.012. Luek, J. L., Harir, M., Schmitt-Kopplin, P., Mouser, P. J., and Gonsior, M. (2018). Temporal dynamics of halogenated organic compounds in Marcellus Shale flowback. Water Research 136, 200–206. doi:10.1016/j.watres.2018.02.055. Luek, J. L., Schmitt-Kopplin, P., Mouser, P. J., Petty, W. T., Richardson, S. D., and Gonsior, M. (2017). Halogenated organic compounds identified in hydraulic fracturing wastewaters using ultrahigh resolution mass spectrometry. Environ. Sci. Technol. 51, 5377–5385. doi:10.1021/acs.est.6b06213. Maguire-Boyle, S. J., and Barron, A. R. (2014). Organic compounds in produced waters from shale gas wells. Environ. Sci.: Processes Impacts 16, 2237–2248. doi:10.1039/C4EM00376D. Makarova, K. S., Wolf, Y. I., Alkhnbashi, O. S., Costa, F., Shah, S. A., Saunders, S. J., et al. (2015). An updated evolutionary classification of CRISPR-Cas systems. Nat. Rev. Microbiol. 13, 722–736. doi:10.1038/nrmicro3569. Mars, A. E., Kasberg, T., Kaschabek, S. R., Agteren, M. H. van, Janssen, D. B., and Reineke, W. (1997). Microbial degradation of chloroaromatics: use of the meta-cleavage pathway for mineralization of chlorobenzene. J. Bacteriol. 179, 4530–4537. doi:10.1128/jb.179.14.4530-4537.1997. McLaughlin, M. C., Borch, T., and Blotevogel, J. (2016). Spills of hydraulic fracturing chemicals on agricultural topsoil: biodegradation, sorption, and co-contaminant interactions. Environ. Sci. Technol. 50, 6071–6078. doi:10.1021/acs.est.6b00240.

185

Miller, C. S., Baker, B. J., Thomas, B. C., Singer, S. W., and Banfield, J. F. (2011). EMIRGE: reconstruction of full-length ribosomal genes from microbial community short read sequencing data. Genome Biology 12, R44. doi:10.1186/gb-2011-12-5-r44. Mohan, A. M., Hartsock, A., Bibby, K. J., Hammack, R. W., Vidic, R. D., and Gregory, K. B. (2013). Microbial community changes in hydraulic fracturing fluids and produced water from shale gas extraction. Environ. Sci. Technol. 47, 13141–13150. doi:10.1021/es402928b. Mohn, W. W., and Tiedje, J. M. (1992). Microbial reductive dehalogenation. Microbiol. Rev. 56, 482. Moore, M. T., Vinson, D. S., Whyte, C. J., Eymold, W. K., Walsh, T. B., and Darrah, T. H. (2018). Differentiating between biogenic and thermogenic sources of natural gas in coalbed methane reservoirs from the Illinois Basin using noble gas and hydrocarbon geochemistry. Geological Society, London, Special Publications 468, 151. doi:10.1144/SP468.8. Mösche, M. (2004). Anaerobic degradability of alcohol ethoxylates and related non-ionic surfactants. Biodegradation 15, 327–336. Mounier, J., Camus, A., Mitteau, I., Vaysse, P.-J., Goulas, P., Grimaud, R., et al. (2014). The marine bacterium Marinobacter hydrocarbonoclasticus SP17 degrades a wide range of lipids and hydrocarbons through the formation of oleolytic biofilms with distinct gene expression profiles. FEMS Microbiol Ecol 90, 816–831. doi:10.1111/1574-6941.12439. Mouser, P. J., Borton, M., Darrah, T. H., Hartsock, A., and Wrighton, K. C. (2016). Hydraulic fracturing offers view of microbial life in the deep terrestrial subsurface. FEMS Microbiol Ecol 92. doi:10.1093/femsec/fiw166. Nagata, Y., Miyauchi, K., Damborsky, J., Manova, K., Ansorgova, A., and Takagi, M. (1997). Purification and characterization of a haloalkane dehalogenase of a new substrate class from a gamma-hexachlorocyclohexane-degrading bacterium, Sphingomonas paucimobilis UT26. Appl Environ Microbiol 63, 3707–3710. Nell, M., and Helbling, D. E. (2018). Exploring matrix effects and quantifying organic additives in hydraulic fracturing associated fluids using liquid chromatography

186

electrospray ionization mass spectrometry. Environmental Science: Processes & Impacts. doi:10.1039/C8EM00135A. Nevin, K. P., and Lovley, D. R. (2002). Mechanisms for Fe(III) Oxide Reduction in Sedimentary Environments. Geomicrobiology Journal 19, 141–159. doi:10.1080/01490450252864253. Nixon, S. L., Walker, L., Streets, M. D. T., Eden, B., Boothman, C., Taylor, K. G., et al. (2017). Guar gum stimulates biogenic sulfide production at elevated pressures: implications for shale gas extraction. Front. Microbiol. 8. doi:10.3389/fmicb.2017.00679. Noble, R., and Fuhrman, J. (1998). Use of SYBR Green I for rapid epifluorescence counts of marine viruses and bacteria. Aquatic Microbial Ecology 14, 113–118. doi:10.3354/ame014113. Orem, W., Tatu, C., Varonka, M., Lerch, H., Bates, A., Engle, M., et al. (2014). Organic substances in produced and formation water from unconventional natural gas extraction in coal and shale. International Journal of Coal Geology 126, 20–31. doi:10.1016/j.coal.2014.01.003. Oren, A. (2008). Microbial life at high salt concentrations: phylogenetic and metabolic diversity. Saline Syst. 4, 2. doi:10.1186/1746-1448-4-2. Paoletti, A. C., Parmely, T. J., Tomomori-Sato, C., Sato, S., Zhu, D., Conaway, R. C., et al. (2006). Quantitative proteomic analysis of distinct mammalian Mediator complexes using normalized spectral abundance factors. Proc Natl Acad Sci USA 103, 18928. doi:10.1073/pnas.0606379103. Parker, K. M., Zeng, T., Harkness, J., Vengosh, A., and Mitch, W. A. (2014). Enhanced formation of disinfection byproducts in shale gas wastewater-impacted drinking water supplies. Environ. Sci. Technol. 48, 11161–11169. doi:10.1021/es5028184. Patel, A., Noble, R. T., Steele, J. A., Schwalbach, M. S., Hewson, I., and Fuhrman, J. A. (2007). Virus and prokaryote enumeration from planktonic aquatic environments by epifluorescence microscopy with SYBR Green I. Nature protocols 2, 269. Pavlová, M., Klvana, M., Jesenská, A., Prokop, Z., Konecná, H., Sato, T., et al. (2007). The identification of catalytic pentad in the haloalkane dehalogenase DhmA from

187

Mycobacterium avium N85: reaction mechanism and molecular evolution. J. Struct. Biol. 157, 384–392. doi:10.1016/j.jsb.2006.09.004. Petsch, S. T., Berner, R. A., and Eglinton, T. I. (2000). A field study of the chemical weathering of ancient sedimentary organic matter. Organic Geochemistry 31, 475–487. doi:10.1016/S0146-6380(00)00014-0. Pirzadeh, P., Lesage, K. L., and Marriott, R. A. (2014). Hydraulic fracturing additives and the delayed onset of hydrogen sulfide in shale gas. Energy Fuels 28, 4993–5001. doi:10.1021/ef501059k. Ploeg, J. van der, Willemsen, M., Hall, G. van, and Janssen, D. B. (1995). Adaptation of Xanthobacter autotrophicus GJ10 to bromoacetate due to activation and mobilization of the haloacetate dehalogenase gene by insertion element IS1247. Journal of Bacteriology 177, 1348–1356. doi:10.1128/jb.177.5.1348-1356.1995. Posada, D., Darriba, D., Taboada, G. L., and Doallo, R. (2011). ProtTest 3: fast selection of best-fit models of protein evolution. Bioinformatics 27, 1164–1165. doi:10.1093/bioinformatics/btr088. Quevillon, E., Silventoinen, V., Pillai, S., Harte, N., Mulder, N., Apweiler, R., et al. (2005). InterProScan: protein domains identifier. Nucleic acids research 33, W116–W120. Rich, A. L., and Orimoloye, H. T. (2016). Elevated atmospheric levels of benzene and benzene-related compounds from unconventional shale extraction and processing: human health concern for residential communities. Environ Health Insights 10, 75–82. doi:10.4137/EHI.S33314. Roalkvam, I., Drønen, K., Stokke, R., Daae, F. L., Dahle, H., and Steen, I. H. (2015). Physiological and genomic characterization of Arcobacter anaerophilus IR-1 reveals new metabolic features in Epsilonproteobacteria. Front. Microbiol. 6. doi:10.3389/fmicb.2015.00987. Rogers, J. D., Burke, T. L., Osborn, S. G., and Ryan, J. N. (2015). A framework for identifying organic compounds of concern in hydraulic fracturing fluids based on their mobility and persistence in groundwater. Environ. Sci. Technol. Lett. 2, 158–164. doi:10.1021/acs.estlett.5b00090.

188

Rogers, J. D., Thurman, E. M., Ferrer, I., Rosenblum, J. S., Evans, M. V., Mouser, P. J., et al. (2018). Degradation of polyethylene glycols and polypropylene glycols in microcosms simulating a spill of produced water in shallow groundwater. Environ. Sci.: Processes Impacts. doi:10.1039/C8EM00291F. Rognes, T., Flouri, T., Nichols, B., Quince, C., and Mahé, F. (2016). VSEARCH: a versatile open source tool for metagenomics. PeerJ 4, e2584. doi:10.7717/peerj.2584. Rook, J. J. (1977). Chlorination reactions of fulvic acids in natural waters. Environ. Sci. Technol. 11, 478–482. doi:10.1021/es60128a014. Rosenblum, J., Thurman, E. M., Ferrer, I., Aiken, G., and Linden, K. G. (2017). Organic chemical characterization and mass balance of a hydraulically fractured well: from fracturing fluid to produced water over 405 days. Environ. Sci. Technol. 51, 14006– 14015. doi:10.1021/acs.est.7b03362. Russell, A. B., Peterson, S. B., and Mougous, J. D. (2014). Type VI secretion system effectors: poisons with a purpose. Nature Reviews Microbiology 12, 137–148. doi:10.1038/nrmicro3185. Schneiker, S., dos Santos, V. A. M., Bartels, D., Bekel, T., Brecht, M., Buhrmester, J., et al. (2006). Genome sequence of the ubiquitous hydrocarbon-degrading marine bacterium Alcanivorax borkumensis. Nature Biotechnology 24, 997–1004. doi:10.1038/nbt1232. Schuchmann, K., Schmidt, S., Lopez, A. M., Kaberline, C., Kuhns, M., Lorenzen, W., et al. (2015). Nonacetogenic growth of the acetogen Acetobacterium woodii on 1,2- propanediol. Journal of Bacteriology 197, 382–391. doi:10.1128/JB.02383-14. Singer, E., Webb, E. A., Nelson, W. C., Heidelberg, J. F., Ivanova, N., Pati, A., et al. (2011). Genomic potential of Marinobacter aquaeolei, a biogeochemical “opportunitroph.” Appl. Environ. Microbiol. 77, 2763–2771. doi:10.1128/AEM.01866-10. Stamatakis, A. (2014). RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30, 1312–1313. doi:10.1093/bioinformatics/btu033. Straß, A., and Schink, B. (1986). Fermentation of polyethylene glycol via acetaldehyde in Pelobacter venetianus. Appl Microbiol Biotechnol 25, 37–42. doi:10.1007/BF00252510.

