Interactions of Aquificae with Mercury And

Home , Aquifex aeolicus, Aquificae, Chloroflexi (phylum), Hydrogenobacter, Persephonella marina, Thermotoga

MICROBIALLY MEDIATED MERCURY DETOXIFICATION IN GEOTHERMAL

ENVIRONMENTS: INTERACTIONS OF AQUIFICAE WITH MERCURY AND

EVIDENCE FOR PHYLOGENETIC NICHE CONSERVATISM IN YELLOWSTONE

NATIONAL PARK HOT SPRINGS

ZACHARY FREEDMAN

A dissertation submitted to the Graduate School-New Brunswick

Rutgers, The State University of New Jersey

In partial fulfillment of the requirements

For the degree of

Doctor of Philosophy

Graduate Program in Ecology and Evolution

Written under the direction of

Professor Tamar Barkay

And approved by

______

New Brunswick, New Jersey

January 2012

ABSTRACT OF THE DISSERTATION

Microbially Mediated Mercury Detoxification in Geothermal Environments:

Interactions of Aquificae with Mercury, and Evidence for Phylogenetic

Niche Conservatism in Yellowstone National Park Hot Springs

By ZACHARY FREEDMAN

Dissertation Director:

Dr. Tamar Barkay

Geothermal features are generated by leaching of minerals and metals as superheated water flows through cracks and fissures in Earth’s crust. As this water reaches the surface, chemical, pH, and temperature gradients are created that drive life in these environments. Geothermal environments are often enriched with toxic metals, e.g. mercury (Hg), the focus of this dissertation. Resistance to toxic Hg(II) is controlled by the enzyme mercuric reductase (MR), which catalyzes Hg(II) reduction to Hg(0). The gene encoding MR, merA, is part of the mercury resistance (mer) operon, which at minimum includes genes encoding transport, enzymatic, and regulatory functions. The primary objective of my research was to achieve better understanding of biotic transformations that modulate Hg toxicity in geothermal environments. I characterized

Hg-resistance in Aquificae, dominant primary producers in geothermal environments, and investigated the diversity and distribution of Hg-resistance genes in geochemically

diverse hot springs in Yellowstone National Park (YNP), and 23 assemblages on a global scale.

Two strains of Aquificae were obtained; Hydrogenobaculum sp. Y04AAS1

(AAS1) and Hydrogenivirga sp. 128-5-R1-1 (R1-1). Genome sequencing revealed homologous sequences to merA, and alignment of putative Hg-resistance genes, MerA,

MerT (Hg(II) transporter) and MerP (periplasmic scavenger), reveal homology with the mer system of Tn501. Characterization of mer in AAS1 and R1-1 include growth at Hg concentrations >10 μM Hg(II), loss of Hg(II) from the growth medium, validation of

Hg(0) production, and MR enzyme activity; mer induction was not observed, suggesting lack of regulatory function.

Microbial mat biomass was collected from Bijah and Succession Springs, YNP, and environmental merA sequences were obtained from GenBank to determine the ecological controls on Hg-resistant communities in YNP hot springs, and on a global scale. merA assemblages exhibited grouping within each community, and total sequence pool, as indicated by positive net relatedness index and nearest taxon index values, respectively. Cluster analyses reveal different clustering patterns of 16S rRNA and merA gene assemblages from YNP, suggesting unique controls on 16S rRNA and merA gene community structure. Meta-analysis of merA communities from 23 assemblages encompassing 782 environmental sequences reveal clustering based on sample location, suggesting that geography structures Hg-resistant communities.

iii

Acknowledgements

I would like to extend my deepest thank you first and foremost to my major professor, Dr. Tamar Barkay, who always made time to discuss ideas, give advice and criticism, all while motivating me with her support and encouragement. I am extremely thankful to the members of my dissertation committee: Dr. Tamar Barkay, Dr. Eric Boyd,

Dr. John Dighton, Dr. Max Haggblom, Dr. Anna-Louise Reysenbach, and Dr. Costantino

Vetriani for their guidance and support during the generation and completion of my dissertation work. I am very thankful to Drs. Theodore Chase, Max Haggblom, Anna-

Louise Reysenbach, and Costantino Vetriani for their good advice and making their laboratory resources available to me, and to Drs. Eric Boyd and John Dighton for advice and guidance in forming my dissertation project.

I am grateful to past and present members of the Barkay lab; Aspa, Riqing,

Sharron, Melitza, Kim, Chu-Ching, Yanping, Heather, and Kritee for their help, direction, critiques, and lively discussions over the years. Thank you to the many undergraduate students who performed research in the Barkay lab, and contributed to research presented in this dissertation; Maribeth Armenio, Anthony DiBattista, Hunter

Hao, and Tzeh Keong Foo. Also to Chengsheng Zhu for his work with Bayesian analysis of MerA presented in chapter 2.

I want to thank and Yitai Liu and Gilbert Flores in the Reysenbach lab for guiding me on how to properly culture the strains of Aquificae included in this dissertation. I would like to extend my gratitude to Ileana Perez-Rodriguez for invaluable advice on media preparation, and a good friend and colleague who always kept the bar high.

I would have been lost without the guidance of Marsha Morin, Eileen Glick,

Arleen Nebel, Kathy Maguire, and Jesse Maguire, whose far-reaching institutional knowledge and assistance made my time at Rutgers far more enjoyable. My most sincere thank you to other faculty members, in Lipman Hall, ENR and elsewhere who have provided much help and assistance; Drs. Bill Belden, Jeff Boyd, Joan Ehrenfeld, Doug

Eveleigh, Karl Kjer, Julie Lockwood, Rebecca Jordan, Peter Morin, John Reinfelder,

Peter Smouse, Gavin Swiatek, and Nathan Yee. Also a big thank you to my friends which made the roller-coaster of graduate school an enriching and fun time; Aabir, Allie,

A. J., Andrea, Ben, Blake, Brandon, Brian, Charlie, Curtis, David, Dom, Elena, Holly,

Faye, Isabel, James, Jess, Josie, Norah, Orion, Robbie, Tiff, Sean, and Wes.

I want to acknowledge funding sources that made this dissertation possible; the

Yellowstone National Park Research Coordination Network, National Science

Foundation GK-12 teaching fellowship, Robert A. and Eileen S. Robison Award,

Department of Ecology and Evolution Small Grants, and the Graduate School for travel assistance.

Last, but most definitely not least, I want to extend my deepest and most sincere gratitude to my family and loved ones that have provided a great amount of unconditional support throughout my time in graduate school. To my parents, Stephen and Eileen who have taught me what it means to not only work hard, but to do it with dedication and enthusiasm. To my brother Noah, who will never call me Dr., and whose time in Los

Angeles, Chicago, and road trips to Vegas have provided great retreats over the last 6 years. And to Kate, whose never-ending caring, patience, and understanding has kept me strong and made life far more enjoyable for the past year and a half.

Dedication

I want to dedicate this thesis to my parents, Eileen and Stephen, who always believed I could succeed if I put my mind to it, even despite what they may have been told at parent/teacher conferences. To my brother Noah, who always seems to have the right words at the right time, and to my girlfriend Kate, for all her support, love, and patience.

Table of Contents

Pages

Abstract ...... ii

Acknowledgments ...... iv

Dedication ...... vi

List of tables ...... viii

List of figures ...... ix . Chapter 1 - Introduction ...... 1 - 15

Chapter 2 – Characterization of mer-mediated mercury resistance in Hydrogenobaculum sp. Y04AAS1 and Hydrogenivirga sp. 128-5-R1-1...... 16 - 55

Chapter 3 – Diversity and distribution of merA in geothermal environments and on a global scale: novel insights into the ecological structure of mercury resistant communities...... 56 - 93

Chapter 4 – Summary and Conclusions ...... 94 - 108

References ...... 109 - 112

vii

List of Tables Page

Table 2.1: Disassociation constants used in MINEQL+ modeling. Adapted from Crespo-Medina et al. (32)………………………………………………………...20

Table 2.2: Primers sets used in qPCR of merA and gyrA in strains AAS1 and R1-1.....26

Table 2.4 Results of MINEQL+ modeling of Hg speciation in Aquificae culture 2- media with thiosulfate (S2O3 ) and hydrogen (H2) as sole energy sources…………….36

Table 2.5: Reduction of Hg(II) to Hg(0) by Aquificae cultures……………………….42

Table 3.1: Geochemical parameters of study sites in two YNP hot springs selected for this study……...... 61

Table 3.2: merA-specific PCR primer sets. Adapted from Wang et al. (156)...... 61

Table 3.3: merA PCR primer sets used to obtain sequences included in the global meta-analysis....………………………………………………...... 68

Table 3.4: PCR Amplification of merA from YNP mat DNA extracts ……………….70

Table 3.5: merA clone library composition……………………...………………...…..71

Table 3.6: Phylogenetic diversity metrics for gene assemblages recovered from mat microbial communities in Bijah (B) and Succession (S) springs, YNP...... 78

Table 3.7: Phylogenetic diversity metrics for gene assemblages recovered from microbial communities...... ……………...... 82

viii

List of Figures Page Figure 1.1: The Mercury Cycle. Adapted from Lin et al. (78)………………………..8

Figure 1.2: The Microbial Hg Detoxification System………………………………....9

Figure 1.3: Bayesian inferred phylogenetic reconstruction of MerA deduced amino acid sequences. Adapted from Barkay et al.(8)………………………………...12

Figure 2.1: Bayesian inferred phylogenetic reconstruction of MerA deduced amino acid sequences………………………………………………………………...... 29

