MICROBIAL DIVERSITY AND METABOLIC POTENTIAL OF THE SERPENTINITE SUBSURFACE ENVIRONMENT By Katrina Irene Twing A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Microbiology and Molecular Genetics Doctor of Philosophy 2015 ABSTRACT MICROBIAL DIVERSITY AND METABOLIC POTENTIAL OF THE SERPENTINITE SUBSURFACE ENVIRONMENT By Katrina Irene Twing Serpentin ization is the hydrous alteration of mafic rocks to form serpentine minerals and magnetite. The reactions of this alteration result in elevated pH of the surrounding fluids, abiotic generation of H 2, CH 4 (and other organic molecules), and depletion of dissolved inorganic carbon. Thus, serpentinization has implications for the origin of life on Earth and possibly Mars and other planetary bodies with water. The microbial diversity of continental serpentinit e systems consistently shows communities that are dominated by two major taxa microaerophilic Betaproteobacteria and anaerobic Clostridia. Previous studies relied on few samples collected from natural springs or seeps, meaning that the flow path of fluid s from the subsurface process of serpentinization was unknown. The Coast Range Ophiolite Microbial Observatory (CROMO), a set of wells drilled into the actively serpentinizing subsurface environment in northern California, was established in northern Calif ornia to gain a better understanding of the habitability and microbial functions within the serpentinite subsurface environment. This dissertation represents a culmination of microbiological investigations into the serpentinite subsurface environment at C ROMO to identify the microbial inhabitants of subsurface fluids, rocks, and in situ colonization experiments using molecular methods and high - throughput sequencing. The CROMO wells represent a broad range of geochemical gradients and pH and the concentrati ons of carbon monoxide and methane have the strongest correlation with microbial community composition. The most extreme ly high pH wells were inhabited exclusively by a single operational taxonomic unit (OTU) of Betaproteobacteria and a few OTUs of Clostri dia, while more moderate pH wells exhibited greater diversity. Genes involved in the metabolism of hydrogen, carbon monoxide, and carbon fixation were abundant in the extreme pH fluids, while genes for metabolizing methane were exclusively in the moderate pH wells. The subsurface environment is an amalgamation of fluids and rocks, and as such, studying fluids alone only gives half the story. CROMO represents the first drill campaign into the continental serpentinite environment and the microbial diversity of serpentinite cores to a depth of 45 meters below surface suggests that specific geological features harbor different microbial communities. Archaea, previously undetected at CROMO, dominated cores containing magnetite - bearing serpentine, while bacteria were more abundant in layers containing clay particles. Additionally, organisms involved in the cycling of nitrogen and methane were found associated with core materials, indicating core - associated communities may have strong biogeochemical roles within th e serpentinite subsurface environment. Given that microbial communities appear to vary with geological composition and that serpentinite fluids are a challenging habitat for life, depleted in inorganic carbon and electron acceptors, microbe - mineral interac tions within the serpentinite subsurface environment through the use of in situ colonization devices to see if communities were able to utilize inorganic carbon in calcite or ferric iron as a terminal electron acceptor from magnetite. In the highest pH wel l, calcite led to an increased abundance of Clostridia and Deinococcus, while magnetite led to an increase in diversity, including Alphaproteobacteria, Gammaproteobacteria, and Actinobacteria, suggesting further that mineralogical composition of solids wit hin the subsurface impact community composition. The data discussed here further our understanding of life associated with serpentinite fluids and minerals within the subsurface environment. iv ACKNOWLEDGEMENTS My scientific education and research has been a great journey, along which I have received advice and mentorship from many peo ple I would like to thank now. A huge t hank you to my advisor and mentor, Dr. Matt Schrenk for his unwavering support and guidance over the last few years. While in his laboratory , I have always felt like I am a valued and contributing member of a team. I would also like to thank the rest of CROMO team, D r. Dawn Cardace, Dr. Tori Hoehler, Dr. Tom McCollom, Mike Kubo, and Dan Carnevale for great discussion and guidance through this interdisciplina ry project and for a memorable field exp erience . Thank you to Christopher Thornton and Alex Hyer for technical a ssistance and guidance on data analysis. Thank you to Drs. Melitza Crespo - Medina and Dani Morgan - Smith for being great lab mates and even better friends. That you to Dr. Billy Brazelton for his thoroughly enjoyed working wit h and learning from him and am eagerly look ing forward to the next stage of my career with him as a postdoctoral researcher in his laboratory . I would like to acknowledge the NASA Astrobiology Institute for funding the establishment of CROMO, where the majority of my dissertation research took place, and the Deep Carbon Observatory for research funding and community networking opportunities. Thank you to my dissertation guidance committee, including Drs. Matt Schrenk, Kazem Kashefi, Ashely Shade, Warren Wood, and Tori Hoehler, for their insightful and constructive input on my res earch project and dissertation. Having transferred to MSU during the middle of my degree, I greatly appreciated the welcoming guidance of my committee and Dr. Bob Hausinger and Ms . Betty Miller in navigating my new department and community. Thanks to these people, I fe lt truly welcome in the MMG department and consider it an honor to now be an alumna of it. v Thank you to the strong female educators and scientists who have mentored me throughout my scientific education . I had a strong interest in the biological sciences from an early age , thanks to my middle and high school science teachers, Ms. Laubach and M s. Brunette , who introduced my inquisitive mind to the scientific process. W hile perusing my B.A. at Clark University, Dr. Deborah Robertson mentored me through my first laboratory experience and introduced me to the wonders of marine microbiology . During my M.S. research at the University of Delaware, my advisor Dr. Barbara Campb ell helped me become an independent researcher . I owe a very big t hank you to Dr. Jennifer Biddle for being a great friend and mentor over the years and particularly for introducing me to Matt and setting me on this exciding journey into subsurface micro biology . Finally, I would like to thank m y incredibly supportive family who ha ve always believed in me and helped me achieve my dreams . Thank you to my mother, Ms. Lori Twing, for encouraging me to always do my best and reach for the stars. Her unwaverin g belief that I can achiev e anything I put my mind helped to mold me into the dedicated scientist I am today. I would like to t hank both my father Mr. Jeff Twing and stepfather Mr. Mike Marrone for encouraging me to look at challenges from different angles in order to find the best solution. A big t hank you to my loving brother, Mr. Peydon Twing, for always being my biggest cheerleader and best friend . A very special thank you to my love and my p artner, Mr. Eric Gardner , for standing b eside me through both the good and stressful times of my Ph .D. education and supporting my dreams for the future . vi TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ..................... v i ii LIST OF FIGURES ................................ ................................ ................................ ....................... x KEY TO ABBREVIATIONS ................................ ................................ ................................ ..... x i CHAPTER 1 Introduction ................................ ................................ ................................ ......... 1 Serpentinization ................................ ................................ ................................ ........................... 1 Habitability and Challenges of the Extreme Environment ................................ .......................... 2 Lost City Hydrothermal Field ................................ ................................ ................................ ...... 4 Continental Ophiolites ................................ ................................ ................................ ................. 6 Coast Range Ophiolite Microbial Observatory ................................ ................................ ........ 10 R EFERENCES ................................ ................................ ................................ ......................... 14 CHAPTER 2 Serpentinizing f luids from a subsurface observatory harbor extremely low diversity microbial communities adapted to high pH ................................ .............................. 19 Abstract ................................ ................................ ................................ ................................ ..... 19 Introduction ................................ ................................ ................................ .............................. 20 Results and Discussion ................................ ................................ ................................ ............. 22 Sampling Site and Geochemistry ................................ ................................ ....................... 22 Bacterial 16S rRNA Gene Diversity and Community Composition ................................ ... 2 6 Biogeochemical Relationships ................................ ................................ ........................... 3 2 Metabolic Potential ................................ ................................ ................................ ............ 34 Conclusion s ................................ ................................ ................................ ............................... 39 Materials and Methods ................................ ................................ ................................ ............ 4 1 Site Description and Sample Collection ................................ ................................ ............ 4 1 Geochemistry ................................ ................................ ................................ ..................... 4 2 Microbial Cell Counts ................................ ................................ ................................ ........ 42 DNA Extraction ................................ ................................ ................................ .................. 4 3 16S rRNA Tag Sequencing and Data Analysis ................................ ................................ .. 4 3 Metagenomic Sequencing and Data Analysis ................................ ................................ .... 4 4 Statistical Analyses ................................ ................................ ................................ ............ 4 5 Bio energetic Calculations ................................ ................................ ................................ .. 45 R EFERENCES ................................ ................................ ................................ ......................... 4 8 CHAPTER 3 Microbial diversity of serpentinite cores from the Coast Range Ophiolite Microbial Observatory ................................ ................................ ................................ ................ 5 3 Abstract ................................ ................................ ................................ ................................ ..... 5 3 Introduction ................................ ................................ ................................ .............................. 5 4 Results and Discussion ................................ ................................ ................................ ............. 5 6 Drill Core Depth Profiles ................................ ................................ ................................ .. 5 6 DNA Yield, Organic Carbon, and 16S rRNA Gene Abundance ................................ ........ 5 8 Core - Enriched Bacterial Communities ................................ ................................ .............. 60 Core - Enriched Archaeal Communities ................................ ................................ .............. 6 6 Biogeochemical Correlations ................................ ................................ ............................ 6 8 Conclusion s ................................ ................................ ................................ ............................... 70 vii Materials and Methods ................................ ................................ ................................ ............ 7 4 Site Description and Sample Collection ................................ ................................ ............ 7 4 DNA Extraction ................................ ................................ ................................ .................. 7 4 Quantitative - PCR ................................ ................................ ................................ ............... 7 5 16S rRNA Gene Sequencing and Data Analysis ................................ ................................ 7 5 Mineralogical and Geochemical Analyses ................................ ................................ ........ 7 6 Statistical Analyses ................................ ................................ ................................ ............ 7 6 R EFERENCES ................................ ................................ ................................ ......................... 7 7 CHAPTER 4 Microbe - mineral interactions in the ser pentinite subsurface environment . 8 2 Abstract ................................ ................................ ................................ ................................ ..... 8 2 Introduction ................................ ................................ ................................ .............................. 8 2 Results and Discussion ................................ ................................ ................................ ............. 8 4 In Situ Colonization Devices and CROMO Wells ................................ .............................. 8 4 16S rRNA Gene Abundance ................................ ................................ ............................... 8 5 Bacterial Community Composition ................................ ................................ .................... 8 5 Mineral - Enriched Bacteria ................................ ................................ ................................ 8 9 Conclusion s ................................ ................................ ................................ ............................... 9 6 Materials and Methods ................................ ................................ ................................ ............ 9 7 In Situ Colonization Devices ................................ ................................ .............................. 9 7 DNA Extraction ................................ ................................ ................................ .................. 9 7 Quantitative - PCR ................................ ................................ ................................ ............... 9 8 16S rRNA Gene Sequencing and Data Analysis ................................ ................................ 9 8 R EFERENCES ................................ ................................ ................................ ....................... 100 CHAPTER 5 Summary and Future Perspectives ................................ ................................ 10 4 Summary and Future Perspectives ................................ ................................ ........................... 10 4 R EFERENCES ................................ ................................ ................................ ....................... 10 8 APPENDICES ................................ ................................ ................................ ............................ 1 10 A PPENDIX A Example DESeq commands for identifying group specific sequences . 11 1 R EFERENCES ................................ ................................ ................................ ............... 1 1 9 APPENDIX B Bioinformatic analyses for Crespo - Medina et al ., 2014 ......................... 1 21 REFERENCES ................................ ................................ ................................ ............... 1 26 viii LIST OF TABLES Table 2.1 Environmental and geochemical parameters assoc iated with samples collected in August 2012 ................................ ................................ ................................ .............. 24 Table 2.2 correlations analysis ................................ ................................ ................................ . 25 Table 2.3 Summary of significant correlations (p - value < 0. 05) between top OTUs (making up > 25% of the sample in which they are most abundant) and environmental parameters ................................ ................................ ................................ ................ 30 Table 2.4 Microbial communities at CROMO are dominated by bacteria ............................... 31 Table 2.5 Summary of significant correlations between gene hits and OTUs .......................... 36 Table 2.6 Bioenergetic calculations for CROMO groundwaters ................................ ............. 36 Table 3.1 Soil and core sample names and 16S rRNA sequence abundances .......................... 5 9 Table 3.2 Relative abundance of individual core - enriched bacterial taxa within samples ...... 6 4 Table 3.3 Relative abundance of individual core - enriched archaeal taxa within samples ...... 6 9 Table 3.4 Classification of lithostratigraphic units and samples belonging to them ............... 70 Table 4.1 Geochemical parameters of well fluids measured at in situ device depl oyment and retrieval ................................ ................................ ................................ ..................... 8 4 Table 4.2 DESeq comparisons for determining fluid - and mineral - specific taxa .................... 8 9 Table 4.3 Relative - abundance of fluid - enriched bacterial taxa in different sample types ....... 9 4 Table 4.2 Relative - abundance of magnetite - enriched bacterial taxa in different sample types ................................ ................................ ................................ ................................ ... 94 Table 4.3 Relative - abundance of calcite - enriched bacterial taxa in different sample types .... 94 ix Table B .1 Percent identity between end point sequences from March microcosms, most abundant OTUs from tag sequence data, and reference sequences ....................... 1 2 4 Table B.2 Percent identity between two representative end point sequences from the August microcosms, most abundant OTUs from tag sequence data, and reference sequences ................................ ................................ ................................ ............... 1 2 4 x LIST OF FIGURES Figure 1.1 Conceptual model of biogeochemical environment at marine serpentinite sites, such as the Lost City Hydrothermal Field ................................ ................................ .......... 5 Figure 1.2 Conceptual model of biogeochemical environment at continental serpentinite sites . 9 Figure 1 .3 Map of northern California, depicti ng the location of CROMO .............................. 11 Figure 2. 1 Microbial community structure ................................ ................................ ................. 2 7 Figure 2.2 Inter - well community composition comparison from August 2012 .......................... 28 Figure 2.3 Network diagram of significant correlations between OTUs and environmental variables ................................ ................................ ................................ .................... 32 Figure 2.4 Heatmap of most abundant OTUs and geochemical parameters across all samples ................................ ................................ ................................ ................................ ... 33 Figure 2.5 Abundance of protein - encoding genes in metagenomes ................................ ........... 35 Figure 3.1 Depth profile of CSW1.1 ................................ ................................ ........................... 5 6 Figure 3.2 Depth profile of QV1.1 ................................ ................................ .............................. 5 7 Figure 3.3 DESeq comparisons between CSW core and CROMO fluid samples ...................... 6 1 Figure 3.4 Relative abundance of core - , soil - , and fluid - enriched bacterial sequences ............ 6 2 Figure 3.5 Bacterial community structure of unique sequences ................................ ................ 6 3 Figure 3.6 Relative abundance of core - , soil - , and fluid - enriched archaeal sequences ........... 6 6 Figure 3.7 Archaeal community structure of unique sequences ................................ ................. 6 7 Figure 3.8 Correlation network of bacteria and lithostratigraphic units ................................ .. 71 xi Figure 3.9 Negative correlation between iron concentration and DNA yield for select taxa .... 7 1 Figure 3.10 Correlation network of archae a and lithostratigraphic units ................................ . 7 2 Figure 4.1 In situ colonization devices and schematic of deployment down well ...................... 8 6 Figure 4.2 16S rRNA gene copy abundance per gram of substrate, as determined by q - PCR . 8 7 Figure 4.3 Rarefaction curve of unique sequences from in situ colonization device and well fluid sequencing ................................ ................................ ................................ ........ 8 8 Figure 4.4 Community similarity dendrogram of unique sequences from in situ colonization devi ces, well fluids, and core material ................................ ................................ ...... 90 Figure 4.5 Community diversity bar chart for CROMO fluids and in situ colonization devices ................................ ................................ ................................ ................................ ... 92 Figure 4.6 Relative abundances of fluid - , glass - , magnetite - , and calcite - enriched bacteria taxa per each sample in a ll four well s ................................ ................................ .............. 9 5 Figure B.1 Diversity of 16S rRNA gene OTUs (97% sequence similarity) from fluids collected at CROMO in March 2013 was analyzed using mothur ................................ ......... 1 21 Figure B.2 Proportion of bacterial classes present in fluids collected at CROMO in March 2013 ................................ ................................ ................................ ......................... 1 2 2 Figure B.3 Diversity of OTUs (97% similarity) belonging to the classes Betaproteobacteria and Clostridia within CSW1.1, N08B, and CSWold ................................ ............... 1 2 3 xii KEY TO ABBREVIATIONS ANAMMOX anaerobic ammonium oxidation ANOSIM analysis of similarity BLASTP basic local alignment search tool protein C/N carbon to nitrogen ratio CH 4 methane CO carbon monoxide CO 2 carbon dioxide cooS carbon monoxide dehydrogenase (anaerobic) CORK Circulation Obviation Retrofit Kit coxL carbon monoxi de dehydrogenase, large chain (aerobic) CROMO Coast Range Ophiolite Microbial Observatory CSW Core Shed Well CVA Cabeço de Vide Aquifer DAPI - diamidino - 2 - phenylindole DIC dissolved inorganic carbon DNA deoxyribonucleic acid DO dissolved oxygen DOE Department of Energy EA - IRMS elemental - analysis isotope - ratio mass spectrometry H 2 - hydrogen IODP International Ocean Discovery Program JGI Joint Genome Institute xiii LCHF Lost City Hydrothermal Field LCMS Lost City Me thanosarcinales Mbs meters below surface mcrA methyl coenzyme - M reductase MG - RAST Metagenomics Rapid Annotation using Subsystem Technology mxaF methanol dehydrogenase ORP oxidation reduction potential OTU operational taxonomic unit Pfam protein family pmoA particulate monooxygenase qPCR quantit ative polymerase chain reaction RDP Ribosomal Database Project rRNA ribosomal ribonucleic acid RuBisCO r ibulose - 1,5 - bisphosphate carboxylase/oxygenase SIMPRO F similarity profile analysis TOC total organic carbon TN total nitrogen QV Quarry Valley VAMPS Visualization and Analysis of Microbial Population Structures XRD x - ray diffraction XRF x - ray fluorescence 1 CHAPTER 1 Introduction Serpentinization Serpentinization is a widespread geochemical process, involving the water - rock alteration of ultramafic rocks, rich in the mineral olivine (fo r sterite and fayalite) and originating within the 1. 1). Fe 2 SiO 4 + 5Mg 2 SiO 4 + 9 H 2 O Mg 3 Si 2 O 5 (OH) 4 + Mg(OH) 2 + 2 Fe ( O H) 2 ( Eq. 1 .1 ) fayalite + fo r sterite + water serpentine + brucite + iron (II) hydroxide The reaction of ferrous iron (Fe 2+ ) in the resulting iron (II) hydroxide oxidizing to ferric iron (Fe 3+ ) result s in the production of the mineral magnetite and release of hydrogen gas (Equation 1. 2) . 6Fe(OH) 2 2Fe 3 O 4 + 2H 2(aq) + 4H 2 O (Eq. 1. 2) iron (II) hydroxide magnetite + hydrogen gas + water At temperatures below 150ºC, this process results in high pH (>10) and the released of H 2 that, under certain circumstances, can react with CO 2 or CO to abiotically produce methane and small - chain hydrocarbons through Fischer - Tropsch Type rea ctions (E quation 1. 3 ; 11, 27, 34 ) . n ( CO 2 ( aq ) ) + H 2 C n H n + n H 2 O (Eq. 1. 