189

Stringfellow, W. T., Domen, J. K., Camarillo, M. K., Sandelin, W. L., and Borglin, S. (2014). Physical, chemical, and biological characteristics of compounds used in hydraulic fracturing. Journal of Hazardous Materials 275, 37–54. doi:10.1016/j.jhazmat.2014.04.040. Strong Lisa C., Gould Trevor, Kasinkas Lisa, Sadowsky Michael J., Aksan Alptekin, and Wackett Lawrence P. (2014). Biodegradation in waters from hydraulic fracturing: chemistry, microbiology, and engineering. Journal of Environmental Engineering 140, B4013001. doi:10.1061/(ASCE)EE.1943-7870.0000792. Struchtemeyer, C. G., and Elshahed, M. S. (2012). Bacterial communities associated with hydraulic fracturing fluids in thermogenic natural gas wells in North Central Texas, USA. FEMS Microbiol. Ecol. 81, 13–25. doi:10.1111/j.1574-6941.2011.01196.x. Thacker, J. B., Carlton, D. D., Hildenbrand, Z. L., Kadjo, A. F., and Schug, K. A. (2015). Chemical analysis of wastewater from unconventional drilling operations. Water 7, 1568–1579. doi:10.3390/w7041568. Thurman, E. M., Ferrer, I., Blotevogel, J., and Borch, T. (2014). Analysis of hydraulic fracturing flowback and produced waters using accurate mass: identification of ethoxylated surfactants. Anal. Chem. 86, 9653–9661. doi:10.1021/ac502163k. Tour, J. M., Kittrell, C., and Colvin, V. L. (2010). Green carbon as a bridge to renewable energy. Nature Materials 9, 871. Tran-Din, K., and Gottschalk, G. (1985). Formation of d(-)-1,2-propanediol and d(-)-lactate from glucose by Clostridium sphenoides under phosphate limitation. Arch. Microbiol. 142, 87–92. doi:10.1007/BF00409243. Trifunović, D., Schuchmann, K., and Müller, V. (2016). Ethylene glycol metabolism in the acetogen Acetobacterium woodii. J. Bacteriol. 198, 1058–1065. doi:10.1128/JB.00942- 15. T. Stringfellow, W., and Kay Camarillo, M. (2019). Flowback verses first-flush: new information on the geochemistry of produced water from mandatory reporting. Environmental Science: Processes & Impacts. doi:10.1039/C8EM00351C.

190

Tummings, S., Panescu, J., Daly, R. A., Wrighton, K. C., and Mouser, P. J. (2018). Draft Genome sequences of Marinobacter strains recovered from Utica Shale-produced fluids. Genome Announc. 6, e00155-18. doi:10.1128/genomeA.00155-18. U.S. Energy Information Administration (EIA) (2016a). Hydraulic fracturing accounts for about half of current U.S. crude oil production - Today in Energy -. Available at: https://www.eia.gov/todayinenergy/detail.php?id=25372 [Accessed October 7, 2017]. U.S. Energy Information Administration (EIA) (2016b). Hydraulically fractured wells provide two-thirds of U.S. natural gas production - Today in Energy -. Available at: https://www.eia.gov/todayinenergy/detail.php?id=26112 [Accessed October 16, 2018]. U.S. Energy Information Administration (EIA) (2016c). Shale gas production drives world natural gas production growth. Available at: https://www.eia.gov/todayinenergy/detail.php?id=27512 [Accessed March 9, 2019]. U.S. Energy Information Administration (EIA) (2018a). Appalachia, Permian, Haynesville drive U.S. natural gas production growth. Available at: https://www.eia.gov/todayinenergy/detail.php?id=36934 [Accessed January 20, 2019]. U.S. Energy Information Administration (EIA) (2018b). Hydraulically fractured horizontal wells account for most new oil and natural gas wells - Today in Energy. Available at: https://www.eia.gov/todayinenergy/detail.php?id=34732 [Accessed October 16, 2018]. U.S. Energy Information Administration (EIA) (2018c). The United States exported more natural gas than it imported in 2017 - Today in Energy. Available at: https://www.eia.gov/todayinenergy/detail.php?id=35392 [Accessed November 13, 2018]. U.S. Environmental Protection Agency (2015). Toxic and priority pollutants under the clean water act. Available at: https://www.epa.gov/eg/toxic-and-priority-pollutants-under- clean-water-act [Accessed January 28, 2019]. U.S. Environmental Protection Agency (2016). Hydraulic Fracturing for Oil and Gas: Impacts from the Hydraulic Fracturing Water Cycle on Drinking Water Resources in the United States (Final Report). (No. EPA/600/R-16/236F). Available at: https://cfpub.epa.gov/ncea/hfstudy/recordisplay.cfm?deid=332990 [Accessed January 18, 2019].

191 van Pée, K.-H., and Unversucht, S. (2003). Biological dehalogenation and halogenation reactions. Chemosphere 52, 299–312. doi:10.1016/S0045-6535(03)00204-2. Vaysse, P.-J., Prat, L., Mangenot, S., Cruveiller, S., Goulas, P., and Grimaud, R. (2009). Proteomic analysis of Marinobacter hydrocarbonoclasticus SP17 biofilm formation at the alkane-water interface reveals novel proteins and cellular processes involved in hexadecane assimilation. Research in Microbiology 160, 829–837. doi:10.1016/j.resmic.2009.09.010. Veenagayathri, K., and Vasudevan, N. (2011). Ortho and meta cleavage dioxygenases detected during the degradation of phenolic compounds by a moderately halophilic bacterial consortium. Int Res J Microbiol 2, 406–414. Veil, J. (2015). U.S. Produced water volumes and management practices in 2012. Oklahoma, City, OK: Ground Water Protection Council Available at: http://www.gwpc.org/sites/default/files/Produced%20Water%20Report%202014- GWPC_0.pdf. Vidic, R. D., Brantley, S. L., Vandenbossche, J. M., Yoxtheimer, D., and Abad, J. D. (2013). Impact of shale gas development on regional water quality. Science 340, 1235009. doi:10.1126/science.1235009. Vigneron, A., Alsop, E. B., Lomans, B. P., Kyrpides, N. C., Head, I. M., and Tsesmetzis, N. (2017). Succession in the petroleum reservoir microbiome through an oil field production lifecycle. The ISME Journal 11, 2141–2154. doi:10.1038/ismej.2017.78. Vikram, A., Lipus, D., and Bibby, K. (2014). Produced water exposure alters bacterial response to biocides. Environ. Sci. Technol. 48, 13001–13009. doi:10.1021/es5036915. Vikram, A., Lipus, D., and Bibby, K. (2016). Metatranscriptome analysis of active microbial communities in produced water samples from the Marcellus Shale. Microb Ecol 72, 571– 581. doi:10.1007/s00248-016-0811-z. Voordouw, G. (2008). “Impact of nitrate on the sulfur cycle in oil fields,” in Microbial Sulfur Metabolism (Springer, Berlin, Heidelberg), 296–302. doi:10.1007/978-3-540-72682- 1_23.

192

Voordouw, G., Grigoryan, A. A., Lambo, A., Lin, S., Park, H. S., Jack, T. R., et al. (2009). Sulfide remediation by pulsed injection of nitrate into a low temperature Canadian heavy oil reservoir. Environ. Sci. Technol. 43, 9512–9518. doi:10.1021/es902211j. Wagener, S., and Schink, B. (1988). Fermentative degradation of nonionic surfactants and polyethylene glycol by enrichment cultures and by pure cultures of homoacetogenic and propionate-forming bacteria. Appl. Environ. Microbiol. 54, 561–565. Warner, N. R., Christie, C. A., Jackson, R. B., and Vengosh, A. (2013). Impacts of shale gas wastewater disposal on water quality in western Pennsylvania. Environ. Sci. Technol. 47, 11849–11857. doi:10.1021/es402165b. Weigold, P., El-Hadidi, M., Ruecker, A., Huson, D. H., Scholten, T., Jochmann, M., et al. (2016). A metagenomic-based survey of microbial (de)halogenation potential in a German forest soil. Scientific Reports 6, 28958. doi:10.1038/srep28958. Weinbauer, M. G., Beckmann, C., and Höfle, M. G. (1998). Utility of green fluorescent nucleic acid dyes and aluminum oxide membrane filters for rapid epifluorescence enumeration of soil and sediment bacteria. Appl. Environ. Microbiol. 64, 5000–5003. Weingarten, M., Ge, S., Godt, J. W., Bekins, B. A., and Rubinstein, J. L. (2015). High-rate injection is associated with the increase in U.S. mid-continent seismicity. Science 348, 1336. doi:10.1126/science.aab1345. Weljie, A. M., Newton, J., Mercier, P., Carlson, E., and Slupsky, C. M. (2006). Targeted profiling: quantitative analysis of 1H NMR metabolomics data. Anal. Chem. 78, 4430– 4442. doi:10.1021/ac060209g. Wiesner, W., Pée, K. H. van, and Lingens, F. (1988). Purification and characterization of a novel bacterial non-heme chloroperoxidase from Pseudomonas pyrrocinia. J. Biol. Chem. 263, 13725–13732. Wirsen, C. O., Sievert, S. M., Cavanaugh, C. M., Molyneaux, S. J., Ahmad, A., Taylor, L. T., et al. (2002). Characterization of an autotrophic sulfide-oxidizing marine Arcobacter sp. that produces filamentous sulfur. Appl. Environ. Microbiol. 68, 316–325. Wiśniewski, J. R., Zougman, A., Nagaraj, N., and Mann, M. (2009). Universal sample preparation method for proteome analysis. Nature Methods 6, 359.

193

Wrighton, K. C., Thomas, B. C., Sharon, I., Miller, C. S., Castelle, C. J., VerBerkmoes, N. C., et al. (2012). Fermentation, hydrogen, and sulfur metabolism in multiple uncultivated bacterial phyla. Science 337, 1661–1665. Wuchter, C., Banning, E., Mincer, T., Drenzek, N. J., and Coolen, M. J. (2013). Microbial diversity and methanogenic activity of Antrim Shale formation waters from recently fractured wells. Front. Microbiol. 4. doi:10.3389/fmicb.2013.00367. Xia, Z., Idowu, I., Marvin, C., Thomas, P. J., Johnson, W., Francisco, O., et al. (2019). Identification of halogenated polycyclic aromatic hydrocarbons in biological samples from Alberta Oil-Sands Region. Chemosphere 215, 206–213. doi:10.1016/j.chemosphere.2018.10.050. Xu, C., Couvillion, S. P., Sontag, R. L., Isern, N. G., Maezato, Y., Lindemann, S. R., et al. (2019). MetFish: a metabolomics platform for studying microbial communities in chemically extreme environments. bioRxiv, 518647. doi:10.1101/518647. Xun, L., Topp, E., and Orser, C. S. (1992). Purification and characterization of a tetrachloro-p- hydroquinone reductive dehalogenase from a Flavobacterium sp. J. Bacteriol. 174, 8003– 8007. Yeh, E., Garneau, S., and Walsh, C. T. (2005). Robust in vitro activity of RebF and RebH, a two-component reductase/halogenase, generating 7-chlorotryptophan during rebeccamycin biosynthesis. Proc. Natl. Acad. Sci. U.S.A. 102, 3960–3965. doi:10.1073/pnas.0500755102. Zhang, J., Xin, Y., Cao, X., Xue, S., and Zhang, W. (2014). Purification and characterization of 2-haloacid dehalogenase from marine bacterium Paracoccus sp. DEH99, isolated from marine sponge Hymeniacidon perlevis. J. Ocean Univ. China 13, 91–96. doi:10.1007/s11802-014-2357-3.

194

Appendix A. Supporting Information for Chapter 2

195

199

Table3A1. Biolog plate reader carbon source testing results 196

Figure12A2. FracFocus Disclosure for Utica-S1

197

Figure13A3. FracFocus Disclosure for Utica-S4

198

Figure14A4. FracFocus Disclosure for Marcellus-MIP3H

199

Table4A1. Salinity curve results for Marinobacter and Arcobacter

200

Table5A2. Pulled metagenome genes meeting bitscore and identity cutoffs

Arcobacter Well 3H Date 12202015 2172016 Day after fracturing began 34 93 Gene copies belonging to Arcobacter 3421 5455 Total genes in MG timepoint 93381 44088 Total genes graded "A-C" 80465 35502 # gene copies (bitscore > 200, Gene Name Abbreviation EC Number KEGG Number identity > 45.9%) periplasmic nitrate reductase A napA 1.7.99.- K02567 1 1 periplasmic nitrate reductase B napB n/a K02568 1 1 nitrite reductase cytochrome c nrfA 1.7.2.2 K03385 1 1 cytochrome c nitrite reductase small subunit nrfH n/a K15876 1 1 nitronate monooxygenase NMO 1.13.12.16 K00459 1 1 nitroalkane oxidase NAO 1.7.3.1 K19823 0 0 sulfur oxidizing protein soxA soxA n/a K17222 1 1 sulfur oxidizing protein soxB soxB n/a K17224 1 1 sulfane dehydrogenase subunit soxC soxC n/a K17225 1 1 cytochrome C soxD n/a K08738 1 1 sulfur oxidizing protein soxX soxX n/a K17223 1 1 sulfur oxidizing protein soxY soxY n/a K17226 1 1 sulfur oxidizing protein soxZ soxZ n/a K17227 1 1 tetrathionate reductase subunit A ttrA n/a K08357 1 1 tetrathionate reductase subunit B ttrB n/a K08358 1 1 tetrathionate reductase subunit C ttrC n/a K08359 0 0 sulfide:quinone reductase sqr 1.8.5.4 K17218 0 0 thiosulfate/3-mercaptopyruvate sulfurtransferase sseA 2.8.1.1/2.8.1.2 K01011 0 0 type VI secretion system VgrG n/a K11891 1 1 Hcp n/a K11892 1 1 Lip/vasD n/a K11903 1 1 IcmF/impL/vasKn/a K11904 1 5