Figure 2.2: Alignment of putative Mer proteins, including MerA, MerT, and “MerP” from AAS1 and R1-1, and the well characterized Mer proteins of Tn501...... 32

Figure 2.3 Putative mer operons from Hydrogenobaculum sp. Y04AAS1, Hydrogenivirga sp. 128-5-R1-1 with the Tn501 mer operon from Pseudomonas aeruginosa as a reference……………………………………………………………....33

Figure 2.4: Growth with S2O3 and H2 for strain R1-1, AAS1, and H1…..……….……37

Figure 2.5 Mercury resistance profiles of Hydrogenobaculum sp. Y04AAS1 grown with S2O3 and H2, and Hydrogenivirga sp. 128-5-R1-1 grown with S2O3…...…38

Figure 2.6: 203Hg(II) remaining in the medium during growth of strains AAS1 and R1-1 using S2O3 and H2 as electron donors……...... ………………….....39

Figure 2.7: Reduction of Hg(II) to Hg(0) by strain AAS1 and R1-1………………….41

Figure 2.8: Effect of temperature on specific MR activity of AAS1 and R1-1……….43

Figure 2.9: merA induction fold for strain R1-1, and AAS1.………………………...45

Figure 2.10: Effect of growth in presence or absence of Hg on MR activities in crude cell extracts of strains AAS1 and R1-1………………………………..……...46

Figure 2.11: merA induction fold of Aquificae cultures compared with previously characterized strains HB27 and Tn501……………….....……………………...……...54

Figure 2.12 Effect of temperature on MerA activities of strains AAS1, and R1-1 compared with previously characterized Thermus thermophilus HB27 and Escherichia coli Tn501...... ….55

Figure 3.1: Selected geothermal springs in Yellowstone National Park. Adapted from Wang et al. (156)………………………………………………...……..59

Figure 3.2: Rarefaction analysis of 16S rRNA gene and merA clone libraries..…...... 72

Figure 3.3: Bayesian phylogram of 16S rRNA gene sequences recovered from microbial communities of Bijah Spring 1, and 2, and Succession Spring 1, and 2 are shown for each unique OTU (97% similarity)…………………...... 74

Figure 3.4: Bayesian phylogram of MerA clones. Representative clones for sites Bijah Spring 1, 2 and Succession Spring 1, 2 representing each unique OTU (99% sequence similarity)...... 76

Figure 3.5: Phylogeny distance based cluster analysis of 16S rRNA and merA gene sequences recovered from Bijah and Succession springs as determined using hierarchical cluster analysis and UniFrac...... ………………………………...80

Figure 3.6: Non-metric multidimensional scaling (NMDS) plot of merA gene assemblages listed in table 3.7...... 83

x 1

Chapter 1 - Introduction

I) Geothermal Environments

Geothermal environments are regions in which thermal energy radiated from Earth’s core provide heat and energy, which may sustain life. These environments include thermal soils, steam vents (fumaroles), mud pots, geysers, hot springs, as well as deep-sea hydrothermal vents. Geothermal waters are generally far from thermodynamic equilibrium as compared to Earth surface waters, which results in defined chemical and thermal gradients that drive a suite of abiotic and microbially mediated reactions (82).

Yellowstone National Park and its Geothermal Features

Yellowstone National Park (YNP) is a truly remarkable environment. The park is home to over 10, 000 hot springs, fumaroles, and mud pots that comprise half of the world’s geothermal features as well as two-thirds of the world’s geysers. In addition to this, Yellowstone is also one of the only essentially undisturbed geothermal areas left in the world (46). In other thermally active regions of the world such as Iceland and New

Zealand, much of the geothermal environs have been altered for human use (135). YNP is set on a high volcanic plateau near a region of active crustal extension (46). It is thought that the Yellowstone region has been volcanically active for 2.2 million years.

Although the region has been continuously reshaped throughout it’s existence, it has been dramatically altered by three major caldera-forming eruptions that occurred about 2.1,

1.3, and 0.4 million years ago, and many of these eruptions consisted of rhyolitic lava flows (46).

The major focus of this project is on one type of thermal feature, the hot spring, which is formed when water seeps through layers of porous rock. The water then

2 continues through cracks and fissures caused by volcanic activity until reaching a depth of over 3,000 meters where it is superheated by geothermal energy reaching temperatures in excess of 400° C (45). Because this superheated water is less dense then the cooler water seeping down, it begins to rise; following cracks, fissures and weak spots in the

Earth’s crust until it meets the surface. In the presence of rhyolite, the high temperatures of the water solubilize silica. This solubilized silica precipitate to form geyserite or sinter near the surface. If the underground system of cracks and fissures creates pockets of highly pressurized water at the surface, this pressure is periodically released and a geyser is formed. If not, the water will create a hot spring (45).

The chemistry of hot spring outflow is affected by physical gradients, inorganic chemical reactions and biological activity. These hot water systems are dominated by solutes (chloride, silica) that stay in solution during boiling as well as major cations

(Ca2+, Mg2+, Na+, K+) (100). Another type of geothermal feature are vapor dominated hot springs which are created when hot gasses mix with the surface waters to form hot, acidic

pools that have a high relative abundance of H2S, CO2, NH3, N2, and Hg (100). When hot water or steam reaches the surface, several gasses can be released due to high partial

pressures and super-saturation compared to atmospheric conditions. CO2 and H2S are two of the dominant gasses discharged by spring water as it meets the surface (162). CO2 loss can lead to a rise in spring pH as well as an increase in calcite precipitation. H2S oxidation ultimately leads to the creation of sulfuric acid that causes most of the acidic waters in Yellowstone. Fe and As are also found in spring water and can highly effect the spring chemistry (57).

Deep-Sea Hydrothermal Vents

Deep-sea hydrothermal vents are another type of geothermal environment, normally associated with spreading centers along mid-ocean ridge systems. These unique systems were initially discovered along the Galapagos Rift (79). Later, more vents were found along the East Pacific Rise (145), Mid-Atlantic Ridge (121), and Lau Basin (44) among others as recently discussed by Shock and Canovas (138). Hydrology differs in these vents as compared to hot springs, as seawater is drawn in to the Earth’s crust, becomes superheated at temperatures greater than 400 °C, and reacts with crustal basalts resulting in superheated seawater saturated with reduced inorganic compounds (30, 36).

As the hot water rises and reaches the seafloor where it may interact with ambient seawater (~2 °C), precipitation of reduced inorganic compounds, minerals, and metals occurs which drives the metabolism of the organisms that inhabit these unique environments. Hydrothermal vents are characterized by the temperature of the released hydrothermal fluid; diffuse flow vents exhibit temperatures ranging from ~2 to 60 °C and focused flow vents or “black smokers” can reach temperatures as high as 350 °C (72, 89).

Geothermal Environments and Mercury

Geothermal features are known geologic sources of Hg to the environment (22,

74, 156). YNP geothermal areas have extremely high fluxes of Hg; Roaring Mountain exhibited fluxes in Hg emissions of over 2000 ng/m2 hr, while other sites ranged 200% over time from the average surface to air Hg flux (1). Elevated levels of total Hg (THg) have been measured in the biomass of microbial mats in Yellowstone hot springs (22,

74). Furthermore, microbial mats have been proposed as mechanisms for bioaccumulation of Hg in fly larvae and birds (22). THg in hot spring water was found to

4 be variable, ranging from 600 ng/L to less then 10 ng/L. Methyl mercury (MeHg) was also found in hot spring waters in low levels, however both THg and MeHg were shown to bioaccumulate in the microbial mat (22, 74). Hg has also been found in the soils of

YNP; distribution of Hg in YNP soils is spotty with Hg concentrations 4 orders of magnitude higher in proximity to geothermally active areas as compared to background soil Hg concentrations (109). How these high concentrations of Hg affect life in YNP is largely unknown.

II) Microbial Communities in Geothermal Environments

Yellowstone National Park may be home to grizzly bears, bison, and proghorn, but perhaps the most interesting and unique of Yellowstone’s inhabitants are much smaller. Yellowstone’s thermal features are home to an array of thermophilic microorganisms that have evolved over time to thrive in this unique extreme environment. These bacteria are particularly interesting because their habitats may resemble volcanic environments that were thought to exist on early Earth (88). Many of the microbes found in Yellowstone’s hot springs belong to lineages whose origins are found near the root of the “tree of life”, and who may have played a role in altering conditions on Earth (16, 88). The study of hot springs has yielded industrial benefits.

For example, thermostable proteins found in thermophilic bacteria have been proven useful in research applications (e.g., Taq DNA polymerase from Thermus aquaticus).

Hot springs have also been suggested as models for extraterrestrial life (88).

Our understanding of bacterial diversity has flourished with the study of geothermal systems through both enrichment (20, 35, 51) and culture-independent

5 techniques (17, 68, 82). Two of the most ancient bacterial lineages, Thermus (25) and

Aquifex (34) were discovered in hot springs.

Many types of metabolism are observed in Yellowstone hot springs. Near the spring source, chemolithoautotrophic Bacteria and Archaea have been shown to dominate the microbial communities present (17, 88, 158). Common reduced chemical species that

drive chemoautotrophy in hot springs include H2, H2S, As(III) and Fe(II). As the water cools away from the source, chemoheterotrophic or mixotrophic microbes can be observed (88, 158). As the water further cools to more moderately thermophilic/mesophilic temperatures (below 52° C), eukaryotic organisms such as

Zygogonium sp. and Cyanidium sp. may begin to flourish and create dense mats through photosynthetic metabolism (22).