3 ) 2 Serpentinite fluids are typically enriched in Ca 2+ ions, released from Ca - silicate minerals associated with olivine in ultramafic rocks (15) . At high pH, the predominant dissolved inorganic carbon (DIC) species is carbonate (CO 3 2 - ), which readily precipitates out of solution with the Ca + ions in the fluids, producing calcium carbonate and effectively scrubbing all DIC from the fluids (2, 22). Thus, serpentinization creates a highly reduced, high pH environment enriched in H 2 and methane and depleted in DIC . On the early Earth, before the crust had fully differentiated, ultramafic rocks were likely pervasive , and serpentinization may have been even more common than it is today (41 ). It has been speculated that abiogenic organic molecules formed in alkaline environments, such as those derive d from serpentinization, were involved in prebiotic chemistry ( 36, 37 ) and could have led to the molecules necessary for the development of life on Earth . On Mars, mineralogical data from remote sensing indicate the presence of olivine ( 26 ) and serpentinit e ( 14 ), providing a strong possibility that serpentinization has taken place there as well. Furthermore , possible methane fluxes into the Martian atmosphere ( 33, 51 ) indicate that serpentinization may be continuing under the Martian surface today. These s imilarities indicate that an understanding of microbiology and habitability within the serpentinite environment not only informs our understanding of present - day subsurface life, but also can have implications for the origin of life here on Earth and elsew here in the universe. Habitability and Challenges of the Extreme Environment The habitability of an environment, or its potential to host life , can be modeled by calculating the balance of energy produced by an environment and the energy required by life forms to grow and function ( 19 ). Organisms can exist in different levels of activity: growth, 3 where cells are ac tively replicating and thriving; maintenance, where cells are supporting only the most basic metabolic functions ; or survival, a state of hibernation where cells are not metabolically active. Each of these activity levels requires a different amount of energy for sustenance ( 18, 34 ). Thermodynamic models have indicated that the serpentinite environment provides enoug h potential energy , in the form of hydrogen (H 2 ) and methane (CH 4 ), to be habitable for microbial growth (29). There are additional challenges to life in the serpentinite environment beyond the supply of energy required for life. The high pH of the environ ment not only challenges the use of a proton - motive force to generate cellular energy (ATP) (23, 39), but it also leads to the depletion of available DIC in the system. At high pH, the most abundant form of DIC is carbonate ion , which readily precipitates out of solution as calcium carbonate mineral s ( 1, 22 ). Furthermore, the energy supply in the form of electron donors (i.e. H 2 and CH 4 ) is only bioavailable if there are electron acceptors present to be oxidized, and thus allow for microbial metabolism to p roceed. While the serpentinite environment is rich in electron donors, in the form of H 2 and CH 4 , there is a depletion of corresponding electron acceptors in the environment for microbi a l growth and metabolism. Some microbes are able to obtain energy from inorganic mineral phases in their environment , through a process called chemolithotrophy (40). A recent study looking at the diversity of microbes from the glacial environment cultured in situ on various mineral substrates has helped elucidate the microbe - mineral interactions w ithin the system by identifying minerals that enriched for unique microbial communities (30). While the serpentinite subsurface environment is rich in electron donors (which can serve as potential energy sources ) , it lacks ino rganic carbon and electron acceptors necessary for microbial life. Some of these chemical 4 constituents are found in mineral forms in the serpentinite environment, such as carbonate ions in calcite and ferric iron in magnetite, and could potentially be uti liz ed by chemolithotrophic microorganisms. Lost City Hydrothermal Field The most well - characterized serpentinite ecosystem to date is the Lost City Hydrothermal Field (LCHF), which was discovered in 2000 on the Atlantis Massif at 30ºN near the Mid - Atlant ic Ridge (21, 22). The vent represents a 1.5 million year old serpentinite system that has been actively venting for at least 30,000 years (16). LCHF is characterized by large calcium carbonate chimneys extending up to 60 m from the seafloor and venting h ot (40 - 90°C), high pH (9 - 11) fluids containing elevated concentrations of hydrogen (> 14 mM), methane (1 - 2 mM), and short chain organic molecules (21, 22). Isotopic analysis indicates that the most abundant organic molecules at LCHF, such as methane and f ormate, are of abiogenic origin (24, 35). The Lost City also exhibits increased concentrations of acetate relative to background seawater, which is thought to be of microbial origin , either as a product of autotrophy or the byproduct of the heterotrophic breakdown of larger organic molecules (24). Whether mic robes are able to utilize mantle - derived carbon is a major question in a system where organic molecules are abiotically generated. Lang et al . (25) demonstrated that up to 50% of the carbon in biomass from actively venting chimney samples was mantle - derived, indicating that microbes in the serpentinite environment may be utilizing the carbon from the deep Earth. The deep, highly reducing fluids from serpentinization mixing with ocean water, containing oxidants and CO 2 , results in a chemical disequilibrium that microorganism in the environment can exploit for energy. The high abundance of methane at the LCHF has clearly impacted the microbial communities at the vents, as the target gene for methanogenesi s and 5 anaerobic methane - oxidation, methyl - coenzyme M reductase ( mcrA ), was detected in chimney samples (22). There are distinct differences in community composition between young, actively venting chimneys and inactive or extinct chimneys. The anoxic inter ior of the actively venting chimneys was dominated by a single archaeal phylotype capable of metabolizing methane, termed the Lost City Methanosarcinales (LCMS; 6, 38). Active chimneys also contained methylotrophic Gammaproteobacteria (3). Meanwhile, older , inactive chimneys were dominated by anaerobic methane - oxidizing archaea from the ANME - 1 group (3, 5). While the LCMS exhibited little diversity among 16S rRNA genes, metagenomic analysis displayed unprecedented abundance and diversity among transposase g enes, which are involved in gene duplication and transfer (4) and were experimentally shown to exhibit diverse physiology with regards to the production and oxidation of methane (6). Hydrogen - oxidizing Betaproteobacteria of the order Burkholderiales and p otentially Figure 1.1 Conceptual model of biogeochemical environment at marine serpentinite sites, such as the Lost City Hydrothermal Field. Reproduced from Schrenk, et al. , 2013 . 6 fermentative and/or sulfate - reducing Clostridia were also identified at the LCHF (5, 7). Additionally, genes required for carbon - fixation, both via the Calvin - Benson - Bassham pathway (RuBisCO) and the Wood - Ljungdahl pathway were detected in samples from the Lost City ( 5 ). Figure 1.1 depicts a conceptual hydrogeological model that provides a transport mechanism to sustain life at the LCHF, where oxic seawater (rich in electron acceptors) mixes with highly reducing serpentinite fluids rich in electron donors , providing a zone of high energy potential for microbial life (Fig. 1.1; 39). Both chemical and biological data from the Lost City Hydrothermal Field suggest that microorganisms in this ecosystem are capable of metabolizing the geochemical products of serpentinization ( 3, 25 ) . Continental Ophiolites Serpentinization takes place when mantle - derived ultramafic rocks are exposed to the surface along slow and ultraslow spreading ridges ( 22 ), as is evident with the LCHF. Through tectonic processes, such as obduction, the ocean crust and mantle associated with it are emplaced onto the continents crust. Obduction, allows for the rela tively easy access to the mantle and its geochemical weatheri ng processes. The Tablelands Ophiolite in Newfoundland, Canada, is a 500 million year old serpentinite, characterized by high pH springs, enriched in H 2 ( 0.5 mM ) and CH 4 ( 18.8 µ M ) ( 45 ). M etagenomic analysis of one of the end member springs yielded some in sights into the metabolic potential and extremely low biodiversity of the continental subsurface environment ( 7 ). Genes encoding enzymes involved in both the oxidation and production of H 2 , [NiFe] - hydrogenase and [FeFe] - hydrogenase, respectively, were iden tified within the sample, indicating that organism within the community have the genetic potential not only to consume H 2 gas, but produce it as well. Phylogenetic analysis of the [NiFe] - hydrogenase identified the gene as 7 belonging to members of the Burkho lderiales. Furthermore, the largest contig formed from the sequence data contained operons for the [NiFe] - hydrogenase, as well as carbon monoxide dehydrogenase ( coxL ) and RuBisCo ( 7 ) and the coxL gene bore strong homology to Hydrogenophaga pseudoflava (order Burkholderiales). The presence of these genes on the same contig, combined with the fact that H. pseudoflava is a facultative anaerobe, capable of autotrophic growth on H 2 and carbon monoxide ( 50), are evidence that these microbes may be inhabitin g a mixing zone, where H 2 and O 2 are both present. Additionally, microcosm experiments from the Tablelands found carbon monoxide - oxidation to take place in samples with an abundance of Hydrogenophaga (32) . Phylogenetic analyses of the [FeFe] - hydrogenase i ndicate that it comes from members of the Clostridiales, who are known to be anaerobic fermenters. A recent interdisciplinary study at Tablelands using correlation networks to determine relationships between microbiology and geochemical factors found that Hydrogenophaga was likely to inhabit H 2 - rich transition zones, while a member of the Firmicutes, related to Clostridia, was more likely to inhabit the anoxic end - member fluids (8). The data indicate that the ecosystem is made up of two main players: aerobi c H 2 - fueled, autotrophic Burkholderiales inhabiting the oxygen mixing - zone and anaerobic H 2 - producing Clostridia /Firmicutes , that may be autotrophically fermenting the abiogenic hydrocarbons pro duced by serpentinization, in end - member fluids ( 7, 8 ). Cabe ç o de Vide is a high pH (11.4), serpentinite - influenced spring in Portugal , which has been studied for its microbiology (46, 47, 48). The s pring has extremely low microbial biomass, with cell abundances of roughly 6 10 2 cells mL - 1 ( 46, 48 ) . Similar to findings at the Tablelands Ophiolite (7, 8), 16S rRNA gene clone libraries and tag sequencing analyses indicate that the bacterial community is dominated by Clostridia, some of which are closely related to species 8 identified in other high - hydro gen, subsurface ecosystems (12, 48). The sample showed relatively high bacterial diversity , compared to other serpentinite sites (8) , which included the Betaproteobacterium Hydrogenophaga , and low archaeal diversity, including methane - oxidizers of the ANME - 1 group (48). Analysis of functional genes targeting carbon - fixation pathways indicated that the Calvin - Benson - Bassham cycle was the only pathway used. Additionally, genes for sulfur - and methane - oxidation were identified (48). In recent years the microb iology at The Cedars, a site of serpentinization located in Northern California on the Franciscan Subduction Complex, has been explored. The Cedars hosts highly reducing ( - 656 to - 585 mV), high pH fluids (pH 11 - 12) of non - marine origin (31). Studying two s prings at The Cedars over three years, researchers observed stable microbial communities strongly correlated with the source of spring fluids ( 43 ). The deeper fluids, characterized by pH 11.9, redox potential of - 700 mV , contained Clostridia, Chloroflexi, members of the candidate - phylum OD - 1 and Euryarchaea (43). In contrast, the shallower spring, thought to be a mixing zone of deeply sourced serpentinite fluids and surface groundwater, were dominated by Betaproteobacteria closely related to the genus Hydro genophaga (43). These findings are consistent with previous studies of the serpentinite system, suggesting that Clostridia (and other taxa) live in the deeper more end - member serpentinite fluids and Betaproteobacteria live at anoxic/oxic interfaces (7, 8, 39 ). A novel Betaproteobacterium closely related to Hydrogenophaga was isolated from The Cedars and given the proposed genus Serpentinomonas ( 44 ). The three strains of Serpentinomonas (A1, B1, and H1) a re all alkilaphilic (optimum pH of 11) and autotrophic with growth on hydrogen, oxygen, and calcium carbonate ( 44 ). The strains differ in their electron acceptor utilizatio n, with B1 and H1 using nitrate and A1 using thiosulfate (44). According to 9 genomic data, all three str ains contain the CO 2 - fixation gene, RuBisCO and the [NiFe] - hydrogenase, needed for hydrogen - oxidation, while only strain A1 possesses carbon monoxide dehydrogenase, coxL ( 44 ) . Comparisons of the 16S rRNA genes of the strains to previously published studies of continental serpentinites suggests that Betaproteobacteria previously described as Hydrogenophaga at CVA (48) and Tablelands (8) are 99 - 100% identical to Serpentinomonas strains H1 and A1, respectively (44). Microbiological studies of continental serpe ntinites from around the world show s trik ing similarities in community composition, dominated by anaerobic Clostridia and microaerophilic, hydrogen - oxidizing Betaproteobacteria (Fig. 1.2; 39). A member of the Comamonadaceae, within the Betaproteobacteria, has dominated 16S rRNA gene libraries from serpentinite fluids around the world (7, 8, 13, 44, 48). This organism, initially described as Hydrogenophaga - like (8, 43, 48), was isolated from the Cedars by Suzuki et al (44) who proposed that it belongs to a Figure 1.2 Conceptual model of biogeochemical environment at continental serpentinite sites. Question marks represent yet resolved aspects of the system. Reproduced from Schrenk, et al. , 2013 . 10 novel genus, Serpentinomonas (44). Betaproteobacteria seen at CVA and the Tablelands have >99% sequence identity across the 16S rRNA gene with the newly proposed Serpentinomonas (44). According to metagenomic evidence from the Tablelands (7) and genomic and isolate information from the Cedars (44), this organism is capable of aerobic H 2 - oxidation and carbon fixation, via the Calvin - Benson - Bassham (CBB) pathway. Genes for carbon monoxide oxidation have been observed in both metagenomes (7) and genomes (44) and CO - oxidation has taken place in microcosms dominated by this organism (32). These complimentary data from multiple serpentinite sites support that one of the dominant members of microbial communities from serpentinite fluids is an aerobic, H 2 - oxidizing Betaproteobacterium with potential for CO 2 - fixation and CO - oxidation. Less is known about the Clostridia found in serpentinite samples because they have yet to be isolated. Microcosms from CROMO have yielded Clostridia closely related to Dethiobacter alkaliphilus, a S facultative autotroph capable of H2 - oxidation and sulfate - reduction (13, 42). Evidence from the Tablelands suggests that the Clostridia from the serpentinite subsurface environment are capable of H 2 - production, based on the phylogeny of [FeFe] - hydrogenases found in metagenomes (7). Coast Range Ophio lite Microbial Observatory Prior studies of continental serpentinites have sampled surface seeps that represent an interface between end - member fluids and the atmosphere ( 7, 43, 48 ) and therefore may not represent the actual conditions encou ntered deep within the subsurface. Furthermore, these samples may be indicative of life existing in intermediate high - energy zones, which form as the volatile - rich fluids from serpentinite systems mix with oxidized surface waters, as opposed to the highly reducing environments where serpentinization takes place. The Coast Range Ophiolite (CRO) is a 155 - 170 million year old ophiolite located in northern California, containing 11 numerous calcium - hydroxide rich s prings, indicating serpentinizing activity below t he surface ( 1) and was chosen as the site for a microbial observatory, created with the express ed purpose of studying life and habitability within the continental serpentinite environment. The Coast Range Microbial Observatory allow s direct access and study of the actively serpentinizing subsurface environment. CROMO was established at the UC - Davis McLaughlin Nature Reserve in Lower Lake, CA (Fig 1. 3 ; 9) in August 2011 and consists of three groups of wells located within a one - mile ra dius of each other: the Core Shed Wells (CSW), Quarry Valley (QV), and the N - wells. The CSW and QV wells were drilled using clean drilling techniques in 2011 to enable Figure 1.3 Map of northern California, depicting the location of CROMO. Serpentinit es are highlighted in blue and the hexagon represents the CROMO site at the McLaughlin Natural Reserve in Lower Lake, CA. Reproduced from Cardace et al ., 2013. 12 subsequent monitoring of the microbial communities and associated geochemistry within the serp entinite subsurface over time (9). CSW consists of five wells, drilled to depths of 9 - 27 meters. QV consists of three wells, drilled to depths of 15 - 23 meters. There are two main wells, CSW1.1 and QV1.1, which were drilled to 31 m and 45 m, respectively, i n a manner that yielded core material (9). The N - wells, ranging in depth from 14 - 40 meters, depth from 14 - 40 meters, represent previously existing wells that were drilled in 1983 for environmental monitoring of mining impacts on groundwater, not with the specific purpose of monitoring microbiology and organic geochemistry (9). The establishment of CROMO allows for the geological, geochemical, and microbiological exploration of ser pentinite core material, as well as fluids from the serpentinite environment over time. One study looking at the formation temperatures of methane through the use of methane isotopologues suggested that the methane at CROMO was not of in situ origin, poten tially sourced from deeper, higher temperature fluids (49). Meanwhile, methane from the nearby Cedars site showed clear evidence of being microbially generated (49). Microcosm experiments from CROMO well fluids, which resulted in the enrichment of dominan t community members from the wells, suggested that the microbial communities are not nutrient limited and favored the addition of organic carbon (acetate) or inorganic sources (13). The main taxon grown in the microcosms were Betaproteobacteria of the fam ily Comamonadaceae, which exhibited 100% sequence identity of the 16S rRNA gene to Serpentinomonas strain B1, isolated from The Cedars (13, 44). The other organism grown in the microcosms, particularly those enriched with the sulf ur compounds , was Dethioba cter of the class Clostridia, which shared >99% sequence identity to clones from CVA (13, 48). These data are consistent with the model of the continental serpentinite subsurface environment hosting Clostridia and Betaproteobacteria (39). 13 One of the main objectives in establishing the Coast Range Ophiolite Microbial Observatory was to help elucidate microbial communities within the serpentinite subsurface environment, associated both with fluids and minerals (9) . The research discussed in this dissertation directly investigates photosynthesis - independent microbial communities within the serpentinite subsurface environment by analyzing sequence data from both fluids and cores from CROMO. By combining 16S rRNA amplicon sequencing, metagenome s, and geochemical analyses to study serpentinite fluids from wells across geochemical gradients, the fluid - associated microbial communities, their metabolic potential, and geochemical drivers were determined. 16S rRNA gene analyses of CROMO core materials allowed for the determination of distinct core - associated microbial communities within the serpentinite subsurface environment. Finally, the use of in situ colonization experiments has allowed for further understanding microbe - mineral interactions at CROMO. Th e research presented here aids in our understanding of the serpentinite subsurface environment and habitability globally. 14 REFERENCES 15 REFERENCES 1. Barnes I, JR ( 1971 ) Calcium - magnesium carbonate solid solutions from Holocene conglomerate cements and travertines in the Coast Range of California. Geochmi Cosmochim Ac 35 : 699 - 718 . 2. Barnes I, Oman, and Yugoslavia. Geochmi Cosmochim Ac 42:144 - 145. 3. Brazelton WJ, Schrenk MO, Kelley DS, Bar os s JA (2006) Methane - and sulfur - metabolizing communities dominate the Lost Ci ty Hydrothermal Field ecosystem. Appl Environ Mircob 72 : 6257 - 6270. 4. Brazelton W J , Baross JA (2009) Abundant transposases encoded by the metagenome of a hydrothermal chimney biofilm. ISME J 3 :1420 - 1424. 5. Brazelton W J, et al. (2010) Archaea and bact eria with surprising microdiversity show shifts in dominance over 1,000 - year time scales in hydrothermal chimneys. P roc N atl A ca S ci 107 : 1612 - 1617. 6. Brazelton WJ , Mehta MP, Kelley DS, Baross JA (2011) Physiological differentiation within a single - species biofilm f ueled by serpentinization. mBio 2 : doi:10.1128/mBio.00127 - 11 . 7. Brazelton WJ , Nelson B , Schrenk MO (2012) Metagenomic evidence for H 2 oxidation and H 2 production by serpentinite - hosted subsurfa ce microbial communities. Front Microbiol 2 : doi: 10.3389/fmicb.2011.00268 . 8. Brazelton WJ, Morrill PL, Szponar N, Schrenk MO (2013) Bacterial communities associated with subsurface geochemical processes in continental serpentinite springs. Appl Environ Microb 79:3906 - 3916. 9. Cardace D, et al. (2013) Establishment of the Coast Range ophiolite microbial observatory (CROMO): drilling objectives and preliminary outcomes. Scientific Drilling 16:45 - 55. 10. Carnevale, DC (2013) Carbon sequestration potential of the Coast Range Ophiolite in 11. Charlou JL, Donval JP, Fouquet Y, Jean - Baptiste P, Holm N (2002) Geochemistry of high H 2 and CH 4 vent fluids issuing from ultramafic rocks at the Rainbow hydrothermal field (36 º Chem Geol 191:345 - 359. 12. Chivian D, et al. (2008) Environmental genomics reveals a single - species ecosystem deep within Earth. Science 322 :275 - 278. 16 13. Crespo - Medina M, et al. (2014) Insights into environmental controls on microbial communities in a continental serpentinite aquifer using a microcosm - based approach. Front Microbiol 5:604. 14. Ehlmann BL , Mustard JF, Murchie SL ( 2010 ) Geologic setting of s erpentine dep osits on Mars. Geophys Res Lett 37 : doi:10.1029/2010GL042596. 15. Frost BR, Beard JS (2007) On silica activity and serpentinization. J Petrol 48:1351 - 1368. 16. Früh - Green GL, et al. (2003) 30,000 years of hydrothermal activity at the Lost City Vent Field. Science 301:495 - 498. 17. Herzog H (2009) Carbon dioxide capture and storage. In : Economics and Politics of Climate Change (Oxford University Press, Oxford, UK) pp. 263 - 283. 18. Hoehler TM (2004) Biological energy requirements as quantitative boundary conditions for life in the subsurface. Geobiology 2 :205 - 215. 19. Hoehler TM ( 2007 ) An en ergy balance concept for habitability. Astrobiology 7:824 - 838. 20. K elemen PB, Matter J (2008) In situ carbonation of peridotite for CO 2 storage. Proc Natl Aca Sci 150:17295 - 17300. 21. Kelley DS, et al . ( 2001 ) An off - axis hydrothermal vent field near the Mid - Atlantic Ridge at 30°N. Nature 412 : 145 - 149 . 22. Kelley DS , et al . ( 2005 ) A serpentinite - hosted ecosystem: The Lost Cit y Hydrothermal Field. Science 307: 1428 - 1434. 23. Krulwich TA, Sachs G, Padan E (2011) Molecular aspects of bacterial pH sensing and homeostasis Nature Rev Microbiol 9:330 - 343. 24. Lang S Q , et al . ( 2010 ) Elevated concentrations of formate, acetate and dissolved organic carbon found at the Lost City hydrothermal field. Geochmi Cosmochim Ac 74 : 941 - 952. 25. Lang S Q , et al . ( 2012 ) Microbial utilization of abiogenic carbon and hydrogen in a serpentinite - hosted system. Geochmi Cosmochim Ac 92 : 82 - 99. 26. Longhi J , Knittle E, Holloway JR, Wänke H (1992) The bulk composition and internal structure of Mars. In Mars Univer sity of Arizona . Press, Tucson) pp. 184 208. 27. Matter J, Kelemen PB (2009) Permanent storage of carbon dioxide in geological reservoirs by mineral carbonation. Nature Geosciences 2:837 - 841. 17 28. McCollom TM, Seewald JS (2001) A reassessment of the potential for reduct ion of dissolved CO 2 to hydrocarbons during serpentinization of olivine. Geochmi Cosmochim Ac 65 : 3769 - 3778. 29. McCollom TM (2007) Geochemical constrains on sources of metabolic energy for chemolithotrophy in ultramafic - hosted deep - sea hyd rothermal systems. Astrobiology 7 : DOI: 10.1089/ast.2006.0119 30. Mitchell AC, Lafreniere MJ, Skidmore ML, & Boyd ES (2013) Influence of bedrock mineral composition on microbial diversity in a subglacial environment. Geology G34194.1 DOI:10.1130/G34194.1 31. Morrill PL, et al . ( 2013 ) Geochemistry and geobiology of a present - day serpentinization site in California: The Cedars. Geochmi Cosmochim Ac 109:222 - 240. 32. Morrill PL, et al. (2014) Investigations of potential microbial methanogenic and carbon monoxide utilization pathways in ultra - basic reducing springs associated with present - day continental serpentinization: the Tablelands, NL, CAN. Front Microbiol 5:613. 33. Mumma MJ , et al . ( 2009 ) Strong release of methane on Mars in northern summer 2003. Science 323 : 1041 - 1045. 34. Price PB & Sowers T (2004) Temperature dependence of metabolic rates for microbial growth, maintenance, and survival. PNAS 101 :4631 - 4636. 35. Proskurowski G , et al . ( 2008 ) Abiogenic hydrocarbon production at Lost City Hydrothermal Field. Science 319 : 604 - 607. 36. Russell MJ ( 2003 ) The importance of being alkaline. Science 302 : 580 - 581 37. Russell MJ, Hall AJ, Martin W (2010) Serpentinization as a source of energy at the origin of life. Geobiology 8:355 - 371. 38. Schrenk MO , Kelley D S, Bolton SA , Baross JA ( 2004 ) Low archaeal diversity linked to subseafloor geochemical processes at the Lost City Hydrothermal Field, Mid - Atlantic Ridge. Environ Microbiol 6 : 1086 - 1095. 39. Schrenk MO, Brazelton WJ, Lang SQ (2013) Serpentinization, carbon and deep life. Rev Mineral G eochem 75:575 - 606. 40. Shock EL (2009) Minerals as energy sources for microorganisms. Econ Geol 104: 1235 - 1248. 41. Sleep NH, et al. (2004) H2 - rich fluids from serpentinization: Geochemical and biotic implications. PNAS 101:12818 - 12823. 18 42. Sorokin DY, Tourova TP, Mussmann M, Muyzer G (2008) Dethiobacter alkaliphilus gen. nov. sp. nov. and Desulfovibrio alkaliphilus gen. nov. sp. nov.: two novel representatives of reductive sulfur cycle from soda lakes. Extremophiles 12: 431 - 439. 