DotU/impK/ompA/vasFn/a K11906 0 0 ClpV/vasG n/a K11907 1 1 PpkA n/a K11912 0 0 Fha1 n/a K11913 0 0

PppA/stp1 n/a K11915 0 0 protease many 3.4.21.- many 1 1 phospholipase many 3.1.1.-,3.1.4.- 1 0 Continued

201

Table A2 continued. Marinobacter Well 3H Date 12202015 Day after fracturing began 11 Gene copies belonging to Marinobacter 34149 Total genes in MG timepoint 93381

Total genes graded "A-C" 80465

# gene copies (bitscore > 200, identity > Gene Name Abbreviation EC Number KEGG Number 45.9%)

K0001,K00121,K04072,K1 1440,K13951,K13952,K139 alcohol dehydrogenase many EC 1.1.1.1,1.1.1.-53,K13954,K13980,K18857 8 (general) aldehyde dehydrogenase many EC 1.2.1.- many 14 aldehyde dehydrogenase n/a 1.2.1.3 K00128,K00149,K14085 3 betaine dehydrogenase betB 1.2.1.8 K00130,K14085 15 GB transport n/a 3.6.3.32 K02000 7 choline dehydrogenase n/a 1.1.99.1 K00108 2 alkane sulfonate monooxygenase ssuD 1.14.14.5 K04091 3 alkane oxygenase alkB 1.14.11.33 K03919,K10859 0 K00496,K07425,K17687, alkane 1 monooxygenase n/a 1.14.15.3 K17688 1 benzene, toluene initial cleavage "phenol hydroxylase"1.14.13.- many 2 catechol monooxygenase catA 1.13.11.1/1.13.99.1K00469 3 phenol 2-monooxygenase pheA1 1.14.13.7/1.14.14.20K03380 1 (S)-2-haloacid dehalogenase L-DEX 3.8.1.2 K01560 6 glyoxylate reductase glcDEF 1.1.1.26 K00104,K11472,K11473 10

K00049,K00090,K12972,K glyoxylate reductase (NADP+) 1.1.1.79 18121 5 glycerate dehydrogenase hprA 1.1.1.29 K00018 7 nitrate reductase narGHIJ 1.7.99.4 K00370,K00371,K00374 23 nitrite reductase nirS 1.7.2.1 K00368/K15864 6 nitric oxide reductase norBC 1.7.2.5 K04561/K02305 6 nitrous oxide reductase nosZ 1.7.2.4 K00376 4 Continued

202

Table A2 continued. # gene copies (bitscore > 200, identity > Gene Name Abbreviation EC Number KEGG Number 45.9%) periplasmic nitrate reductase A napA 1.7.99.- K02567 0 periplasmic nitrate reductase B napB n/a K02568 0 nitrite reductase cytochrome c nrfA 1.7.2.2 K03385 0 cytochrome c nitrite reductase small subunit nrfH n/a K15876 0 nitronate monooxygenase NMO 1.13.12.16 K00459 4 nitroalkane oxidase NAO 1.7.3.1 K19823 0 sulfate adenylyl transferase cysNCD 2.7.7.4 K00955/K00956/K00957 14 adenylyl sulfate kinase cysC 2.7.1.25 K00860 9 phosphoadenosine phosposulfate reductase cysH 1.8.4.8 K00390 8 assimilatory sulfite reductase cysJI 1.8.1.2 K00380/K00381 6 nitrilase n/a 3.5.5.1 n/a 1 K01427,K01428,K01429,K urease n/a 3.5.1.5 01430,K14048 0 urea carboxylase n/a 6.3.4.6 K01941 19 allophanate hydrolase n/a 3.5.1.54 K01457 5 peroxidase many 1.11.1.- many 3 type VI secretion system VgrG n/a K11891 9 Hcp n/a K11892 2 Lip/vasD n/a K11903 3 IcmF/impL/vasKn/a K11904 1 DotU/impK/ompA/vasFn/a K11906 2 ClpV/vasG n/a K11907 2 PpkA n/a K11912 0 Fha1 n/a K11913 0 PppA/stp1 n/a K11915 0 amidase n/a 3.5.1.4 K01426 1 protease many 3.4.21.- many many phospholipase many 3.1.1.-,3.1.4.- many many

203

Appendix B. Supporting Information for Chapter 3

Supplemental Methods

Produced fluid metabolite analysis: Ethylene glycol, propylene glycol, acetaldehyde, propionaldehyde, acetate, propionate, ethanol, n-propanol, and acetone were measured in the Utica-Point Pleasant natural-gas well produced fluids by LC-MS according to a previously established method (Xu et al., 2019). Additional confirmation of surfactant metabolites (with the exception of propylene glycol, acetaldehyde, and propionaldehyde) was performed using 1H NMR.

Carbon analysis Non-purgeable organic carbon (NPOC) was measured using a TOC/TN analyzer equipped with autosampler (TOC-V CSN/TNM-1/ASI-V, Shimadzu, Kyoto, Japan) in borosilicate vials baked at 450 °C for 2 hours prior to use.

Ion and elemental analysis Br and Cl were analyzed using a Dionex ICS-2100 ion chromatograph (Thermo-Scientific Waltham, MA) after dilution (100-1000) in DI water. Mg, Ca, Sr were measured using a Perkin-Elmer Optima 4300 DV inductively coupled plasma (ICP) optical emission spectrometer (Waltham, MA), after dilution (100-1000) in 2% nitric acid (trace metal grade). Li was measured on a ThermoFinnigan Element 2 ICP Sector Field Mass Spectrometer. Samples were diluted in trace metal grade nitric acid (2%) spiked with Indium (10 ppb). pH and conductivity were measured using a Hach DR-900 Colorimeter.

NMR Analysis Utica-Point Pleasant produced fluid samples were sent to the Pacific Northwest National Laboratory for metabolite analysis by NMR as previously reported (Borton et al., 2018).

204

Samples were diluted by 10% (vol/vol) with 5 mM 2,2-dimethyl-2-silapentane-5- sulfonate-d6 as an internal standard. Spectra were obtained with a Varian Direct Drive 600- MHz NMR spectrometer equipped with a 5-mm triple resonance salt-tolerant cold probe. Chenomx NMR Suite 8.3 was used to process and analyze all 1D 1H NMR spectra with quantification based on spectral intensities relative to the internal standard. Metabolites were determined by matching the chemical shift, J-coupling, and intensity information of experimental NMR signals against the NMR signals of standard metabolites in the Chenomx library. The 1D 1H spectra were collected following standard Chenomx data collection guidelines (Weljie et al., 2006), using a 1D NOESY pre-saturation (TNNOESY) experiment with 65,536 complex points and at least 512 scans at 298 K. Further, 2D spectra (including 1H–13C heteronuclear single-quantum correlation spectroscopy, 1H-1H total correlation spectroscopy) were acquired on most of the fluid samples, aiding in the 1D 1H assignments of ethylene glycol, acetate, and ethanol.

Mixing model construction The major anion chemistry of flowback and produced fluid samples from this Utica natural- gas well is similar to other Appalachian Basin formational brines with a distinctively high Br/Cl molar ratio (>5x10-3), indicative of ~25 fold evaporated Silurian-age seawater (Warner et al., 2013). Further mineral weathering from the use of chemical additives during the natural-gas well completion process is evidenced in the enrichment of calcium (Ca/Cl >0.13) and depletion of Mg (Mg/Cl >0.03) above evaporated seawater concentrations, consistent with produced waters from other unconventional low-permeability formations in the Appalachian Basin (Warner et al. 2014). Mg, Ca, Sr, and Br were strongly related to chloride in formation waters (Spearman’s >0.94), suggesting a conservative mixing model

(% formation(t)=1-(C(t)-C(input))/(C(input)-C(day 204)) was appropriate to discern dilution relative to biotransformation effects in our system. This analysis assumed lake/tank input water and day 204 formation water as two end members in our conservative mixing model.

Phylogenetic tree construction The pduC gene from Acetobacterium woodii DSM 1030 was compared against the Utica metagenomes available in IMG/MER using IMG’s built in blastp program with an e-score

205

cutoff of 1e-5 (Chen et al., 2017; Huntemann et al., 2016). These genes were then compared against the non-redundant protein database (Altschul et al., 1990) and the top hit was recorded as the gene name as follows: MetagenomeDay_TaxonomyHit_BitScore_PercentIdentity. These genes, along with the pduC genes from Acetobacterium woodii DSM 1030 and 8 Halanaerobium congolense isolates, were aligned in Geneious v. 8.1.9 (http://www.geneious.com) using MUSCLE with max. 10 iterations. Then, a RAxML tree was constructed with the following parameters: GAMMA GTR, rapid bootstrapping and search for best ML tree, and 100 bootstrap replicates, and visualized in iTOL (Letunic and Bork, 2016).

Microbial isolation and culturing Halanaerobium congolense WG10 (H. congolense WG10) was isolated from produced fluid from the Utica-Point Pleasant natural-gas well at 140 days after fracturing occurred using yeast extract-peptone-dextrose streaked plates (YPD, ATCC medium 1245). Individual colonies were selected and transferred to liquid YPD medium and were examined via microscopy and 16S rRNA gene sequencing to confirm isolation. Growth experiments were performed in sterile borosilicate glass tubes suitable for anaerobic culturing (18x150 mm, Bellco 223 Glass, Vineland, NJ) with butyl rubber stoppers and metal clamps, using salt water dissolved media (SWDM) containing the following: 1 g L-1 -1 -1 -1 -1 - NH4Cl, 10 g L MgCl2•6H2O, 0.1 g L CaCl2•2H2O, 1 g L KCl, 100 g L NaCl, 0.5 g L 1 cysteine, 1% DL mineral solution and 1% DL vitamin solution, amended after autoclaving with 10 mM D-Glucose, 0.2% NaHCO3 solution, and 0.003% phosphate solution (KH2PO4 and K2HPO4).

Acetaldehyde and alcohol analysis Samples from cultures were analyzed using a TRACE 1300 Gas Chromatograph (GC) with flame ionization detection (FID) (Thermo-Fisher Scientific, Waltham, MA) equipped with a Zebron ZB-WAXplus GC Capillary Column (30 m x 0.25 mm x 0.25 µm) according to a previously established method (Luek and Hanson). For limits of detection and quantification, see Supplemental Table 3.

206

ATP analysis ATP concentrations were determined by extracting 100 µL of sample and aliquoting into a well of a white 96 well plate containing 100 µL of BacTiterGlo Microbial Cell Viability Assay Reagent. All instructions for the reagent preparation were followed based on the manufacturer’s manual (Promega, Madison, WI). A calibration curve containing ATP from 10-7 to 10-3 µM was generated and used to quantitate ATP in samples. All luminescence measurements were performed using a BioTek Synergy HTX Multi-Mode Reader (BioTek, Winooski, VT) within 1 hour of sampling and blanks were generated using autoclaved media.

Initial experiment testing degradation of PEG and EG Prior to the main growth experiment, a preliminary experiment was performed to assess the ability of H. congolense WG10 to grow on polyethylene glycol and ethylene glycol using the same experimental setup and SWDM recipe except with 20 mM D-Glucose, and amendments of PEG-400 (polyethylene glycol with an average molecular weight of 400 g/mol, 2mM based on avg. MW, Alfa Aesar, Haverhill, MA) and ethylene glycol (2 mM, Fisher Scientific, Waltham, MA). Optical density readings were performed as reported above. Samples were extracted in singlet at t=0, t=56, and t=120 for GC and LC-MS analysis of EG and PEG, respectively.