The tendency of a lineage to retain ancestral ecological traits and environmental distribution is known as niche conservatism (163). Evidence for niche conservatism can be obtained by observable defined patterns in the distribution of phylotypes within assemblages along a physicochemical gradient (phylogenetic structure) (59, 160).

Geothermal springs in YNP are geochemically diverse and exhibit physical and chemical gradients within (82) and among (100, 139) geothermal features. Many YNP springs have also been shown to have elevated Hg concentrations (22, 74, 156). These conditions provide an ideal environment in which the effect of physical/chemical parameters on the phylogenetic structure of Hg-resistant communities can be observed.

Additionally, the aforementioned physical and chemical gradients create strong selective pressures that may result in the presence of an assortment of ecological niches, with some

6 locations supporting the growth of certain Hg-resistant microorganisms and limiting the growth of others.

There has been much research regarding the effect of abiotic factors on hot spring microbial communities. Mathur et al. (88) showed that mineral chemistry and metabolic potentials played a dominant role in shaping acidic hot spring communities, while King et al. (74) attributed this role to spring temperature and pH. Mineral chemistry can be affected by pH and temperature, however, so these observations may not be mutually exclusive. More recently, the effects of abiotic factors on the distribution of Hg resistant microbes and mer genes in hot springs have been studied. Wang et al. (156) researched the effect of abiotic factors on the distribution of merA in YNP hot springs. Temperature, pH, and concentrations of sulfide, total Hg and MeHg in microbial mats all affected merA distribution in hot springs, however these trends were not well supported statistically leading to the study that is presented in chapter 3. In this study I investigated the degree to which Hg-resistant communities were structured by the physical and chemical gradients in YNP geothermal hot springs.

III) Microbial interactions with Hg – Sources, Function and Evolution.

Mercury (Hg) is a highly toxic heavy metal, and the only metal that is liquid at room temperature. Mercury has historically been used in barometers, thermometers, fluorescent lamps, and dental amalgam. Since the 1960’s, Hg has gained much attention due to human health concerns centered around the bioaccumulation of MeHg in fish, since Hg is the only metal known to increase in concentration in all trophic levels of the aquatic food chain (22, 84).

Mercury is found in three oxidation states, 0, +1, and +2. Natural concentrations of Hg in the Earth’s crust range from 21 ppb in the lower crust to 56 ppb in the upper crust where Hg is found in elemental form as well as in HgS type minerals such as cinnabar, metacinnabar and hypercinnabar (10). Anthropogenic sources account for at least half of Hg emissions to the atmosphere (24): these inputs include burning of fossil fuels, disposal of products such as light fixtures, batteries, and electrodes, as well as emissions from industrial buildings and landfills (10). Geothermal features are known sources of Hg to the environment, with Hg being present in both gaseous and aqueous forms (74). Recent UN estimates show that approximately one-third to one-half of Hg emissions to the atmosphere come from geothermal sources (24). Furthermore, geothermal areas have been show to produce the 2nd highest area-averaged Hg fluxes to the atmosphere lagging only behind some mine-waste sites that emit greater than 10,000 ng Hg/m2 * hr (1). Water discharged from YNP geothermal features can reach Hg concentrations of nearly 600 ng/L, however most features tested have Hg concentrations less than 100 ng/L (74).

Figure 1.1: The Mercury Cycle. Blue and gray arrows indicate biotic, and abiotic transformations, respectively. The boxed transformation, reduction of Hg(II) to Hg(0) by MR indicated the transformation which is the focus of this study. Adapted from Lin et al. (78).

Metal resistance is common among microorganisms and is critical for the impact of metals in the environment. Microbes must have been exposed to toxic heavy metals since the oxygenation of the biosphere and have evolved diverse mechanisms to live in the presence of high concentrations of toxic metal ions (141). These mechanisms, such as efflux, intra- or extra-cellular precipitation, and enzyme-mediated transformations control intracellular concentrations of heavy metal ions that may be inhibitory to physiological functions and may form unspecific complex compounds in the cell (99). In

Bacteria (8, 10, 157) and Archaea (142), Hg resistance is an important part of the Hg geochemical cycle (Figure 1.1). An elaborate system of Hg-detoxification, the mercury resistance (mer) system (Figure 1.2), facilitates survival at elevated Hg concentrations.

Figure 1.2: The microbial Hg detoxification (Mer) system. Mer enzymatic, transport and regulatory functions are encoded by the mer operon, whose coordinated activities result in Hg(II) reduction and MeHg degradation.

The mer operon (merTPCAD) codes for a group of proteins and regulatory elements that are involved in the sensing, transport, and reduction of mercuric ions

(Hg[II]) (Figure 1.2) (131). This operon is central to the widespread Hg resistance system found in both Bacteria and Archaea (10, 131). The best characterized mer operon is found in Gram negative bacteria where Hg(II) is bound in the periplasm by the protein

MerP which transfers Hg to the inner membrane transport protein MerT (131). Two cysteine residues in MerP displace anions such as Cl-, a common Hg ligand, and thus binds Hg(II). How Hg is transferred with MerT into the cytosol is not well understood, but once there, Hg is transferred to the central protein of the mer system, the enzyme mercuric reductase (MR). MerA is the protein that forms an active MR enzyme comprised of a MerA homodimer; this enzyme is what catalyzes the reduction of Hg(II) to Hg(0). The reduction of Hg(II) to Hg(0) uses NADPH or NADH as a reductant (10,

157). Additionally, some mer operons may include organomercurial lyase (MerB) which degrades organomercurial compounds such as phenylmercury, ethylmercury, and methlymercury (10). mer operons that encode for MerB are considered to provide broad spectrum mercury resistance (131). MerR, a metal responsive regulatory protein, regulates expression of the mer operon. MerR binds to the promoter/operator region of mer together with RNA polymerase. In the absence of Hg, this tripartite complex prevents the initiation of transcription. When Hg is present in the cytoplasm, it binds to

MerR causing a conformational change of the transcription initiation complex leading to the alignment of the operator region with the transcription start site and expression of the operon (10).

Much of our existing knowledge of these resistance mechanisms has arisen from research focused on metal contamination from the perspective of environmental and human health concerns (26, 31, 97, 105). Although, a cosmopolitan distribution of metal resistant microorganisms and an abundance of environments that are enriched with metals of geological origin suggest evolution of metal ion resistance prior to industrial release of metal contaminants (8, 33, 156).

Functional mer operons have been mostly described in the Actinobacteria,

Firmicutes, and among the Alpha -, Beta- and Gamma-proteobacteria (8), however the only thermophilic bacterium for which a functional mer operon has been characterized is

Thermus thermophilus HB27 (157). The phylum Aquificae, the dominant primary producers in many geothermal environments (144), is the deepest branching bacterial lineage (83) in which merA homologs were found in the genomes of Hydrogenobaculum sp. Y04AAS1 (AAS1) and Hydrogenivirga sp. 128-5-R1-1 (R1-1) (119). Phylogenetic

11 reconstructions clearly established the basal position of the Aquificae loci in the MerA tree, suggesting that merA originated in an ancestor common to deep branching thermophilic bacteria (Figure 1.3) (8). It is not known whether these homologs encode for an active Hg resistance system, which prompted research presented in chapter 2 investigating if merA homologs in representatives from the phylum Aquificae encode for

MR-mediated Hg resistance.

4 T. Barkay, K. Kritee, E. Boyd and G. Geesey

FigureFig. 2. 1.3:Bayesian Bayesian inferred phylogenetic inferred reconstruction phylogenetic of MerA deduced reconstruction amino acid sequences. of Full-length MerA MerA deduced were truncated amino by the acid elimination of NmerA sequences prior to alignments. Nodes are labelled with blue, green, yellow circles and no circles indicating > 90, > 80, sequences.> 70 and < 70% Full posterior-length probability MerA values respectively;were truncated red beaded lines by indicatethe elimination thermophilic lineages; of bar NmerA indicates 1 sequences substitution per 10 positions. The hatched branch at the root of the MerA clade indicates a discontinuity in branch length introduced to conserve space. priorSequences to alignments. of dihydrolipoamide Nodes dehydrogenase are labeled deduced from withlpdA homologuesblue, green, in the genomes yellow of Magnetospirillum circles and magneticum no circlesAMB-1, indicatingPseudomonas > ﬂuorescens90, > 80,Pf0-1 > and 70Thermus and thermophilus< 70% posteriorHB27 were used probability as outgroups. values, respectively; red beaded lines indicate thermophilic©2010SocietyforAppliedMicrobiologyandBlackwellPublishingLtd, lineages; bar indicates 1 substitutionEnvironmental per 10 positions. Microbiology The hatched branch at the root of the MerA clade indicates a discontinuity in branch length introduced to conserve space. Sequences of dihydrolipoamide dehydrogenase deduced from lpdA homologues in the genomes of Magnetospirillum magneticum AMB- 1, Pseudomonas fluorescens Pf0-1 and Thermus thermophilus HB27 were used as outgroups (8).

IV) The Phylum Aquificae

Close to the root of the Bacterial 16S rRNA gene tree, lies the Aquificae, a phylum that has been impressing scientists as early as the late 1800’s. W.A. Setchell noticed filamentous life growing up to 89° C, “the filamentous types occurred in such masses and in connection with such high temperatures as to make them noticeable” (118).