43. Suzuki S, et al. ( 2013) Microbial diversity in The Cedars, an ultrabasic, ultrareducing, and low salinity serpentinizing ecosystem. Proc Natl Acad Sci USA 110:15336 - 15341. 44. Suzuki S, et al. (2014) Physiological and genomic features of highly alkaliphilic hydrogen - utiliz ing Betaproteobacteria from a continental serpentinizing site. Nat Commun 5:3900. 45. Szponar N, et al. (2012) Geochemistry of a continental sit of serpentinization in the Tablelands Oph iolite, Gros Morne National Park: a Mars analogue. Icarus doi:10.1016 / j.icarus.2012.07.004 . 46. Tiago I , Chung AP, Ver í ssimo A ( 2004 ) Bacterial diversity in a nonsaline alkaline environment: Heterotrophic aerobic populations. Appl Environ Mircob 70 : 7378 - 7387. 47. Tiago I , Mendez V, Pires C , Morais PV , Ver í ssimo A . ( 2006 ) Chimaereicella alkaliphila gen. nov., sp. nov., a Gram - negative alkaliphilic bacteriu m isolated from a nonsaline alkaline groundwater. Syst Appl Microbiol 29 : 100 - 108. 48. Tiago I, Veríssimo A (2013) Microbial and functional diversity of a subterrestrial high pH groundwater associated to serpentinization. Environ Microbiol 15:1687 - 1706. 49. Wang DT, et al. (2015) Nonequilibrium clumped isotope signals in microbial methane. Science 348:428 - 431. 50. Willems A , et al . ( 1989 ) Hydrogenophaga, a new genus of hydrogen - oxidizing bacteria that includes Hydrogenophaga flava comb. nov. (formerly Pseudomonas flava ), Hydrogenophaga palleronii (formerly Pseudomonas palleronii ), Hydrogenophaga pseudoflava (formerly Pseudomonas pseudofla va Pseudomonas caboxydoflava and Hydrogenophaga taeniospiralis (formerly Pseudomonas taeniospiralis ). Int J Syst Bacteriol 39 : 319 - 333. 51. Zahnle K , Freedman RS , Catling DC ( 2011 ) Is there methane on Mars? Icarus 212 : 493 - 5 19 CHAPTER 2 Serpentinizing fluids from a subsurface observatory harbor extremely low diversity microbial communities adapted to high pH 1 Abstract Serpentinization is a widespread geochemical process associated with the aqueous alteration of ultramafic rocks, resulting in abundant fuels (H 2 and CH 4 ) , but potentially challenging conditions for life, including high pH and limited availability of terminal electron acceptors and low concentrations of dissolved inorganic carbon. As a consequence, past studies of ser pentinites have reported low cellular abundances and microbial diversity. E stablishment of the Coast Range Ophiolite Microbial Observatory well network has allowed , for the first time, a comparison of microbial communities and physicochemical parameters di rectly within a serpentinization - influenced subsurface aquifer. Samples were collected from seven wells and subjected to a range of analyses, including solute and gas chemistry, microbial diversity by 16S rRNA gene sequencing, and metabolic potential by me tagenomics, in an attempt to elucidate what factors drive microbial activities in serpentinite habitats. The strongest geochemical influences on biodiversity in these samples were pH and the concentrations of methane and carbon monoxide. A single operation al taxonomic unit (OTU) of Betaproteobacteria and a few OTUs of Clostridia almost exclusively inhabited fluids exhibiting the most serpentinized character. M etagenomes from these extreme samples contain ed abundance s of sequences encoding proteins associate d with hydrogen metabolism, carbon monoxide oxidation, and carbon 1 The work described i n this chapter is currently in submission to the ISME Journal for publication : K.I. Twing, D. Cardace, M.D. Kubo, W.J. Brazelton, T.M. Hoehler, T.M. McCollom, and M.O. Schrenk. ( In Submission ) Serpentinizing fluids from a subsurface observatory harbor extremely low diversity microbial communities adapted to high pH. 20 fixation. Samples from more moderate pH wells were characterized by higher bacterial diversity and the presence of genes required for the metabolism of methane . These data contribute to a growing body of evidence that serpentinite environments are characterized by limited taxonomic diversity, are fueled by hydrogen and carbon metabolisms, and may significantly contribute to subsurface fluxes of these important elements. Int roduction I t is one or more steps removed from photosynthetic sources of carbon and energy and is instead influenced by geological sources ( 1, 2, 3) . However, given the inhe rent lack of accessibility , direct sampling of subsurface environments has been limited. In continental settings, researchers have use d caves (4, 5) , mines (6, 7) , or springs (8, 9, 10, 11) , as windo ws into the subsurface environment. These features grant access to an otherwise inaccessible environment , but they represent opportunistic sampling at locations where the subsurface environment interacts with the surface. In the present study, wells were drilled directly into serpentinization - influenced aquifers of the Coast Range Ophiolite, a portion of ancient seafloor in northern California, USA, to sample microbial communities in serpentinizing rocks and groundwaters. This observatory represents the first investigation of microbial communities with direct acc ess to the range of conditions in the serpentinizing subsurface (12). Serpentinization is a widespread geochemical process involving the aqueous alteration of peridotite to serpentine minerals, resulting in an abundance of potential reductants , in the form of hydrogen, methane, and small organic molecules ( 13, 14) . The geochemical process also releases hydroxyl ions, which creates extremely high pH fluids. At high pH, bicarbonate and carbonate are the dominant species of dissolved inorganic carbon (DIC), an d the latter can precipitate out 21 of solution as carbonate minerals when in the presence of divalent cations, such as Ca 2+ , commonly found in serpentinite fluids. Thus, serpentinized fluids are characteristically low in DIC, particularly dissolved CO 2 . Comp ared to the abundance of reductants, in the form of H 2 and CH 4 typically found in these system, there are a lack of corresponding oxidants, probably limiting the potential for microbial metabolism. Thus the subsurface serpentinite environment is characteri zed by challenges to life such as extreme pH (> 10), limited availability of dissolved carbon, and a lack of potential terminal electron acceptors. The best - characterized serpentinite - hosted microbial ecosystem to date is the Lost City Hydrothermal Field, located 15 km from the Mid - Atlantic Ridge ( 15). The tall carbonate chimneys at Lost City are dominated by methane - cycling archaea in the anoxic chimney interiors (16) and by methanotrophic and sulfur - oxidizing bacteria in the chimney exteriors ( 17) . More recently, researchers have started exploring life within continental serpentinite environments by using natural springs, such as the Tablelands Ophiolite in Newfoundland, Canada (8, 9) and The Cedars site in northern California ( 11, 18) , or previously esta blished wells, such as the Cabeço de Vide Aquifer (CVA) in Portugal ( 19) , to gain access to the serpentinite subsurface. In these previous studies of continental serpentinite sites, microbial communities were dominated by the bacterial taxa Betaproteobacte ria and Firmicutes (20). Metagenomic surveys of the Tablelands Ophiolite suggest that subsurface serpentinite communities are dominated by Firmicutes in the deep, anoxic source - waters and microaerophilic H 2 - oxidizing Betaproteobacteria at the shallow, oxi c/anoxic interface (8, 9, 20). Microcosm experiments from the Coast Range Ophiolite Microbial Observatory (CROMO), the location of this study, have indicated that Betaproteobacteria closely related to Hydrogenophaga pseudoflava and Clostridia (phylum Firmi cutes) closely related to Dethiobacter alkal i philus , are 22 stimulated by small organic molecules that are expected to be available in the serpentinite environment (21). Furthermore, recently published genomes of cultivated isolates of the proposed genus Serpentinomonas , which are most closely related to the genus Hydrogenophaga , are consistent with a role for these organisms in oxic/anoxic interfaces in serpentinizing systems (18). This study combines high - throughput 16S rRNA gene sequencing, metagenomic analyses, and geochemical monitoring across a range of physicochemical conditions in order to relate patterns in microbial diversity and metabolic potential to underlying geochemical parameters and processes in the serpentinite subsurface environment. This work improves our understanding of physiological adaptation and ecology in these ubiquitous ecosystems and facilitates our integration of these systems into models of carbon cycling. Results and Discussion Sampling Site and Geochemistry Fluids were collec ted from seven wells within the Coast Range Ophiolite Microbial Observatory ( CROMO ) , drilled for the purpose of monitoring biogeochemistry and microbial community dynamics with temporal and spatial resolution (12). The wells, ranging in depth from 9 - 27 meters, are located within a 1.5 km radius and exhibit a wide range of geochemical char acteristics. To tease apart which environmental parameters contribute to microbial community composition, geochemical and microbiological data from these seven wells were compared . The geochemical data associated with the well fluids from August 2012 are summarized in Table 2. 1. Samples from wells CSW1.1 and QV1.1 are characterized by extremely high pH (12.2 and 11.5, respectively) and low redox potential (Table 2. 1 ). These wells are depleted in 23 dissolved inorganic carbon (DIC), containing one to two order s of magnitude less DIC than a well with circumneutral pH, CSW1.4 (Table 2. 1). The elevated pH and low DIC at CSW1.1 and QV1.1 are typical of actively serpentinizing environments and are considered to be representative of end - member fluids. CSW1.1 also exh ibited higher concentrations of H 2 and organic acids relative to the other wells (Table 2.1). Meanwhile, wells CSW1.4 and QV1.2 exhibited circumneutral pH, low conductivity and high er DIC (Table 2.1), suggesting they contain a significant component of non - serpentinized groundwater. Nevertheless, these wells have elevated concentrations of H 2 and CO comparable to the wells with higher pH, indicating some input from serpentinization. CSW1.2 (pH 9.3), a moderately high pH well, defined here as pH 8 - 10, contai ned the highest concentration of methane (1.6 mM; Table 2.1) and appears to represent a mixing zone between serpentinite end - member fluids and surface waters. Unsurprisingly , many of the environmental parameters of the system wer e correlated with one another (Table 2.2) , as determined by pairwise Pearson correlation analyses of samples in Table 2. 1 . pH wa s negatively correlated with oxidation - reduction potential (ORP), carbon monoxide , dissolved oxygen ( DO ), and DIC , and it was positively correlated wi th depth and organic acid concentration (Table 2.2). The concentrations of the organic acids acetate, formate, propionate, and butyrate wer e positive ly correlated with one another and they were negatively correlated with ORP (Table 2.2 ). Carbon monoxide concentration was positively correlated with ORP (i.e. positively correlated with a more positive ORP value) and DO, and it was negativel y correlated with conductivity and H 2 concentration ( Table 2.2 ). The concentration of methane was not significantly cor related with any other environmental parameter s . 24 Table 2.1 Environmental and geochemical parameters associated with samples collected in August 2012. CSW1.1 CSW1.2 CSW1.3 CSW1.4 CSW1.5 QV1.1 QV1.2 Depth (mbs) 19.5 19.2 23.2 8.8 27.4 23.0 14.9 Temp (°C) 17.2 18.5 16.9 15.2 16.2 17.9 18.4 pH 12.2 9.3 10.1 7.9 9.7 11.5 7.9 ORP (mV) - 284 - 32 - 83 - 35 - 121 - 155 - 30 DO (mg/L) 0.05 0.41 0.06 1.05 0.03 0.03 0.03 Conductivity (µ S/cm) 5200 3710 4500 1560 4220 2068 1655 DIC (µM) 253 ± 8 605 ± 268 172 ± 16 5046 ± 531 545 ± 13 96 ± 2 979 ± 32 Dissolved H 2 (µM) 0.289 ± 0.004 0.140 ± 0.001 0.283 ± 0.018 0.271 ± 0.013 0.138 ± 0.020 0.075 ± 0.001 0.076 ± 0.009 Dissolved CH 4 (mM) 0.524 ± 0.132 1.625 ± 0.055 0.969 ± 0.529 0.002 ± 0.0003 1.266 ± 0.032 0.301 ± 0.021 0.303 ± 0.030 Dissolved CO (µM) 0.089 ± 0.002 0.158 ± 0.001 0.115 ± 0.013 0.187 ± 0.005 0.124 ± 0.008 0.142 ± 0.004 0.150 ± 0.008 Acetate (µM) 70.79 ± 1.26 < 1.55 < 1.55 < 3.04 < 3.04 10.20 ± 0.33 < 2.01 Formate (µM) 15.74 ± 0.99 < 1.39 < 1.39 < 1.79 < 1.79 < 1.39 < 2.23 Propionate (µM) 3.49 ± 0.003 < 0.01 < 0.01 < 0.01 < 0.01 0.16 ± 0.05 0.22 ± 0.01 Butyrate (µM) 20.99 ± 0.45 < 1.11 < 1.11 < 2.75 < 2.75 5.97 ± 0.30 < 1.89 Microbial cells (cells/mL) 1.8 × 10 5 6.6 × 10 5 2.3 × 10 5 1.0 × 10 5 3.9 × 10 5 1.0 × 10 6 9.5 × 10 5 25 Table 2.2 Significant correlations between environmental parameters according to Pearson s correlation analysis. Correlation R p - value Depth pH + 0.79 < 0.001 DO - 0.59 0.01 DIC - 0.59 0.01 Temp H 2 - 0.54 0.02 pH Butyrate + 0.77 < 0.001 Formate + 0.73 < 0.001 Propionate + 0.66 0.004 Acetate + 0.64 0.005 ORP - 0.87 < 0.001 CO - 0.75 < 0.001 DIC - 0.58 0.01 DO - 0.55 0.02 ORP CO + 0.87 < 0.001 Butyrate - 0.94 < 0.001 Acetate - 0.92 < 0.001 Propionate - 0.89 < 0.001 Formate - 0.88 0.03 Conductivity - 0.63 < 0.001 H 2 - 0.53 0.006 DO DIC + 0.80 < 0.001 CO + 0.65 0.004 Conductivity H 2 + 0.77 < 0.001 Formate + 0.69 0.002 Propionate + 0.67 0.003 Acetate + 0.66 0.004 Butyrate + 0.59 0.01 CO - 0.80 < 0.001 DIC CO + 0.57 0.02 H2 Formate + 0.66 0.004 Propionate + 0.64 0.005 Acetate + 0.61 0.009 Butyrate + 0.55 0.02 CO - 0.60 0.01 CO Acetate - 0.82 < 0.001 Formate - 0.81 < 0.001 Propionate - 0.81 < 0.001 Butyrate - 0.80 < 0.001 Formate Acetate + 0.99 < 0.001 Propionate + 0.99 < 0.001 Butyrate + 0.97 < 0.001 Acetate Propionate + 0.99 < 0.001 Butyrate + 0.99 < 0.001 Propionate Butyrate + 0.98 < 0.001 26 Bacterial 16S rRNA Gene Diversity and Community Composition Bacterial diversity was assessed in fluids collected in August 2012 from the seven CROMO wells. Three nearby previously established wells (N08A, N08B, N08C) were also included in this analysis in order to address po tential concerns of post - drilling stabilization within the new wells. Environmental sequences of 16S rRNA gene amplicons were obtained with an Illumina MiSeq platform, yielding between 78,000 and 226,000 merged paired - end sequences per sample, for a total of 4,354,377 16S rRNA sequences in this study. These sequences were clustered into 11,454 OTUs at a 97% sequence similarity threshold, and only 71 of these OTUs comprised greater than 1% of the sequences in any of the samples analyzed. All diversity analyses in this study were conducted with OTUs, instead of relying solely on taxonomic annotations, in order to avoid the biases and limitat ions inherent to database - dependent analyses that are magnified when studying poorly characterized microbial communities. Field replicates of samples were collected and analyzed in parallel and we re statistically indistinguishable from one another , as dete rmined by a SIMPROF test of the whole bacterial community similarities among all samples ( Fig. 2. 1) . The community compositions of samples from different wells were clearly distinct from each other (ANOSIM, R = 0.96, p - value = 0.001) . Alpha diversity of t he samples, as measured by the inverse Simpson diversity index, decreased with increasing pH (Fig. 2.2). The wells with the highest pH, CSW1.1 and QV1.1, exhibited extremely low diversity, containing almost exclusively Betaproteobacteria and Firm icutes (Fig. 2.2 ), which is consistent with previous reports of communities in high - pH serpentinite fluids ( 9, 11, 19) . CSW1.1 was dominated by a single betaproteobacterial OTU (OTU001), classified as a member of the family Comamonadaceae and 100% identica 27 Figure 2. 1 Microbial community structure. A) Rarefaction analysis of 16S rRNA amplicon sequences. Multiple samples from the same well represent field replicates. B) Community similari ty dendrogram calculated from Bray - Curtis index. Samples connected by red lines are not distinguishable from one another by a SIMPROF test and ANOSIM analysis indicated that there is a significant difference in community composition betwe en wells (R = 0.9, p - value<0.05 ) (calculated with MatGat) to strain B1 from the proposed genus Serpentinomonas isolated from The Cedars serpentinite site (18 ; Table 2.3 ). The second most abundant OTU in CSW1.1 (OTU018) comprised 12.5 ± 5.8% of the sequences from that well, was classified as Thermoanaerobacterales SRB - 2 , and exhibited 99% sequence identity to a Clostridia clone from a well in Cabeço de Vide (CVA) in Portugal (19 ; Table 2.3 ). The third OTU detected in CSW1.1 (OTU002) accounted for only 1% of the sequence s from that well, was classified as Dethiobacter , and shared 100% sequence identity to a clone from CVA (19 ; Table 2.3 ), a microcosm isolate from CROMO ( 21) , and a clone from the deep groundwater site at The Cedars 28 . Figure 2.2 Inter - well community composition comparison from August 2012. OTUs were formed at the 97% similarity level in mothur using the average - neighbor algorithm. Bars indicate individual OTUs and are color - - proteobacteria (red), Firmicutes (blue), - proteobac - - proteobacteria (teal), and other classes of bacteria (brown). Other (purple bars) represents any OTUs making up < 1% of all sample s in the dataset. Samples are organized by pH (ANOSIM: R = 0.66, p - value < 0.05). T he numbers in parentheses beside the sample names represent the inverse - Simpson diversity index. 29 ( 11) . The remaining 29.0 ± 5.3% of the CSW1.1 microbial community was made up of rare species, defined as OTUs comprising less than 1% of the total sequences in any sample QV1.1 was dominated by three Clostridia OTUs (OTU003, OTU002, and OTU007) that together account for 47.1 ± 13.4% of the bacterial community (Fig . 2.2 ). Both OTU003 (classified as Thermoanaerobacterales SRB - 2) and OTU002 (classified as Dethiobacter ) exhibited 99 - 100% sequence identity to clones from CVA in Portugal (19 ; Table 2.3 ). The same single betaproteobacterial OTU001 made up 28.5 ± 12.9% of the QV1.1 community, and the remaining bacterial taxa were rare , and accounted for 2 4 . 3 ± 1 . 6 % of the community (Fig. 2.2 ). A time - series analysis of samples collected from QV1.1 over the course of a year indicate that community composition within wells is relatively constant over time, with no sign ificant difference between time points (R=0.2, p - value=0.902). While Betaproteobacteria made up a large proportion of all samples above neutral pH, the diversity and composition of the Betaproteo bacteria shifted with pH (Fig. 2.2 ). As expressed above, OTU001 made up 42.7 ± 17.6% of the extreme ly high pH wells. However, in samples Azonexus hydrophilus ; Table 2 .3 ), OTU005 (classified as Methylophilaceae) and OTU004 (classified as Comamonadaceae and 100% identical to Alicycliphilus denitri ficans ; Table 2 .3 ) as the dominant taxa. Clostridia, which accounted for up to 64% of the bacteria in the highest pH fluids, were also found in the moderately high pH wells. Dethiobacter OTUs made up 13.9 ± 10.1% of samples with a pH 9.5 - 11.0. Erysipelot richia (another class of the phylum Firmicutes), which was enriched in subsurface conditions at the Tablelands Ophiolite ( 9 ) , made up 8% and 2% of CSW1.5 (pH 9.7) and CSW1.2 (pH 9.3), respectively, but was not detected in any other CROMO samples. OTUs clas sified as Thermoanaerobacterales SRB - 2 were detected in N08A 30 Table 2.3 Summary of significant correlations (p - value < 0.05) between top OTUs (making up > 25% of the sample in which they are most abundant) and environmental parameters. OTU Variable R Corr* Strength Max Sample Max Abundance (% of sample) Class Order Family Closest Relative (NCBI accession number) % Identity OTU001 Butyrate 0.88 + S CSW1.1 61.7 Betaproteobacteria Burkholderiales Comamonadaceae Serpentinomonas B1 ( AP014569.1 ) 1 100% pH 0.86 + S Acetate 0.86 + S Propionate 0.83 + S Formate 0.82 + S Conductivity 0.63 + M Hydrogen 0.54 + M ORP 0.91 - V S CO 0.86 - S OTU004 Methane 0.81 + S CSW1.2 38.4 Betaproteobacteria Burkholderiales Comamonadaceae Alicycliphilus denitrificans (NR_074585.1) 2 100% DO 0.53 + M OTU003 Depth 0.54 + Mo QV1.1 38.2 Clostridia Thermoanaerobacterales SRB2 CVCloAm3Ph15 (AM778006) 3 99.6% OTU008 pH 0.61 - M QV1.2 38.0 Betaproteobacteria Rhodocyclales Rhodocyclaceae Azonexusv Hydrophilus (EF158391.1) 4 100% OTU002 Depth 0.67 + M QV1.1 26.5 Clostridia Clostridiales Syntrophomonadaceae CVCloAm2Ph135 (AM777954) 3 100% OTU006 Methane 0.60 + M CSW1.5 21.9 Clostridia Clostridiales Syntrophomonadaceae CVCloAm2Ph135 (AM777954) 3 98.2% OTU018 Formate 0.91 + V S CSW1.1 20.6 Clostridia Thermoanaerobacterales CVCloAm3Ph98 (AM778028) 3 99.1% Propionate 0.91 + V S Acetate 0.90 + V S Butyrate 0.88 + S Conductivity 0.64 + M Hydrogen 0.60 + M pH 0.60 + M ORP 0.85 - S CO 0.74 - S Correlation relationship is defined by correlation coefficient: > 0.9 = very strong (VS) , 0.7 - 0.9 = strong (S) , 0.5 - 0.7 = moderate (M) This is a subset of the data used to make the correlation network (Fig. 2.3) * Corr = correlation relationship Percent identity, as determined by MatGat ( 29 ) Closest relative references: 1 Suzuki et al , 2014; 2 Oosterkamp et al. , 20 11 ; 3 Tiago and Veríssimo, 2013; 4 Chou et al ., 2008 31 and QV1.2, as well as CSW1.1 and QV1.1. No Firmicutes OTUs were detected in the wells with pH less than 9 (Fig. 2.2 ). In addition to Betaproteobacteria and Firmicutes, the moderately high pH wells also contained Bacteroidetes as well as Alpha - , Delta - , and Ga mmaproteobacteria (Fig. 2.2 ). The circumneutral wells contained a greater complement of rare taxa and many taxa that were not prese nt in the high pH wells (Fig. 2.2) . The main taxa found within the highest pH samples, predominately Betaproteobacteria and C lostridia, are consistent with those found at other serpentinite sites (9, 11, 19). These bacterial 16S rRNA gene data suggest a core microbial community found within the serpentinite - influenced fluids, that shifts as the pH and serpentinite end - member wat er decreases. No archaea were detected in any of the 16S rRNA amplicon libraries, which were created with the universal primers targeting the V 4 region of the 16S rRNA gene used by the DOE Joint Genome Institute (22) . To further investigate the potenti al presence of archaea in CROMO f luids, the relative abundance of archaea w as also assessed by count ing archaeal sequences in the metagenomic datasets (description of metagenomic data below). The number of metagenomic sequences classified as archaea by MG - RAST (23) did not exceed 1% of the total seq uences in any sample (Table 2.4 ). Furthermore, none of these archaeal metagenomic reads included a 16S rRNA gene. These combined data indicate that archaea, if present, are not major contributors of the serpentin ite subsurface microbiome at CROMO, and it is therefore not surprising that archaea Table 2.4 Microbial communities at CROMO are dominated by bacteria . Relative abundance metagenomic sequence reads assigned to different domains of life via the M5NR data base in MG - RAST (Meyer et al , 2008) . Domain CSW1.1 QV1.1 CSW1.3 QV1.2 Bacteria 98.3 97.9 98.8 99.1 Archaea 0.2 1.1 0.7 0.2 Eukaryotes 1.2 0.8 0.3 0.5 Viruses 0.1 0.0 0.0 0.1 Unassigned 0.2 0.1 0.1 0.1 32 were not detected via amplicon sequencing with the universal primers. Biogeochemical Relationships One of the main goals of this study was to identify the geochemical drivers of microbial community composition within the serpentinite subsurface environment. A combination of pH, carbon monoxide, and CH 4 concentrations could explain 83% of bacterial community composition variability across wells (as determined by BEST in Primer - 6; 24). Therefore, we visualized pairwise Pearson correlations among these three envir onmental parameters and the relative abundances of all associated OTUs with a correlation network diagram (Fig. 2 .3 ). Figure 2 .3 Network diagram of significant correlations between OTUs and environmental variables i dentified in BEST analysis as accounting for 83% of community composition (p - value = 0.001) . OTU node size is relative to the maximum abundance of the OTU across the samples. Node color represents the - proteo - - proteobacteria (teal), Actinobacteria (yellow), and Bacteroidetes (pink). Nodes represent OTUs with a relative abundance > 1% of any sample, while OTUs making up > 10 % of any sample are labele d with OTU IDs . Positive and negative correlations are represented with black and blue lines, respectively. 33 pH was positively correlated with the two OTUs most abundant in CSW1.1 and negatively correlated with OTU008, classified as Azonexus hydrop hilus (Table 2.3) and the dominant betaproteobacterial OTU in wells with a pH below 10 (Fig. 2.3). The top OTUs from CSW1.1, OTU001 (classified as Comamonada ceae) and OTU018 (classified as Thermoanaerobacterales SRB - 2) , wer e negatively correlated with carb on monoxide concentration. Except for the betaproteobacterial OTU033, all other OTUs that positively correlate with carbon monoxide concentration belong to the alphaproteobacterial order Sphingomonadaceae (Fig. 2 .3 ). Among the OTUs positively correlated with the abundance of Figure 2.4 Heatmap of most abundant OTUs and geochemical parameters across all samples . Dendrogram at the top represents community similarity between samples and red lines indicate no statistical difference between field replicates, as determined by SIMPROF (38). 