PEGs analysis in initial laboratory cultures: Polyethylene glycols were analyzed using an LC-ToF-MS at Colorado State University, equipped with an Agilent 1100 series LC, reverse phase C8 column (150 mm × 4.6 mm, with 3.5 μm particle size; Zorbax Eclipse

XDB-C8), coupled to an Agilent G3250AA MSD TOF system equipped with electrospray ionization. Separation conditions: injection volume was 5 μL, and the flow rate was 0.8 mL/min. Mobile phase composition was a constant 30% 0.1 % formic acid (A) and 70% acetonitrile (B) for the first 5 min, followed by a linear gradient to 90% B for 5 - 15 min, then B held at 90% for 15 - 20 min. The MSD TOF was operated in positive ion mode with: voltage 4,000 V; nebulizer pressure 45 psig; drying gas (N2) flow 10 L/min; gas temperature 325 °C; fragmentor voltage 190 V; skimmer voltage 45 V; octopole RF 300 V. Accurate mass measurements were obtained using an automated calibrant delivery

207

system via low flow of a calibrating solution (calibrant solution A, Agilent Technologies, Inc.), containing the internal reference masses (purine at m/z 121.0509 and HP-921 at m/z 922.0098). Accurate mass spectra were recorded across the range 100 - 3,000 m/z.

Ethylene glycol analysis in initial laboratory cultures: Gas chromatography (GC) coupled to flame ionization detection (FID) was used to analyze ethylene glycol in the initial experiment as mentioned above at the OSU Noble Gas Laboratory. A Thermo-Fisher Scientific Trace-Ultra 1300 gas chromatograph (Thermo Fisher Scientific, Waltham, MA) was used with a Restek liner 5mm Straight w/ Wool and Zebron ZB-WAXplus PEG GC Capillary Column (30 m x 0.25 mm x 0.25 µm). Ethylene glycol was analyzed using the following GC parameters: Oven profile 80 °C to 200 C at 8 °C/min, 1.2 mL/min carrier gas (helium), 1 µL splitless injection with 140 °C injector, FID at 250 °C. Washes of 50:50 methanol:DI water were utilized between each injection.

Supplemental Text

Initial biotic and killed losses In the culture samples, the same concentrations of the Revert Flow additive were used in biotic, abiotic, and killed controls, yet an immediate decrease in AEO response in biotic and killed controls (likely due to initial sorption to biomass) was observed relative to the abiotic controls. Thus, to fully capture losses, the AEO data presented in Figure 3 were generated using the C0 of the abiotic controls for each ethoxylate/propoxylate species (Figure 3, Supplemental Table 6).

Other biodegradation pathway mechanisms

In the laboratory culture samples, we detected C10 AEOs as well as C8 AEOs, but the C10 AEOs were often below detection limits and were thus left out of our analysis. We searched accurate mass spectra for AEOs and PPGs with carboxylate and aldehyde end groups which are formed during aerobic biodegradation (Luek and Hanson; Rogers et al., 2018) and were unable to find any at any time point, making it unlikely that this biodegradation mechanism is utilized by Halanaerobium.

208

Supplemental Figures

Figure15B1. Chromatogram of PEGs (0, EOx) and alkyl polyethoxylates (AEOs) (Cx,EOy) in Utica-Point Pleasant natural-gas well produced fluid at 86 days after hydraulic fracturing. n6,X refers to linear C6 AEOs which are distinguished from branched C6 AEOs

209

Figure16B2. A) Chromatogram of detected PPGs and AEOs adducts in culture samples at t=0. B) Chromatogram zoom-in of C8 AEO ethoxymers.

210

Figure17B3. Maximum-likelihood phylogenetic tree of pduC genes from Utica-Point Pleasant natural-gas well metagenomes and isolates. Bootstrap values (0-100%) are denoted by blue circles.

211

Figure B3 continued.

Continued

212

Figure B3 continued.

213

Figure18B4. Metabolite trends in produced fluids.

Left y-axes and blue colored lines and markers correspond to C2 species (e.g. EG, acetaldehyde, acetate, ethanol), right y-axes and red colored lines and markers correspond to C3 species (e.g. PG, propionaldehyde, propionate, n-propanol) with the exception of 1 acetone in black in D), which is on the right C3 y-axis scale. Metabolites verified by H NMR are denoted by markers in the boxes below each plot corresponding to metabolite type and time point.

214

Figure19B5. Two end member conservative mixing models for strontium, chloride, and lithium-chloride ratios over 120 days assuming input fluid and day 204 as end members.

Figure20B6. Growth of H. congolense WG10 on glucose alone (left) and glucose + revert flow (right). Optical densities (left y-axis) for biotic samples (in triplicate), killed and abiotic controls (duplicate) are reported alongside ATP measurements (right y-axis) in biotic samples (triplicate) at various points during the growth cycle.

215

Figure21B7. Growth curves and EG and PEG trends during H. congolense WG10 growth A) Growth curve from preliminary experiments culturing H. congolense WG10 with 20 mM glucose and either ethylene glycol (EG) or polyethylene glycol (PEG-400). B) Trends in ethylene glycol (EG) and polyethylene glycol (PEG) during the growth phase of H. congolense WG10.

216

Figure22B8. Average z-score in proteins associated with glycol-metabolism surfactant treated cultures (Revert Flow, RF) and glucose controls (GC). Gray boxes indicate an absence of the given protein in a treatment. Black outlines denote statistical significance (p<0.1). Pathways are color coded as follows: methylglyoxal bypass (orange), diol dehydratase (blue), cobalamin cofactor synthesis (red), glycolysis (green). Numbers correspond to protein numbers in Figure 3.4; stars denote the protein used of duplicate copies in Figure 3.4.

217

Supplemental Tables

Table6B1. Frequency of use as disclosed in the FracFocus database, adapted from Rogers et al. 2015*Frequency is dependent on structure.

Frequency of Chemical Use Ethylene glycol 7.1 Propylene glycol 49.7 Polyethylene glycol 14.4 (PEGs) Polypropylene glycol 0.31 (PPGs) Alkyl polyethoxylate *0.002-19.0% (AEOs)

Table7B2. Listed constituents of industrial Revert Flow used for culturing experiments

Revert Flow Components Name CAS # % DB-964 (PEG-PPG co-block polymer) 9003-11-6 10-30 Alcohol C6-C12, ethoxylated 68439-45-2 10-30 Isopropyl alcohol 67-63-0 7-13 Orange terpenes 8028-48-6 5-10

Table8B3. Detection limits, linear range, and R2 values for GC and IC organic acid data. (*LOD/LOQ is given as the lowest calibration curve point)

Species Method Linear Range R2 LOD/LOQ (µM) (n=5) (µM) Acetaldehyde GC-FID 31.25-2500 0.9973 17.8/54.0 Ethanol/Propanol GC-FID 31.25-2500 0.9993 31.3* Acetate IC 92.1-1476 0.9926 92.1* Propionate IC 114-914 0.9965 114*

218

Table9B4. Surfactant data from the Utica-Point Pleasant natural-gas well produced fluids including retention times, rate constants, and LOD/LOQs

Pseudo- Retention Molecular C/C0 C/C0 First Order Rate Fit Species time Day 86 Day 89 Day 94 Day 105 Day 126 Day 154 Day 175 Day 204 Weight LOD LOQ Rate (R2) (min.) Constant PEGEO3 150.0894 5.5 0.019366 0.043314 152000 101000 146000 136000 126000 71400 62000 50400 0.0089 0.8505 PEGEO4 194.1156 5.5 0.003487 0.008842 31200000 32900000 31000000 27700000 31800000 23400000 17000000 19700000 0.005 0.7961 PEGEO5 238.1419 6.2 0.007917 0.016721 35800000 37500000 33700000 36200000 36400000 29200000 28300000 27400000 0.0026 0.8261 PEGEO6 282.1682 7 0.018954 0.039791 27400000 30800000 30100000 29600000 28300000 24300000 22900000 20900000 0.003 0.8859 PEGEO7 326.1944 7.8 0.024924 0.044925 15900000 17600000 15400000 16300000 16900000 12000000 12500000 10200000 0.004 0.8277 PEGEO8 370.2205 8.7 0.049797 0.076914 7090000 6740000 6950000 6050000 6660000 5100000 4520000 3770000 0.0051 0.9305 PEGEO9 414.2467 9.5 0.116608 0.196855 3580000 3600000 3270000 3060000 2690000 2240000 1960000 1660000 0.0065 0.9948 PEGEO10 458.273 10.2 0.307917 0.453858 1080000 1010000 1170000 884000 636000 635000 639000 344000 0.0085 0.8793 PEGEO11 502.2992 10.9 0.555347 0.733412 469000 608000 422000 416000 293000 329000 233000 121000 0.0106 0.8754

222 C4EO2 162.1258 10.1 0.01775 0.035206 1230000 1290000 1540000 1860000 1920000 1450000 1330000 1580000 -0.0004 0.0115

C4EO3 206.152 11.5 0.008341 0.022109 4560000 5160000 5880000 6320000 6350000 5170000 5060000 5910000 -0.0003 0.0162

C4EO4 250.1783 12.8 0.00792 0.021418 3590000 4110000 4430000 4840000 5450000 4280000 3660000 3920000 0.0007 0.0455 C4EO5 294.2045 13.8 0.003348 0.008575 1270000 1660000 1690000 1880000 1920000 1510000 1310000 1170000 0.023 0.3047 C4EO6 338.2307 14.7 0.001676 0.002461 260000 255000 225000 245000 253000 184000 164000 118000 0.006 0.8781 nC6EO3 234.1834 19.1 0.004538 0.010243 218000 278000 196000 246000 227000 147000 165000 273000 0.011 0.0436 nC6EO4 278.2096 19.9 0.002203 0.004923 522000 635000 431000 431000 428000 223000 221000 341000 0.0065 0.5735 nC6EO5 322.2358 20.5 0.002455 0.004729 300000 383000 281000 251000 253000 67300 92800 95800 0.0131 0.7694 C6EO2 190.157 15.5 0.001893 0.003341 365000 255000 150000 98000 50100 26900 17300 6830 0.0307 0.9662 C6EO3 234.1833 16.6 0.000306 0.00053 2710000 4560000 3610000 2090000 1090000 545000 364000 173000 0.0263 0.9713 C6EO4 278.2096 17.4 0.000202 0.000404 5040000 11700000 7420000 4530000 2420000 1160000 793000 360000 0.0264 0.9496 C6EO5 322.2357 18.2 0.000118 0.000273 11700000 9230000 5780000 3320000 1820000 659000 474000 194000 0.0333 0.976 C6EO6 366.262 18.8 0.00046 0.001189 4370000 4230000 2440000 1230000 647000 131000 129000 44000 0.0391 0.9672 C6EO7 410.2882 19.4 0.001812 0.004982 1570000 1550000 867000 395000 225000 6430 17100 6460 0.0508 0.8945 C6EO8 454.3144 19.9 0.002206 0.005619 515000 463000 235000 55800 50100 359 1590 481 0.0641 0.8688 C6EO9 498.3406 20.3 0.003618 0.006424 90100 74900 31300 2280 8850 278 296 279 0.0511 0.8074 C8EO3 262.2147 23 0.000325 0.000548 1160000 1000000 524000 217000 131000 11200 16900 6360 0.0453 0.9296 C8EO4 306.2408 23.5 0.000558 0.001105 1390000 1220000 533000 146000 130000 914 4440 2210 0.0606 0.835 C8EO5 350.2669 23.9 0.000878 0.001516 558000 543000 190000 9410 50700 332 531 419 0.0657 0.8162 C8EO6 394.2932 24.3 0.002039 0.00337 135000 114000 20600 802 11600 188 199 219 0.0542 0.7107 C8EO7 438.3194 24.6 0.014076 0.024748 20100 18900 1410 238 1590 191 195 204 0.0339 0.5626

219

Table10B5. PPG and AEO species detected in culture samples with R2 values, LOD/LOQs (n=16), and intraday variability

Approx. % retention C/C0 C/C0 Species Observed m/z Major adduct R2 intraday time LOD LOQ deviation (min) PPG4 EIC 273.1678±0.01 1 [M+Na]+ 0.9992 0.0262 0.0785 PPG5 EIC 331.2097±0.01 1.1 [M+Na]+ 0.9977 0.0081 0.0227 PPG6 EIC 389.2515±0.01 1.3 [M+Na]+ 0.9992 0.0035 0.0086 PPG7 EIC 447.2934±0.01 1.6 [M+Na]+ 0.9995 0.0017 0.0041 PPG8 EIC 505.3353±0.01 2.1 [M+Na]+ 0.9721 0.0003 0.0007 PPG9 EIC 558.4212±0.01 2.8 [M+NH4]+ 0.9999 0.0007 0.0017 PPG10 EIC 616.4630±0.01 3.7 [M+NH4]+ 0.9999 0.0002 0.0005 C8EO2 EIC 236.2226±0.01 4.6 [M+NH4]+ 0.9911 0.0429 0.1251 C8EO3 EIC 280.2487±0.01 4.7 [M+NH4]+ 0.9953 0.0029 0.0077 C8EO4 EIC 324.2750±0.01 4.7 [M+NH4]+ 0.9996 0.0015 0.0042 C8EO5 EIC 368.3012±0.01 4.8 [M+NH4]+ 0.9986 0.0014 0.0037 C8EO6 EIC 412.3269±0.01 4.8 [M+NH4]+ 0.9944 0.0010 0.0027 C8EO7 EIC 456.3531±0.01 4.8 [M+NH4]+ 0.9902 0.0008 0.0021 C8EO8 EIC 500.3793±0.01 4.9 [M+NH4]+ 0.9923 0.0018 0.0044 C8EO9 EIC 544.4055±0.01 4.9 [M+NH4]+ 0.9651 0.0002 0.0006 C8EO10 EIC 588.4317±0.01 4.9 [M+NH4]+ 0.998 0.0003 0.0007 C8EO11 EIC 632.4580±0.01 4.9 [M+NH4]+ 0.9998 0.0005 0.0016 C8EO12 EIC 676.4842±0.01 4.9 [M+NH4]+ 0.9992 0.0002 0.0005 PPG 0.9963 0.0058 0.0167 7.81% sum C AEO 8 0.9932 0.0049 0.0139 11.79% sum