Although it is mostly accepted that Aquificae represent the deepest branching bacterial lineage, recently the evolutionary placement of the Aquificae has been challenged through the use of ribosomal proteins as evolutionary markers (18). Griffiths and Gupta

(53) also show a late branching of the phylum Aquificae, diverging prior to the

Proteobacteria, using presence or absence data of signature sequences consisting of conserved insertions or deletions. Much of the current literature focuses on identification and characterization of new members within the Aquificae (2, 16, 20, 27, 43, 51, 66, 103,

104, 149), however, to date little research has focused on their ecology.

Historically, seven genera have fallen within the phylum Aquificae: Aquifex,

Hydrogenobacter, Hydrogenobaculum, Thermocrinis, Hydrogenothermus, Persephonella and Sulfurihydrogenibium (2). More recently the number of genera has increased to 11 to include Hydrogenivirga, Balnearium, Desulfurobacterium, Thermovibrio, and

Venenivibrio (NCBI). Members of the Aquificae have been isolated from hot springs (2,

35), hydrothermal vents (50), and other high temperature environments such as subsurface gold mines (149). Aquificae are primary producers of biomass in high temperature environments (65) and have been shown to be dominant organisms in such environments (91, 144). Geothermal environments suits their chemolithoautotrophic way

14 of life well, especially some Aquificae’s ability to oxidize hydrogen, a common gas in some hot springs, by the knallgas reaction:

2H2 + O2 -> 2H2O

Due to the role of Aquificae in primary productivity and dominance in Yellowstone hot springs in addition to thermodynamic modeling, it was hypothesized that the entire energy economy of YNP is hydrogen based (144).

At the time my dissertation project was developed (May, 2007), the genomes of two members of the Aquificae were shown to have a homolog of the merA gene in their chromosome (119). Hydrogenobaculum sp. Y04AAS1 was isolated from Obsidian Pool in YNP and is known to obtain energy through the oxidation of hydrogen or reduced sulfur compounds. This aerobic organism is a motile rod with a microaerophilic oxygen requirement and an optimal growth temperature of 58° C (117). Hydrogenivirga sp. 128-

5-R1-1 was isolated from a hydrothermal vent in the Lau basin at a depth of 2200 m.

Cells of Hydrogenivirga sp. 128-5-R1-1 are motile rods with that can respire microaerophilically or anaerobically oxygen requirement. This organism exhibits a chemolithoautotrophic metabolism and can use reduced sulfur compounds or hydrogen as an electron donor (117). More recently, the sequenced genome of Hydrogenobacter thermophilus TK-6 revealed a merA homolog (165), however this strain was not included in my study.

Study Scope and Objectives:

The main goal in this study was to gain a better understanding of microbially mediated Hg(II) reduction to Hg(0) among thermophilic deep branching Bacteria through characterization of the proteins and enzymes that are encoded by their mer operons and to place this function within the ecology of microbial mats in geothermal environments.

These goals were achieved through i) the characterization of mer mediated Hg-resistance in the phylum Aquificae, and ii) an analysis of the effects of physical and chemical parameters on structuring Hg-resistant communities in YNP hot springs. In addition, I applied the analytical tools that were developed for the community study (task ii) to a collection of 23 merA gene assemblages from diverse environments to examine the effect of geography on the global distribution of merA.

Specific Objectives:

1) To characterize the mechanism of Hg-resistance in the Aquificae

Hydrogenobaculum sp. Y04AAS1 and Hydrogenivirga sp. 128-5-R1-1 whose

genome sequence revealed genetic homologs to mer genes;

2) To study the diversity and distribution of microbial communities and merA genes

in microbial mats from geochemically diverse YNP hot springs with elevated Hg

concentrations;

3) To identify the degree to which Hg-resistant communities are structured by

physical and chemical parameters in i) YNP hot springs, and ii) 23 merA

assemblages from diverse environments.

Chapter 2: Characterization of mer-mediated mercury resistance in Hydrogenobaculum sp. Y04AAS1 and Hydrogenivirga sp. 128-5-R1-1.

Introduction

Microbes must have been exposed to toxic heavy metals since the beginning of life on Earth and have evolved diverse mechanisms to live in the presence of high concentrations of toxic metal ions (141). These mechanisms, such as efflux, intra- or extra-cellular precipitation, and enzyme-mediated transformations control intracellular concentrations of heavy metal ions that may be inhibitory to physiological functions and form unspecific complex compounds in the cell (99). While much of our existing knowledge of these resistance mechanisms has arisen from research focused on metal contamination from the perspective of environmental and human health concerns (13, 26,

31, 97, 105), a cosmopolitan distribution of metal resistant microorganisms and an abundance of environments that are enriched with metals of geological origin suggest evolution of metal ion resistance prior to industrial release of metal contaminants (8, 33,

156).

Mercury (Hg) is a potent neurotoxic substance and the most toxic heavy metal to microorganisms due to its high affinity to sulfur (98). Globally distributed Hg (10) is toxic to humans and wildlife mostly due to the accumulation of methyl-mercury (MeHg) in aquatic and terrestrial food webs (22). Microbial activities are central in modulating environmental Hg toxicity and mobility. Resistance to inorganic Hg (Hg[II]) is controlled by the activities of the enzyme mercuric reductase (MR), an NAD(P)H dependent flavin oxidoreductase which catalyzes the reduction of Hg(II) to the elemental form, Hg(0). The gene encoding MR, merA, is part of the Hg resistance (mer) operon,

17 which is widespread among both Bacteria and Archaea (8, 10, 142), allowing these organisms to survive in the presence of elevated Hg concentrations (10, 11). At minimum, Hg resistance systems are comprised of transport, enzymatic, and regulatory functions. MerT is involved in the transport of thiolated mercuric mercury (Hg[II]) into the cytoplasm for reduction by MR, although it is known that Hg may enter cells lacking specific transport proteins (Barkay, 2003). MerR acts as a regulator of the mer operon, binding to the operator/promoter (O/P) region of the operon to repress transcription in the absence of Hg(II). If present, Hg(II) forms a complex with MerR-mer O/P prompting the operator DNA to unwind, inducing transcription of the operon’s functional genes (10,

55).

A recent body of literature supports the hypothesis that microbial resistance to Hg evolved in geothermal environments where microbial life has perhaps been exposed to

Hg since the beginning of life on Earth (8, 122, 154). Culture independent techniques have been used to amplify Hg resistance genes from hot springs in Yellowstone National

Park (YNP) (156) and Coso Hot Springs, CA (142), and Hg resistant organisms have been previously isolated from geothermal environments (28, 131, 154). The large scale sequencing of microbial genomes has resulted in an increased availability of merA sequences and allowed for a robust analysis of gene evolution, further supporting an origin and early evolution of Hg resistance among thermophilic microbes from geothermal environments (8).

Functional mer operons have been mostly described in the Actinobacteria,

Firmicutes, and among the Alpha-, Beta- and Gammaproteobacteria (8), however the only thermophilic bacterium for which a functional mer operon has been characterized is

Hydrogenobaculum sp. Y04AAS1 (AAS1) and Hydrogenivirga sp. 128-5-R1-1 (R1-1)

(119). Strain AAS1 was isolated from Obsidian Pool, YNP, and R1-1 from the Eastern

Lau Spreading Center, South Pacific (119), both of which are geothermal environments similar to those where elevated Hg concentrations were reported (32, 74). Phylogenetic reconstructions clearly established the basal position of the Aquificae loci in the MerA tree (8) suggesting that merA originated in an ancestor common to deep branching thermophilic bacteria. However, it is not known whether these homologs encode for an active Hg resistance system. Here we report on the activity and characteristics of the Hg resistance systems of these two Aquificae strains representing chemotrophic primary producers in many geothermal environments.

Materials and Methods

Bacterial Strains and Growth Conditions

Strain Hydrogenobaculum sp. Y04AAS1 (AAS1), Hydrogenivirga sp. 128-5-R1-

1 (R1-1), and Persephonalla marina str. Ex-H1 (H1) were kindly provided by Dr. Anna-

Louise Reysenbach. All growth media were prepared under a CO2 headspace.

Microaerophilic conditions where then created by the post autoclaving addition of O2 (to

4% v/v), and the tubes were pressurized with either H2 or CO2 after inoculation. Strain

AAS1 was cultivated at 55ºC in modified DSMZ 743 medium (Boone’s Medium #5) as described by Shima et al. (137), which contained the following (L -1): 5 g elemental

sulfur, 1 g (NH4)2SO4, 2 g Na2S2O3.5H2O, 1 g K2HPO4.3H2O, 1 g NaCl, 0.3 g

MgSO4.7H2O, 1 mg FeSO4.7H2O, 1 mg CaCl2.2H2O, 0.06 mg NiSO4.6H2O, and 0.5 mL of a trace-element stock solution (adapted from Ferguson and Mah, (41)). Prior to autoclaving, the pH was adjusted to 4.5 with 6 M HCl. S(0) is present as a potential

electron donor, however was omitted from growth tests to only allow for growth on S2O3 or H2. Strains R1-1 and H1 were grown at 70ºC in modified MSH medium (Boone’s

Medium #2) (19) which contained the following (L -1): 29 g NaCl, 2 g NaOH, 0.5 g KCl,

1.36 g MgCl2.6H2O, 7 g MgSO4.7H2O, 2 g Na2S2O3.5H2O, 0.4 g CaCl2.2H2O, 0.2 g

NH4Cl, 0.3 g K2HPO4.3H2O and 10 ml of the same trace-element stock solution. Prior to autoclaving, the pH was adjusted to 6.0 with 6 M H2SO4. In both types of growth media,

S2O3 was omitted when H2 was provided as the sole electron donor, whereas no H2 was added when S2O3 was used. Since strain R1-1 cannot grow on H2 as an electron source, this strain was only tested in thiosulfate-amended medium. Unless specifically described, all culture maintenance and growth experiments were performed in a final volume of 5 mL in 26 mL Balch tubes (Bellco, Vineland, NJ) fitted with crimp-sealed Teflon- stoppers.