34 methane were B etaproteobacter ia OTU0 0 4 , most abundant in CSW1.2, and Clostridia OTU006 , most abundant in CSW1.5 (Table 2.3, Fig. 2.3). Additionally, five Gammaproteobacteria O TUs. The most abundant OTUs in CSW1.1, OTU001 (classified as Comamonadaceae) and OTU018 (classified as Thermoanaerobacterales SRB - 2 ), were also positively correlated with conductivity and the concentrations of organic acids and H 2 , and were neg atively cor related with ORP ( Table 2 .3 ). Two of the most abundant Clostridia OTUs (OTU002 and OTU003), both dominant in QV1.1, were significantly corr elated only with well depth ( Table 2 .3 ). While Table 2 .3 and the discussion above denote the sample in which each OTU was most abundant, it should be mentioned that many of those top OTUs were found in multiple samples, though at lower abundances ( Fig. 2 .4 ). Metabolic Potential To elucidate whether microbes within the serpentinite subsurface environment are capable of metabolizing the geochemical products of serpentinization, specifically hydrogen, methane, and carbon monoxide, assembled metagenomes from four of the wells were se arched for sequences predicted to encode proteins diagnostic of targeted metabolic processes. The protein targets were: [Fe - Fe] - hydrogenase and [Ni - Fe] - hydrogenase for hydrogen metabolism, ribulose - 1,5 - bisphosphate carboxylase/oxygenase ( RuBisCo ) for autot rophic CO 2 fixation via the Calvin - Benson - Bassham cycle, carbon monoxide dehydrogenase specific for aerobic (c oxL ) and anaerobic ( cooS ) carbon monoxide oxidation, particulate methane monooxygenase ( pmoA ) and methanol dehydrogenase ( mx aF ) for bacterial aero bic methane oxidation, and methyl coenzyme - M reductase ( mcrA ) for archaeal methanotrophy or methanogenesis. The numbers of matching proteins were normalized to total kilobases per metagenome and are reported as normalized 35 Figure 2.5 Abundance of protein - encoding genes in metagenomes . Numbers of matching proteins are normalized to Mb per metagenome. Proteins searched for include coxL (A), cooS (A), RuBisCo ( B ) , Fe - Fe hydrogenase ( C ), Ni - Fe hydrogenase ( C ), pmoA (D), mxaF (D) , and mcrA (D). coverage (Fig. 2.5 ). The only significant correlation between the abundance of protein targets and any of the geochemical parameters was a positive correlation between RuBisCo and the concentration of the organic acid butyrat e(R=0.99, p - value=0.0 02 ). OTUs with significant correlations to gene targets are summarized in Table 2.5 . The high pH of the serpentinite environment limits the availability of CO 2 , potentially leading to the use of carbon monoxide as an inorganic carbon source in carbon fixati on. To assess carbon monoxide metabolism at CROMO , two forms of carbon monoxide dehydrogenase were searched for in the metagenomes: coxL and cooS were used to identify aerobic and anaerobic carbon monoxide oxidation, 36 Table 2.5 Summary of significant correlations between gene hits and OTUs. Gene OTU Max Sample Max Abundance Phylum Class Order Family Genus cooS Otu007 QV1.1B 17.1 Firmicutes Clostridia pmoA Otu006 CSW1.3A 11.3 Firmicutes Clostridia Clostridiales Syntrophomonadaceae Dethiobacter pmoA Otu017 CSW1.3B 7.61 Firmicutes Erysipelotrichia Erysipelotrichales Erysipelotrichaceae Erysipelothrix pmoA Otu083 CSW1.3B 2.06 Firmicutes Erysipelotrichia Erysipelotrichales Erysipelotrichaceae Erysipelothrix pmoA Otu186 CSW1.3B 1.04 Proteobacteria Gammaproteobacteria Xanthomonadales Xanthomonadaceae Arenimonas Table 2.6 Bioenergetic calculations for CROMO groundwaters : total energy available (Joules of energy per kg of modeled CROMO groundwater) based on the limiting reactant for various metabolic reactions and overall Gibbs Energy (kJ per electron transferred) for the reaction. Metabolic Reaction Limiting Reactant Energy Available (Joules of energy/kg water) Gibbs Energy (kJ/electron transferred) QV1.2 CSW1.3 QV1.1 CSW1.1 QV1.2 CSW1.3 QV1.1 CSW1.1 Aerobic hydrogen oxidation H 2 (aq) + 0.5O 2 (aq) = H 2 O(l) H 2 1.6x10 - 2 6.0x10 - 2 1.6x10 - 2 6.1x10 - 2 - 100 - 110 - 100 - 110 Aerobic carbon monoxide oxidation CO(aq) + 0.5O 2 (aq) = CO 2 (aq) CO 3.6x10 - 2 3.0x10 - 2 4.0x10 - 2 2.5x10 - 2 - 120 - 130 - 140 - 140 Aerobic methane oxidation CH 4 (aq) + 2O 2 (aq) = CO 2 (aq) + 2H 2 O(l) O 2 3.7x10 - 1 7.7x10 - 1 3.9x10 - 1 6.6x10 - 1 - 97 - 100 - 100 - 100 Anaerobic oxidation of methane CH 4 (aq) +SO 4 2 - =HCO 3 - + HS - (aq) + H 2 O(l) SO 4 2 - 1.0x10 - 4 1.1x10 - 3 1.0x10 - 4 2.6x10 - 3 - 8.1 - 8.4 - 8.1 - 11 Hydrogenotrophic methanogenesis CO 2 (aq) + 4H 2 (aq) = CH 4 (aq) + 2H 2 O(l) H 2 4.0x10 - 4 1.6x10 - 3 1.0x10 - 4 2.0x10 - 4 - 4.5 - 5.0 - 1.6 - 2.4 Hydrogenotrophic methanogenesis CO(aq) + 3H 2 (aq) = CH 4 (aq) + H 2 O(l) H 2 1.5x10 - 3 7.1x10 - 3 1.7x10 - 3 7.3x10 - 3 - 13 - 15 - 14 - 15 Methanogenesis by carbon monoxide disproportionation 4CO(aq) + 2H 2 O(l) = CH 4 (aq) + 3CO 2 (aq) CO 6.7x10 - 3 6.7x10 - 3 1.0x10 - 2 6.6x10 - 3 - 33 - 41 - 50 - 52 37 respectively. Both forms of the gene were most abundant in the QV1.1A metagenome (CO = 0.142 µM; Table 2.1) with the anaerobic form of the gene having an abundance more than fifty - times that of the aerobic form (Fig. 3A). The coxL was detected in all the metagenomes, except for CSW1.1AC at low abundance. There is evidence of aerobic carbon monoxide oxidation at other sites of serpentinization, including the presence of coxL in the Tablelands metagenome (7) and the detection of carbon monoxide oxidation during microcosm experiments with fluids from the Tablelands (23 ) . While c oxL has been detected in one Serpentinomonas genome (strain A1; 16) , it was not seen in Serpentinomonas strain B1 (18), the strain that shares 100% sequence identity to OTU001 in this study. Sequences encoding the RuBisCo enzyme of the Calvin - Benson - Bassham cycle were detected in all four metagenomes, but it was most abundant in the CSW1.1 metagenome ( Fi g. 2.5B ). The RuBisCo gene has been detected in the serpentinite environment previously and showed sequence similarity to Betaproteobact eria ( 8, 19 ) . Furthermore, the genomes of all three Serpentinomonas strains contain the genes encoding for RuBisCo and other genes required for CO 2 fixation (18). Hydrogenase sequences were present in all four metagenomes, and the abundance of both hydroge nases was highest in the extreme - pH wells compared to moderate - pH wells (Fig. 2.5C). The [Fe - Fe] - hydrogenase gene was most abundant in the QV1.1 metagenome with a six - fold increase in gene hits compared to the CSW1.1 metagenome (Fig. 2.5C). An abundance of [Fe - Fe] - hydrogenases closely related to Clostridia was also detected in the Tablelands metagenome (8). The [Ni - Fe] - hydrogenase gene was more abundant in the CSW1.1 metagenome, and [Ni - Fe] - hydrogenase sequence s from Betaproteobacteria were present in the Tablelands metagenome ( 8) and identified in the Serpentinomonas genomes ( 18) . 38 We searched for pmoA and mxaF , the first and second proteins in the methane oxidation pathway of aerobic methanotrophs, within the CROMO metagenomes. Both pmoA and mxaF were det ected in the moderate pH wells, with a higher abundance of mxaF hits (Fig. 2.5 D), which is consistent with previous findings that pmoA is absent in some methanotrophic bacteria (26). Neither pmoA nor mxaF was detected in the CSW1.1 or QV1.1 metagenomes (F i g. 2.5 D). No mcrA sequences were detected in any of the metagenomes, which is not surprising given the low abundance of archaea in the metagenomes ( Table 2.4 ). The lack of genomic evidence for the presence of methanogens, as well as the energetic unfavorability of methanogenesis to occur in the fluids at CROMO, points to a non - in situ source of the methane at CROMO. This interpretation is consistent with methane isotopologue data that may suggest a higher - temperature (deeper) origin of methane at CROMO ( 27) . To assess the chemical energy available to support metabolic activity in the system, thermodynamic models were used to estimate the molar Gibbs energy c hange and potential total energy yield for several possible metabolic reactions in fluids from CSW1. 1, CSW1.3, QV1.1, and QV1.2 ( Table 2.6 ). The calculations focused in particular on chemolithoautotrophic metabolisms that might be supported by products of serpentinization. Consistent with metagenomic data, the calculations suggest that the most energetically favorable metabolisms across all wells are hydrogen oxidation, aerobic carbon monoxide oxidation, an d aerobic methane oxidation ( Table 2.6 ). Although anaerobic methanotrophy and methanogenesis are energetically favorable, the energy yields from these reactions may be too low to support metabolic activities (Table 2.6 ). The small amount of energy available from these reactions may explain the absence of archaea and the mcrA gene in our samples. 39 These metagenomic results indicate metabolic differences among the microbial communities at CROMO. CSW1.1 contained the greatest abund ance of RuBisCo (Fig. 2.5 B) and was dominated by Serpentinomas strain B1 (OTU0 01), whose genome contains a diversity of carbon utilizing genes. Given the limited carbon availability in a serpentinite environment, a diversity of strategies for obtaining carbon can allow for a more opportunistic life style. Furthermore, it has been pr oposed that the Betaproteobacteria within the serpentinite environment inhabit mixing zones between the end - member serpentinite fluids and oxic surface conditions (8, 20). Therefore, microorganisms in this niche may be able to access chemicals provided fro m meteoric input. Conversely, QV1.1 contained the most protein sequences associated with the oxidation of carbon monoxide and hydrogen (which are likely to be geochemical products of serpentinization) as well as the greatest abundance of Clostridia, organi sms thought to be inhabitants of end - member serpentinite fluids ( 8, 11, 20) . The metagenomes from the more moderate pH wells, CSW1.3 (pH 10.1) and QV1.2 (pH 7.9), contained more sequences associated with the metabolism of methane, which could have to do with the less challenging nature of these fluids and access to oxidants provided by the mixing with oxic surface waters. Conclusions Despite the abundance of potential energy sources (i.e. H 2 and CH 4 ) within the serpentinite environment, a lack of electron acceptors, as well as extreme pH and depletion of DIC, makes it a challenging habitat for life. While point - source studies have noted that the serpentinite environment is generally dominated by Betaproteobacteria and Firmicutes ( 9, 11, 19, 20), the capabi lity to sample across a gradient of geochemical parameters at CROMO allowed us to assess how these groups and others are distributed with respect to the measured environmental 40 conditions. The wells considered to be most representative of the serpentinite e nd - member fluids, based on pH, ORP, and DIC concentrations, exhibit extremely low diversity. Roughly 75% of the microbial community in CSW1.1 (pH 12.2) and QV1.1 (pH 11.5) were made up of two and four OTUs, respectively, all within the Betaproteobacteria a nd Clostridia (Fig. 2.2 ). Microbial communities within the lower pH wells were still dominated by Betaproteobacteria and C lostridia, but different OTU s within these classes. T he remarkable similarities in community composition , including 100% sequence iden tity of the 16S rRNA gene for many of the most abundant OTUs, between samples obtained from the cleanly drilled CROMO wells and the near - surface samples collected from the Cedars ( 11) and the Tablelands Ophiolite ( 9) support two main conclusions: 1) Clostr idia are dominant inhabitants of the serpentinite subsurface environment and 2) Betaproteobacteria live at oxic/anoxic interfaces associated with sites of serpentinization . The current study, using a comprehensive database of genomic and geochemical data, highlights the occurrence of Betaproteobacteria and Clostridia in the serpentinizing subsurface and the roles that high pH and associated parameters play in shaping the taxonomic and functional diversity of these communities (Fig. 2.3, Table 2 .3 ). Microbial communities in end - member wells are more likely to encode proteins associated with hydrogen metabolism, carbon monoxide oxidation, and carbon fixation compared to those in more moderate wells (Fig. 2.5 ). Meanwhile, methane was most strongly co rrelated with the community compositions of the moderate pH wells (Fig. 2.5 ). These data suggest that there are different ecological niches within the serpentinite subsurface environment, with hydrogen metabolism and carbon monoxide oxidation taking place in the most end - member fluids and methane oxidation occurring in mixing zones between serpentinite end - member fluids and surface waters. The findings of this study add 41 to a growing body of evidence that serpentinizing subsurface aquifers are dominated by c arbon - fixing, hydrogen - oxidizing Betaproteobacteria and anaerobic Clostridia. These data serve as a foundation for future studies investigating the activities of subsurface microbial populations and their influence on the flux of hydrogen and carbon from s ubsurface environments. Materials and Methods Site Description and Sample Collection Samples were collected from seven wells at the Coast Range Ophiolite Microbial Observatory (CROMO), located at the UC - Davis McLaughlin Nature Reserve in Lower Lake, CA, wh ich consists of three groups of wells located within a 1.4 km radius of each other: the Core Shed Wells (CSW), Quarry Valley (QV), and the N - wells. The CSW and QV wells were drilled using clean drilling techniques in 2011 to enable subsequent monitoring of the microbial communities and associated geochemistry within the serpentinite subsurface over time (1 0 ). CSW consists of five wells, drilled to depths of 9 - 27 meters. QV consists of three wells, drilled to depths of 15 - 23 meters. The N - wells, ranging in d epth from 14 - 40 meters, represent previously existing wells that were drilled in 1983 for environmental monitoring of mining impacts on groundwater, not with the specific purpose of monitoring microbiology and organic geochemistry (10). Well f luid s were c ollected using positive displacement Teflon bladder pumps ( Geotech Environmental Equipment, Denver, CO) and pumped through a YSI 3059 flow cell fitted with a YSI 556 multiprobe (Yellowsprings, OH), which measured water temperature, specific conductance, pH , dissolved oxygen (DO) and oxidation - reduction potential (ORP) once the DO measurement stabilized at a minimum value. The outlet of the YSI flow cell was fitted with silicon tubing, allowing for the collection of fluids for geoche mical and microbiological samples. 42 In the field, samples for dissolved gas analyses ( CH 4 , CO, and H 2 ) were extracted from the fluids and aqueous phase samples ( DIC and organic acid s) were collected anaerobically, as previously described ( 19) . Prior to fluid collection, the peristaltic pump tubing was flushed for three minutes with well water to remove any contaminants from previous wells. Fluids (ranging between 0.5 to 10 L, depending on the well) cartridge (Millipore, Billerica, MA) using an inline peristaltic pump for DNA analyses. Field r eplicate samples , ranging between two to eight filters per well, were collected. Sterivex cartridges were flash frozen with liquid nitrogen and stored at - until DNA extraction. For microbial cell quantification, replicate samples of 50 mL of fluids were preserved at a final concentration of publically available data generated from this project can be found at: http:/ /cromo.arc.nasa.gov. Geochemistry Dissolved gas es (H 2 , CH 4 , and CO) were extracted into an inert gas phase of known volume and analyzed for CH 4 via a SRI 8610C GC - FID and dissolved H 2 and carbon monoxide via a Trace Analytical RGA3 Reduced Gas Analyzer. D IC was measured by acidifying a known volume of well fluid within a sealed vial, and analyzing the concentration of liberated CO 2 in the headspace by GC - FID ( SRI 8610 . Organic acid samples were analyzed by HPLC wit h UV/VIS detection , following derivatization with 2 - nitrophenylhydrazide (34) . All sample vials were analyzed with duplicate injections. Microbial Cell Counts Fluids preserved for cell counts were filtered through a 0.2 µm filter. The cells were stained with 1 - diamidino - 2 - phenylindole (DAPI) and were counted by 43 epifluorescence microscopy using appropriate filter sets according to previously published protocols ( 35, 36) . DNA Extraction DNA extractions from sterivex filters were performed by lysis via freeze/thaw cycles and lysozyme/Proteinase K treatment and purified with phenol - chloroform extractions, precipitation in ethanol, and further purification with QiaAmp (Qiagen, Hilden, Germany) columns according to the manufacturer's instructions for purification of genomic DNA , as described previously in Brazelton et al ., 2013 (8) . 16S rRNA Tag Sequencing and Data Analysis Samples were submitted to the DOE Joint Genome Institute (JGI) for 16S rRNA amplicon sequencing of the V4 region on an Illumi nia M iSeq instrument, as described in Caporaso et al . (20). Sequence reads were aligned to the SILVA SSURef alignment (v119) , and taxonomic classification was assigned using mothur ( 27, 37) . Sequences were clustered into operational taxonomic units (OTUs) at the 3% distance threshold using the cluster. split command and the average - neighbor clustering algorithm in mothur ( 28). Beta diversity (between sample) of the microbial communities was assessed by calculation of the Bray - Curtis index and displayed in a dissimilarity dendrogram with statistical significance determined by the SIMPROF test run in Primer - 6 ( 22, 38) . Alpha diversity (within sample) was asse ssed by rarefaction analysis and the Simpson diversity index, which was calculated after subsampling the data to the sample with the fewest sequences (77,580) . Sequence identit y of reads belonging to the top OTUs compared with 16S rRNA sequences from other serpentinite studies (8, 9, 17) was performed using MatGAT with the default settings ( 29) . The 16S rRNA sequence data are 44 publ ically available in the NCBI S equence Read Archive under the accession number SRA280854 . Metagenomic Sequencing and Data Analysis Samples were submitted to JGI for metagenomic sequencing on an Illumina HiSeq2000 instrument, as described in Hawley et al . (26) . These data are publically available in the JGI IMG/M database ( www.img.jgi.doe.gov ) under the project IDs: 1021918 , 1021921 , 1021924 , and 1021927 ; and in the MG - RAST database ( www.metagenomics.anl.gov ) under the following sample IDs: 4569549.3, 4569550.3, 4569551.3, and 4569552.3. Metagenomes were searched for protein - encoding genes indicative of metabolic pathways of interest using similar methods as those described by Brazelton et al ( 7 ). Representative protein sequences of the genes - of - interest were used as queries in BLASTP ( 30) searches against the assembled metagenomes. Unique metagenomic sequence reads that matched at least on e query were quantified as reads per kilobase and normalized to total kilobases in each metagenome. The queries for each BLASTP search were representative ('seed') protein sequences obtained from Pfam ( 31) for the large subunits of [Fe - Fe] hydrogenase (PF02906) and [Ni - Fe] hydrogenase (PF00374), particulate methane monooxygenase pmoA (PF02461), methanol dehydrogenase mxaF ( PF01011), methyl co - enzyme - M reductase A mcrA (PF02249), and the large subunit of the RuBisCo gene (PF00016) . The nickel - dependent carbon monoxide dehydrogenase gene, cooS (TIGR01702) , was obtained from the TIGRFAM database (39) . Representative sequences for the large subunit of carbon monoxide dehydrogenase, coxL , were obtained from Cunli ffe et al . (40) . The assembled metagenomes were annotated with Prokka ( 41), and these annotations were compared with the results of the BLASTP searches described above. Specifically, any BLASTP hits with annotations that did not match the protein target we re considered to be false positives. None of 45 the BLASTP searches had false positive rates greater than 20% , except for coxL . Therefore, the normalized coverages for coxL reported in Figure 2.5 were obtained by searching for annotations matching UniProt ID P19913 , corresponding to coxL in Hydrogenophaga pseudoflava . Th e coverage (as reads per kilob ase) for the contigs associated with these annotations were summed and normalized to total kilobases in each metagenome. Statistical Analyses C orrelation network a nalyses were constructed from statistically significant pairwise Pearson correlations among environmental variables and sequence data ( 32) and visualized in Cytoscape v 2.8.3 ( 33) . A matrix containing environmental data and relative OTU (97% similarity) ab undance for each sample was used as input for pairwise Pearson correlation analysis computed with the rcor.test function in the R package lmt ( 42) . The false - discovery rate (q - value) was computed f or the distribution of Pearson p - values to account for mult iple tests. Pairwise correlation s with both p - and q - values of <0.05 were considered significant and included in network analyses. Network models of significant correlations were created using Cytoscape v2.8.3 ( 33). The ANOSIM test using a Bray - Curtis similarity index was used to test whether individual environmental parameter categories had significant effects on the community composition of samples ( 22) . To statistically determine which combinations of numerical env ironmental variables best described the community composition variation within the dataset, the BEST analysis was pe r formed in PRIMER - 6 ( 22, 38) . Bioenergetic Calculations r ) calculations identify which reactions are thermodynamically favor able (i.e., exergonic) , and therefore good candidates to supply energy for microbial 46 r ), which is the sum of the standard Gibbs r º) and a term that reflects the chem ical composition of the r is taken into account; interstitial fluid chemistries differ across the four modeled wells r also differs. Measured concentrations of interstitial fluids are transformed into in situ activities, which are used directly r , using the thermodynamic modeling package of EQ3/6 ( 43) . r can be calculated at in situ conditions using the expression ( Equation 2. 1): r r º + RT lnQ r ( Eq. 2. 1) r r º is the standard Gibbs energy, R and T represent the gas constant and temperature in Kelvin, respectively, and Q r stands for the activity product, r º can be determined at the appropriate temperature and pressure for the aqueous species and minerals can be calculated using established equations of state ( 44 and references therein ). Q r , the activity product, can be computed from environmental data as shown in ( Equation 2. 2 ): Q r i i,r ( Eq. 2. 2 ) where a i represents the activity of the i i , r represents the stoichiometric reaction coefficient. Values of a i are generated from concentration data (Table 2. 1) and activity coefficients, using the geochemical speciation code EQ3 ( 43) . In this code, activity coefficients are calculated using a variant (B - dot equation) of the Extended Debye - Hückel activity coefficient formalism ( 45) , with reference to the SUPCRT92 ( 46) thermody namic database. For comparative purposes, values of r are standardized to the amount of energy per mole of electron transferred (44). As an estimate of the amount of metabolic energy available in 47 the environment at the time of sampling, the total pote ntial energy yield for each reaction was calculated by multiplying r by the concentrations of the limiting reactant in the reaction (see 46, 47) . The result is an estimate of the amount of metabolic energy that was available from the reaction at the time of sampling if all of the limiting reactant was consumed in the specified reaction. 48 REFERENCES 49 REFERENCES 1. Whitman WB, Coleman DC, Weibe WJ (1998) Prokaryotes: the unseen majority. Proc Natl Acad Sci USA 95:6578 - 6583. 2 microbial abundance and biomass in subseafloor sediments. Proc Natl Acad Sci USA 109:16213 - 16216. 3 . Northrup DE, Lavoie KH (2001) Geomicrobiology of caves: a review . Geomicrobiol J 18:199 - 222. 4 . Engel AS, Stern LA, Bennett PC (2004) Microbial contribution to cave formation: new insights into sulfuric acid speleogenesis. Geology 32:369 - 372. 5 . Onstott TC, et al. (2003) Indigenous and contaminant microbes in ultrade ep mines. Environ Microbiol 5:1168 - 1191. 6 . Chivian D, et al. (2008) Environmental genomics reveals a single - species ecosystem deep within Earth. Science 322:275 - 278. 7 . Brazelton WJ, Nelson B, Schrenk MO (2012) Metagenomic evidence for H 2 oxidation and H 2 production by serpentinite - hosted subsurface microbial communities. Front Microbiol 2:268. 8 . Brazelton WJ, Morrill PL, Szponar N, Schrenk MO (2013) Bacterial communities associated with subsurface geochemical processes in continental serpentinite spr ings. Appl Environ Microb 79:3906 - 3916. 9 . Suzuki S, et al. (2013) Microbial diversity in The Cedars, an ultrabasic, ultrareducing, and low salinity serpentinizing ecosystem. Proc Natl Acad Sci USA 110:15336 - 15341. 1 0 . Cardace D, et al. (2013) Establishment of the Coast Range ophiolite microbial observatory (CROMO): drilling objectives and preliminary outcomes. Scientific Drilling 16:45 - 55. 1 1 . McCollom TM, Seawald JS (2007) Abiotic synthesis of organic compounds in d eep - sea hydrothermal environments. Chem Rev 107:382 - 401. 1 2 . Proskurowski G, et al. (2008) Abiogenic hydrocarbon production at Lost City Hydrothermal Field. Science 319:604 - 607. 1 3 . Kelley DS, et al. (2005) A serpentinite - hosted ecosystem: the Lost City Hydrothermal Field. Science 307:1428 - 1434. 50 1 4 . Schrenk MO, Kelley DS, Bolton SA, Baross JA (2004) Low archaeal diversity linked to subseafloor geochemical processes at the Lost City Hydrothermal Field, Mid - Atlantic Ridge. Environ Microbiol 6:1086 - 1095. 1 5 . Brazelton WJ, Schrenk MO, Kelley DS, Baross JA (2006) Methane - and sulfur - metabolizing microbial communities dominate the Lost City hydrothermal field ecosystem. Appl Environ Microbiol 72:6257 - 6270. 1 6 . Suzuki S, et al. (2014) Physiological and genomic features of highly alkaliphilic hydrogen - utilizing Betaproteobacteria from a continental serpentinizing site. Nat Commun 5:3900. 1 7 . Tiago I, Veríssimo A (2013) Microbial and functional diversity of a subterrestrial high pH groundwater associated to serp entinization. Environ Microbiol 15:1687 - 1706. 18 . Schrenk MO, Brazelton WJ, Lang SQ (2013) Serpentinization, carbon and deep life. Rev Mineral Geochem 75:575 - 606. 19 . Crespo - Medina M, et al. (2014) Insights into environmental controls on microbial commun ities in a continental serpentinite aquifer using a microcosm - based approach. Front Microbiol 5:604. 2 0 . Caporaso JG, et al. (2011) Ultra - high - throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. ISME J 6:1621 - 1624. 2 1 . Meyer F, et al. (2008) The metagenomics RAST server: a public resource for the automatic phylogenetic and functional analysis of metagenomes. BMC Bioinformatics 9:386. 2 2 . Clarke KR, Gorley RN (2006) PRIMER v6: User Manual/Tutorial. PRIMER - E, Plymouth. 2 3 . Morrill PL, et al. (2014) Investigations of potential microbial methanogenic and carbon monoxide utilization pathways in ultra - basic reducing springs associated with present - day continental serpentinization: the Tablelands, NL, CAN. Front Microbiol 5: 613. 