Table11B6. Relative standard deviation (n=3) for the sum of PPGs and AEOs at C/C0=0.1 to C/C0=1

Species C/C0=0.10 C/C0=0.25 C/C0=0.50 C/C0=1.00 PPG 9.01% 10.00% 10.06% 7.81% C8AEO 12.35% 9.73% 12.51% 11.79%

220

Table12B7. Reported Cfinal/C0 values from Figure 3

Species Biotic Abiotic Killed PPG4 0.92 1.07 1.04 PPG5 0.88 1.14 0.98 PPG6 0.74 1.10 0.83 PPG7 0.62 1.12 0.81 PPG8 0.57 1.13 0.75 PPG9 0.53 1.19 0.66 PPG10 0.57 1.14 0.69 C8EO2 0.31 0.50 0.26 C8EO3 0.12 0.57 0.22 C8EO4 0.13 0.58 0.23 C8EO5 0.17 0.61 0.28 C8EO6 0.20 0.64 0.31 C8EO7 0.23 0.66 0.34 C8EO8 0.27 0.68 0.38 C8EO9 0.29 0.65 0.39 C8EO10 0.31 0.63 0.39 C8EO11 0.32 0.62 0.40 C8EO12 0.32 0.59 0.36

Table13B8. BLASTp results from Halanaerobium congolense WG10 to metagenomes

Days after Assembled fracturing pduC pduD pduE pduP FeADH Genome Size 86 1 0 0 0 1 19799876 87 2 0 0 0 0 16549434 88 1 1 1 1 1 22175348 92 1 0 0 1 1 12577384 94 6 2 2 2 4 39979368 96 4 2 2 2 3 33789609 98 5 1 1 3 3 37366011 112 2 2 2 3 2 20821072 126 5 2 4 4 4 13409114 140 5 1 4 3 3 15421198 154 5 1 4 3 3 15488382 175 5 2 3 3 3 12211726 204 0 0 0 0 0 15192315

221

Table14B9. Relevant geochemical measurements from the Utica-Point Pleasant well fluid through time. NPOC (Non-purgeable organic carbon). Gray cells indicate samples where data was not measured.

Days after Concentration (mg/L) Hydraulic Conductivity Sr Cl Mg Ca Br Li NPOC pH Fracturing (mS/cm) Lake 1.84 8.21 111 231 0.04 Tank 1.83 8.54 113 235 0.04 Chem 1.57 116 235 0.04 86 993 66400 1330 9670 608 167.7 5.34 110.2 87 1030 67800 1270 9410 622 32.7 161.3 5.56 101.3 88 1030 64200 1260 8620 595 28.7 143.7 5.73 102.4 89 1020 64400 1250 8700 598 26.9 142.2 5.71 104.1 91 1070 64000 1270 8710 618 25.4 144.2 5.71 99.2 92 1260 63300 1400 9480 613 25.8 137.2 5.69 101.7 94 1320 65500 1400 9520 649 34.9 129.1 5.71 106.5 96 1330 65600 1400 9500 660 26.4 127.5 5.69 104.8 98 1360 67500 1450 9830 673 25.7 127.6 5.68 108.3 105 1570 71600 1530 10600 751 27.6 112.9 5.65 117.3 112 1790 77300 1650 11700 835 29.1 130.4 5.73 122.5 126 1870 81300 1720 12200 888 32.5 100.0 5.93 130.4 140 2170 92500 1960 14200 1033 36.6 97.84 5.89 146.4 154 2310 95200 1990 14500 1068 82.29 6.63 167.9 175 2320 93500 1960 14500 1063 36.4 6.44 177.9 204 2770 104000 2360 16900 1278 41.8 73.10 6.22 189.4

222

Table15B10. Presence (red) or absence (gray) of key genes in all Halanaerobium isolates from the Utica-Point Pleasant natural-gas well

Genome Day Fe Accession Genome Name pduC pduD pduE pduP Isolated ADH Number 88 Ga0073306 H. congolense UTICA-W4B2 89 Ga0073307 H. congolense UTICA-W4B3 140 Ga0073276 H. congolense WG1 140 Ga0073277 H. congolense WG2 140 Ga0073280 H. congolense WG5 140 Ga0073281 H. congolense WG6 140 Ga0073282 H. congolense WG7 140 Ga0073283 H. congolense WG8 140 Ga0073284 H. congolense WG9 140 Ga0073285 H. congolense WG10

Captions for Additional Supplemental Files

File B1 - Proteomic results for Halanaerobium congolense WG10 comparing surfactant amended cells and glucose controls.

223

Supplemental References

Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990). Basic local alignment search tool. J. Mol. Biol. 215, 403–410. doi:10.1016/S0022- 2836(05)80360-2.

Huntemann, M., Ivanova, N. N., Mavromatis, K., Tripp, H. J., Paez-Espino, D., Palaniappan, K., et al. (2015). The standard operating procedure of the DOE-JGI Microbial Genome Annotation Pipeline (MGAP v.4). Standards in Genomic Sciences 10, 86. doi:10.1186/s40793-015-0077-y.

Letunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for the display and annotation of phylogenetic and other trees. Nucleic Acids Research 44, W242–W245. doi:10.1093/nar/gkw290.

224

Weljie, A. M., Newton, J., Mercier, P., Carlson, E., and Slupsky, C. M. (2006). Targeted profiling: quantitative analysis of 1H NMR metabolomics data. Anal. Chem. 78, 4430–4442. doi:10.1021/ac060209g.

Xu, C., Couvillion, S. P., Sontag, R. L., Isern, N. G., Maezato, Y., Lindemann, S. R., et al. (2019). MetFish: a metabolomics platform for studying microbial communities in chemically extreme environments. bioRxiv, 518647. doi:10.1101/518647.

225

Appendix C. Supporting File for Chapter 3

Please see file available with this document on OhioLink.

226

Appendix D. Supporting Information for Chapter 4

Supplemental Methods

Geochemical analyses

Bromide and chloride were analyzed using a Dionex ICS-2100 ion chromatograph (Thermo-

Scientific, Waltham, MA). Non-purgeable organic carbon (NPOC) was measured using a

TOC/TN analyzer equipped with autosampler (TOC-V CSN/TNM-1/ASI-V, Shimadzu, Kyoto,

Japan) in borosilicate glass vials baked at 450 °C for 2 hours prior to use. Samples for all geochemical measurements were diluted in MilliQ water as needed.

Cell counts

Microbial cell counts were performed using an epifluorescent stain, filtration, and microscopy as described previously(Evans et al., 2018). Briefly, 2X SYBR Gold (Life Technologies, Carlsbad,

CA, United States) in TE buffer was used to stain microbial cells, with sample (1-15 mL) filtration onto 0.2 µm black PTCE filters (Sterlitech, Kent, WA, United States), followed by addition of SlowFade reagent (Life Technologies, Carlsbad, CA, United States). Dilutions were performed per sample such that there were 10-100 cells per counting field. Filters were viewed by a Labomed Lx500 epifluorescent microscope under a 40X air objective with 480 nm excitation. For cell counts, 20 random fields were selected per filter. Cell counts for Marcellus-1 were quantified using a Guava EasyCyte flow cytometer (EMD Millipore) as previously reported(Daly et al., 2016).

227

Supplemental Text

Amino acid sequences for relevant genes >Daly_T328_ORIG_scaffold_660_6_BLAST:ferredoxin(db=KEGG_evalue=9.3e- 185_bit_score=651.7_identity=82.9) MMSDFLNPVTTQTWVNGRHQVRCVKVIQETWDVRTFCFMAEQPVLYFFKPGQFVTLE LEIDGEQVMRSYTISSSPSVPYSFSISVKRLPGGVVSNWLHDNLKAGDELAVHGPVGNFN IIDHAADKVLMLSGGVGITPLMSMTRWLFDTNASVDLAFVHSARSPKDIIFHRELAHIFS RIPEFKLHIVCERSDELCEAWAGFRGYLSQEMLELMAPDFMEREIFCCGPTPYMNAVKRI LRDNGFDMSRYHEESFGATPLDVQEDVLELAEQAEAQAEELDVADMYSVEFSATGKSV RVQPGETVHAAAAKLGLHIPKACGMGICGTCRVPLSSGQVEMDHNGGITDEDVAEGYI LSCCSKPTGNVVVDF*

>S-1-Day17_scaffold_409_6_RBH:FAD/NAD(P)-binding_oxidoreductase(db=KEGG) MTTDETMRPLLVTHKEFIATDTVLFELKSPDSLPLARFTAGAHIAVQVPNGAMRHYSLC SNPDEVDHYQLAVKREADGRGGSQSLVDEVNAGDTLMAGTPSNLFGLSDKAKSFILVA GGIGITPMRAMIHSLQAEGLRSFKLYYLTRTPESAAFLNELRSPELAGSVVVHHTHGDPT KTFDLWSVFEKPQAGVHVYCCGPKRLMDEVKDMTGHWPSSAVHFESFGADTKPHADD KPFEVELAKAGKTLLVPANRSILDALRDHGIRVPSSCESGTCGSCKVRLLKGEADHRDL ALLPEEQEDHIMVCVSRAKSDTLVLDL*

>S-1- Day17_scaffold_461_11_RBH:hypothetical_protein_n=1_Tax=Caldimonas_manganoxidan s_RepID=UPI0003709B6D(db=UNIREF) MSTAPASLRVKVARKAVEAEGICSFELVSADGRPLPAFSAGSHIDVHLPDGLVRQYSLC NDAAETHRYLIGVLKDPATRGGSRTMHEQVAEGDELTISPPRNHFALAHDAKRHLLLA GGIGVTPILCMAERLAATGGVFHMHYCTRSRARTAFASRIEASAFAAQVQFHHDDGPPE QQLQLDAVLAGADDHTHLYVCGPKGFMDAVLGAARARGWPESRLHYEFFSAEPVSTE GDRAFEVVLASSGRVVPVPADQSVVKALAACGVDVMVSCEQGVCGTCLTRVLEGEID HRDVYLTPEERAAGDQFTPCCSRAKSARLVLDL*

228

>S-1- Day17_scaffold_5110_2_RBH:ferredoxin;_K03863_vanillate_monooxygenase_[EC:1.14.13. 82](db=KEGG) MSTLNLRVQAIRRQAEGIHAFELVRPDGGDLPEVQAGAHVDVHLPGGLVRSYSLAGDP AERCRWTLGVLREKAGRGGSAALHDKVKVSDVLAVGEPRNAFALVPGARHSVLLGGG IGITPLKAMAHTLRDRGESFELHYCARTPGHAAFLAELQALVPAGRLHLHFDQGDPARG LDMAGLLRVPADGTHLYYCGPAGFMQACADASAHWPAGTVHCEHFKPPANTASDLPA GGFEVRLQRQGITVPVAPSQSIVQALEAAGQAVPTSCLSGLCGACKVGYLEGEVEHNDY ILSDDEKARCLTLCVSRARRSRA

>S-1- Day17_scaffold_516_6_RBH:ferredoxin;_K03863_vanillate_monooxygenase_[EC:1.14.13.8 2](db=KEGG) MSSAQTLQVRVARKATEALDIVTLELVATDGSALPAFGAGAHIDVQLPGGITRQYSLCN DPKETHRYLIGVLRDPASRGGSLAVHDRVKEGDVLQISTPKNHFPLAHDAKKSLLLGGG IGITPILCMAERLANTGAAFEMHYATRSAERTAFRERIAKSSFADQVAFHFDDGAAEQKL DLARLLVKPEAGTHLYVCGPKGFMDAVLNTARAQGWPEDQLHYEFFGATVEKSDSDA SFEVKLASSGRIVMVPKDKTVTQALAEAGVEIMMSCEQGVCGTCLTRVLEGVPDHKDS YLTPEEQAANDQFLPCCSRAKTPQLVLDL*