Growth Measurements

Growth was determined by measuring the optical density at 660 nm or by direct counts by Acridine Orange staining. For AAS1 direct counts, to prevent staining of a medium precipitate, cells were first fixed on a 0.22 micron polycarbonate filter (General

Electric, Feasterville-Trevose, PA) and immersed in 200 μL 0.1% Acridine Orange for 30 seconds, at which time the remainder of the stain was filtered through by vacuum pump

(Gast Manufacturing, New York City, NY). Preps were visualized using an Olympus

BX60 microscope with and oil-immersion objective lens (Uplan F1 100X/1.3), and counted using Olympus Microsuite Basic (Ver. 3.2) (Olympus Corp, Center Valley, PA).

Modeling of Hg Speciation

Reduced sulfur species are known to affect Hg bioavailability (15). In order to determine the speciation of Hg in both types of growth media, the chemical equilibrium speciation model MINEQL+ (version 4.5) (130) was used. Input parameters were obtained from the MINEQL+ and the National Institute of Standards and Technology databases (87). Further dissociation constants used as input parameters in the modeling are shown in table 2.1. Hg speciation in each growth medium was modeled with 5-60 μM

HgCl2.

Table 2.1: Disassociation constants used in MINEQL+ modeling. Species LogC LogK

Hg(OH)2 -22.6 0 -4 H2(S2O3)3Hg(OH) (-4) -5.41 36 -2 H2(S2O3)2Hg(OH) (-2) -5.21 34.5 HgClOH (aq) -20 9.85 -2 HgCl4 (-2) -16.9 20.8 HgCl+ (+1) -23.9 12.8 HgCl2 (aq) -18.1 18.9 -1 HgCl3 (-1) -17.4 19.9 + HgHCO3 (+1) -27.7 21.6 -2 Hg(CO3)2 (-2) -27.7 20.9 HgCO3 (aq) -25.1 17.4 Hg+2 (+2) -30.1 6.23 HgOH+ (+1) -26.8 2.67 -1 Hg(OH)3 (-1) -30.4 -14.6 HgSO4 (aq) -30.5 7.69 Adapted from Crespo-Medina et al. 2009 (32).

Resistance to Hg(II) During Growth

Mid-log phase cultures of strains AAS1, R1-1 and H1 were diluted 100 fold into

fresh growth medium to a final volume of 5 mL at an OD660nm of 0.010. Post inoculation,

HgCl2 was added to each tube to a final concentration of 0, 5, 10, 20 and 40 μM. All treatments were performed in triplicates and the tubes were arranged randomly in a test tube rack and incubated in the dark without shaking at each organism’s optimal growth temperature. Growth was monitored every 4-8 hours until commencement of stationary

phase. The effect of Hg on growth was expressed as cell density at a given HgCl2

concentration as a percent of cell density of the no HgCl2 control, according to the following equation:

[Cells/mL(t1) – Cells/mL(t0)]Hg % resistance = 100 * [Cells/mL(t1) – Cells/mL(t0)]No Hg

T1 was chosen as the point where the no HgCl2 reference culture approached late exponential growth, just prior to stationary phase; at this point, the direct counts were

9 9 6.95 * 10 and 3.72 * 10 cells/mL for strain AAS1 growing on S2O3 and H2, respectively, and 1.71 * 109 cells/mL for strain R1-1. This cell density represented the latest measurement at which all cultures were still growing and where a clear pattern of the effect of Hg on growth rate could be distinguished.

Loss of Hg(II) during culture growth

Fresh growth media were inoculated with mid-log phase cultures of strains AAS1

203 or R1-1 to an OD660nm of 0.010 and cultures were spiked with 5 and 10 μM HgCl2

(specific activity 0.5-0.12 nCi µM Hg) for cultures grown on H2 and S2O3 respectively.

The isotope was kindly provided by Christie Bridges (Mercer University, GA). Strain

R1-1 and AAS1 were incubated in the dark with no shaking at 70 and 55° C, respectively. 203Hg remaining in the growth media was monitored by removing 250 μL aliquots from the growing cultures every 4-8 hours to 3 mL Scinti-Verse scintillation fluid (Thermo Fischer Scientific, Waltham, MA) and samples were counted in a

Beckman LS 6500 liquid scintillation counter (Beckman-Coulter, Brea, CA). Growth

was measured in parallel cultures containing unlabeled HgCl2 at similar concentrations by either direct cell counts or optical density as described above.

Maximum apparent Hg loss rates were calculated using the equation below (128):

(Hg - Hg ) Fmol Hg(II) lost/h/cell = t1 t2 (t2 – t1)*N

Where N = the average number of cells counted at t1-t2, Hg = fmol Hg(II) lost between t1 and t2 as calculated from the concentration of Hg that remained in the growth medium at

t1 and t2, and t1 and t2 are the times (h) between which the loss of Hg from the medium was the greatest. All data presented indicate the difference in the rate of Hg(II) loss of the growing culture subtracted from the uninoculated control. Significance (P<0.05) of differences in Hg(II) loss rates for each treatment was calculated using Student’s t-test.

Production of Hg(0) by growing cultures

The production of Hg(0) by the Aquificae cultures was determined to verify whether or not the Hg(II) that was lost from growth media was reduced and volatilized,

203 hallmarks of the activity of all mer systems (10). The amount of HgCl2 in each culture at T0 was determined as described above. Once cultures entered stationary phase, i.e.,

OD660nm=0.2 and 0.06 for AAS1 grown on S2O3 and H2 respectively, and OD660nm=0.08 for

R1-1, the culture headspace was flushed with sterile air for 40 minutes to drive Hg(0) that accumulated in the headspace during culture growth into a Hg trapping solution

consisting of 0.75% KMnO4, 0.40% K2S2O3, 3.5% HNO3, and 5% H2SO4. At the conclusion of the 40 minutes, the remaining growth medium was acidified to 1.75% HCl and mixed by vortexing to free any 203Hg sorbed to the glass wall. 250 μL aliquots were removed from the stationary phase culture and the trapping solution for scintillation counting from which the remaining Hg(II) and Hg(0) was monitored as described above.

Crude cell extract mercuric reductase (MR) assays

Mid-log phase cultures of strains AAS1 (OD660nm=0.15-0.18 and 0.04-0.05 for cultures grown on S2O3 and H2, respectively), and R1-1 (OD660nm=0.07-0.09) were diluted

100 fold into 250 mL medium in rubber capped 2 Liter pyrex media bottles (Corning,

Lowell, MA, USA) with a microaerophilic headspace as described under “growth conditions”. Cultures were grown to mid-log phase (O.D. as described above) at which time cells were harvested by centrifugation, and cell free lysates were prepared as described by Vetriani et al. (154). Cells were harvested and cooled at mid-log phase by centrifugation for 10 minutes at 5,750 x g at 4ºC in a pre-refrigerated Sorvall RC-5B centrifuge (Thermo Scientific, Waltham, MA). Pelleted cells were washed in phosphate- buffered saline, weighed, and stored at -20ºC until used for further analysis. Typical yields of such post-harvest biomass were 0.04 g, 0.08 g, and 0.04 g for strain R1-1, and

AAS1 grown on S2O3 and H2, respectively. Cells were re-suspended to a concentration of

200 mg mL-1 (wet weight) in a buffer consisting of 20 mM sodium phosphate (pH 7.5),

0.5 mM EDTA, and 1 mM β-mercaptoethanol, and were lysed by intermittent sonication using a Misonex S-4000 sonicator (Misonex, Newtown, CT) for a total of 3 minutes on ice. MR assays were performed in 80 mM sodium phosphate buffer (pH 7.4) with 1 mM

β-mercaptoethanol, 200 μM NAD(P)H, 50 μM HgCl2 to a final volume of 800 μL (47).

Mercury dependent NAD(P)H oxidation was monitored as the decrease in A340 using a

UV/VIS spectrophotometer (Cary 300, Agilent Technologies, Budd Lake, NJ). Initial

rates of NAD(P)H oxidation were determined while the decrease in A340 over time was linear, usually during the first 10 seconds of each assay; each assay at a given cell extract

concentration was performed once with and once without HgCl2 and activities of at least

3 different extract concentrations were tested per extract. Specific Hg dependent

NAD(P)H oxidation rates, expressed as Units (U) mg protein-1, in cell free extracts were

then calculated by subtracting the slope of the curve obtained in absence of HgCl2 from

that observed following HgCl2 addition. A unit of MR activity was defined as 1 μM of

NAD(P)H oxidized min-1. Protein concentrations in crude cell extracts were determined using the Bradford assay (Bio-Rad Laboratories Inc., Hercules, CA).

Possible induction of mer in AAS1 and R1-1 was determined by comparing the

MR activities of cultures grown with and without HgCl2 using the methods described above. Media contained HgCl2 to a final concentration of 5 μM for AAS1 grown on H2

and 10 μM for AAS1 and R1-1 grown on S2O3 respectively, the highest subtoxic concentration tolerated by each culture. An additional 5 or 10 μM HgCl2 spike was

added at mid-log phase (representative OD660nm described above) following which cultures were incubated for 2 doubling times prior to cells harvesting.