2 4 . Lau E, Fisher MC, Steudler PA, Cavanaugh CM. (2013) The methanol dehydrogenase gene, mxaF , as a functional and phylogenetic marker for proteobacterial methanotrophs in natural environments. PLoS ONE 8:e56993. 2 5 . Wang DT, et al. (2015) Nonequilibrium clumped isotope signals in microbial methane. Science 348:428 - 431. 2 6 . Hawley, ER, et al. (2014) Metagenomic analysis of microbial consortium from nature crude oil that seeps into the marine ecosystem offshore South ern California. Stand Genomic Sci 9:1259 - 1274. 51 2 7 . Pruesse E, et al. (2007) SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Res 35:7188 - 7196. 28 . Schloss PD, Westcott S L (2011) Assessing and improving methods used in operational taxonomic unit - based approaches for 16S rRNA gene sequence analysis. Appl Environ Microb 77:3219 - 3226. 29 . Campanella JJ, Bitincka L, Smalley J (2003) MatGAT: an application that generates simil arity/identity matrices using protein or DNA sequences. BMC Bioinformatics 4:29. 3 0 . Altschult SF, et al. (1997) Gapped BLAST and PSI - BLAST: a new generation of protein database search programs. Nucleic Acids Res 25:3389 - 3402. 3 1 . Finn RD, et al. (2008) The Pfam protein families database. Nucleic Acids Res 36:D281 - D288. 3 2 . Fuhrman J, Steele J (2008) Community structure of marine bacterioplankton: patterns, networks, and relationships to function. Aquat Microb Ecol 53:69 - 81. 3 3 . Shannon P, et al. (2003) Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res 13:2498 - 2504. 34. Albert D, Martens C (1997) Determination of low - molecular - weight organic acid concentrations in seawater and pore - water samples via HPLC. Mar Chem 56:27 - 37. 35 . Hobbie JE, Daley RJ, Jasper S (1977) Use of nucleopore filters for counting bacteria by fluorescence microscopy. Appl Environ Microb 33:1225 - 1228. 36 . Schrenk MO, Kelley DS, Delaney JR, Baross JA (2003) Incidence and diversity of microorganisms within the walls of an active deep - sea sulfide chimney. Appl Environm Microbiol 69:3580 - 3592. 37 . Schloss PD, et al. (2009) Introducing mothur: open - source, platform - independent, community - supported softwar e for describing and comparing microbial communities. Appl Environ Microb 75:7537 - 7541. 38 . Clarke KR (1993) Non - parametric multivariate analyses of changes in community structure. Aust J Ecol 18:117 - 143. 39 . Haft HH, et al. (2001) TIGRFAMs: a protein f amily resource for the functional identification of proteins. Nucleic Acids Res 29:41 - 43. 40 . Cunliffe M, et al. (2008) Phylogenetic and functional gene analysis of the bacterial and archaeal communities associated with the surface microlayer of an estuar y. ISME J 2:776 - 789. 52 41 . Seeman, T (2014) Prokka: rapid prokaryotic genome annotation. Bioinformatics 14:2069 - 2069. 42 . Rizopoulos D (2006) ltm: an R package for latent variable modeling. J Stat Softw 17:5. 43 . Wolery TJ (1992) EQ6: EQ3NR, A Computer Program for Geochemical Aqueous Speciation - Documentation (Version 7.0) Lawrence Livermore National Lab. UCRL - MA - 110662 PT III 44 . Amend JP, Shock EL (2001) Energetics of overall meta bolic reactions of thermophilic and hyperthermophilic Archaea and Bacteria. FEMS Microbiol Rev 25 :175 - 243. 45 . Helgeson HC (1969) Thermodynamics of hydrothermal systems at elevated temperatures and pressures. Am J Sci 267 :729 804. 46 . Johnson JW, Oelkers EH, Helgeson HC (1992) SUPCRT92: A software package for calculating the standard molal thermodynamic properties of minerals, gases, aqueous species, and reactions from 1 to 5000 bar and 0 to 1000C. Computers & Geosciences 18:899 - 947. 47 . McCollom TM, Shock EL (1997) Geochemical constraints on chemolithoautotrophic metabolism by microorganisms in seafloor hydrothermal systems. Geochim Cosmochim Ac 61 :4375 - 4391. 48. Oosterkamp MJ, et al. (2011) Genome sequences of Alicycliphilus denitrificans strains BC and K601. J Bacteriol . 193:5028 - 5029. 49. Chou JH , et al . (2008 ) Azonexus hydrophilus sp. nov., a nifH gene - harboring bacterium isolated from freshwater. Int J Syst Evol Micr 58: 946 - 951. 53 CHAPTER 3 Microbial d iversity of s erpentinite c ores from the Coast Range Ophiolite Microbial Observatory 2 Abstract Serpentinization is a geochemical pr ocess in which ultramafic rocks are transformed to serpentine minerals and magnetite in the presence of water. Th e reactions create high pH fluids with an abundance of hydrogen and methane that can serve as potential energy sources for microbial communities. The subsurface environment is an amalgamation of rock matrix and the groundwaters permeating it, creating a complex habitat for microorganisms, which can be free - living in fluids or particle - attached in biofilms. Fluids from serpentinites have consistently contain ed two dominant taxa of bacteria, Betaproteobacteria and Clostridia. Little is known, h owever, about microbe - mineral interactions within the serpentinite subsurface environment. The Coast Range Ophiolite Microbial Observatory (CROMO), a set of well s drilled into an actively serpentinizing ophiolite, was established with the aim of characterizing the habitability of the serpentinite environment and from it cores were obtained . This study represents the first look at the microbial communities associat ed with cores from the serpentinite subsurface environment . Different core - enriched taxa were correlated with distinct lithostratigraphic zones within the heterogeneous cores, suggesting that mineralogy may impact community composition. Archaea, previously undetected in CROMO fluid samples, were abundant in core samples 2 The work described i n this chapter is being developed for submission to Applied and Environmental Microbiology : K.I. Twing, W.J . Brazelton , M.D. Kubo, M. Tominga, D . Cardace , T.M. Hoehler, T.M. McCollom, and M.O. Schrenk. ( In Prepara tion ) Microbial diversity of serpentinite cores from the Coast Range Ophiolite Microbial Observatory . Appl Environ Microbiol . 54 containing magnetite - bearing serpentinite , while bacteria were more abundant in samples containing clay particles. The taxonomy of sequences significantly enriched in core samples suggest th at serpentinite cores contain an abundance of bacteria and archaea involved in the cycling of nitrogen and methane , indicating that a more complex biogeochemical cycling may be taking place in the serpentinite subsurface environment than can be detected by investigating fluid samples alone. Introduction Serpentinization is the hydrous alteration of ultramafic rocks from the seafloor into magnetite and serpentine minerals (e.g. lizardite, chrysotiles, and antigorite; 14 ). The series of reactions creates h igh pH fluids, containing an abundance of hydrogen and methane. The precise mineral products of the reaction depend on a number of factors, including temperature, pressure, initial rock and mineral composition, and water to rock ratio (24). While this proc ess begins on the subseafloor , sites of serpentinizatio n can be accessed on land where the mantle has been obducted onto the continental crust and serpentinization continues , making it more accessible for research. Fluids from continental serpentinite site s from around the world have yielded surprisingly consistent microbial community composition, containing low diversity and a dominance of Betaproteobacteria and Clostridia ( 3, 4, 10, 41, 44, 45, 46). The combination of interdisciplinary studies, integrating genetic and environmental data (4, 43 ), metagenomics (3), and cultured isolates (10, 43 ), has allowe d for the development of models suggesting that anaerobic Clostridia are truly endemic to the extreme serpentinite subsurface environment, while Betaproteobacteria reside in the more hospitable oxic/anoxic mixing zones ( 41 ). 55 The serpentinite subsurface environment is a rock - hosted system. Ho wever, the majority of samples collected from these systems to - date have been fluid samples. Chimney samples from the Lost City Hydrothermal Field were dominated by methane - cycling archaea of the Methan osarcinales in anoxic chimney interiors (2, 40 ) and sulfur - cycling and methanotrophic bacteria at the exterior of the chimneys (1). In the continental setting, the only serpentine minerals samples examined for microbiology are from the Leka ophiolite in Norway , where minerals from fractures, ranging from 15 - 160 cm deep were studied ( 11 ). Clone libraries of mineral sample s from 155 - 160 cm were dominated by Actinobacteria, Acidobacteria, Firmicutes, Alpha - and Betaproteobacteria, and archaea from the Soil Crenarcheotic Group ( 11 ). Serpen tine soils are characterized by high levels of heavy metals (Ni, Cr, and Co) and low nutrients, which leads them to host very specialized flora ( 48 ). Due to this unique macro - ecology, many of the studies regarding the microbiology of serpentine soils have focused on the rhizosphere (microbes associated with plant roots) and heavy metal tolerant microorganisms (19 , 27, 30, 31, 34 ). The bacterial communities defined in these sources include: Acidobacteria, Actinobacteria, Bacteroidetes, Firmicutes, Gemmatimonadetes, Proteobacteria (Alpha - , Beta - , and Gamma - ), and Ver rucomicrobia. There have not been reports of archaea in serpentine soil. The serpentinite subsurface environment is a complex system, the microbiology of which is just beginning to be understood. The Coast Range Ophiolite Microbial Observatory (CROMO) was established the help address the question of microbe - mineral interactions in the serpentinite environment (7) . This paper represents the microbial investigation, through the use of high - throughput 16S rRNA sequencing, of cores obtained from the serpentini te subsurface to a depth of 45 m and suggests that there are unique communities associated with solid surfaces in the serpentinite subsurface environment. 56 Results and Discussion Drill C ore D epth P rofiles Recovered cores indicate difference s in lithology between the CSW1.1 and QV1.1 (Figs. 3.1 & 3.2), which may result in the variation in core recovery: 46% for CSW1.1 and 64% for QV1.1 (7). At CSW1.1, detrital serpentine soil was observed for 0 - 4 m, followed by 20 m of poor core recovery with a few sections of serpentine gravel in clay (Fig. 3.1A); then, below 24 mbs (meters below surface) , magnetite, serpentine, and altered mafic rocks were found (Fig. Figure 3. 1 Depth profile of CSW1.1 . (A) Lithostratigraphic units, as described in Cardace et al ., 2013; (B) Resistivity data from down - hole logging; (C) DNA yield from DNA extractions, as quantified by fluorometric methods; (D) Bacterial 16S rRNA gene copies, as determined by domain - specific quantitative - PCR; (E) Percent organic carbon and percent ni trogen, as determined by EA - IRMS . N/A: sections with no core recovery. Filled symbols represent samples that yielded 16S rRNA sequences, samples represented by closed symbols failed sequencing. 57 3.1A). In contrast, QV1.1 only contained detrital serp entine soil for the first meter, after which there were various layers of magnetite, serpentine, and chlorite throughout the depth of the core (Fig. 3.2A). QV1.1 contained regions with no serpentine, such as at 31 mbs where magnetite - bearing clay with albi te was found and 34 - 36 mbs where rocks containing albite, quartz, and chlorite were seen (Fig. 3.2A). Core recovery was higher in QV1.1 than in CSW1.1, with only two large sections between 2 - 5 mbs and 11 - 14 mbs where there was no core recovery (Fig. Fig ure 3. 2 Depth profile of QV1.1 . (A) Lithostratigraphic units, as described in Cardace et al ., 2013; (B) Resistivity data from down - hole logging; (C) DNA yield from DNA extractions, as quantified by fluorometric methods; (D) 16S rRNA gene copies, as deter mined by domain - specific quantitative - PCR; (E) Percent organic carbon and percent nitrogen, as determined by EA - IRMS . N/A: sections with no core recovery. Filled symbols represent samples that yielded 16S rRNA sequences, samples represented by closed symbols failed sequencing. 58 3.2A). Resistivity logging was conducte d as part of the post - drilling, down - hole geophysical logging (Figs. 3. 1B & 3. 2B; 7). Due to the instrument limitations, the boreholes were only matrix is comp rised of electrically resistive mineralogy (e.g. clay, sulfides) and/or more lithified, logged to 18 m (Figs. 3. 1B & 3.2B). A broad peak of relatively high resistivity in CSW1.1 at the depth range of 5 - 15 m (Fig. 3.1B) coincides with the low core recovery in this zone (Fig. 3.1A). In QV1.1, there are two sharp spikes in resistivity at 2 and 15 m (Fig. 3.2B), which coincide with magnetite - bearing serpentine (Fig. 3.2A). High resistivity measurements indicate that the rock l ess porous formations. Other than t he above - mentioned sections, most of the logged interval of both cores is very low resistivity, possibly suggesting where water resides in pore spaces, cracks, and faults within the wells (Figs. 3.1B & 3.2B). DNA Y ield , O rganic C arbon, and 16S rRNA G ene A b undance In order to conserve DNA for downstream analyses, only 2 µL of the 50 µL DNA extract was sacrificed for DNA quantification, making the limit of detection of the fluorometric quantification method 0.1 ng/µL. DNA was extracted from roughly 20 g of c ore material for each sample, making the limit of detection for DNA quantification 0.2 ng/g, which many of the samples fell below (Fig. 3.1C & 3.2C). It should be noted, however, that despite unquantifiable DNA concentrations, many of the samples still wer e amplifiable via domain - specific q - PCR (Figs. 3.1D & 3.2D) and yielded 16S rRNA sequences (Table 3.1). Therefore, amplifiability, as opposed to DNA concentration was used to determine whether a sample was fit for submission to the sequencing facility. Due to the potential of contamination in very low biomass samples, blank control samples were run at every step of the DNA extraction and purification process. The 59 control samples were n either quantifiable by fluorometric methods n or amplifiable via domain - specific q - PCR, and therefore were considered to be negative of any contamination. The total organic carbon (TOC) and total nitrogen (TN) were measured for each core sample by elemental - analysis isotope - ratio mass spectrometry (EA - IRMS; Figs. 3.1E & 3.2E). Other than the top 4 m of CSW1.1, which contained more organic carbon than nitrogen and had a C/N ratio representative of decaying organic matter (11:1), the two elements were below detection throughout the core, except for an increase of TN betw een 19 - 26 mbs (Fig. 3.1E), which is of unknown origin . This high C/N spike at ~20 m depth could indicate fresh biological organic matter, which has yet to be consumed by microorganisms. In QV1.1, the TN remained Table 3.1 Soil and core sample names and 16S rRNA gene sequence abundance s Sample Depth (mbs) DNA Yield (ng/g) Bacterial 16S Sequences* Archaeal 16S Sequences* CSWsoil1 0 67.0 42,570 151,538 CSWsoil2 0 82.1 82,363 236,134 CSW1R 2.4 138.1 246,412 89,017 CSW2R 4.2 2.4 147,623 105,289 CSW16R 15.7 0.4 18 0 CSW17R 17.4 0.3 16 0 CSW10R 19.1 < 0.2 285,703 12,020 CSW 14R 25.6 < 0.2 225,988 0 CSW22R 28.5 < 0.2 9 0 CSW23R 28.9 < 0.2 9 0 CSW24R 29.5 < 0.2 11 0 QVsoil1 0 66.4 44,333 193,560 QVsoil2 0 106.0 83,590 242,101 QV3R 0.3 < 0.2 1,919 2,806 QV7R 6.5 0.2 78,108 4,683 QV11R 9.7 < 0.2 4,635 266 QV13R 15.7 < 0.2 4,568 1,440 QV18R 20.8 < 0.2 9,225 504 QV21R 22.2 < 0.2 10,827 998 QV25R 26.1 < 0.2 8,495 2,486 QV30R 30.8 1.4 25,322 103,561 QV42R 43.3 < 0.2 11,414 130,842 *Sequence counts in bold were successfully sequenced, while those in regular font failed sequencing (i.e. produced < 5000 sequence reads). 60 low below the soil level, while the TOC spiked between 25 - 30 mbs (Fig. 3.2E). The increased TOC with low TN could suggest abiotic input of carbon molecules (as from serpentinization) or older organic matter that has already been processed by microorganisms in the system. There does not appear to be a consistent relationship between DNA concentration, 16S rRNA gene copies (obtained from q - PCR), and successful 16S rRNA sequencing (Figs. 3.1 & 3.2; Table 3.1). The spike in TN (~20 mbs) in CSW1.1 corresponds to the only samples that yielded 16S rRNA sequences at depth from that core, even though samples above and below that were amplified via q - PCR. The TOC spike in QV1.1 coincides with an increase in DNA concentrations and 16S rRNA genes (Fig. 3.2). The resisti vity data (Figs. 3.1B & 3.2B), a proxy for how hard the rock material is, appears to correspond with whether quantifiable, amplifiable, and sequenceable DNA could be obtained from a core sample. This could be due to poor core recovery from highly resistive regions or due to fluid flow in cond uctive zones making them more habitable. Ongoing efforts are working to resolve the complex hydrogeology of the CROMO subsurface, which will help address this hypothesis more fully. Core - E nriched B acteri al C ommunities When dealing with natural samples, it is important to keep in mind that different scales. For instance, this study is focused on describing the core - enriched communi ty of microorganisms in the serpentinite subsurface environment and contains sequences from core - , soil - , and fluid - associated microbes. It would be naïve to think there is not natural flow of microbes between the fluids within the fractures of the rocks a nd vice versa. However, that is (e.g. 37) they rely on removing any sequence that occurs in both the sample (in this case core) and 61 background (in this case fluids and s oils), assuming that the sequence must have come from the background and contaminated the true sample. Instead of using this overly conservative method, which potentially removes true inhabitants of the sample, this study used a statistical method modeled after differential expression in RNA - sequencing studies ( 26 ) to identify which sequences are significantly enriched in a particular type of sample, and therefore more likely to have originated there. In this fashion, core samples were compared to fluid and soil samples to - - - (sequences significantly more abundant in that sample type compared to the others) , as opposed to the no statistically defined habit at (Fig. 3.3) . Figure 3.3 DESeq c omparison between CSW core and CROMO fluid samples. Core samples are depicted in red, fluid samples are depicted in blue. (A) Relative abundance of sequences found in individual core and fluid samples; (B) Sequences enriched in core and fluid samples versus those not statistically enriched in either group (No Pattern); (C) Plot of log fold change of each sequence between the two sample types over the total count (abundance) for each sequence across the data set. 62 The fluid - enriched taxa, which make up 20% of QV1.1 and 50% of CSW1.1 fluid communities (Fig. 3. 4), are made up exclusively of Betaproteobacteria and Clostridia. The identification and abundances of these taxa are consistent with the findings in Chapter 2 ( 46 ), supporting the theory that Betaproteobacteria and Clostridia dominate fluids in the subsurface serpentinite environment (41). The remaining 80 and 50% of the communities, respectively, - statistically significant increase in those taxa in the fluids compared to the soil or core samples. - Acidobacteria, Actinobacteria, Bacter oidetes, Chloroflexi, Gemmatimonadetes, Planctomycetes, Figure 3. 4 Relative abundance of core - , soil - , and fluid - enriched bacterial sequences . 63 Protobacteria, and Verrucomicrobia and made up between 1 - 2% of the soil communities (Fig. 3. 4). The soil - enriched taxa found here are consistent with previous reports of serpentine soil and rhizosphere microbial communities (19, 27, 31 ), with the exception of the Chloroflexi and Planctomyce tes, which were not seen in previous studies. The membe rs of the Chloroflexi identified as soil - enriched belong to the anaerobic, non - phototrophic, filamentous classes ( 16 ). The soil - enriched Planctomycetes sequences belonged to the order Planctomycetales, which contains deeply - branching bacteria capable of an aerobic ammonium oxidation (ANAMMOX; 13) and they have been isolated from a variety of soil environments (6), though not serpentine soils. The remaining 99% of the soil sequences were not significantly enriched in the soils, compared to the cores and fluids . Figure 3. 5 Bacterial community structure of unique sequences . (A) Rarefaction analysis of unique 16S rRNA amplicon sequences. (B) Community simila rity dendrogram calculated from M orisita - Horn dissimilarity index. 64 Table 3.2 - Relative abundance of individual core - enriched bacterial taxa within samples . Phylum Class CSW1R CSW2R CSW10R CSW14R QV7R QV18R QV21R QV25R QV30R QV42R Actinobacteria Actinobacteria 0 0.003 0.112 0.163 0.201 3.39 5.43 2.92 0.019 7.97 Proteobacteria Deltaproteobacteria 2.52 0.532 1.27 1.81 0 0 0 0 0 0 Acidobacteria Acidobacteria 2.35 0.128 1.33 1.51 0 0 0 0 0 0 Firmicutes Bacilli 0 0.003 0.215 0.157 0.024 0.367 3.78 0.024 0.008 0.228 Nitrospirae Nitrospira 1.43 1.26 0.653 0.647 0 0 0 0 0 0 Proteobacteria Betaproteobacteria 1.37 0.062 0.715 0.833 0 0 0 0 0 0 Proteobacteria Gammaproteobacteria 1.39 0.008 0.633 0.799 0 0 0 0 0 0 Acidobacteria Subgroup_22 0.829 0.169 0.184 0.320 0 0 0 0 0 0 Spirochaetae Spirochaetes 0.504 0.033 0.268 0.363 0 0 0 0 0 0 Chlorobi Ignavibacteria 0.425 0.025 0.422 0.251 0 0 0 0 0 0 Proteobacteria Alphaproteobacteria 0.391 0.033 0.136 0.251 0 0 0 0 0 0 WS3 0.313 0.055 0.198 0.205 0 0 0 0 0 0 Chloroflexi Anaerolineae 0.399 0.083 0.080 0.199 0 0 0 0 0 0 Gemmatimonadetes Gemmatimonadetes 0.226 0.055 0.162 0.217 0 0 0 0 0 0 Chloroflexi Ktedonobacteria 0.221 0.002 0.142 0.131 0 0 0 0 0 0 Acidobacteria Holophagae 0.069 0.083 0.282 0.036 0 0 0 0 0 0 OD1 0.177 0 0.054 0.108 0 0 0 0 0 0 Planctomycetes Phycisphaerae 0.171 0.005 0.091 0.066 0 0 0 0 0 0 Chloroflexi Dehalococcoidia 0.101 0.010 0.111 0.069 0 0 0 0 0 0 Chloroflexi 0.074 0 0.059 0.074 0 0 0 0 0 0 TM6 0.079 0.001 0.097 0.019 0 0 0 0 0 0 Planctomycetes Pla4_lineage 0.101 0.001 0.038 0.029 0 0 0 0 0 0 Chloroflexi P2 - 11E 0.058 0.0007 0.061 0.042 0 0 0 0 0 0 Chloroflexi JG30 - KF - CM66 0.032 0 0.028 0.099 0 0 0 0 0 0 Chloroflexi Thermomicrobia 0.030 0.0007 0.034 0.042 0 0 0 0 0 0 Chloroflexi TK10 0.024 0.0007 0.019 0.021 0 0 0 0 0 0 Thermotogae Thermotogae 0.024 0.001 0.031 0.008 0 0 0 0 0 0 Chloroflexi S085 0.008 0.010 0.010 0.016 0 0 0 0 0 0 Unclassified Bacteria 0.240 0.078 0.173 0.175 0.068 0.033 0.037 0.059 0 1.15 TOTAL 13.6 2.64 7.60 8.66 0.293 3.79 9.25 3.01 0.028 9.35 65 - - 14% of the bacterial communities and no soil - enriched taxa were detected (Fig. 3. 4). The QV core samples, on the other hand, exhibited different community patterns. QV7R (4.2 mbs) a nd QV30R (30.8 mbs) contained mostly soil - enriched sequences, while the deepest sample, QV42R contained 10% core - enriched sequences and no soil - enriched sequences. The core samples from the middle of QV (20 - 26 mbs) contained mostly core - enriched sequences with ~1% soil - enriched sequences. These results are consistent with the results depicted in the community similarity dendrogram, suggesting that there are differences in the communities in CSW cores and QV cores and that QV7R and QV30R are more similar to soil samples than other core samples (Fig. 3. 5). A list of the core - enriched taxa can be found in Table 3. 2. Among them are Chloroflexi and OD - 1, which are suggested to be representative of the serpentinite subsurface environment at The Cedars ( 44 ), but not detected in CROMO fluids (Chapter 2; 46 ). The CSW core samples contained an abundance of nitrogen cycling core - enriched ta xa, including nitrite - oxidizing Nitrospira ( 12 ) and Nirtrospinaceae (23) and a mmonium - oxidizing Nitrosomonadaceae ( 32 ). CSW cores also contained iron - oxidizing Betaproteobacteria of the family Gallionellaceae ( 15 ). The most abundant core - enriched sequence in QV cores belonged to the Mycobacteriaceae , a taxon that has also been detected in deep subsurface samples from the Homestake Go ld Mine in South Dakota ( 35 ) and chromium contaminated sites ( 20 ). The only sequences found previously in serpentine mineral samples were from clone libraries (11) and determined to be core - enriched in this study are H 2 - oxidizing Actinobacteria ( 42 ) of the family Mycobacterium, abundant in cores from QV1.1 and nitrite - oxidizing Nitrospira (12) found in CSW1.1 cores. 66 Core - E nriched A rchaeal C ommunities Between 25 - 40% of the soil archaeal communities were made up of soil - enriched sequences (Fig. 3. 6), which exclusively belonged to the Soil Crenarcheotic Group. None of the fluid samples from CSW1.1 or QV1.1 submitted for 16S archaeal sequencing yielded sequences. This is consistent with findings from Chapter 2, where no archaeal sequences were detected when u niversal 16S rRNA primers were used and metagenomic datasets contained only 1% archaeal sequences ( 46 ). Contrary to findings at other serpentinite sites, which have detected Methan osarcinales (44) and archaea from the SAGMEG - 2 and ANME - 1 groups (45), to da te no archaea have been detected in the serpentinite fluids at CROMO. While 18 core samples were submitted for 16S rRNA archaeal sequencing, only four of them were successfully sequenced. As with bacteria, the success of sequencing does not correla te with q - PCR results or DNA yield per g rock (Fig. 3. 1 & 3. 2). From CSW, only the shallowest Figure 3.6 Relative abundance of core - , soil - , and fluid - enriched archaeal sequences . 67 samples , CSW1R (2.4 mbs) and CSW2R (4.2 mbs) yielded sequences. While from QV, only the deepest samples, QV30R (30.8 mbs ) and QV42R (43.3 mbs) were successfully sequenced. Between 45 - 50% of CSW1R, CSW2R, and QV42R consisted of core - enriched archaeal sequences (Fi g. 3. 6), which contain a diverse group of Euryarchaea, Crenarchaea, and Thaumar chaea (Table 3. 3). The most abundant core - enriched archaeal taxa were from the Miscellaneous Crenarchaeotic Group, making up roughly 50% of the core - specific archaeal sequences in each of the three samples (Table 3. 3). Almost 17% of the archaeal community from QV42R is composed of methanogens (Methanomicrobia, Methanococci, Methanobacteria; Table 3. 3), while these same taxa cumulatively made up less than 0.5% of the CSW1R and CS W2R communities and were not detected in QV30R. The microbial communities of the anoxic interiors of chimneys at the Lost City Hydrothermal Field were almost exclusively composed of Methanosarcinales ( 2, 40 ) , an order that made up almost 11% of the archaeal community in QV42R. Similar to the bacterial Figure 3.7 Archaeal community structure of unique sequences . (A) Rarefaction analysis of 16S rRNA amplicon sequences. (B) Community similarity dendrogram ca lculated from Morisita - Horn dissimilarity index. 