>S-1-Day17_scaffold_94_10_BLAST:oxidoreductase(db=KEGG_evalue=6.2e- 68_bit_score=263.5_identity=48.1) MSADIEVTVAGLRAVARDVLAVELRHKSGQPLPGAQAGAHIDLALGNGLVRQYSLVNA MGQPGMDRYVVAVGWDANSRGGSVWIHEKLKVGQSLRVSAPRNLFEMVSEHRRVLL LAGGIGVTPIYAMAQSCAQKGMPFELWASARSAPRLAYLDELKALAADRLHLHFDDEQ GGPMNLVERLQTQRWDAVYACGPAPMLDALTAATAHWQPGSVRMERFKGVEAPAGE RQPFELLLQRSGLHTTVNAHESVLEAMERLGVDYPWSCREGICGTCEAPVLEGEVKHLD YVLSPAEHAEQRRMMVCVSRCGSGRLVLDI*

>S-1-Day17_scaffold_124_5_BLAST:vanillate_O- demethylase_oxidoreductase(db=KEGG_evalue=8.5e-97_bit_score=359.4_identity=56.3)

229

MSSHLLQALVFSIRYEARDIVSVELRPATTDVLFPSVEPGAHIDLHLTTGLVRSYSLTNPG ESDRYVIAVLKDRASRGGSRYVHENLRVGQIIAISAPRNHFYLHEEARLSVFLAGGIGITP LVAMLKHLSALGRQAHLIYCARSRQDAAFVAEIKGLVANSKKNCLTSHFHFDDEQGAP PDITRLLRSFPPDAHFYSCGPGRMLEAYEKACAQLGYKNVHMERFNAMEVQEGQAIAP EGYEVELRRSNKTIHVPPGTTLLDALLTAGCEVESSCREGLCGSCETRVLSGEVEHRDSIL TKEERAANKSMMICVSSCRSGTLVLDA*

>S-1-Day17_scaffold_409_1_RBH:ferredoxin(db=KEGG) MVTKTLQALIFQMRHEAPGIVGVELRPVPPAAAFPAVEAGAHIDLHLGNGLVRSYSLVN PGETHRYAVAVLNDRHSRGGSRFVHEQLRVGQTIAIGAPRNHFRLDEAAPRSVLLAGGI GITPVYAMLRRLAALGREAQLVYCARSRSEAAFLAEIEALVETHDSRLSLRCHFDNEQG GPPDLERLLADHPPGTHFYCCGPGPMLDAYERACERLGQQNVHLERFAATPTTAPAIPA SGYTVELRRSGRTVQVPPGVTLLDALIEAGLSPDHSCREGVCGACETKVISGDIDHRDQL LSKQERAANKSMMICVSSCRSGCLVLDA*

>S-1-Day17_scaffold_611_5_RBH:ferredoxin(db=KEGG) MNESSFHVVIDAVRDVASGIREFTFRRADGQPFPLYSAGSHVVVSVPAGERPRRNPYSLL GDPNERMSWRIAVRRQEPSRGGSAWLHEQARVGDRLDITAPMNLFPLISTARSHLLVAG GIGVTPILSQARELARRGADFEVHYAWRSPAHAAYVAELEALAPGKVHHYDESQGVRI DFKTLFAGRRLGTHFYICGPTPMVSAAMEGGAAMGWPATNLHSERFASAEPGEPFDVL LARSGQRITVPADLSLLEALEQCGAPVNALCRGGACGQCETPVLSAEGELLHHDVYLSD ADKQAGKLVMPCVSRFKGACLTLDL*

>S-1-Day17_scaffold_91_8_BLAST:oxidoreductase_fad/NAD(P)-binding_domain- containing_protein(db=KEGG_evalue=6e-106_bit_score=389.8_identity=58.8) MAPIYLPAMSLPLLQARVFNLRYEARDIISVELRPASPEVDFPFTEAGSHIDLHLGNGLIR SYSLTNPGESQRYVVAVLKDRKSRGGSLYVHEQLRVGQVISISAPRNNFRLDEDACQNV LLAGGIGITPLYAMLKRLSALGRRTHLIYCARSRSDAAFVEDIQALVSDSCDGALSSNFH FDEERGVAPDLVGLLGGFSPETHFYCCGPGPMLEAYESACEKLGYTNVHVERFAAKPA LTNQVADPVGYTVELKKSGKTVQVPPGTRLLDALLSAGCKVEFSCREGVCGACETRVIS GEVEHRDSILTQKERAANKSMMICVSSCKSGTLVLDA*

230

>S-1- Day1_scaffold_4013_2_RBH:putative_oxygenase_electron_transfer_component(db=KEG G) MSAMIEVQVTDVRQLTPVVREFSFRAIDGSLPVFSCGSHVQVVLPIEGRTLRNAYSLLGD PRETGVYRIAVRLQEGSRGGSRYLHEQVRVGDRLQIGAPSNLFALHSQARHHILVAGGI GITPFMAYLAELEAQGASFELHYAYRSGLTDAYADELRERLGERFHGYDAAAGQRLSC ADLFAGKPLGSHLYVCGPQGLLDELKALAAAHGWSAGRLHWEAFAAPEPGLPFTVELA RSGQRLQVPGDHSLLEALEAAGVEVPNLCRGGVCGQCTTRYLSGEVEHRDHYLDEQQR GAALMPCVSRGGCSGTLLLDL*

>S-1-Day1_scaffold_1070_21_RBH:fold;_non-heme_chloroperoxidase(db=KEGG) MNPITRTLAMPLLAIALQPAFAAQAAQPAKPGSAAVVAHTASTITTADGVQLYYKDWG PKDGPVVTFSHGWPLSSDSWESQMLFLASEGYRVVAHDRRGHGRSSQPWEGNDMDHY ADDLAAVIDALDLQDVTLVGFSTGGGEVARYIGRHGTGRVKKAVLVSAVPPMMLRTE DNPDGLPQEVFDGIRKASLEDRAQLYMDLASGPFYGFNRPGAKVSQGLIDNWRAQGM QAGHKNTYDSIAAFSATDFREDLKKFDVPTLVIHGDDDQIVPLDISGRASAAQIKGAKLI VYPGAPHGLTDTHKARFNQDLLDFLRK*

>S-1-Day1_scaffold_17013_2_RBH:non- heme_chloroperoxidase;_K00433_chloride_peroxidase_[EC:1.11.1.10](db=KEGG) MNSITRTFAFSLLAIALQPAFAAEAAQPVKPGSSVVAAQTASTITTADGVQLYYKDWGP KDGPVVTFSHGWPLSSDSWESQMMFLASQGYRVVAHDRRGHGRSSQPWEGNDMDHY ADDLAAVIEALDLQDVTLVGFSTGGGEVARYIGRHGTGRVKKAVLVSAVPPMMLKTA DNPGGLPLEVFDGIRKASLEDRAQLYLDLASGPFYGFNRPGAKVSQGLIDNWRAQGMQ AGHKNTYDSIAAFSATDFREDLKKFDVPTLVIHGDDDQIVPLDSSGKASAALIEGAQLIV YPGAPHGLTDTHKERFNNDLLAFLKE*

>S-1-Day1_scaffold_6370_3_RBH:hypothetical_protein(db=KEGG) MNNTTRTFAASMLAITLQAGVTAHAAQPETPAQIAAAATLDGNYVVTADGVRLYYKD WGPKDGPVVTFSHGWPLSSDSWESQMIFLADQGYRVVAHDRRGHGRSSQPWDGNDM

231

DHYADDLAAVIEALDLDDITAVGFSTGGGEVARYIGRHGTDKVKKAALISAVPPLMLKT ADNPGGVPLEVFDGIRKASLENRSQLYLDIASGPFFGFNRPGAKVSQGLIDSWWAQGMQ AGHKNTYDSIAAFSATDFREDLKKFDVPTLVIHGDDDQIVPLDISGKASAAQIKDAQLIV YPGAPHGLTDTHKDRFNQDLLNFLKN*

>3H_02032016_scaffold_24865_0001_BLAST:reductive_dehalogenase(db=KEGG_evalue= 8.5e-57_bit_score=225.3_identity=67.8) VDASIVGMILSYYIRSLGYEARNHMDANYLLMPVLVAQDAGLGEIGRNTMLTTKEYGS RVRLGVVTTNLDLEVDEPISFGLEDFCKICKNCAHTCPSNSISHETEKETFGRLNWGIEQE SCYVKWRYYGTDCGLCINSCPFSQEME

>3H_02032016_scaffold_5978_0001_RBH:reductive_dehalogenase(db=KEGG) VDERDTMFARVNYKKGSKAYKDYYKRNPDKKSIDDSIRTRPNLCEEGTMTYNELNSPM ASCAFSFLEDIRHLCEGNVNPEKVNCNKKIVTKRIKGFAKQYGAKLVGITELKDYHFYT HRGRHEENYGEKVKLNHKYGIVFAVEMNKDMINRGPMLAEVVETSKCYVDTAIIGMM LSYYIRSLGYDARNHMDANYLLMPVLVARDAGLGEIGRNTMLTTKEYGSRIRLGVVTT NLELETDNPISFGLEDFCKICNNCGFTCPSQSISHESEKETNGRLNWSIEQETCYIKWRYY GTDCGICIASCPFSQEMESIRKVDTFKGKYKLIGRTVKEFREKYGKKPFVPGNPVWMR*

>3H_02032016_scaffold_7781_0001_BLAST:reductive_dehalogenase(db=KEGG_evalue=2 .4e-105_bit_score=387.9_identity=52.0) LKGWIFMKRIDERDTMFARMSYKKGTEEYEDYYKRNLDKKELDDELRSRPNIGEEGTM AYHPIHAPIANAGFEFLGDIKKYADGEPNAKKVEVEPEIITKKIKKIVKYFGADLVGVTK MREEYYYSHRGREPETYGKEVTDFHEYGIVFAVEMDRDMINRAPQAEEVIEVTKGYIKA AIIGMWLSYYIRGLGYEARNHMDGNYLVVAPLVAQEAGLGELGRNGILITKKYGQRVR LGVVTTNISLIPDEKNEFGIKEFCKVCGKCANTCPGKAIPKDDMEEIDGHTRWRIEQEKC YTMWRSLGTDCGICLSTCPFSQEVPYKLVDKMKGSKEIMNQILERYDKKHGIRPYIRKP LDLLK*

>3H_06082016_scaffold_7059_0001_RBH:reductive_dehalogenase(db=KEGG) KGWIFMKRIDERDTMFARMSYKKGTEEYEDYYKRNLDKKELDDELRSRPNIGEEGTMA

232

YHPIHAPIANAGFEFLGDIKKYADGEPNAKKVEVEPEIITKKIKKIVKYFGADLVGVTKM REEYYYSHRGREPETYGKEVTDFHEYGIVFAVEMDRDMINRAPQAEEVIEVTKGYIKAA IIGMWLSYYIRGLGYEARNHMDGNYLVVAPLVAQEAGLGELGRNGILITKKYGQRVRL GVVTTNISLIPDEKNEFGIKEFCKVCGKCANTCPGKAIPKDDMEEIDGHTRWRIEQEKCY TMWRSLGTDCGICLSTCPFSQEVPYKLVDKMKGSKEIMNQILERYDKKHGIRPYIRKPL DLLK*

>3H_06082016_scaffold_8758_0001_BLAST:reductive_dehalogenase(db=KEGG_evalue=3 .2e-84_bit_score=317.0_identity=62.8) HKYGIVFAVEMNKDMINRGPMLAEVVETSKCYVDTAIIGMMLSYYIRSLGYDARNHM DANYLLMPVLVARDAGLGEIGRNTMLTTKEYGSRIRLGVVTTNLELETDNPISFGLEDFC KICNNCGFTCPSQSISHESEKETNGRLNWSIEQETCYIKWRYYGTDCGICIASCPFSQEME SIRKVDTFKGKYKLIGRTVKEFREKYGKKPFVPGNPVWMR*