The temperature profiles of MR activities for strains AAS1 and R1-1 were determined as described above using extract of cultures grown without Hg. Cell extracts and test buffer were pre-incubated at the assay temperature for 10 minutes prior to the addition of 1-4 μL of extract to assay buffer. Significance (P<0.05) of differences in MR activities for each treatment was calculated by Student’s t-test.

Quantitation of merA Transcripts

Cultures of R1-1 and AAS1 were grown in 125 mL serum bottles (Wheaton

Scientific, Millville, NJ) with 25 mL medium and incubated at 70 and 55° C respectively.

When the culture approached mid log phase (target OD660 values as described above),

HgCl2 was added to a final concentration of 1 μM, at which time, 1.5 mL of each culture was immediately harvested as the T0 sample. Similar aliquots were removed at 5, 10, 15,

20, 30, 60, 80, 100, and 120 minutes post Hg addition, harvested by centrifugation for 1 minute at 13,000 x g, and immediately frozen at -80° C for RNA extraction and DNAse treatment. Control cultures to which Hg was not added were included to determine to background levels of merA transcription. Total RNA from strains AAS1 and R1-1 were extracted using the TRIzol reagent (Invitrogen, Carlsbad, CA). Extracted RNA samples were quantified by a Biophotometer Plus spectrophotometer (Eppendorf, Hauppauge,

NY) and were diluted to 20 μg mL-1 prior to DNAse treatment with the TURBO-DNA free kit (Applied Biosystems, Carlsbad, CA). Both methods were performed according to manufacturer’s instructions. cDNA was synthesized using the High-Capacity cDNA RT kit (Applied Biosystems, Carlsbad, CA) using a GeneAmp PCR System 9700 (Applied

Biosystems) thermocycler.

To establish whether induction of mer activities occurred in the two Aquificae strains, the change in transcription level of merA produced at different time intervals

post-HgCl2 addition was monitored using methods adapted from Wang et al. (157).

Transcript abundance was determined by quantitative real time PCR (qPCR) using cDNA synthesized as described above. Transcription levels of gyrA, encoding the α subunit of

DNA gyrase, were determined in order to normalize merA transcription levels to those of a constitutively expressed cellular gene.

Table 2.2: Primers sets used in qPCR of merA and gyrA in strains AAS1 and R1-1.2 Start Tm Amplicon Name2 Sequence (5’ to 3’) Position3 (°C) Length HBgyrA-F 441 58 TGTCATGGAGCCTCAGGTTCT 66 HBgyrA-R 506 57 ATGCCTGTAGTACCGTTGCAAA HBmerA-F 441 59 ATGCGGCAGGAGATTGTGTT 71 HBmerA-R 506 58 GCTGCTATCCCTCCTTCCATAG HVgyrA-F 14 57 ACAGGTATTGCTGTTGGACTTTCA 107 HVgyrA-R 120 55 TCCTCAACAGTTGCATTTGGAA HVmerA-F 1049 59 AGAGCCTCGGGCTTGATAGG 70 HVmerA-R 1118 59 AGAAACTCGTTCACCTTCACGAA 1PCR reactions for all primers included consisted of an initial denaturation stage of 90o C for 10 minutes, then 45 cycles of 90o C for 15 seconds followed by 1 minute at at 55o C. Upon completion, a melt curve was performed to verify identity of the amplification products. 2All primer sets shown were designed for this study; primers are referred in text to with prefix RT-. 3Nucleotide numbering for each primer set is according to the relative nucleotide position within the merA or gyrA locus in the genomes of AAS1 and R1-1. Target accession numbers are given in Materials and Methods.

For strain AAS1, primers RT-HBmerA(f/r) and HBgyrA(f/r) were used to amplify

71 and 66 bp regions of merA and gyrA-specific products respectively. For R1-1, RT-

HVmerA(f/r) and HVgyrA(f/r) were used to produce 70 merA and 107 bp gyrA-specific products. Primers were designed using default parameters within Primer Express

(version 3.0), (Applied Biosystems, Carlsbad, CA) using merA and gyrA locus sequences from the genomes of AAS1 (gyrA locus HY04AAS1_0371; merA locus

HY04AAS1_1213) and R1-1 (gyrA locus HG1285_00715; merA locus HG1285_05690).

The Power SYBR Green PCR Master Mix (Applied Biosystems, Carlsbad, CA) was used in all qPCR reactions. Amplifications were performed in triplicate for both merA and gyrA transcripts using the StepOne Plus PCR machine running StepOne Software

(version 2.1) (Applied Biosystems). qPCR conditions included an initial denaturation stage of 90° C for 10 minutes, then 45 cycles of 90° C for 15 seconds followed by 1 minute at 55° C for all primer sets used. Upon completion, a melt curve was performed to verify identity of the amplification products. A control without reverse transcriptase for each RNA extract was included to test for the presence of DNA contamination that might have affected mRNA quantitation.

Induction folds were calculated by the comparative Ct method (108), with Ct defined as the PCR cycle at which the sample fluorescence increased significantly above background, as determined by the software. The relative expression levels of the target gene, merA, versus that of the reference gene, gyrA, were used to calculate induction levels following the equation:

+ Hg (merA/gyrA) ΔCt = - Hg (merA/gyrA)

Where ΔCt is the ratio between the ratios calculated for threshold cycles observed for

merA and gyrA transcripts for cultures grown with or without Hg. Such ΔCt were obtained for each time point following the addition of HgCl2 to exposed cultures.

Results

Identification of putative mer operons in the genomes of Hydrogenobaculum sp.

Y04AAS1 and Hydrogenivirga sp. 128-5-R1-1

Using the amino acid sequence of MerA from Tn501 as a query in tBLASTn searches of all sequenced microbial genomes identified two ORF’s, HY04AAS1_1213, and

HG1285_05690 as merA homologs in Hydrogenobaculum sp. AAS1 and

Hydrogenovirga sp. R1-1, respectively (8). An updated phylogeny that included all

MerA homologs in sequenced microbial genomes confirmed the basal position of the

Aquificae sequences in the MerA phylogeny (Figure 2.1). The alignment block of bacterial and archaeal MerA sequences corresponding to positions 8-472 of Streptomyces lividans (P30341) of 97 selected sequences out of a total of 284 gene homologs available as of April 2011. These homologs were representative of all major branches in a MerA phylogeny, were used in the reconstruction. The two MerA homologs of

Hydrogenobaculum sp. (AAS1) and Hydrogenivirga sp. (R1-1) clustered together with a third Aquificae MerA homolog from the recently sequenced genome of Hydrogenobacter thermophilus TK-6 (4) and a MerA homolog of Deferribacter desulfuricans SSM1 in a sister position to all Archaeal sequences with accompanying posterior probability values of 100 (Figure 2.1). This cluster shared a common, likely bacterial, ancestor with a large cluster consisting of the remaining bacterial MerA sequences. Thus, the present MerA phylogeny (Figure 2.1) is consistent with our previous suggestions that merA of the

Aquificae represents an early lineage, that merA originated in a bacterial ancestor, and that it was horizontally transferred from such a bacterial ancestor to the Archaea (8).

Archaea Alphaproteobacteria Betaproteobacteria/Gammaproteobacteria Other Firmicutes Actinobacteria Deinococcus/Thermus Bacteroidetes Aquificae

Figure 2.1: Bayesian inferred phylogenetic reconstruction of MerA deduced amino acid sequences. Full-length MerA were truncated by the elimination of NmerA sequences prior to alignments. Nodes are labeled with black, gray, green, yellow circles and no circles indicating > 90, > 80, > 70 and > 60 and < 50 % posterior probability values respectively. Red beaded lines indicate thermophilic lineages; bar indicates 3 substitutions per 10 positions. Sequences of dihydrolipoamide dehydrogenase deduced from lpdA homologues in the genomes of Magnetospirillum magneticum AMB-1, Pseudomonas fluorescens Pf0-1 and Thermus thermophilus HB27 were used as outgroups.

The amino acid sequences of both Aquificae loci contained signature motifs known to be required for MerA activity, including the redox-active site (identical in both

Aquificae sequences – GGTCLNRGCIPSK corresponding to residues 133 to 145 in

MerA of Tn501), the vicinal CC pair at the carboxy-terminus (C565-C566), and two tyrosine residues; one approximately 60 residues downstream of the redox active site, and one approximately 20 residues upstream of the vicinal CC pair (Figure 2.2) (10). The two proteins however differ in size; 464 amino acid residues in Hydrogenobaculum sp. AAS1 and 545 in Hydrogenivirga sp. R1-1. This is due to the presence in the later, but not in the former of NmerA, the ~ 70 amino acids N terminal extension heavy-metal-associated

(HMA) domain that is a part of more than half of all MerA sequences (8). The identity of this domain as NmerA is supported by the presence of the sequence GMTCEHC, 8 amino acids downstream from the N terminus. This sequence corresponds to a typical metal binding motif, xMxCxxC, common to many metal chaperone proteins (120) and a part of all NmerA domains (10). In both Aquificae strains, two ORF’s upstream of the putative merA code for proteins that bear homology to Mer transport functions. The merA proximal ORF’s, loci YP_002121875 (79 AA) and ZP_02177251 (92 AA) in the genomes of strains AAS1 and R1-1, respectively, may encode for MerP as they include the signature metal binding motif sequence, GMTCxxC; GMTCEHC in strain AAS1 and

GMTCKVC in strain R1-1 (Figure 2.2). In the proteobacterial MerP, there is a ~19 residue Sec-type signal sequence whose role is to direct MerP to the periplasm where it is removed by an energy dependent process (10, 71). Such a motif may be present in MerP of strain R1-1, although it shares no homology with the signal peptide of Tn501 (Figure

2.2). However, twin-arginines (corresponding to position 5-6 in the Tn501 MerP)

31 followed by 9 hydrophobic amino acids (AVLAALILV) followed by one positively charged amino acid (K) may suggest a twin-arginine translocation (Tat) leader peptide.