68 findings, the QV30R archaeal community looks much more like the soil samples than other core samples (Fig. 3. 7). Almost 30% of QV30R was made up of the soil - enriched Soil Crenarcheotic Group, meanwhile le ss than 0.5% of the community was made up of core - enriched taxa (Fig. 3.6). Biogeochemical C orrelations The abundance of core - enriched bacterial and archaeal sequences were compared with XRD (mineralogy) , EA - IRMS (carbon and nitrogen abundance and stable isotope composition) , and core lithology data to determine what might be driving community composition within the core samples. The lithostratigraphic zones of the cores (Figs. 3.1A & 3.2A; Table 3.4) had an impac t on the microbial communities. With the exception of detrital serpentine soils ( LS_1 ; table 3.4) , which ha d significant correlations with both bacterial and archaeal sequences, the lithostratigraphic regions appeared to have correlations with members of o ne domain or the other. Within the detrital serpentine soil (LS_1) , there was an abundance of bacteria, from the phyla Acidobacteria, Nitrospirae, Chloroflexi (class Anaerolineae), and Proteobacteria (class Deltaproteobacteria), and archaea, from the Misce llaneous Crenarchaeal Group, Thermoplasmata, and Thaumarchaeal Group C3 (Figs. 3.8 & 3.9A). Serpentine gravel in a clay matrix (LS_2; Fig. 3.1A), had a positive correlation with percent nitrogen and bacterial members of the Acidobacteria, Proteobacteria (B eta - , Alpha - , and Delta - ), Chlorobi, and Gemmatimonadetes (Fig. 3.8B). Magnetite - bearing serpentine (LS_3), contained a diversity of archaeal groups, including the methanogenic orders Methanosarcinales, Methanomicrobiales, and Methanococcales (Fig. 3.9B). Magnetite - bearing serpentine with clay, albite, and quartz (LS_4) and was positively correlated to C/ N ratio and Actinobacteria and Firmicutes (Fig. 3.8C). 69 Table 3.3 Relative abundance of individual c ore - enriched archaeal taxa within samples . Phylum Class CSW1R CSW2R QV30R QV42R Miscellaneous Crenarchaeotic Group 29.3 17.9 0 20.1 Euryarchaeota Thermoplasmata 14.4 12.4 0.141 11.1 Euryarchaeota Methanomicrobia 0.281 0.137 0 10.6 Thaumarchaeota Group C3 4.59 0.720 0 1.56 Euryarchaeota Methanococci 0.077 0.014 0 3.81 Thaumarchaeota Marine Group I 1.58 2.20 0.140 0.057 Thaumarchaeota Soil Crenarchaeotic Group 0.182 2.83 0 0.429 Thaumarchaeota AK31 0.005 2.96 0 0 Euryarchaeota Methanobacteria 0.039 0.050 0 2.36 Thaumarchaeota Marine Benthic Group A 0.369 1.53 0 0.100 Thaumarchaeota South African Gold Mine Group 1 0.017 1.23 0.019 0 Thaumarchaeota Marine Benthic Group B 0.001 0.019 0 0.967 Unclassified Thaumarchaeota 0.026 0.863 0 0.005 Aenigmarchaeota Deep Sea Euryarchaeotic Group 0.415 0.081 0 0.209 Woesearchaeota 0.183 0 0 0.144 Euryarchaeota Archaeoglobi 0 0.024 0 0.211 Thaumarchaeota AK56 0 0.169 0 0 Crenarchaeota Thermoprotei 0 0.061 0 0.084 Thaumarchaeota AK59 0 0.111 0 0 Diapherotrites 0 0 0 0.103 Thaumarchaeota pSL12 0.035 0 0 0 Euryarchaeota Thermococci 0 0 0 0.032 Miscellaneous Euryarchaeotic Group 0 0 0.028 0.031 Parvarchaeota 0 0 0.025 0 Aigarchaeota Terrestrial Hot Spring Group 0 0.015 0 0 Unclassified Aenigmarchaeota 0 0 0 0.008 Unclassified Archaea 0.466 0.149 0 0.280 TOTAL 52.063 43.635 0.354 52.169 70 Table 3.4 Classification of lithostratigraphic units and samples belonging to them . LS Classification Lithostratigraphic Units* Samples LS_1 Detrital serpentine soil CSW1R, CSW2R LS_2 Serpentine gravel in clay CSW10R LS_3 Magnetite - bearing serpentine CSW14R, QV7R, QV25R, QV40R LS_4 Magnetite - bearing serpentine w/clay, albite, quartz QV18R, QV21R LS_5 Magnetite - bearing serpentine w/clay QV30R *Lithostratigraphic units were defined in Cardace et al ., 2013 and depicted in Figs. 1 & 2. Magnetite - bearing serpentine with clay (LS_5), which was found in the deep outlier sample QV30, contained the archaeal groups Thermoplasmata and Marine Group I (Fig. 3.9C). These data suggest that the geological ma ke up of a sample has strong influence on the microbiology living within it and that different geological horizons favor bacteria over archaea and vice versa. A diverse group of bacterial phyla exhibited significant negative correlations with iron concentration (from XRF data) and significant positive correlations with DNA yield, indicating an inverse relationship between the two parameters ( Fig. 3. 10). This inverse relationship could have to do with the adsorption of nucleic acids to iron clay mine rals ( 17, 47 ). While iron oxide clays can decrease the yield of high molecular weight DNA, it is not suspected that they alter the community composition, as the DNA adsorption is non - selective with regard to nucleotide patterns ( 9). Given the potential for iron oxides, such as magnetite, to adsorb nucleic acid particles, it is possible using a DNA extraction method to specifically address this issue, such as the one developed by Hurt et al. (17) would yield different results. Conclusions The method s employed in this study successfully allowed for the identification of core - , soil - , and fluid - enriched taxa from CROMO samples. The fluid - and soil - enriched sequences were consistent with previous studies of continental serpentinite fluids (3, 10, 41, 44 , 45, 46 ) and serpentine rhizospheres (19 , 27, 30, 31, 34) , respectively . Archaea were previously undetected at 71 Figure 3. 8 Correlation network of bacteria and lithostratigraphic units . The size of the bacterial nodes is proportional to the relative abundance of the sequence in the sample where it is most abundant (ranging from 0.01 - 1% of the community). Figure 3. 9 Negative correlation between iron concentration and DNA yield for select taxa . Positive correlations in black and negative c orrelations in black. The size of the bacterial nodes is proportional to the relative abundance of the sequence in the sample where it is most abundant (ranging from 0.01 - 1% of the community). 72 CROMO (46), however this study found them to be abunda nt and diverse within core samples. The core - enriched sequences identified in this study represent a diversity of archaea and bacteria, not previously within the serpentinite environment, containing taxa important in the cycling of nitrogen - containing mole cules and methane. Contrary to other sites of continental serpentinization ( 28, 45 ), fluids at CROMO contain elevated nitrogen species , particularly ammonium ( 10 ). The findings here indicate that bacteria and archaea associated with cores aid in the subsurface nitrogen cycling, through ammonium - and nitrite - oxidation within the serpentinite subsurface environment , which could supply this essential nutrient to the fluid - associated community . Figure 3. 10 Correlation network of archaea and lithostra tigraphic units . The size of the bacterial nodes is proportional to the relative abundance of the sequence in the sample where it is most abundant (ranging from 0.01 - 2% of the community). 73 While methanogens have not previously been detected in fluid sample s from CROMO ( 46 ), methane isotopologue data from the site suggests that the methane at CROMO is from a combination of thermogenic and microbial sources ( 49). Here methane - cycling archaea associated with core samples from QV were detected . These data suggest that particle - associated microbial communities are contributing to the cycling of nitrogen and methane within the serpentinite subsurface environment and that more complex biogeochemical cycles exist than can be seen from studying fluids alone. The correlation analyses suggest a relationship between the lithostratigraphic data and the core - enriched taxa. Of particular intrigue wa s the s hift from bacterial - to archaeal - dominated communities in the regions with magnetite - bearing serpentine both with out cla y (LS_3) and with clay (LS_5), though not in magnetite - bearing serpentine with clay, albite , and quartz ( LS_4; Table 3. 1; Fig. 3. 9) . These data suggest that the specific mineralogy of this region may favo r bacterial taxa over archaeal. With in th e subsurface, fluids and solid - substrates interact naturally, and therefore, both must be thought of as co mponents of the environment. P re vious studies have focused solely on the microorganisms dete cted within fluid samples and neglected the components of the communities associated with subsurface solids. The data presented here, combined with the study of serpentinite fluids from CROMO (46), provide a more comprehensive depiction of the serpentinite subsurface environment than has been seen in previous studies (4, 44, 45) and indicates that core - associated communities are distinct within different lithologies and could be contributing to cryptic biogeochemical cycling within the serpentinite subsurfa ce environment. 74 Materials and Methods Site Description and Sample Collection The Coast Range Ophiolite Microbial Observatory (CROMO) was established in August 2011, when eight wells were drilled into an actively serpentinizing subsurface environment at the UC Davis Donald and Sylvia McLaughlin Natural Reserve in Lower Lake, Califor nia. Details of the drilling operations can be found in Cardace et al (7). Bri efly, two main wells, CSW1.1 and QV1.1, were drilled 1.4 km apart to depths of 31 m and 45 m , respectively, using HQ wireline coring with an inner diameter of 63.5 mm . To mitigat e the potential of contamination from drill fluids used during drilling, as has been seen in ocean drilling studies (21), clean water ( filtered through 0.1um filter and ozonated) was used as drilling fluid and fluorescent microbead tracers were included to track contamination of exterior portions of the core . Cores from these wells were cataloged and preserved for complimentary mineralogical, geochemical, and microbiological analyses. Microbiological samples were wrapped in combusted aluminum foil, placed i n sterile bags, frozen with liquid nitrogen and stored at - 80ºC until DNA extraction. DNA Extraction Thawed core and soil material was homogenized using autoclaved and ethanol sterilized percussion mortars and ceramic mortars and pestles. DNA was extracted from 2 x 10 g of each homogenized core using the MoBio PowerMaxSoil Kit (Carlsbad, CA) , per the manufacturers instructions. The resulting DNA suspensions were pooled from replicate extractions and concentrated in an Amicon Ultra - 2 Centrifugal Filter Unit with Ultracel - 30 membrane (Millipre, Darmstadt, Germany) to a volume of 50 µL. Filtering of the well fluids and DNA extractions of the filters were conducted as described previously (4). DNA was quantified using the High 75 Sensitivity regents for a Qubit® 2.0 Fluorometer (Life Technologies, Grand Island, NY, USA), which has a detection limit of 0.1ng/µL DNA . Quantitative - PCR Copies of the 16S rRNA gene were determined by quantitative polymerase chain reacti on (q - PCR) using domain - specific primers targeting the V6V4 hypervariable region of the 16S rRNA gene for archaea and bacteria, respectively (43). Samples were run on a BioRad C - 1000 thermo - cycler with a q - PCR module using the SsoAdvanced SybrGreen Assay ( Biorad, Hercules, CA). 16S rRNA Gene Sequencing and Data Analysis Purified DNA samples from core, soils, and well fluids were submitted to the Josephine Bay Paul Center at the Marine Biological Laboratory for sequencing of the V4V5 region of the 16S rRNA g ene on an Illuminia MiSeq instrument as part of the Census of Deep Life project (29). The paired - - processing quality control for removal of low quality reads and chimera checking (18). The samples yiel ded be tween 8 , 493 and 246 , 345 merged, quality - filtered sequences each . Any sample that yielded less than 5,000 sequences was considered a failed sequencing run and not included the analysis. Using Mothur (39), the sequence data was clustered into unique sequences and assigned taxonomy compared by alignment to the SILVA database (v119 for bacteria and v123 for archaea; 31). Significant differences in the abundances of unique sequences between g roups of samples were tested with the aid of the Phyloseq (25) and DESeq (22) packages in R. A complete list of R commands for on example comparison of two groups of samples is provided in Appendix A. For DE Seq comparison s , samples were pooled into categor - or - 76 outliers (Fig. 5) and s between categories and were instead analyzed separately as ind ividual core samples against the fluids and soils. - both fluids and soils. Mineralogical and Geochemical Analyses Subsamples of each core were analyzed to obtain total nit rogen (TN) and total organic carbon (TOC), as well as 13 C of TOC by elemental analysis isotope ratio mass spectroscopy (EA - IRMS) . Core samples were acidified via the adapted capsule method ( 5 ) prior to analysis on a Carlo Erba 1108 elemental analyzer coup led to a ThermoFinnigan Delta Plus XL isotope ratio mass spectrometer. Descriptions of the core lithostratigraphy, down - hole logging, X - ray diffraction (XRD) , and X - ray fluorescence (XRF) methods can be found in refs 7 and 8 . Statistical Analyses A matrix containing environmental geological and geochemical data and relative abundances of core - enriched sequences for each core sample was used as input for pairwise mt (36). The false - discovery rate (q - value) was computed f or the distribution of Pearson p - values to account for multiple tests. Pairwise correlation s with both p - and q - values of <0.05 were considered significant and included in network analyses. Network models of significant correlations were created using Cytoscape v2.8.3 (38). 77 REFERENCES 78 REFERENCES 1 . Brazelton WJ, Schrenk MO, Kelley DS, Baross JA (2006) Methane - and sulfur - metabolizing microbial communities dominate the Lost City hydrothermal field ecosystem. Appl Environ Microbiol 72:6257 - 6270. 2 . Brazelton WJ , Mehta MP, Kelley DS, Baross JA (2011 ) Physiological differentiation within a single - species biofilm f ueled by serpentinization. mBio 2 : doi:10.1128/mBio.00127 - 11 . 3 . Brazelton WJ, Nelson B, Schrenk MO (2012) Metagenomic evidence for H 2 oxidation and H 2 production by serpentinite - hosted subsurface microbial communities. Front Microbiol 2:268. 4 . B razelton WJ , Morrill PL, Szponar N , Schrenk MO (2013) Bacterial communities associated with subsurface geochemical processes in continental serpentinite sp rings. A ppl E nviron Microb 79: 3906 - 3916. 5 . Brodie, C. R., M. J. Leng, J. S. L. Casford, C. P. Kendrick, J. M. Lloyd, Z. Yongqiang, and M. I. Bird. 2011. Evidence for bias in C and N concentrations and delta13C composition of terrestrial and aquatic organic materials due to per - analysis acid prepara tion methods. Chemical Geology . 282 : 67 - 83. 6 . Buckley DH, et al. (2006) Diversity of Planctomycetes in soil in relation to soil history and environmental heterogeneity. Appl Environ Microb 72:4522 - 4531. 7. Cardace D, et al. (2013) Establishment of the Coast Range ophiolite microbial observatory (CROMO): drilling objectives and preliminary outcomes. Scientific Drilling 16:45 - 55. 8. Carnevale, DC (2013) Carbon sequestration potential of the Coast Range Ophiolite in 9. Cleaves HJ, et al. (2011) The adsorption of short single - stranded DNA oligomers to mineral surfaces. Chemosphere 88:1560 - 1567. 10. Crespo - Medina M, et al. (2014) Insights into environmental contro ls on microbial communities in a continental serpentinite aquifer using a microcosm - based approach. Front Microbiol 5:604. 11. Daae FL, et al. (2013) Microbial life associated with low - temperature alteration of ultramafic rocks in the Leka ophiolite complex. Geobiology 11:318 - 339. 12. Ehrich S, et al. (1995) A new obligately chemolithoautotrophic, nitrite - oxidizing bacterium, Nitrospira moscoviensis sp. nov. and its phylogenetic relationship. Arch Microbiol 164:16 - 23. 79 1 3. Fuerst JA, et al. (2006) Anammoxosomes of anaerobic ammonium - oxidizing Planctomycetes. Microbiol Monogr 2:260 - 283. 14. Früh - Green GL , et al. ( 2004 ) Serpentinization of oceanic peridotites: Implications for geochemical cycles and biological activity. I n: The Subseafloor Biosphere at Mid - Ocean Ridges. (American Geophysical Union, Washington, D.C.) pgs. 119 - 136. 15. Hallbeck L, Pedersen K ( 2014 ) The family Gallionellaceae . The Prokaryotes, ( Springer - Verlag, Berlin ) pp. 853 - 858 . 16. Hanada S (2014) The phylum Chloroflexi, the Family Chlorofexaceae , and the related phototrophic families Oscillochloridaceae and Roseiflexaceae . In: The Prokaryotes Other Major Lineages of Bacteria and Archaea. (Springer - Verlag, Berlin). 17. Hurt RA , et al. (2014) Improv ed yield of high molecular weight DNA coincides with increased microbial diversity access from iron oxide cemented sub - surface clay environments. PLOS One 9:e102826. 18. Huse SM, et al. (2014) VAMPS: a website for visualization and analysis of microbial population structures. BMC Bioinformatics 15:41. 19. Idris R, et al. (2004) Bacterial communities associated with flowering plants of the Ni hyperaccumulator Thalspi goesingense . Appl Environ Microb 70:2667 - 2677. 20. Katsaveli K, Vayenas D, Tsiamis G , Bourtzis K (2012) Bacterial diversity in Cr(VI) and Cr(III) - contaminated industrial wastewaters. Extremophiles 16:285 - 296. 21. Lever MA, et al. (2006). Trends in basalt and sediment core contamination during IODP Expedition 301. Geomicrobiol J 23:517 - 530. 22. Love MI, Huber W, Anders S ( 2014) Moderated estimation of fold change and dispersion for RNA - Seq data with DESeq2. Genome Bio l 15 :550 . 23. Lücker S, Daims H ( 2014) The family Nitrospinaceae . In: The Prokaryotes, ( Springer - Verlag, Berli n ) pp. 231 - 237 . 24. McCollom TM, Bach W (2009) Thermodynamic constraints on hydrogen generation during serpentinization of ultramafic rocks. G eochim C osmochim A c 73:856 - 875. 25. McMurdie PJ, Holmes S (2013) phyloseq: An R package for reproducible interactive analysis and graphics of microbiome census data. PLOS One 8:e61217. 26. McMurdie PJ & Holmes S (2014) Waste not, want not: why rarefying microbiome data is inadmissible. PLoS Computational Biology , 10 (4), e1003531. doi:10.1371/journal.pcbi.1003531 26. 80 27. Megoni A, et al. (2004) Genetic diversity of bacterial communities of serpentein soil and of rhizosphere of the nickel - hyperaccumulator plant Alyssum bertolonii . Micr obial Ecol 48:209 - 217. 28. Morrill PL, et al. (2013) Geochemistry and geobiology of a present day serpentinization site in California: the Cedars. Geochim Cosmochim Acta 109:222 - 249. 29. Morrison HG, Grim SL, Vineis JH, Sogin ML (2013) 16S amplicon fusi on primers and protocol for Illumina platform sequencing. Figshare. Available online at: http://dx.doi.org/10.6084/m9.figshare.833944 30. Oline DK (2006) Phylogenetic comparisons of bacterial communities from serpentine and nonserpentine soils. Appl Environ Microb 72:6965 - 6971. 31. Pal A, Wauters G, Paul AK (2007) Nickel tolerance and accumulation by bacteria from rhizosphere of nickel hyperaccumulators in serpentine ecosystem of Andaman, India. Plant Soil 293:37 - 48. 32. Prosser JI, Head IM, Stein LY (2014) The family Nitrosomonadaceae . In: The Prokaryotes, ( Springer - Verlag, Berlin ) pp. 901 - 918 . 33. Pruesse E, et al. (2007) SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compati ble with ARB. Nucleic Acids Res 35:7188 - 7196. 34. Rajkumar M, Prasad MNV, Freitas H, Ae N (2009) Biotechnological applications of serpentine soil bacteria for phytoremediation of trace metals. Crit Rev Biotechnol 29:120 - 130. 35. Rastogi G, et al. (2010) Microbial and mineralogical characterization of soils collected from the deep biosphere of the former Homestake Gold Mine, South Dakota. Microb Ecol 60:539 - 550. 36. Rizopoulos D (2006) ltm: an R package for latent variable modeling. J Stat Softw 17:5. 37. Salter SJ, et al. (2014) Reagent and laboratory contamination can critically impact sequence - based microbiome analyses. BMC Biology 12:87. 38. Shannon P, et al. (2003) Cytoscape: a software environment for integrated models of biomolecular intera ction networks. Genome Res 13:2498 - 2504. 39. Schloss PD, Westcott SL (2011) Assessing and improving methods used in operational taxonomic unit - based approaches for 16S rRNA gene sequence analysis. Appl Environ Microb 77:3219 - 3226. 81 40. Schrenk MO , Kelley D S, Bolton SA , Baross JA ( 2004 ) Low archaeal diversity linked to subseafloor geochemical processes at the Lost City Hydrothermal Field, Mid - Atlantic Ridge. Environ Microbiol 6 : 1086 - 1095. 41. Schrenk MO, Brazelton WJ, Lang SQ (2013) Serpentinization, ca rbon, and deep life. Rev Mineral Geochem 75:575 - 606. 42. Schwartz E , et al. ( 2006 ) The H 2 - metabolizing prokaryotes. In: The Prokaryotes, (Springer, New York, NY) pp. 496 - 563. Proc. Natl. Acad. Sci. USA. 103:12115 - 12120. 44. Suzuki S, et al. (2013) Microbial diversity in The Cedars, an ultrabasic, ultrareducing, and low salin ity serpentinizing ecosystem. Proc Natl Acad Sci USA 110:15336 - 15341. 45. Tiago I, Veríssimo A (2013) Microbial and functional diversity of a subterrestrial high pH groundwater associated to serpentinization. Environ Microbiol 15:1687 - 1706. 46. Twing KI, et al. (In Submission ) Serpentinizing fluids from a subsurface observatory harbor extremely low diversity microbial communities adapted to high pH. ISME J . 47. Vermeer AWP, van Riemsdijk WH, Koopal LK (1998) Adsorption of humic acid to mineral particals : Specific and electrostatic interactions. Langmuir 14:2810 - 2819. 48. Walker RB (1954) Factors affecting plant growth on serpentine soils. Ecology 35:259 - 266. 49. Wang DT, et al. (2015) Nonequilibrium clumped isotope signals in microbial methane. Science 348:428 - 431. 82 CHAPTER 4 Microbe - mineral interactions in the serpentinite subsurface environment Abstract Serpentinization creates a challenging habitat for microorganisms, with high pH fluids depleted in terminal electron acceptors and dissolved inorganic carbon (DIC). Findings from Chapter 3 suggest that there are specialized microbial communities associated with minerals in the serpentinite subsurface environment compared to fluids. I n situ colonization devices containing mineral substrates were deployed into microbial observatory wells to test whether microbes in the serpentinite subsurface were capable of utilizing inorganic carbon in calcite and ferric iron in magnetite to make up for the lack of DIC and electron acceptors in surrounding fluids . Mineral substrates deployed into the most extreme serpentinite well (CSW1.1, pH 12.5) selected for unique microbial communities, compared with well fluids collected at the start of the expe riment. Devices containing calcite resulted in an abundance of Syntrophomonas (a clostridium also found in starting flu ids) and Deinococcus. Magnetite led to an increase in diversity, including Alphaproteobacteria, Gammaproteobacteria, and Actinobacteria, in addition to the Betaproteobacteria commonly found in CSW1.1 fluids. It is not clear, however, if or how these enriched microbial communities are utilizing the calcite and magnetite. These data suggest that specific mineral composition may impact microbial diversity , particularly in the harshest environmental conditions . Introduction M ost m icrobial metabolisms (catabolic) are based on redox reactions, for which they need both electro n donors and electron acceptors. Some microbes are able to metabolize chemicals found within mineral forms to obtain energy through chemolithotrophy (25 ). Microbes 83 can form complex communities and micro - niches within biofilms on hard surfaces ( 7 ). Given th at the subsurface environment is rock - hosted, there is a great deal of interest in understanding microbe - mineral interactions. Previous studies have yielded successful results with in situ colonization of mineral substrates deployed into the International Ocean D iscovery Program ( IODP ) boreholes in the basaltic ocean crust in borehole packers (otherwise known as circu lation obviation retrofit kits or CORKS; e.g. 9 , 20 , 2 1 ) or as mineral chips in sub - glacial systems (1 7 ), suggesting that mineral composition leads to differences in microbial biomass and community composition. The geochemical process of serpentinization creates a challenging habitat for life, not only in terms of the extreme pH of fluids, but also with limited availability of dissolved inorgani c carbon (DIC) and scarcity of electron acceptors. The high pH shifts the DIC species in fluids to predominantly carbonate, which precipitates out of solution in the presence of abundant Ca 2+ ions, to form calcium carbonate (1, 11 ). C haracterization and g enomic analyses of an isolate of Serpentinomonas , the Betaproteobacterial genus that dominates serpentinite fluids , indicated that the organism is capable of autotrophic growth on calcium carbonate (3 1 ). The abundant H 2 and methane frequently generated during serpentinization are potential electron donors for microbial metabolisms, but this potential energy can only be harnessed if they can be paired with electron acceptors. There is a dearth of electron acceptors in serpentinite fluids, relative to the abundance of electron donors, particularly in the anoxic end - member serpentinite fluids. While not detected in the serpentinite subsurface environment before, iron - reducing bacteria, such as Shewanella putrefaciens are able to reduce ferric iron found in t he mineral magnetite (1 2 ). 