>3H_02172016_scaffold_53_0049_RBH:alpha/beta_hydrolase_fold_protein;_K00433_chlor ide_peroxidase_[EC:1.11.1.10](db=KEGG) MSMITMNDGTQIYYKDWGTGQPIVFSHGWPLNADSWESQMLFLASKGYRCIAHDRRG HGRSSQPWDGNEMDTYADDLSEIIEALDLKSIVLIGFSAGGGEVARYIGRHGTKRVAKA ALIAAVPPLMLKTDANPSGLPIEAFDEIRLGSIADRSQFYKDLASGPFFGANRTGSKVSQG MIDSFWLQGMQAGSKNTFDCIKAFSETDFTEDLKKFDVPTLIIHGDDDQIVPIGAAALAS SKIIKHASLKIYPGAPHGLAYTHKDQLNTDLLAFVQSR*

>MIP3H_04062016_scaffold_766_7_RBH:non- heme_chloroperoxidase;_K00433_chloride_peroxidase_[EC:1.11.1.10](db=KEGG) MEARPNSTGFITTKDGVRIFYKDWGAGQPIVFSHGWPLSADDWDAQMLFFGERGYRVI AHDRRGHGRSDQTWDGNDMDTYADDLRALVLALDLKDAIHVGHSTGGGEVTRYLGR HGADRAARAVLIGAIPPVMVRKESNPEGLPIEVFDGYRAAYLADRPQLYRDVAAGPFY GFNRPGAKVSQGLIDKWWLQAMLGGAKAQYDCIRVFSETDFTEDLKRIDIPVLLMHGD DDQVVPLADSALKGIKLLQKGTLKVYPGLPHGMASTHAEAINADLLAFIRGEDAAARPA EAVLETA*

233

>MIP5H_02172016_scaffold_6184_1_RBH:alpha/beta_hydrolase_fold_protein;_K00433_c hloride_peroxidase_[EC:1.11.1.10](db=KEGG) MSMITMNDGTQIYYKDWGTGQPIVFSHGWPLNADSWESQMLFLASKGYRCIAHDRRG HGRSSQPWDGNEMDTYADDLSEIIEALDLKSIVLIGFSAGGGEVARYIGRHGTKRVAKA ALIAAVPPLMLKTDANPSGLPIEAFDEIRLGSIADRSQFYKDLASGPFFGANRTGSKVSQG MIDSFWLQGMQAGSKNTFDCIKAFSETDFTEDLKKFDVPTLIIHGDDDQIVPIGAAALAS SKIIKHASLKIYPGAPHGLAYTHKDQLN

>Daly_0_ORIG_scaffold_207_40_RBH:chloride_peroxidase;_K00433_chloride_peroxidase _[EC:1.11.1.10](db=KEGG) MSTLNRALTLAALTLATAVTAQAVQAGDTRSAATLGSERSESYVTTKDGVSLYYKDW GPRDGQVVTFSHGWPLNSDSWESQMMFLASKGYRVVAHDRRGHGRSSQPWDGNDMD HYADDLAAVLEALDLKDATLIGFSTGGGEVARYIGRHGTARVKKAVLVASVPPLMLKT ESNPDGVPLEVFEGLRQASLGNRSQLYLDIASGPFFGYNRPGATPSQGLIQSFWVQGMQ AGHKNTYDSIAAFSATDFREDLRKFDVPTLVIHGDDDQIVPIDTSARASASIVKDAELIIY PGAPHGLTDTHKERLNQDLLAFLRK*

234

Supplemental Tables

Table16D1. Metagenomic information and accession numbers

sample Formation timepoint total sequencing (bp) read length reads Location Accession Number Marcellus_1_0 Marcellus input 5,574,662,568 108 51,617,246 NCBI SAMN04417440 Marcellus_1_T7 Marcellus 7 7,079,550,120 108 65,551,390 NCBI SAMN04417539 Marcellus_1_T13 Marcellus 13 6,749,251,992 108 62,493,074 NCBI SAMN04417544 Marcellus_1_T82 Marcellus 82 9,260,494,776 108 85,745,322 NCBI SAMN04417545 Marcellus_1_T328 Marcellus 328 7,114,543,200 108 65,875,400 NCBI SAMN04417546 Marcellus_4_T25 Marcellus 25 uploaded, awaiting public availability Marcellus_4_T34 Marcellus 34 uploaded, awaiting public availability Marcellus_4_T36 Marcellus 36 uploaded, awaiting public availability Marcellus_4_T79 Marcellus 79 uploaded, awaiting public availability Marcellus_4_T93 Marcellus 93 uploaded, awaiting public availability Marcellus_4_T142 Marcellus 142 10,912,741,310 151 72,269,810 JGI-IMG 3300013020 Marcellus_4_T205 Marcellus 205 uploaded, awaiting public availability Marcellus_5_T46 Marcellus 46 uploaded, awaiting public availability Marcellus_5_T103 Marcellus 103 6,619,196,740 151 43,835,740 JGI-IMG 3300013018 Marcellus_5_T118 Marcellus 118 8,579,021,210 151 56,814,710 JGI-IMG 3300013016 Marcellus_5_T152 Marcellus 152 8,946,390,620 151 59,247,620 JGI-IMG 3300013021 Marcellus_5_T250 Marcellus 250 8,267,142,790 151 54,749,290 JGI-IMG 3300013019 Marcellus_5_T313 Marcellus 313 8,818,376,142 151 58,399,842 JGI-IMG 3300013017 Utica_3_T1 Utica 1 3,136,131,382 151 20,769,082 JGI-IMG 3300010372 Utica_3_T9 Utica 9 7,079,485,510 151 46,884,010 JGI-IMG 3300009744 Utica_3_T17 Utica 17 4,355,844,452 151 28,846,652 JGI-IMG 3300010374 Utica_3_T23 Utica 23 8,329,763,094 151 55,163,994 JGI-IMG 3300009576 Utica_3_T59 Utica 59 11,846,352,264 151 78,452,664 JGI-IMG 3300009625 Utica_3_T87 Utica 87 7,030,271,892 151 46,558,092 JGI-IMG 3300009574 Utica_3_T122 Utica 122 7,406,509,230 151 49,049,730 JGI-IMG 3300009575

235

Table17D2. V3-V4 region of the EMIRGE reconstructed 16S rRNA for reported taxa

sample_name length relative_abundance sequence_score bp_score identity quality lca_tax_slv Utica-3_Input_FT 1326 0.02 0.989 101 99.17 98 Bacteria;Proteobacteria;Gammaproteobacteria;Pseudomonadales;Pseudomonadaceae;Pseudomonas; Utica-3_Input_RT 1449 0.03 0.999 118 100.00 99 Bacteria;Proteobacteria;Gammaproteobacteria;Pseudomonadales;Pseudomonadaceae;Pseudomonas; Utica-3_Day38 1386 0.18 0.998 104 99.70 99 Bacteria;Proteobacteria;Gammaproteobacteria;Pseudomonadales;Pseudomonadaceae;Pseudomonas; Utica-3_Day38 1232 0.01 0.990 45 98.26 98 Bacteria;Proteobacteria;Gammaproteobacteria;Pseudomonadales;Pseudomonadaceae;Pseudomonas; Utica-3_Day54 1537 0.04 0.999 121 99.34 99 Bacteria;Proteobacteria;Gammaproteobacteria;Pseudomonadales;Pseudomonadaceae;Pseudomonas; Utica-3_Day60 1464 0.02 0.875 111 96.02 87 Bacteria;Proteobacteria;Gammaproteobacteria;Pseudomonadales;Pseudomonadaceae;Pseudomonas; Utica-3_Day96 1410 0.00 0.986 95 99.01 98 Bacteria;Proteobacteria;Gammaproteobacteria;Pseudomonadales;Pseudomonadaceae;Pseudomonas;

Pseudomonas Utica-3_Day96 1290 0.00 0.922 90 94.15 92 Bacteria;Proteobacteria;Gammaproteobacteria;Pseudomonadales;Pseudomonadaceae;Pseudomonas;

Paramaledivibacter Marcellus-4_Day79 1248 0.01 0.980 97 92.03 98 Bacteria;Firmicutes;Clostridia;Clostridiales;Clostridiaceae 4;Paramaledivibacter;

Utica3_InputFT 1488 0.013 0.99 119 95.16 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Acidovorax; Utica3_InputFT 1360 0.012 0.99 116 98.74 98 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Aquabacterium; Utica3_InputFT 1373 0.007 1.00 104 96.40 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae; Utica3_InputFT 1365 0.007 1.00 112 99.44 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Acidovorax; Utica3_InputFT 1515 0.006 1.00 119 97.46 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Oxalobacteraceae;Undibacterium; Utica3_InputFT 1483 0.006 0.98 116 97.78 98 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;uncultured; Utica3_InputFT 1518 0.005 1.00 118 96.90 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Oxalobacteraceae;Undibacterium; Utica3_InputFT 1482 0.005 0.99 116 98.50 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae; Utica3_InputFT 1434 0.005 0.99 114 98.26 98 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Aquabacterium; Utica3_InputFT 1369 0.005 1.00 104 99.12 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Sphaerotilus; Utica3_InputFT 1334 0.005 1.00 117 98.49 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Paucibacter; Utica3_InputFT 1471 0.004 0.99 118 97.82 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Oxalobacteraceae;Massilia;

238 Utica3_InputFT 1483 0.004 0.99 116 96.91 98 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae; Utica3_InputFT 1311 0.004 0.99 116 98.02 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Oxalobacteraceae;Duganella;

Utica3_InputFT 1232 0.003 1.00 114 98.59 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Oxalobacteraceae;

Utica3_InputFT 1411 0.002 1.00 110 97.31 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae; Utica3_InputFT 1391 0.002 0.99 111 94.48 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Burkholderiaceae; Utica3_InputFT 1491 0.002 0.94 120 94.61 94 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Burkholderiaceae;Polynucleobacter; Utica3_InputFT 1480 0.002 0.99 116 96.34 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Oxalobacteraceae; Utica3_InputFT 1394 0.002 0.99 108 96.28 98 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae; Utica3_InputFT 1292 0.001 0.98 110 94.00 98 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Oxalobacteraceae;Undibacterium; Utica3_InputFT 1348 0.001 0.97 117 92.88 97 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Oxalobacteraceae; Utica3_InputRT 1455 0.009 1.00 116 98.14 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Aquabacterium; Utica3_InputRT 1417 0.003 0.94 110 94.98 94 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Aquabacterium;

Burkholderiales Utica3_Day38 1495 0.143 1.00 121 100.00 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Comamonas; Utica3_Day38 1515 0.105 1.00 123 99.93 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Aquabacterium; Utica3_Day38 1517 0.059 0.99 118 100.00 98 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Burkholderiaceae;Limnobacter; Utica3_Day38 1490 0.036 0.98 120 99.70 97 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae; Utica3_Day38 995 0.018 1.00 106 99.50 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Aquabacterium; Utica3_Day54 1436 0.051 0.99 115 98.55 98 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Aquabacterium; Utica3_Day54 1519 0.046 1.00 122 99.65 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Hydrogenophaga; Utica3_Day54 1491 0.039 1.00 120 99.22 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae; Utica3_Day54 1488 0.026 1.00 121 97.90 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Hydrogenophaga; Utica3_Day54 1394 0.027 1.00 107 98.57 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Limnohabitans; Utica3_Day54 1491 0.021 0.96 115 96.07 95 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae; Utica3_Day54 1229 0.012 0.86 107 88.91 86 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales; Utica3_Day54 862 0.013 0.99 90 97.45 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Ramlibacter; Utica3_Day54 1154 0.009 1.00 106 99.22 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Burkholderiaceae;Polynucleobacter; Utica3_Day60 1311 0.016 0.99 102 97.10 98 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Hydrogenophaga; Utica3_Day96 1377 0.002 1.00 100 97.82 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae; Utica3_Day96 1452 0.001 0.91 103 93.10 91 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Delftia; Utica3_Day96 1349 0.001 0.78 73 86.43 78 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae;Aquabacterium; Utica3_Day96 1345 0.001 0.97 86 96.73 96 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Oxalobacteraceae;Massilia; Utica3_Day96 1278 0.000 0.96 103 95.62 96 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Comamonadaceae; Utica3_Day96 394 0.001 1.00 62 99.24 99 Bacteria;Proteobacteria;Betaproteobacteria;Burkholderiales;Oxalobacteraceae;Massilia; 236