Tat protein export signals, initially discovered as an aid in protein transport from the stroma into thylakoids in plants, are now known to be associated with the cross membrane transport of periplasmic proteins in many bacteria (93). The amino terminus of the MerP homolog in strain R1-1 could not be fully identified as a Tat peptide because it lacks the SRRxFLK motif (126). Accordingly, analysis using the TatP server (version

1.0) (http://www.cbs.dtu.dk/services/TatP/) indicates a low probability (C, S, Y scores <

0.14) for the presence of a Tat motif (14). A hydrophobic leader signal is missing in the putative MerP locus of strain AAS1 questioning the transport of its product to the periplasmic space and thus functioning as a MerP.

Tn501 R1-1 AAS1 D E E P P P Y Y Y R R R W W P Y Y I F F L L L I Tn501 R1-1 AAS1 G A V V V I G 60 A A Tn501 R1-1 AAS1 L Y Y Tn501 R1-1 AAS1 V V I G A A L V A G F A M M M F Y Y R S K A S S E E E M M M W L P F F I Tn501 R1-1 AAS1 R S K R Y Y K L V 32 Q L E E E G R R 70 T S K I G G G P F F G S S F L I A A A K L V the terminus the proteins of highlighted in gray. locations The membrane of spanning loop between tyrosines, and the viscinal CC pair in the well characterized Figure 2.2: ! R Y Y Y - - “ MerT MerA N G T S G - - R Tn R1-1 AAS1 Tn R1-1 AAS1 Tn R1-1 AAS1 Tn R1-1 AAS1 - - L L

MerP G Q R R Tn 10 S S A - - Q K K 501 501 501 501 R V V V L - - - - 501 N - - G T L L F L A - - W 11 11 - - 10 A L V T V L

M - - I

- - V proteins - - G G L A A S ” - - I K Alignment of putativeAlignment of - - K - - 10 G M F L A E E E T K the second and third membra - - E R N - - Tn501 R1-1 AAS1 T V L E L F F M L - 80 A K G G G N N S - A A A R F - . A L G M T T V A S A - R I A -

A L L F F L - I I I I I S A 133 - Mer proteins C 20 G G G K L L T A A A - L V . Respective sequences trans from - G G G A A A A S A K K K A L - S A S L R T T T P P 10 I I L A - 20 C C C A S L L L T G A K A A - C C C A A A V L L E A L I I - R A E C C C N N N 20 S A S D V V E I - 50 A L L R R L V F A S L V 90 Mer C C C V L 1 - G of Tn of T A G C C C V G G G I A V 90 A A A - M V E T C C C P P P C C C P S

C C - proteins G G erA I P L T L V F V - I I V F - P E E 501 N T E 30 V L G V P P P I W I S ne spanning domains - Q E K , the L L L F L I P P P S S S A F - . Conserved regions of . Conserved regions of D D N V P S V L L T K K K I 20 - A A 145 - R G R F , including A V S 30 - I V L Y Y conserved I I T E - R R W W S A L F M L F L L L I Q K - N G G G K K T T A 30 V L M I I I I T V 60 L V Q D F V V Tn501 R1-1 AAS1 Y R I A V S E E V L I L V S S S I L F A T V R I T V G A E E A T G G G A A A L L poson Tn poson domains I Q G MerA L K C A S H N N F V V D S K

Q W 40 K K S S 38 M residues in S S S T Tn501 R1-1 AAS1 I V V I Q E E S G - - A T A L F Tn501 R1-1 AAS1 P E T in A E E L

G H - - A S Y Y 31 (top), (top), G G G 30 R L L L W 40 M I I S S

L A T are boxed (1 M M M 501 G V L L S S functional R S P P erT - - T T T Q M N D S F S K K R N 40 L K

C C C - - MerT are included reference. for Numbering indicated 70 L V L L E E E the membrane first spanning I , and the metal in binding motif GMTCxxC I R L E S I I S K E D D H F A S L L L G N N S S S H A V R R R P V V V A V N F S 101 37 P P P C C C

50 K K K (middle), and H Y Y Y L L F K K L L F F I I P P A 100 importance, the redox active site, two downstream D L L F - E E L P P A E E H D A S 3) in P P Q I I S S P P P T T T K K K - G G V A V S 197 T A T K V M Y Y Y L L Y Y Y Tn501 R1-1 AAS1 - G L L F 40 V V V - - G R R R Y W W V S S E - M V V K K K I W W V A 10 P G N N Tn501 R1-1 AAS1 E E E I erT K K K F T T G Y Y S A V V I I A A A W W W 80 F F L L L L L

M I I ( L I I 123 F V G G L L K K K E I I erP S S E G V A S G A V A A L L L E E S ) A L V V C C A I

I . (bottom) from AAS1 and R1 AAS1 (bottom) from V T V 110 G 60 A V V A A A A A Gray E K E Q M S L L F L Y Y Q Q Q domain 3 2 G G N L V V L L V V I T T S I V V V D D F G L V V A A F F F circles S K K P V V V F F L V A N 20 F K 544 W K K Y Y Y G A A V F A K K K and in the cytoplasmic V V A L P P L T T F Y Y D D D D K E T T T G indicate the F F A S S V V I M V T V L L W F Y Y L K S S I T S Y V S S R P P P Y Y Q K K M 10 F L F E E E 120 R R 70 Y Y L K L L L erP G S A E K P V V F F L S S S I position 124 565 N Q L T M R L I I Y Y Y C C C

are R R G 60 74 K K K P E K R L L C C C caboxy Q L L L M V Y F K Q K K A A A - C C A H A A A 1, and F N G Q I E - - s 106 129 A A A 569 V V E Y Y F W

- - in Q Q Q G V V 32 V - - I - - T T S G T V K - - - - -

F F F F V E - - - N 20 F K D 80 A K - - - K K K D D E A L - D D D D G A A L

- V V I C N R 70 K K - K S S K K K T V A Q K K R P P S K S L L L G A K V E I S S S Q E E T I I C C C D V E N K E C C C C C C L V L A A A 90 A A A T L I G Q E C C K K E I P E E A A A Q E K 80 T V I V P S R A K R G D R - K W W A A A A T T A G G G Q D Y S - - Y Y Y R G - G P A S K E 90 S L E V V I G K A Q K S A F - E E - F V - F - - - - I 33

ORF’s encoding for MerT homologs are found upstream of the putative merP genes in both genomes (locus HY04AAS1_1211 in strain AASI, and locus ZP_02177250 in strain R1-1; Figures 2.2, 2.3). Homology includes three hydrophobic inner membrane spanning sequences as determined by TMpred (62), the vicinal C pair in the first membrane spanning sequence, and a C pair located in the cytoplasmic loop between the second and the third membrane embedded sequences (123). No additional nearby ORF’s appeared to contain signature sequences associated with mer functions. Thus, it seems that the mer operons in the two Aquificae strains consist of merT-merP-merA-like genes

(Figure 2.3). tBLASTn using MerR of T. thermophus HB27 (locus TTC0791) did not detect MerR homologs in the genomes of strains AAS1 and R1-1 though several homologs of ArsR, a regulator reported in mer systems of the Archaea

Hypothetical Hypothetical A" merT merP merA protein protein (98 AA) (112 AA) (79 AA) (465 AA) (931 AA) YP_002121873 YP_002121874 YP_002121875 YP_002121876 YP_002121877

216 565 794 2, 188 2, 471 Hypothetical Hypothetical B" merT merP merA protein protein (97 AA) (122 AA) (92 AA) (544AA) (164 AA) ZP_02177249 ZP_02177250 ZP_02177251 ZP_02177252 ZP_002127253

338 710 1000 2, 634 C" merR merT merP merA merD (144 AA) (116 AA) (91 AA) (561 AA) (121 AA)

Figure 2.3 Putative mer operons from Hydrogenobaculum sp. Y04AAS1 (A), Hydrogenivirga sp. 128-5-R1-1 (B) with the Tn501 mer operon from Pseudomonas aeruginosa (C) as a reference. Arrowed boxes indicate each ORF and direction of transcription, with the accession number given directly above each box. Names of putative transcription products and corresponding number of amino acids (aa) are given above the boxes in parenthesis. The numbered line below the ORF represent the genomic nucleotide position starting at 1,123,000 and 618,000 bp for panel A and B respectively.

34 and the Actinobacteria (8, 10) were identified using the ArsR homolog from the sequenced genome of Sulfurihydrogenibium yellowstonense SS-5 (SULYE_0236) as a query.

Mercury resistance in Hydrogenobaculum sp. AAS1, Hydrogenivirga R1-1, and

Persephonella marina Ex-H1

To determine if the mer operons of strains AAS1 and R1-1 are functional, the Hg tolerance profiles of these strains were determined from growth with increasing

concentrations of HgCl2. Such profiles were also determined for P. marina EX-H1 (H1) whose genome does not contain mer gene homologs. This strain was used as a negative control because the technology to create knock out mutants is not yet available for the

Aquificae and thus, obtaining merA- mutants of strains AAS1 and R1-1 was not possible.

The major characteristics of these three Aquificae strains are summarized in Table 2.3.