84 This study explores the potential for microorganisms native to the serpentinite subsurface environment to utilize solid inorganic carbon and the potential electron acceptor ferric iron from mineral phases calcite and magnetite, respectively, to make up for the lack of these necessary chemicals in serpentinite fluids. In situ colonization devices containing mineral substrates were incubated in four wells spanning a range of geochemical gradients at the Coast Range Ophiolite Microb ial Observatory (CROMO; 5) over the course of a year and the resulting microbial communities were analyzed by 16S rRNA gene sequencing. Results and Discussion In S itu C olonization D evices and CROMO W ells Colonization devices containing mineral substrates included borosilicate glass as an inert substrate for control, magnetite (as a source of ferric iron) , or calcite (as a source of carbon ) were deployed near the bottom (~15 mbs) of the CSW1.1, CSW1.3, QV1.1, and QV1.2 wells f or one year to assess the microbe - mineral interactions within the well. DNA was extracted from the mineral substrates and used for 16S rRNA gene quantitative - PCR and amplicon sequencing. For experimental replication, the devices were deployed in triplicate , in alternating order (Fig. 4.1). The bladder pump used to sample well fluids sits right above the bottom of the well and the in situ devices were deployed directly above it (Fig. 4.1D). Since the pump pulls water from above, Table 4.1 - Geochemical para meters of well fluids measured at in situ device deployment and retrieval . Well Depth (mbs) Sampling Trip pH ORP (mV) Temp (°C) Conductivity (µS/cm 3 ) DO (mg/L) CSW1.1 19.5 Aug. 2013 12.4 - 254 15.8 4460 0.10 July 2014 12.3 - 297 16.3 4455 0.43 CSW1.3 23.2 Aug. 2013 10.3 - 234 17.2 4610 0.03 July 2014 10.3 - 328 15.6 4597 0.19 QV1.1 23.0 Aug. 2013 11.6 - 92 16.4 2610 0.15 July 2014 11.5 - 239 17.2 2841 0.27 QV1.2 14.9 Aug. 2013 9.1 - 73 17.1 2800 0.22 July 2014 9.4 - 168 16.9 3042 0.20 85 the devices were deployed two full days after sampling the wells in August 2013, so as to allow for fluid recharge in the well. A preliminary time - series study of in situ colonization at CROMO suggested that one year was optimum for microbial growth (data not shown). Table 4.1 summarizes the environmental parameters of the wells waters at the beginning (referred to henceforth as 2013 fluids) and end (referred to henceforth as 2014 fluids) of the experiment of the one - year colonization experiment , which are consistent with prior measurements at of these wells at CROMO (Chapter 2; 3 3 ). 16S rRNA G ene A bundance The 16S rRNA gene abundances of microbial communities from the in situ devices were determined by bacteria - specific quantitative - PCR (Fig. 4.2). There were significantly more 16S rRNA gene copies per g substrate in the glass - treatment than on the magnetite or calcite - treatments in CSW1.1 (F 2,6 = 6.35, p - value = 0.03) and fewer gene copies on magnetite than either glass or calcite in QV1.1 (F 2,6 = 6.93, p - value = 0.03). No other significant differences in gene copy abundance were seen either within wells or between wells (Fig. 4.2), likely due to the high variation between replicates ( e.g. CSW1.3 calcite or QV1.2 magnetite). Bacterial C ommunity C omposition Across all wells, magnetite yielded communities with greater diversity than did calcite (Fig. 4.3). Glass, on the other hand, yielded communities of variable diversity, with the most diverse glass samples coming from QV1.1 and QV1.2 and the least diverse glass samples coming from CSW1.1 (Fig. 4.3). Microbial communities associated with the in situ colonization devices were more similar to communities from CROMO well fluids than from CROMO cores or soils (Fig. 4.4). Overall, community co mposition was affected to a much greater extent by which well the colonization devices were deployed in than by the specific mineral matrix within the device, 86 Figure 4.1 In situ colonization devices and schemat ic of deployment down well . A) Open devices filled with borosilicate glass beads for inert substrate control. B) Fully assembled devices prior to autoclave sterilization. C) Devices tied together with Teflon fishing line being deployed into QV1.1. D) Diagram of down - well deployment. The in s itu devices were tied together with Teflon fishing line and secured to the top of the well - head and deployed just above the bladder pump, which sits at the bottom of the well. 87 Figure 4.2 16S rRNA gene copy abundance per gram of substrate, as d etermined by q - PCR . Mineral - substrates are color - coded: glass, green; magnetite, orange; calcite, purple. Significant variations in 16S rRNA gene copy number, as determined with ANOVA, are represented with asterisks . with all the communities from QV1.2 and CSW1.3 treeing together, respectively (Fig. 4.4). Samples from the extreme pH wells, CSW1.1 (pH 12.4) and QV1.1 (pH 11.6) were interspersed with each other in the community similarity dendrogram (Fig. 4.4). The exceptions to this are two of the magneti te treatments CSW1.1 , which were more similar to communities on magnetite from CSW1.3 than other CSW1.1 or QV1.1 communities (Fig. 4.4). With regard to mineral substrates, glass and calcite tree together and separately from magnetite for all of the well. These data suggest that, similar to the findings in from CROMO fluids , fluid source (and pH) has an impact on microbial community composition (33 ) , with mineral substrate being secondary. Similar to previous studies of the serpentinite subsurface, and as shown for CROMO in Chapter 2 (3 3 ), fluids were dominated by Comamonadaceae (Betaproteobacteria) and Clostridia 88 Figure 4.3 Rarefaction curve of unique sequences from in situ colonization device and well fluid sequencing sequences (Fig. 4.5). CSW1.1 f lu ids from 2013 (start of the experiment) contained an abundance of Betaproteobacteria and QV1.1 fluids contained a mix of different Clostridia species (Fig. 4.5; 3 3 ). Similarly, the lower pH wells were more diverse than the high pH wells (Fig. 4.5). In CSW1 .1, there is a shift in community composition between 2013 fluids (start of the experiment) and 2014 fluids (end of the experiment) , namely the spike in Deinococcu s - Thermus sequences, that can perhaps be explained by sloughing off from glass and calcite colonization devices or the stimulated growth of those organisms within the fluids due to the presence of calcite . Similarly, QV1.1 end fluids contained Bacteroidetes sequences that were not apparent in the starting fluid community. The community composit ion on the mineral substrates from CSW1.1 look distinct 89 from starting fluids . In contrast, the fluid and mineral samples from other wells (CSW1.3, QV1.1, QV1.2) share similar community composition at the family level (Fig. 4.5 ) . Mineral - E nriched B acteria Due to the difficulty in determining the source of sequences in comparative datasets, the statistical method of DESeq (1 3 ), modeled after differential ex pression in RNA studies (1 6 ) , was used to determine sequences that were statistically enriched in a pa rticular sample type compared to the others (Table 4.2) . Given the potential for transfer of organisms between fluids at the end of the experiment and the mineral substrates, either from microbes being knocked off the in situ colonization devices or stimul ation of microbial growth within fluids by the addition of minerals , only sequences from fluids from the start of the experiment, collected in August 2013, were used Communities associated with mineral substrates incubated in CSW1.1 were the most divergent from well fluids compared to the other three wells investigated (Fig. 4.5) and therefore more sequences were determined to be enriched for different minerals within that well (F ig. 4.6). Table 4.2 DESeq comparisons for determining fluid - and mineral - specific taxa . Enriched Community Sample Group 1 Sample Group 2 CSW1.1 Fluid start CSW1.1 2013 fluids CSW1.1 2014 fluids CSW1.1 2013 fluids CSW1.1 in situ devices (pooled) CSW1.1 2013 fluids CSW1.1 glass devices CSW1.1 2013 fluids CSW1.1 magnetite devices CSW1.1 2013 fluids CSW1.1 calcite devices Glass CSW1.1 glass devices CSW1.1 2013 fluids CSW1.1 glass devices CSW1.1 magnetite devices CSW1.1 glass devices CSW1.1 calcite devices Magnetite CSW1.1 magnetite devices CSW1.1 2013 fluids CSW1.1 magnetite devices CSW1.1 glass devices CSW1.1 magnetite devices CSW1.1 calcite devices Glass CSW1.1 calcite devices CSW1.1 2013 fluids CSW1.1 calcite devices CSW1.1 glass devices CSW1.1 calcite devices CSW1.1 magnetite devices * Comparisons were repeated for all four wells 90 Figure 4.4 Community similarity dendrogram of unique sequences from in situ colonization devices, well fluids, and core material . The sorclass dissimilarity index was used to make the de ndrogram in Mothur (2 4 ). 91 The fluid - enriched taxa in CSW1.1 are consistent with previous findings from the serpentinite environment (3, 4, 29, 31) and that well, in particular (Chapter 2; 7, 32), with a do minance of Comamonadaceae and Clostridia (Table 4.3). For the fluid samples (and specific sampling dates) described i n Chapter 2, these taxa made up greater than 70% of the microbial communities from CSW1.1 ( 3 3 ), however only 20% of the community was ident - study (Fig. 4.6A). This could either be due to a shift in community structure or, more likely, sequences belonging to these major groups were also found in in situ devices and therefore not - Magnetite - enriched ta xa made up 25% of the magnetite - filled in situ devices (Fig. 4.6A) and contained an abundance of Comamonadaceae sequences (Table 4.4). It should be noted that t hese are not the same Comamonadaceae that make up the fluid - enriched sequences, as they were not detected in the well fluids (Table 4.4). Additionally, members of the Sporichthyaceae (class Actinobacteria) and LD12 - freshwater - group of the SAR11 were detect ed (Table 4.4). None of the taxa enriched in the magnetite in situ devices are known to be capable o f iron reduction (1 9 , 3 6 ). The calcite - enriched microbial communities consisted of Synthrophomonadaceae (class Clostridia) and Deinococci (Table 4. 5) and made up 50 - 70% of the communities within calcite devices, but also 10 - 20% of glass devices and 50 - 70% of fluids from the end of the experiment (Fig. 4.6A). Given the abundance of these organisms in fluids from July 2014, it is possible that they na turally occurred in the fluids and were just trapped in the in situ devices, instead of being enriched by them. However, given that neither taxa has been detected in appreciable abundances in CROMO fluids before and the stability of microbial communities a t CROMO over time (Chapter 2; 3 3 ), it is more likely that they were enriched by the addition of the calcite substrate 92 Figure 4.5 Community diversity bar chart for CROMO fluids and in situ colonization devices . Taxonomy is defined at the family level and chart was made with data from VAMPS ( 10 ). 93 or slo ugh ed off into the fluids during sampling. Members of the Syntrophomonadaceae c an oxidize carbon monoxide ( 26, 27 , 28 ) and are generally found in syntrophic relationships with methanogens (2 8 ). It should be noted that the DNA from in situ devices was only sequenced for bacterial 16S rRNA genes and therefore archaea were not looked for in this study, so whether or not a syntrophic relationship was taking place in this instance cannot be addressed . Deinococcus - Thermus have been isolated from arid deserts and exhibit adaptations to resist challenges, such as UV - radiation and desiccation (1 4 ). While it is not clear what their role in the serpentinite system is, they have been previously detected in l ow abundances (0.5 - 1.5%) of 16S rRNA clone and amplicon libraries from fluids (3 2 ) and carbonate structures (2, 3 5 , 2 2 ) within the serpentinite environment. It should be noted that both of these org anisms are known spore - formers and their abundance in the in situ devices and 2014 fluids compared to 2013 fluids suggests either their grow within the samples or that the minerals help trap rare spore and recalcitrant microbes within the system . Fluid - enriched sequences from CSW1.3 belonged to the Comamonadaceae and Gammaproteobacteria (Table 4.3), consistent with findings in Chapter 2 (3 3 ). No calcite - enriched taxa were detected in CW1.3 (Fig. 4.6B) , a well with pH 10.3 fluids containing low DIC (172 µM; 33) . The magnetite - enriched bacterial commun ity made up ~3% of the magnetite in situ devices communities (Fig. 4.6B) and contained a variety of Actinobacter, Firmicutes, and Proteobacteria, however none of the taxa made up more than 1% of the communities. The Comamonadaceae sequence enriched in magn etite samples from CSW1.1 was also enriched in this sample, bu t much less abundant (0.3% versu s 14% of the magnetite - enriched communities; Table 4.5). The microbial communities from fluids and in situ devices from the QV1.1 showed 94 Table 4.3 - Relative - abu ndance of fluid - enriched bacterial taxa in different sample types Well Phylum Class Order Family 2013 2014 Glass Magnetite Calcite CSW1.1 Proteobacteria Betaproteobacteria Burkholderiales Comamonadaceae 42 ± 0.44 8.4 ± 1.6 0.40 ± 0.2 0.12 ± 0.09 0.33 ± 0.28 Firmicutes Clostridia Thermoanaerobacterales SRB2 0.98 ± 0.07 0.04 ± 0.02 0.03 + 0.01 0 0.02 ± 0.01 CSW1.3 Proteobacteria Gammaproteobacteria Xanthamonadales Xanthomonadaceae 1.4 ± 1.2 0.94 ± 1.3 0.02 ± 0.006 0 0.03 ± 0.01 Proteobacteria Betaproteobacteria Burkholderiales Comamonadaceae 0.71 ± 0.51 0.57 ± 0.71 0.01 ± 0.002 0.01 ± 0.01 0.02 ± 0.004 * No fluid - enriched taxa were found in QV1.1 or QV1.2. ** Values represent average and standard deviation of relative abundance for that treatment in the specified well. Table 4.4 Relative - abundance of magnetite - enriched bacterial taxa in different sample types Well Phylum Class Order Family 2013 2014 Glass Magnetite Calcite CSW1.1 Proteobacteria Betaproteobacteria Burkholderiales Comamonadaceae 0.01 ± 0.01 0 0.04 ± 0.01 14 ± 6.5 0.02 ± 0 Actinobacteria Actinobacteria Frankiales Sporichthyaceae 0 0 0 1.8 ± 1.1 0.001 ± 0 Proteobacteria Alphaproteobacteria SAR11 LD12 freshwater group 0 0 0 1.2 ± 1.1 0 QV1.2 Proteobacteria Betaproteobacteria Burkholderiales Oxalobacteraceae 0 0 0.09 ± 0.03 6.2 ± 0.54 0.07 ± 0.04 Actinobacteria Actinobacteria Propionibacteriales Propionibacteriaceae 0 0 0.01 ± 0.01 2.5 ± 2.8 0.01 ± 0.01 Proteobacteria Betaproteobacteria Burkholderiales Comamonadaceae 0.02 ± 0.01 0.02 ± 0.01 0.02 ± 0.01 1.2 ± 0.23 0.02 ± 0.01 Proteobacteria Alphaproteobacteria Rhizobiales Methylobacteriaceae 0 0 0.01 ± 0 1.1 ± 0.31 0.01 ± 0.01 Proteobacteria Betaproteobacteria Burkholderiales Burkholderiaceae 0 0 0.01 ± 0 1.0 ± 0.08 0.01 ± 0.01 * No magnetite - enriched taxa were found in CSW 1. 3 or QV1. 1 . ** Values represent average and standard deviation of relative abundance for that treatment in the specified well. Table 4.5 Relative - abundance of calcite - enriched bacterial taxa in different sample types Well Phylum Class Order Family 2013 2014 Glass Magnetite Calcite CSW1.1 Deinococcus - Thermus Deinococci Deinococcales Trueperaceae 0.001 ± 0.002 33 ± 5.9 5.6 ± 1.5 0.26 ± 0.08 25 ± 7.4 Firmicutes Clostridia Clostridiales Syntrophomonadaceae 0.001 ± 0.002 32 ± 5.8 5.3 ± 1.4 0.25 ± 0.08 24 ± 7.1 * No calcite - enriched taxa were found in CSW 1. 3, QV1. 1, or QV1.2 . ** Values represent average and standard deviation of relative abundance for that treatment in the specified well. 95 Figure 4.6 Relative abundances of fluid - , glass - , magnetite - , and calcite - enriched bacterial taxa per each sample in all four wells . Calcite - enriched taxa were only detected in CSW1. 1, while magnetite - enriched taxa were detected in CSW1.1, CSW1.3, and QV1.2. No mineral - or fluid - enriched taxa were determined in QV1.1. similar composition s (Fig. 4.5) and therefore no taxa were determined to be statistically enriched in one type of sa mple compared to the others (Fig. 4.6C). The fluid - enriched taxa in QV1.2 constituted 1% of the fluid samples (Fig. 4.6D) and contained members of the Bacteroidetes, Actinobacteria, and Betaproteobacteria. Magnetite - enriched taxa from this well made up 20% of the communities from magnetite - containing in situ devices (Fig. 4.6D) and consisted of various families within the Betaproteobacteria, Actinobacteria, and methylotrophic Alphaproteobacteria (Table 4.4). 96 Conclusion s The purpose of this study was to explore microbe - mineral interactions within the serpentinite subsurface environment on environmentally relevant mineral substrates. Both calcite and magnetite are prevalent throughout CROMO cores (6). Additionally , we aimed to assess the microbiological capabilities to access inorganic carbon (from calcite) and terminal electron acceptors (in the form of ferric iron from magnetite) from mineral sources, given that such vital solute s are depleted in serpentinite fluids. The m ineral - substrate in situ colonization devices yielded deployed into four different wells at CROMO yielded different microbial communities, depending on the inoculum source (i.e. well fluids; Figs. 4.5 & 4.6) , with the most drastic results coming from CSW1. 1, the most extreme serpentinite well . The addition of calcite to CSW1.1 resulted in the significant enrichment of Deinococcus - Thermus sequences, previously detected at serpentinite sites in low abundances (2, 2 2 , 3 1 , 3 3 ) and the addition of magnetite in t he same well resulted in enrichment of Comamonadaceae, common inhabitants of the serpentinite the serpentinite subsurface environment. The results from more moderate CROMO wells were less clear, indicating that enrichment of select communities on minerals in this environment may be dependent on the start inoculum (i.e. well fluids). CSW1.1 fluids are depleted in DIC and exhibit extreme pH (Table 4.1) and therefore the mineral substrates may have offered rare taxa a refuge from the harsh environment. The lack of evident relationships between the mineral - enriched taxa and the mineral substrates (e.g. known ability to fix CO 2 from carbonate or iron - reduction of magnetite) raises the question of why s pecific taxa were enriched by the addition of certain mineral substrates. These data , along with the findings from CROMO cores ( 34 ), suggest that the mineralogy of the subsurface environment can impact community composition, potentially offering rare taxa refuge from challenging environmental. 97 Materials and Methods In Situ Colonization Devices To assess the growth of microorganisms on solid surfaces within the serpentinite subsurface environment, in situ colonization devices containing mineral substrates were developed and deployed into select wells. The in situ colonization devices were made of ethanol - washed, 1 inch Schedule 80 PVC tubing cut into 1.5 inch pieces and capped with 143 µm Teflon mesh (Small Parts Inc, Logansport, IN), allowing for the trans port of fluids and small particles within the device (Fig. 4.1B) . The devices were filled with ~7.5 g of 500 µm diameter borosilicate glass beads (Sigma - Aldrich, St. Louis, MO), magnetite (Fe 3+ 2 Fe 2+ O 4 Natural Science, Rochester, NY), or calcite (CaCO 3 , which were crushed and sieved to insure uniform mineral size . Post - assembly, the devices were wrapped in aluminum foil and autoclaved. The devices were deployed in triplicate into the CROMO wells CSW1.1 (pH 12.4), CSW1.3 (pH 10.3), QV1.1 (pH 11.6), and QV1.2 (pH 9.1) in August 2013. The devices were tied together with Teflon fishing line, with 10 cm between devices, and deployed to a depth of 15 mbs. The devices were left undisturbed in the wells for one yea r, until recovery in July 2014. Upon retrieval, the in situ devices were placed into individual sterile Whirlpak bags (Nasco, Fort Atkinson, WI) , flash frozen in liquid nitrogen, and stored at - 80 ° C until DNA extraction. Well fluids were collected during b oth deployment and retrieval, as previously described ( 8 ). DNA Extraction DNA was extracted from mineral substrates using the MoBio PowerMaxSoil Kit (Carlsbad, CA), per the manufacturers instructions. The resulting DNA was concentrated in an Amicon Ultra - 2 Centrifugal Filter Unit with Ultracel - 30 membrane (Millipore, Darmstadt, 98 Germany) to a volume of 50 µL. Filtering of the well fluids and DNA extractions of the filters were conducted as described previously (4). DNA was quantified using the High Sensitiv ity regents for a Qubit® 2.0 Fluorometer (Life Technologies, Grand Island, NY, USA), which has a detection limit of 0.1ng/µL DNA. Quantitative - PCR Copies of the 16S rRNA gene were determined by quantitative polymerase chain reaction (q - PCR) using domain - specific primers targeting the V6V4 hypervariable region of the 16S rRNA gene of bacteria (2 9 ). Samples were run on a BioRad q - PCR instrument using the SsoAdvanced SybrGreen Assay (Biorad, Hercules, CA). The statistical package StatPlus (AnalystSoft, Inc.) was used in Excel 2011 to perform one - way ANOVA tests, to determine if the different mineral substrates in the in situ devices, and two - way ANOVA test, to determine if the interaction between well and mineral substrates in microcosms, resulted in a signif icant difference in 16S rRNA gene abundance (from q - PCR data). 16S rRNA Gene Sequencing and Data Analysis Purified DNA samples from in situ device minerals, CROMO well fluids, microcosm minerals, and microcosm fluids were submitted to the Josephine Bay P aul Center at the Marine Biological Laboratory for sequencing of the V4V5 region of the 16S rRNA gene on an Illuminia MiSeq instrument as part of the Census of Deep Life project (1 8 ). The paired - end reads were - processing quality control for removal of low quality reads and chimera checking ( 10 ). The in situ device and microcosm samples yielded between 15,753 and 277,722 merged, quality - filtered sequences each. Any sample that yielded less than 5,000 sequences was considere d a failed sequencing run and not included the analysis. 99 Community diversity bar charts were generated using the Visualization and Analysis of Microbial Populations Structures (VAMPS) interface, which assigns taxonomy of se quences against SILVA and RDP (1 0 ). The charts presented here were formed using Family - level taxonomic assignment and represent relative abundances of sequences. Using Mothur (2 4 ), the sequence data was clustered into unique sequences and assigned taxonomy compared by alignment to the SILVA database v119 (2 3 ). Rarefaction analyses and creation of a community similarity dendrogram, using the sorclass dissimilarity index, were performed in Mothur (2 4 ). Significant differences in the abundances of unique sequences between groups of sample s were tested with the aid of the Phyloseq (1 5 ) and DESeq (1 3 ) packages in R. More details about this method can be found in Chapter 3 ( 34 ) and Appendix A. 100 REFERENCES 101 REFERENCES (1971) Calcium - magnesium carbonate solid solutions from Holocene conglomerate cements and travertines in the Coast Range of California. Geochmi Cosmochim Ac 35 :699 - 718. 2. Brazelton WJ, et al. (2010) Archaea and bacteria with surprising microdiversity show shifts in dominance over 1,000 - year time scales in hydrothermal chimneys. Proc Natl Aca Sci 107:1612 - 1617. 3. Brazelton WJ, Nelson B, Schrenk MO (2012) Metagenomic evide nce for H 2 oxidation and H 2 production by serpentinite - hosted subsurface microbial communities. Front Microbiol 2: doi: 10.3389/fmicb.2011.00268. 4. Brazelton WJ, Morrill PL, Szponar N, Schrenk MO (2013) Bacterial communities associated with subsurface g eochemical processes in continental serpentinite springs. Appl Environ Microb 79:3906 - 3916. 5. Cardace D, et al. (2013) Establishment of the Coast Range ophiolite microbial observatory (CROMO): drilling objectives and preliminary outcomes. Scientific Dri lling 16:45 - 55. 6. Carnevale, DC (2013) Carbon sequestration potential of the Coast Range Ophiolite in 7. Costerton JW, et al. (1994) Biofilms, the customized microniche. J Bacteriol 176:2137 - 2142. 8 . Crespo - Medina M, et al. (2014) Insights into environmental controls on microbial communities in a continental serpentinite aquifer using a microcosm - based approach. Front Microbiol 5:604. 9 . Edwards KJ, McCollom TM, Konishi H, Buseck PR. (2003) Seafloor bioalteration of sulfide minerals: results from in situ incubation studies. Geochmi Cosmochim Ac 67:2843 - 2856. 10 . Huse SM, et al. (2014) VAMPS: a website for visualization and a nalysis of microbial population structures. BMC Bioinformatics 15:41. 1 1 . Kelley DS, et al. (2005) A serpentinite - hosted ecosystem: The Lost City Hydrothermal Field. Science 307:1428 - 1434. 1 2 . Kostka JE, Nealson KH (1995) Dissolution and reduction of magnetite by bacteria. Environ Sci Technol 29:2535 - 2540. 1 3 . Love MI, Huber W, Anders S (2014) Moderated estimation of fold change and dispersion for RNA - Seq data with DESeq2. Genome Biol 15 :550. 102 14 . Makarova KS, et al. (2001) Genome of the extremely rad iation - resistant bacterium Deinococcus radiodurans viewed from the perspective of comparative genomics Microbiol Mol Biol Rev 65:44 - 79. 1 5 . McMurdie PJ, Holmes S (2013) phyloseq: An R package for reproducible interactive analysis and graphics of microbiome census data. PLOS One 8:e61217. 1 6 . McMurdie PJ & Holmes S (2014) Waste not, want not: why rarefying microbiome data is inadmissible. PL oS Computational Biology , 10 (4), e1003531. doi:10.1371/journal.pcbi.100353126. 1 7 . Mitchell AC, Lafrenière MJ, Skidmore ML, Boyd ES (2013) Influence of bedrock mineral composition on microbial diversity in a subglacial environment. Geology DOI: 10.1130 /G34194.1 1 8 . Morrison HG, Grim SL, Vineis JH, Sogin ML (2013) 16S amplicon fusion primers and protocol for Illumina platform sequencing. Figshare. Available online at: http://dx.doi.org/10.6084/m9.figshare.833944 1 9 . Normad P (2006) The families Frankiaceae, Geodermatophilaceae, Acidothermaceae, and Sporichthyaceae. In: The Prokaryotes. 3:669 - 681. 20 . Orcutt BN, et al. (2010) Colonization of subsurface microbial observatories deployed in young ocean crust. ISME J 5:692 - 703. 2 1 . Orcutt BN, Wheat CG, Edwards KJ (2010) Subseafloor ocean crust microbial observatories: development of FLOCS (Flow - through Osmo Colonization System) and evaluation of borehole construction materials. Geomicrobiol J 27:143 - 157. 2 2 . Postec A, et al. (2015) Microb ial diversity in a submarine carbonate edifice from the serpentinizing hydrothermal system of the Prony Bay (New Caledonia) over a 6 - year period. Front Microbiol 6:857. 2 3 . Pruesse E, et al. (2007) SILVA: a comprehensive online resource for quality checke d and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Res 35:7188 - 7196. 2 4 . Schloss PD, Westcott SL (2011) Assessing and improving methods used in operational taxonomic unit - based approaches for 16S rRNA gene sequence analysis. Appl Environ Microb 77:3219 - 3226. 2 5 . Shock EL (2009) Minerals as energy sources for micoorganisms. Economic Geology 104:1235 - 1248. 103 2 6 . Slepova TV, et al. (2006) Carboxydocella sporoproducens sp. nov., a novel anaerobic CO - utilizing/H2 - producing thermophilic bacterium for a Kamchatka hot spring. Int J Syst Evol Micro 56:797 - 800. 2 7 . Sokolova TG, et al. (2002) Carboxydocella thermautotrophica gen. nov., sp. nov., a novel anaerobic CO - utilizing thermophile from Kamchatkan hot spring. Int J Syst Evol Micro 52:1 961 - 1967. 2 8 . Sokolova TG, et al. (2009) Diversity and ecophysiological features of thermophilic carboxydotrophic anaerobes. FEMS Microbiol Rev 68:131 - 141. 2 9 biosph Proc. Natl. Acad. Sci. USA. 103:12115 - 12120. 