Table18D3. Isolate genes used for reference gene database

Gene Isolate NCBI Accession Reference (Wiesner Non-heme chloroperoxidase Pseudomonas pyrrocinia P25026.3 et al., 1988) Lechevalieria (Yeh et Tryptophan 7-halogenase WP_030469177.1 aeocolonigenes al., 2005) Xanthobacter (Ploeg et Haloacetate dehalogenase WP_012115074.1 autotrophicus DSM 432 al., 1995) (Kurihara Burkholderia sp. FA1 Q1JU72.1 et al., 2003) (Zhang et (S)-2-haloacid dehalogenase Paracoccus sp. DEH99 ACX54860.1 al., 2014) (Pavlová Haloalkane dehalogenase Mycobacterium avium WP_003872427.1 et al., 2007) (Nagata et Sphingomonas japonicum WP_013040256.1 al., 1997) (Jesenská Mycobacterium KXN95962.1 et al., tuberculosis 2005) Ortho-chlorophenol reductive Desulfitobacterium (Jugder et WP_072770830.1 dehalogenase chlororespirans al., 2016) Desulfitobacterium (Jugder et WP_015261904.1 dichloroeliminans al., 2016) Ortho-bromophenol reductive (Jugder et Nitratireductor pacificus WP_008593084.1 dehalogenase al., 2016) Meta-chlorophenol reductive Desulfitobacterium (Jugder et AAL87752.1 dehalogenase hafniese sp. DCB-2 al., 2016) Ethane/Methane reductive Dehalobacter sp. (Jugder et WP_021315176.1 dehalogenase UNSWDHB al., 2016) Propane reductive Dehalogenimonas (Jugder et WP_013217756.1 dehalogenase lykanthroporepellens al., 2016) Chlorobenzene reductive Dehalococcoides (Jugder et WP_012882535.1 dehalogenase mccartyi sp. CBDB1 al., 2016) Dehalococcoides (Jugder et Ethene reductive dehalogenase WP_012882535.1 mccartyi al., 2016) Tetrachlorohydroquinone Sphingobium (Xun et WP_037457245.1 reductive dehalogenase chlorophenolicum al., 1992)

237

Table19D4. Reference gene database amino acid bit-score and identity (%) cutoffs

Gene Bitscore Identity Halogenases Non-heme chloroperoxidase 200 45 Tryptophan 7-halogenase 200 30 Hydrolytic dehalogenases Haloacetate dehalogenase 100 30 (S)-2-haloacid dehalogenase 200 50 Haloalkane dehalogenase 100 30 Reductive dehalogenases Ortho-chlorophenol 125 30% Ortho-bromophenol 200 30% Meta-chlorophenol 200 40% Ethane/Methane 100 30% Propane 100 25% Chlorobenzene 100 30% Ethene 100 30% Tetrachlorohydroquinone 200 50%

238

Table20D5. Gene annotations and blastp taxonomy Blast % in Reference Blast Bit Well_Day Scaffold_Gene_Annotation Blast Taxonomy Identical ID in bins? Classification EMIRG Gene Type Score Sites E 16S? Daly_T328_ORIG_scaffold_660_6_BLAST :ferredoxin(db=KEGG_evalue=9.3e- Halomonas a,b Marcellus1_328185_bit_score=651.7_identity=82.9) taeanensis 769.229 100 Yes- HalomonadaceaeHalomonadaceae n/a S-1- Day17_scaffold_409_6_RBH:FAD/NAD(P a,b Utica3_54 )-binding_oxidoreductase(db=KEGG) Malikia granosa 646.736 97.8 No Burkholderiales Yes

S-1- Day17_scaffold_461_11_RBH:hypothetical_

241 protein_n=1_Tax=Caldimonas_manganoxida Zhizhongheella ns_RepID=UPI0003709B6D(db=UNIREF) a,b Utica3_54 caldifontis 637.876 96.6 No Burkholderiales Yes S-1- Day17_scaffold_5110_2_RBH:ferredoxin;_ K03863_vanillate_monooxygenase_[EC:1.14 a,b Utica3_54 .13.82](db=KEGG) Hydrogenophaga 617.461 99 No Burkholderiales Yes S-1- Day17_scaffold_516_6_RBH:ferredoxin;_K 03863_vanillate_monooxygenase_[EC:1.14.1 a,b Utica3_54 3.82](db=KEGG) Curvibacter delicatus 628.246 93.8 No Burkholderiales Yes S-1- Day17_scaffold_94_10_BLAST:oxidoreduct ase(db=KEGG_evalue=6.2e- a,b Utica3_54 68_bit_score=263.5_identity=48.1) Hydrogenophaga 536.569 86.1 No Burkholderiales Yes

Continued

239

Table D5 continued.

3H_02032016_scaffold_24865_0001_BLAS T:reductive_dehalogenase(db=KEGG_evalu Natronincola e=8.5e-57_bit_score=225.3_identity=67.8) c,d,e,f,g,h Marcellus4_79 ferrireducens 270.011 84.2 No None No 3H_02032016_scaffold_5978_0001_RBH:re Natronincola c,d,e,f,g,h Marcellus4_79 ductive_dehalogenase(db=KEGG) ferrireducens 619.002 80.9 No None No

3H_02032016_scaffold_7781_0001_BLAST :reductive_dehalogenase(db=KEGG_evalue= 2.4e-105_bit_score=387.9_identity=52.0) c,d,e,f,g,h Marcellus4_79 Paramaledivibacter 549.666 72.8 No ParamaledivibacterYes 3H_06082016_scaffold_7059_0001_RBH:re c,d,e,f,g,h Marcellus4_205ductive_dehalogenase(db=KEGG) Paramaledivibacter 550.051 72.8 No ParamaledivibacterYes 3H_06082016_scaffold_8758_0001_BLAST :reductive_dehalogenase(db=KEGG_evalue= Natronincola 242 c,d,e,f,g,h Marcellus4_2053.2e-84_bit_score=317.0_identity=62.8) ferrireducens 395.586 82.6 No None No

S-1- Day1_scaffold_4013_2_RBH:putative_oxyg enase_electron_transfer_component(db=KE a Utica3_Day38 GG) Pseudomonas 611.683 96.5 No Pseudomonas Yes

S-1- Day17_scaffold_124_5_BLAST:vanillate_O- demethylase_oxidoreductase(db=KEGG_eva lue=8.5e-97_bit_score=359.4_identity=56.3) a Utica3_Day54 Curvibacter delicatus 474.552 70.6 No Burkholderiales Yes

Continued

240

Table D5 continued. S-1- Day17_scaffold_409_1_RBH:ferredoxin(db a Utica3_Day54 =KEGG) Hydrogenophaga 652.899 99.4 No Burkholderiales Yes S-1- Day17_scaffold_611_5_RBH:ferredoxin(db a Utica3_Day54 =KEGG) Derxia gummosa 439.884 68.3 No Burkholderiales Yes S-1- Day17_scaffold_91_8_BLAST:oxidoreducta se_fad/NAD(P)-binding_domain- containing_protein(db=KEGG_evalue=6e- a Utica3_Day54 106_bit_score=389.8_identity=58.8) Curvibacter delicatus 563.148 84 No Burkholderiales Yes Daly_0_ORIG_scaffold_207_40_RBH:chlor ide_peroxidase;_K00433_chloride_peroxidas NHCP Marcellus1_Inpute_[EC:1.11.1.10](db=KEGG) Pseudomonas 641.728 99.1 Yes- Pseudomonas n/a n/a NCP_3H_02172016_scaffold_53_0049_RB H:alpha/beta_hydrolase_fold_protein;_K004 33_chloride_peroxidase_[EC:1.11.1.10](db= NHCP Marcellus4_93 KEGG) Acetobacterium 568.54 100 No None No

243 NCP_MIP3H_04062016_scaffold_766_7_R BH:non- heme_chloroperoxidase;_K00433_chloride_ NHCP Marcellus4_142peroxidase_[EC:1.11.1.10](db=KEGG) Phenylobacterium 498.049 82.5 No None No NCP_MIP5H_02172016_scaffold_6184_1_ RBH:alpha/beta_hydrolase_fold_protein;_K 00433_chloride_peroxidase_[EC:1.11.1.10]( NHCP Marcellus5_103db=KEGG) Acetobacterium 549.28 100 No None No NCP_S-1- Day1_scaffold_1070_21_RBH:fold;_non- NHCP Utica3_Day38 heme_chloroperoxidase(db=KEGG) Pseudomonas 641.343 99.1 No Pseudomonas Yes NCP_S-1- Day1_scaffold_17013_2_RBH:non- heme_chloroperoxidase;_K00433_chloride_ NHCP Utica3_Day38 peroxidase_[EC:1.11.1.10](db=KEGG) Pseudomonas 645.58 99.7 No Pseudomonas Yes NCP_S-1- Day1_scaffold_6370_3_RBH:hypothetical_p NHCP Utica3_Day38 rotein(db=KEGG) Pseudomonas 648.277 100 No Pseudomonas Yes

241

Supplemental References

Evans, M. V., Panescu, J., Hanson, A. J., Welch, S. A., Sheets, J. M., Nastasi, N., et al. (2018). Members of Marinobacter and Arcobacter influence system biogeochemistry during early production of hydraulically fractured natural gas wells in the appalachian basin. Front. Microbiol. 9, 2646. doi:10.3389/fmicb.2018.02646.

Jesenská, A., Pavlová, M., Strouhal, M., Chaloupková, R., Těšínská, I., Monincová, M., et al. (2005). Cloning, biochemical properties, and distribution of Mycobacterial haloalkane dehalogenases. Appl. Environ. Microbiol. 71, 6736–6745. doi:10.1128/AEM.71.11.6736-6745.2005.

Kurihara, T., Yamauchi, T., Ichiyama, S., Takahata, H., and Esaki, N. (2003). Purification, characterization, and gene cloning of a novel fluoroacetate dehalogenase from Burkholderia sp. FA1. Journal of Molecular Catalysis B: Enzymatic 23, 347–355. doi:10.1016/S1381-1177(03)00098-5.

Nagata, Y., Miyauchi, K., Damborsky, J., Manova, K., Ansorgova, A., and Takagi, M. (1997). Purification and characterization of a haloalkane dehalogenase of a new substrate class from a gamma-hexachlorocyclohexane-degrading bacterium, Sphingomonas paucimobilis UT26. Appl Environ Microbiol 63, 3707–3710.

242

Pavlová, M., Klvana, M., Jesenská, A., Prokop, Z., Konecná, H., Sato, T., et al. (2007). The identification of catalytic pentad in the haloalkane dehalogenase DhmA from Mycobacterium avium N85: reaction mechanism and molecular evolution. J. Struct. Biol. 157, 384–392. doi:10.1016/j.jsb.2006.09.004.

Ploeg, J. van der, Willemsen, M., Hall, G. van, and Janssen, D. B. (1995). Adaptation of Xanthobacter autotrophicus GJ10 to bromoacetate due to activation and mobilization of the haloacetate dehalogenase gene by insertion element IS1247. Journal of Bacteriology 177, 1348–1356. doi:10.1128/jb.177.5.1348-1356.1995.

Wiesner, W., Pée, K. H. van, and Lingens, F. (1988). Purification and characterization of a novel bacterial non-heme chloroperoxidase from Pseudomonas pyrrocinia. J. Biol. Chem. 263, 13725–13732.

Xun, L., Topp, E., and Orser, C. S. (1992). Purification and characterization of a tetrachloro-p-hydroquinone reductive dehalogenase from a Flavobacterium sp. J. Bacteriol. 174, 8003–8007.

Yeh, E., Garneau, S., and Walsh, C. T. (2005). Robust in vitro activity of RebF and RebH, a two-component reductase/halogenase, generating 7-chlorotryptophan during rebeccamycin biosynthesis. Proc. Natl. Acad. Sci. U.S.A. 102, 3960–3965. doi:10.1073/pnas.0500755102.

Zhang, J., Xin, Y., Cao, X., Xue, S., and Zhang, W. (2014). Purification and characterization of 2-haloacid dehalogenase from marine bacterium Paracoccus sp. DEH99, isolated from marine sponge Hymeniacidon perlevis. J. Ocean Univ. China 13, 91–96. doi:10.1007/s11802-014-2357-3.

243

Appendix E. Supporting Information for Chapter 5

Figure23E1. Rarefaction curves for observed OTUs at the lowest sequencing depth (18,054 sequences)

Table21E1. Constituents in disclosed chemical additives used in the disposal well facility treatment process

Chemical additive Intended purpose Components

Scale-Clear A250 Dissolve carbonate, iron Phosphoric acid (15-40%),

sulfide, iron oxide scales hydroxyacetic acid (15-30%),

hydrochloric acid (1-10%),

methyl alcohol (10-30%)

Alpha 3207 Corrosion inhibition Organic acid amine salts,

isopropanol, water

244