30 . Suzuki S, et al. (2013) Microbial diversity in The Cedars, an ultrabasic, ultrareducing, and low salinity serpentinizing ecosystem. Proc Natl Acad Sci USA 110:15336 - 15341. 3 1 . Suzuki S, et al. (2014) Physiological and genomic features of highly alkaliphilic hydrogen - utilizing Betaproteobacteria from a continental serpentinizing site. Nat Commun 5:3900. 3 2 . Tiago I, Veríssimo A (2013) Microbial and functional diversity of a su bterrestrial high pH groundwater associated to serpentinization. Environ Microbiol 15:1687 - 1706. 3 3 . Twing KI, et al. (In Review) Serpentinizing fluids from a subsurface observatory harbor extremely low diversity microbial communities adapted to high pH. Proc Natl Acad Sci USA. 34. Twing KI , et al. (In Prep aration ) Microbial diversity of serpentinite cores from the Coast Range Ophiolite Microbial Observatory. Appl Environ Microbiol . 3 5 . Woycheese KM, et al. (2015) Out of the dark: transitional subsurface - to - surface microbial diversity in a terrestrial serpentinizing seep (Manleluag, Pangasianan, the Philippines). Front Microbiol 6:44. 3 6 . Zaremba - Niedzwiedzka K, et al. (2013) Single - ce ll genomics reveal low recombination frequencies in freshwater bacteria of the SAR11 clade. Genome Biology 14:R130. 104 CHAPTER 5 Summary and Future Perspectives Prior to this research, our understanding of life in the continental subsurface of active serpentinite environment s was limited to a few studies based on opportunistic sampling of natural springs a nd seeps ( 1, 2, 10). While these studies yielded surprisingly similar results that helped develop a model of the system (9), it can be diff icult to constrain what is driving microbiological trends from relatively few samples in time and space . The establishment of CROMO, a long - term microbial observatory within an actively serpentinizing subsurface environment , presently consists of two main wells with end - member characteristic fluids . In addition of the two main wells, it has a set of six monitoring wells, drilled to varying depths across geochemical gradients, which allowed for the identification of geochemical drivers (pH, carbon monoxide, and methane concentrations) of the abundant taxa (Chapter 2; 11). The most extreme pH wells exclusively contained a single OTU of Betaproteobacteria and a few OTUs of Clostridia, in addition to genes involved in carbon fixation and the cycling of methane a nd carbon monoxide. Wells with more moderate pH (9 - 11) contained a higher diversity of bacteria and genes for the aerobic oxidation of methane. These data suggest that there are different ecological niches within the serpentinite subsurface environment, wi th hydrogen metabolism and carbon monoxide oxidation taking place in the most end - member fluids and methane oxidation occurring in mixing zones between serpentinite end - member fluids and surface waters (Chapter 2; 11). CROMO represents the first drill camp aign into the continental serpentinite environment and cores of CSW1.1 and QV1.1 permitted microbiological exploration of subsurface serpentinite rocks for the first time. The taxa that appear to be specifically enriched in cores 105 relative to fluids suggest that there may be cryptic chemical cycling (e.g. nitrogen and methane) taking place on subsurface rocks, which may not be evident from fluid samples alone (Chapter 3, 12). Differences in lithostratigraphic zonation within the homogenous cores led to diffe rences in community composition, with archaea (previously undetected at CROMO; Chapter 2, 11) dominating magnetite - bearing serpentine minerals and diverse bacteria dominating clay - containing layers (Chapter 3, 12). Given that unique microorganisms are asso ciated with core materials in comparison to fluids at CROMO, the aim of Chapter 4 was to further explore microbe - mineral interactions through in situ experimentation down well. The results from in situ colonization on mineral substrates within the pH 12.5 well suggest that magnetite and calcite may select for different microbial taxa . These data suggest that core - associated communities are distinct within different lithologies and could be contributing to cryptic biogeochemical cycling within the serpentini te subsurface environment and that specific minerals within those lithologies may be driving community composition (Chapter 3; 12; Chapter 4). Bioinformatic methods, such as DESeq (7), have long been used in to determine the differential expression of RNA data, however, they have not previously been employed on environmental DNA data (8). By statistically assessing differential relative abundance s of 16S rRNA sequences, taxa that were particularly enriched in a given sample type, such as fluids, cores, or mineral - substrates, could be assessed without having to remove all overlapping sequences from the datasets in an effort to control for contamina tion. This method allowed for the identification of mineral - enriched taxa , which could be further explored to help decipher what bio geo chemical cycling may be taking place in distinct geological features within the subsurface environment. 106 Despite all of t he great technological advances in molecular biology and sequencing technologies over the last decade, it is still essential to culture and characterize microorganisms to fully un derstand their metabolic capabilities. There are ongoing culturing efforts to isolate relevant organisms from the serpentinite subsurfac e environment at CROMO and else where . T he data presented here can help inform those efforts by defining what the environmentally - relevant taxa are, as was seen in the study by Crespo - Medina, et al . (4), and identifying geochemical parameters that control their abundance , which may be the key to getting them to grow in the laboratory. The Coast Range Microbial Observatory represents the first drill campaign into the continental subsurface serpentini te environment and allow s for d irect access and study of the actively serpentinizing subsurface environment . With few exception s ( 3, 4, 1 3 ), the research presented here represents some of the first data to come out of the CROMO project and makes a major co ntribution to understanding life within serpentinite fluids and minerals . As CROMO is a large, interdisciplinary project, as more geochemical , geolog ical , and hydrolog ical data from the site become available, it will yield a better understanding of the connectivity of the wells and the system as a whole, which will likely clarify some of the questions still remaining about the distribution and constraints on life within the serpentinite subsurface environment. In the coming months, two large - scale drilling projects will take place to further explore the extent of life within the serpentinite subsurface environment. The International Ocean Discovery Program (IODP) Expedition 357 will drill into the Atlantis Massif, the locatio n of the LCHF, to investigate the extent, diversity, and activity of microbes living within the serpentinizing ocean crust (5). In Oman, the Samail Ophiolite will be drilled to investigate life in the continental subsurface environment there (6). As CROMO was the first drill campaign in the 107 continental serpentinite environment (3), the work completed there, both in terms of the methodology and findings can serve to inform both of these upcoming drilling projects. A great deal of troubleshooting in this proj ect was put into determining the best nucleic acid extraction method for getting DNA out of serpentinite cores, which will help inform molecular methodologies used on upcoming drilling projects. Furthermore, t he data described in Chapter 3 regarding the mi crobial diversity of serpentinite cores can serve as a comparison to rock - hosted organisms found in other continental and marine sites of serpentinization, such as Oman and the Atlantis Massif. 108 REFERENCES 109 REFERENCES 1. Brazelton WJ, Nelson B, Schrenk MO (2012) Metagenomic evidence for H 2 oxidation and H 2 production by serpentinite - hosted subsurface microbial communities. Front Microbiol 2:268. 2. Brazelton WJ, Morrill PL, Szponar N, Schrenk MO (2013) Bacterial communities associated with subsurface geochemical processes in continental serpentinite springs. Appl Environ Microb 79:3906 - 3916. 3. Cardace D, et al. (2013) Establishment of the Coas t Range ophiolite microbial observatory (CROMO): drilling objectives and preliminary outcomes. Scientific Drilling 16:45 - 55. 4. Crespo - Medina M, et al. (2014) Insights into environmental controls on microbial communities in a continental serpentinite aqui fer using a microcosm - based approach. Front Microbiol 5:604. 5. Früh - Green G, et al. Serpentinization and life: biogeochemical and tectono - magmatic processes in young mafic and ultramafic seafloor. IODP Proposal. 758 - Full2. 6. Kelemen P, et al. (2013) Scientific drilling and related research in the Samail Ophiolite, Sultanate of Oman. Scientific Drilling 15:64 - 71. 7. Love MI, Huber W, Anders S (2014) Moderated estimation of fold change and dispersion for RNA - Seq data with DESeq2. Genome Biol 15 :550. 8. McMurdie PJ & Holmes S (2014) Waste not, want not: why rarefying microbiome data is inadmissible. PLoS Computational Biology , 10 (4), e1003531. doi:10.1371/journal.pcbi.100353126. 9. Schrenk MO, Brazelton WJ, Lang SQ (2013) Serpentinization, carbon and deep life. Rev Mineral Geochem 75:575 - 606. 10. Suzuki S, et al. (2013) Microbial diversity in The Cedars, an ultrabasic, ultrareducing, and low salinity serpentinizing ecosystem. Proc Natl Acad Sci USA 110:15336 - 15341. 11. Twing KI, et al. (In Review) Serpentinizing fluids from a subsurface observatory harbor extremely low diversity microbial communities adapted to high pH. Proc Natl Acad Sci USA . 1 2 . Twing KI, et al. (In Preparation) Microbial diversity of serpentinite cores from the Coast Range Ophiolite Microbial Observatory. Appl Environ Microbiol . 1 3 . Wang DT, et al. (2015) Nonequilibrium clumped isotope signals in microbial methane. Science 348:428 - 431. 110 APPENDICES 111 APPENDIX A Example DESeq commands for identifying group specific sequences 3 The input files are (all input files must be saved with unix linebreaks): 1. AllCoreArch.final.count_table file from Mothur (4) 2. AllCoreArch.final.taxonomy table from Mothur (4) 3. Arch_SampleInfo.csv - Sample info file with exact same sample names as count table assigning samples to groups ## Edit the taxonomy table from mothur with this script to make it more readable downstream taxonomy_edit.py [filename].taxonomy ## Output : AllCoreArch.final.taxonomy.rename d.txt ## Create phyloseq (2) otus - tax - sample object in R R library(phyloseq) otus = read.table("AllCoreArch.final.count_table", header=TRUE, row.names=1) otus = otu_table(otus, taxa_are_rows=TRUE) otus = subset(otus, select= - c(1:1)) tax = read.table("AllCoreArch.final.taxonomy.renamed.txt", sep=' \ t', header=FALSE, row.names=1) tax = tax_table(as.matrix(tax)) 3 Workflow developed in the laboratory of Dr. William Brazelton at the University of Utah, Department of Biology. 112 sam = read.csv("Arch_SampleInfo.csv") sam = sample_data(sam) row.names(sam) = sample_names(otus) merged = merge_phyloseq(otus,tax,sam) merged ## Create a bar chart of whole - sample taxonomic class ifications. family level. merged_props = transform_sample_counts(merged, function(x) 100 * x/sum(x)) merged_props_glomV6 = tax_glom(merged_ props, "V6") merged_props_glomV6 ## Output: ## phyloseq - class experiment - level object ## otu_table() OTU Table: [ 155 taxa and 8 samples ] ## sample_data() Sample Data: [ 8 samples by 8 sample variables ] ## tax_table() Taxonomy Table: [ 155 taxa by 8 taxonomic ranks ] ## Use number of taxa ( 155 ) in next command library(ggplot2) cbPalette < - c("#999999", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7") png(filename = "predeseq - barplot - glomV6.png", width = 480, height = 480, units = "px", pointsize = 18) 113 plot_bar(merged_props_glomV6, fill="V6") + coord_flip() + ylab("Percent Sequences") + scale_fill_manual(values=colorRampPalette(cbPalette)(155)) + theme(legend.position='none') ## Identify sample - specific taxa with the DESeq package (1): library("DESeq2") deseqd = phyloseq_to_deseq2(merged, ~CSW_vs_woOL) deseqd = DESeq(deseqd) res = results(deseqd, contrast=c("CSW_vs_woOL","CSW","QV")) res = res[order(res$padj), ] write.csv(res,file='deseq - results.csv') ## Plot of positively and negatively enriched sequences ## Make deseq plot with red points as significant png(filename = "plotMA.png", width = 480, height = 480, units = "px", pointsize = 18) plotMA(res, ylim = c( - 3,3)) dev.off() ## red points are "signi ficant at 10% false discovery rate (padj)" - DESeq - vignette.pdf # # S elect the group - specific taxa, find the taxonomy information associated with them, and ## record this information as R objects and write them to files so that we can open them as tables # # First: select significantly up - or down - "regulated" taxa, including very large differences where ## padj=NA: 114 bigpos = res[(res$log2FoldChange > 0.9), ] bigneg = res[(res$log2FoldChange < - 0.9), ] res_no_na = res[order(res$padj, na.last=NA),] sigres = res_no_na[(res_no_na$padj < 0.1),] sigrespos = sigres[(sigres$log2FoldChange > 0),] sigresneg = sigres[(sigres$log2FoldChange < 0),] bigpos_names = rownames(res) %in% rownames(bigpos) bigneg_names = rownames(res) %in% rownames(bigneg) sig_names = rownames(res) %in% rownames(sigres) all_names = bigpos_names | bigneg_names | sig_names other = res[!all_names,] # # Add taxonomy data to the selected results bigpos_tax = cbind(as(bigpos, "data.frame"), as(tax_table(merged)[rownames(bigpos), ], "matrix")) bigneg_tax = cbind(as(bigneg, "data.frame"), as(tax_table(merged)[rownames(bigneg), ], "matrix")) sigrespos_tax = cbind(as(sigrespos, "data.frame"), as(tax_table(merged)[rownames(sigrespos), ], "matrix")) sigresneg_tax = cbind(as(sigresneg, "data.frame"), as(tax_table(merged)[rownames(sigresneg), ], "matrix")) other_tax = cbind(as(other, "data.frame"), as(tax_table(merged)[rownames(other), ], "matrix")) 115 ## Add count table information to the selected results, for future reference in the output ## files big pos_tax_counts = cbind(as(bigpos_tax, "data.frame"), as(otu_table(merged)[rownames(bigpos_tax), ], "matrix")) bigneg_tax_counts = cbind(as(bigneg_tax, "data.frame"), as(otu_table(merged)[rownames(bigneg_tax), ], "matrix")) sigrespos_tax_counts = cbind(as(s igrespos_tax, "data.frame"), as(otu_table(merged)[rownames(sigrespos_tax), ], "matrix")) sigresneg_tax_counts = cbind(as(sigresneg_tax, "data.frame"), as(otu_table(merged)[rownames(sigresneg_tax), ], "matrix")) other_tax_counts = cbind(as(other_tax, "data. frame"), as(otu_table(merged)[rownames(other_tax), ], "matrix")) ## Tables of DESeq results broken down by positive, negative, and other ## Merge all sig results including NAs with huge log2changes, skipping duplicate rows ## Write the results for future reference duprows_pos = rownames(bigpos_tax_counts) %in% rownames(sigrespos_tax_counts) duprows_neg = rownames(bigneg_tax_counts) %in% rownames(sigresneg_tax_counts) allpos = rbind(data.frame(bigpos_tax_counts[!duprows_pos,]),data.frame(sigrespos_tax_count s)) allneg = rbind(data.frame(bigneg_tax_counts[!duprows_neg,]),data.frame(sigresneg_tax_counts)) write.csv(allpos,file='deseq - allpos - results.csv') 116 write.csv(allneg,file='deseq - allneg - results.csv') write.csv(other_tax_counts,file='deseq - other - results.csv') ## Make phyloseq objects with OTU counts and taxonomy for manipulation in R later otus_pos = otus[row.names(allpos), ] otus_neg = otus[row.names(allneg), ] otus_other = otus[row.names(other), ] physeq_pos = merge_phyloseq(otus_pos,tax) physeq_neg = merge _phyloseq(otus_neg,tax) physeq_other = merge_phyloseq(otus_other,tax) ## Rename sample names and merge into a single phyloseq object sample_names(physeq_pos) = paste("pos - ", sample_names(physeq_pos), sep="") sample_names(physeq_neg) = paste("neg - ", sample _names(physeq_neg), sep="") sample_names(physeq_other) = paste("other - ", sample_names(physeq_other), sep="") physeq_all = merge_phyloseq(physeq_neg, physeq_other) ## The next set of commands will merge the counts for the positive, negative, and other ## categories of samples. Edit the first command so that it lists the correct number of ## pos, neg, and other. Check sample_names(physeq_all) to be sure: sample_names(physeq_all) 117 types = c('pos','pos','pos','pos','pos','pos','pos','pos','neg','neg','neg','n eg','neg','neg','neg','neg','oth er','other','other','other','other','other','other','other') sampledata = sample_data(data.frame(Type = types, size = nsamples(physeq_all), replace = TRUE, row.names = sample_names(physeq_all), stringsAsFactors = FALSE)) sam pledata physeq_all_merged = merge_phyloseq(physeq_all,sampledata) physeq_all_merged = merge_samples(physeq_all_merged, 'Type') physeq_all_merged_props = transform_sample_counts(physeq_all_merged, function(x) 100 * x/sum(x)) ## Create Bar charts of taxonomy specific to positive, negative, and other sequences ## Family - level plot: physeq_all_merged_props_glom6 = tax_glom(physeq_all_merged_props, "V6") library(ggplot2) cbPalette < - c("#999999", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7") png(filename = "postdeseq - barplot - glom6.png", width = 480, height = 480, units = "px", pointsize = 18) ## Use the same number of taxa as before ( 155 ) plot_bar(physe q_all_merged_props_glom6, fill="V6") + coord_flip() + ylab("Percent Sequences") + scale_fill_manual(values=colorRampPalette(cbPalette)(155)) + theme(legend.position='none') 118 dev.off() ## Output 6: Tables of results corresponding to bar charts above ## Writ e csv file containing otu counts and taxonomy to a id in interpretation of the bar plot. The ## t() transposes the otu_table so that species are rows, as in the tax table. final_tax_otu_props_glom6 = cbind(as(t(otu_table(physeq_all_merged_props_glom6)), "ma trix"), as(tax_table(physeq_all_merged_props_glom6), "matrix")) write.csv(final_tax_otu_props_glom6,file='final_tax_otu_props_glom6.csv') 119 REFERENCES 120 REFERENCES 1. Love MI, Huber W, & Anders S (2014) Moderated estimation of fold change and dispersion for RNA - Seq data with DESeq2. bioRxiv , 1 21. doi:10.1101/002832 2. McMurdie PJ & Holme s S (2013) phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PloS One , 8 (4), e61217. doi:10.1371/journal.pone.0061217 3. McMurdie PJ & Holmes S (2014) Waste not, want not: why rarefying microbiome data is inadmissible. PLoS Computational Biology , 10 (4), e1003531. doi:10.1371/journal.pcbi.1003531 4. Schloss PD, et al. (2009) Introducing mothur: Open Source, Platform - independent, Community - supported Software for Describing and Comparing Microbial Communities. Appl Environ Microbiol AEM.01541 09. doi:10.1128/AEM.01541 - 09 121 APPENDIX B Bioinformatic a nalyses for Crespo - Medina et al ., 2014 4 Summary of Crespo - Medina et al., 2014 and my bioinformatic contributions In Crespo - Medina et al . (2) , fluids from CROMO wells were incubated in microcosms with various carbon sources and the addition of nutrient and/or electron acceptors. Methane and acetate additions led to the enrichment of a Betaproteobacterium of the family Comamonadaceae. While the a ddition of electron acceptors was not necessary for growth, it did and in serpentinite subsurface systems as a whole, the 16S rRNA sequences from the Figure B.1 Diversity of 16S rRNA gene OTUs (97% sequence similarity) from fluids collected at CROMO in March 2013 was analyzed using mothur ( 7 ). A) Alpha - diversity, as displayed by rarefaction, indicates the greatest within sample diversity in the control sample of N08C and the least within sample diversity in the extreme sample of CSW1.1. Within sample di versity of field replicates and T0 from both N08B and CSWold show similar patterns. B) Beta - diversity, displayed in a community - dissimilarity dendrogram calculated from the Morisita - Horn index, indicates that samples from a single well are more similar to each other than samples from different wells. 4 The work desc ribed here was published in Frontiers in Microbiology in 2014: Crespo - Medina M, Twing KI, Kubo MDY, Hoehler TM, Cardace D, McCollom T, Schrenk MO (2014) Insights into environmental controls on microbial communities in a continental serpentinite aquifer us ing a microcosm - based approach. Front Microbiol 5:604 122 microcosms needed to be compared to environmental sequences. I performed the bioinformatic analyses to assess the bacterial diversity of environmental samples (Fig. B.1, B.2, & B.3) and comparisons of 16S rRNA genes to sequences from other serpentinite sites to determine closest relatives (Tables B.1 & B.2). Bioinformatic Results The microbial diversity at the class level varies by sampling well, with the most extreme well, CSW1.1, exhibiting the lowes t diversity and the moderate well N08C exhibiting high diversity (Fig. B.1 A ). Field replicate samples from the wells N08B and CSWold show nearly identical community composition to one another, while the T0 samples from both those wells, filtered merely day s later immediately before the beginning of the microcosm experiments, Figure B.2 Proportion of bacterial classes present in fluids collected at CROMO in March 2013 . OTUs were made at 97% similarity level and assigned taxonomy from the SILVA database using mothur (Pruesse et al , 2007; Schloss & Westcott, 2011). The CSWold and N08B T0 samples show a slight shift in community composition compared to their field replica te counterparts. 123 Figure B.3 Diversity of OTUs (97% similarity) belonging to the classes Betaproteobacteria and Clostridia within CSW1.1, N08B, and CSWold . The Betaproteobacteria in CSW1.1 and N08B are dominated by a single OTU (OTU001) belongin g to the family Comamonadaceae. The same OTU makes up 45% of the Betaproteobacteria in CSWold. display slight shifts in community composition (Fig. B.1B). The total community of CSW1.1 is 92% of a single OTU (OTU001) of Betaproteobacteria belonging to the family Comamonadaceae (Figs. B.2 and B.3). Forty - four percent of the total community of N08B_T0 is made up of Betaproteobacteria, also all belonging to OT U001. Only 6.5% of the total community of CSWold_T0 is made up of Betaproteobacteria (Fig. B.2) and among them 45% are OTU001, 40% are Hydrogenophaga (also of the family Comamonadaceae) and 15% are an OTU belonging to the family Rhodocyclaceae (Figs. B.2 a nd B.3). Clostridia, which make up 60% of CSWold_T0 total community, are dominated by a single OTU of Dethiobacter (OTU002) (Figs. B.2 and B.3). The same OTU makes up 18% and 124 9% of the clostridial OTUs in CSW1.1 and N08B_T0, respectively (Fig.B.3). Instead 81% of sequences related to Clostridia in CSW1.1 belong to OTU009, which can only be classified to Clostridia belong to a different OTU of unclassified Clostridiales (Fig. B.3). Table B.1 Percent identity between end point sequences from March microcosms, most abundant OTUs from tag sequence data, and reference sequences The [microcosm] sequences from CSWold and N08B were 97.3 and 98.9 similar, respectively, to an environmental sequence from another continental serpentinizing site in Portugal ( 10 ) (Table B.1 ). Two representative sequences [from the August microcosms] , one from CSW1.1 and one from N08B microcosm , with CO 2 and acetate respectively, shared 100% identity to Comamonadaceae bacterium B1 , isolated from the Cedars ( 9; Table B. 2 ). Table B.2 Percent identity between two representative end point sequences from the August microcosms, most abundant OTUs from tag sequence data, and reference sequences 125 16S rRNA Tag Sequencing and Data Analysis Methods DNA samples were submitted (MBL) Josephine Bay Paul Center for 16S rRNA amplicon sequencing of the V4V5 region of the 16S rRNA gene on an Illumina MiSeq instrument ( 3, 4 ). Sequences were subjected to quality control and preprocessing in accor on their VAMPS server ( 5 ). Sequence reads were aligned to the SILVA SSURef alignment (v102) and taxonomic classification was assigned using mothur ( 6 ; 8 ). Sequences were clustered into operational taxonomic units (OTUs) at the 3% distance threshold using the cluster.split command and the average - neighbor clustering algorithm in mothur ( 7 ). Alpha - diversity (within sample) was assessed by rarefaction analysis, while beta - diversity (between sample) of the microbial communities was assessed by calculation of the Morisita - Horn index and displayed in a dissimilarity dendrogram produced with the tree.shared command in mothur ( 7 ). MatGAT ( 1 ) was used to determine the percent sequence identify of the most abundant OTUs compared with 16S rRNA sequences, trimmed to the V4V5 region of the gene, from isolates within this study and from previous studies of the serpentinite subsurface ( 9, 10 ). 126 REFERENCES 127 REFERENCES 1. Campanella JJ, Bitincka L, Smalley J (2003) MatGAT: an application that generates similarity/identity matrices using protein or DNA sequences. BMC Bioinformatics 4:29. 2. Crespo - Medina M, et al. (2014) Insights into environmental controls on microbial communities in a continental serpentinite aquifer using a microcosm - based approach. Front Microbiol 5:604. 3. Huse S M , et al . (2014). Comparison of brush and biopsy sampling methods of the ileal pouch for assessment of mucosa - associated microbiota of human subjects. Microbiome . 2, doi: 10.1186/2049 - 2618 - 2 - 5. 4. Huse SM, et al . (2014). VAMPS: a website for visualization and analysis of microbial population structures. BMC Bi oinformatics . 15, doi: 10.1186/1471 - 2105 - 15 - 41. 5. Morrison HG, Grim SL, Vineis JH, Sogin ML. (2013). 16S amplicon fusion primers and protocol for Illumina platform sequencing. figshare . http://dx.doi.org/10.6084/m9.figshare.833944 6. Pruesse E, et al . (2007). SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acid Research . 35, 7188 - 7196. 7. Schloss PD, et al . (2009). Introducing mothur: open - source, platform - independent, community - supported software for describing and comparing microbial communities. AEM . 75, 7537 - 7541. 8. Schloss PD , Westcott SL (2011). Assessing and improving methods used in operational t axonomic unit - based approaches for 16S rRNA gene sequence analysis. AEM . 77, 3219 - 3226. 9 . Suzuki S, et al. (2014) Physiological and genomic features of highly alkaliphilic hydrogen - utilizing Betaproteobacteria from a continental serpentinizing site. Nat Commun 5:3900. 10 . Tiago I, Veríssimo A (2013) Microbial and functional diversity of a subterrestrial high pH groundwater associated to serpentinization. Environ Microbiol 15:1687 - 1706.