Table 2.3 Comparison of physiological traits of merA+ Aquificae cultures and merA- control Physiological Data from the Aquificales Data Warehouse (http://alrlab.research.pdx.edu/Aquificales) Hydrogenobaculum Hydrogenivirga Persephonella Characteristic sp. Y04AAS1 sp. 128-5-R1-1 marina str. Ex-H1 Obsidian Pool, Eastern Lau East Pacific Rise, Origin Yellowstone Spreading Center, Mid-Ocean Ridge National Park South Pacific merA Homolog + + - Present Optimal Growth 55 70 70 Temperature (oC) Optimal pH 4.5-5 6 6

o 2- o 2- Electron donors S , S2O3 , H2 S2O3 S , S2O3 , H2 Electron O O O , NO - acceptors 2 2 2 3

Carbon Source CO2 CO2 CO2 Growth Medium Boone’s Medium Boone’s Medium Boone’s Medium Used #5a #2b #2b aShima et al. 1993(137) bBoone et al. 1989(19)

Mercury is known to form complexes with media components, which control its bioavailability and therefore toxicity (39). Our study utilized growth media that included

either S2O3 or H2 as a sole energy source. The known affinity of Hg to sulfur (32, 37,

153) suggested that the added HgCl2 may form very different complexes in presence and absence of S2O3. Therefore, I first determined the speciation of Hg that was added to each growth medium used in experiments that tested strains AAS1 and R1-1 level of resistance to Hg. The results of the employed MINEQL+ modeling may explain differences in tolerance to Hg when cells were grown using different e- donors. In the

-1 -1 presence of 10 μmol L HgCl2 and 8 mmol L thiosulfate, the model indicated that all

Hg(II) speciated as negatively charged Hg-thiosulfate complexes, with 84% of Hg(II) as

36 the -2 dithiosulfate complex and the remaining 16% comprised of the -4 trithiosulfate

complex. In the absence of S2O3, Hg speciated mostly as uncharged HgCl2 and

HgCl(OH) complexes in both growth media (Table 2.4). The model suggested no change

-1 in Hg speciation when the concentration of HgCl2 was increased from 2 to 60 μmol L

HgCl2 (data not shown).

Table 2.4 Results of MINEQL+ modeling of Hg speciation in Aquificae culture media with thiosulfate (S2O3) and hydrogen (H2) as sole energy sources. Electron Donor Hg(II) speciation Boone’s Medium #2a Boone’s Medium #5b -2 2- % Hg(S2O3)2 84.1 84.1 S2O3 -4 % Hg(S2O3)3 15.9 15.9

% HgCl2(aq) 61.2 85.3

% HgClOH(aq) 33.1 H2 -1 % Hg(Cl3) 3.3 13.7

% Hg(OH)2 2.4

Modeled media contained 10 μM HgCl2. No Hg-dependent speciation difference was observed from 2-60 μM HgCl2. aBoone et al. 1989(19) bShima et al. 1993(137)

All three Aquificae strains were tested for their ability to grow in elevated levels of Hg. Typically, Hg(II) concentrations higher than 10 μM select for Hg-resistant

microorganisms (28). Using H2 and S2O3 as electron donors, the lowest tested mercury concentration, 2 μM and 5 μM, respectively, completely inhibited the growth of the control strain P. marina Ex-H1 (Figure 2.4). For strain AAS1, sensitivity to Hg(II) was

enhanced in the H2 supplemented medium (Figure 2.5). The IC50 is a measurement of toxicity indicating the concentration required to inhibit 50% of a given biological

process. The IC50 for strain AAS1 was reduced from 13.5 μM when grown on S2O3 to 2.8

μM Hg(II) when grown with H2 as electron donor.. Strain R1-1 exhibited an IC50 of 7.6

μM Hg(II) when grown on S2O3. Strain R1-1 could not be tested with H2 as it cannot use this electron donor.

Hydrogenivirga sp. (R1-1) Hydrogenobaculum sp. (AAS1) Persephonella marina (H1) 1×1010 0.03 2.0×1009 A 8×1009 B C 1.5×1009 L 0.02 L

09 0 m 6

m 6×10 / / 6 s s l l l l 09 D e 1.0×10 e

09 O C C 4×10 0.01 08 5.0×10 2×1009

0.0 0 0.00 0 10 20 30 40 50 0 20 40 60 80 0 10 20 30 40 50 Time (hours) Time (hours) Time (hours) 0 µM 0.04 4×1009 D E 2 µM 0.03 3×1009 L

5 µM 0 6 m / 6 s l D 0.02 l 09 e 10 µM 2×10 O C

20 µM 0.01 1×1009 40 µM 0 0.00 0 20 40 60 0 20 40 60 80 Time (hours) Time (hours)

Figure 2.4: Growth with S2O3 (A-C) and H2 (D-E) are shown for strain R1-1 (A), AAS1 (B, D), and H1 (C, E). Points represent the means of triplicate samples + 1 standard - deviation. Strain R1-1 does not grow by using H2 as an e donor. Symbols depict different concentrations of HgCl2 for which growth curves were obtained.

When grown on S2O3, strain R1-1 exhibits a different dose response curve than

AAS1. Strain R1-1 was more resistant to 5 μM Hg(II) and less resistant at 10 μM Hg(II), with 7.2 and 88.1% growth inhibition respectively, whereas growth of strain AAS1 was

22.8 and 33.6% inhibited at these two Hg(II) concentrations, respectfully, relative to the no Hg(II) control. Results suggested that the mer systems of AAS1 and R1-1 may have conferred resistance to Hg(II) and that the negatively charged Hg-thiosulfate complexes

were less toxic to strain AAS1 than uncharged Hg(II) in the H2 supplemented medium.

AAS1 - H2 AAS1 - S2O3 R1-1 - S O 100 2 3

50 % resistant

0 0 10 20 30 40 50 60 Hg Concentration (µM)

Figure 2.5 Mercury resistance profiles of Hydrogenobaculum sp. Y04AAS1 grown with

S2O3 (⚫) and H2 () and Hydrogenivirga sp. 128-5-R1-1 grown with S2O3 (). Percent growth was calculated from the direct counts of cultures that were grown in the indicated

HgCl2 concentrations relative to counts in Hg free media obtained at 43 hours for strain

R1-1, and time points 47 and 68 hours for strain AAS1 grown with S2O3 and H2, respectively. Standard deviations of 3 replicate incubations, when not seen, are covered by the symbols.

Loss of Hg(II) during growth of strains AAS1 and R1-1

To determine if growth of strains AAS1 and R1-1 was related to the removal of

Hg(II) from growth media, changes in cell counts were related to Hg(II) concentrations during cellular growth to stationary phase (Figure 2.6). Growth experiments were

performed using 5 or 10 μM HgCl2 in media containing H2 or S2O3 as energy sources, respectively. These concentrations inhibited, though did not abolish, growth in either media (Figure 2.5).

10 4×1009 10 A B 2×1009 8 8 3×1009 Cells/mL 09 Cells/mL 6 6 1×10 2×1009 4 4 08 µmol Hg (II) Lost µmol Hg (II) Lost 5×10 1×1009 2 2

0 0 0 0 0 20 40 60 80 0 20 40 Time (hours) Time (hours)

5 2×1009 4 cells)

6 S O C D a 2 3 H 4 2 3 a Heat Killed 09

1×10 Cells/mL 3 2 2 b 5×1008 b µmol Hg (II) Lost 1 1 c 0

0 0 Initial Hg(II)/h/10 Hg Loss (µmol of Rate 0 20 40 60 80 Time (hours) AAS1 R1-1 Figure 2.6: 203Hg(II) remaining in the medium during growth of strains AAS1(A, C) and

R1-1(B) using S2O3 (A, B, respectively) and H2 (C) as electron donors. Symbols: x, cell density; μmol Hg(II) lost from: ●, growing culture; □, heat-killed control; △, uninoculated control. Initial rates of Hg(II) loss (μmol Hg(II)/h/106 cells) are presented

(D) for strain AAS1 and R1-1 in cultures grown on S2O3 (light column), H2 (dark column), and heat-killed controls (diagonal stripes). Different letters indicate statistical significance (P < 0.05). All data represent the mean difference in Hg(II) loss of a growing culture from an uninoculated control of triplicate cultures + 1 std. dev.

Regardless of e- donor, 203Hg(II) was lost from the growth medium well before the

commencement of growth by strains R1-1 and AAS1. In the S2O3 amended medium, significant amounts of 203Hg(II) were lost from the heat killed controls and the uninoculated media that were incubated at the growth temperatures (55ºC for AAS1 and

70ºC for R1-1; P < 0.05). At the commencement of growth, uninoculated controls of strain AAS1 and R1-1 had only 4.8 + 1.3 and 5.4 + 1.4 μmol Hg(II) remaining in the medium respectively, approximately half of the original starting concentration. No such

loss was observed in the H2 supplemented growth medium. Since no loss of Hg(II) was observed when media were incubated at room temperature (not shown), this abiotic loss

40 likely resulted from elevated temperatures. Nevertheless, the live growing cultures of both strains lost a significantly greater proportion of 203Hg(II) than all abiotic controls (P

< 0.05).

Subtracting the activities from uninoculated controls, rates of Hg loss were 3 fold

- greater in strain AAS1 when S2O3 was provided as an e donor than with H2, with initial rates of Hg(II) loss of 3.2 + 0.4, and 1.2 + 0.1 μmol Hg(II) lost/h/106 cells respectively (P

< 0.05). Strain R1-1 lost Hg(II) at a rate of 2.6 + 0.2 μmol Hg(II) lost/h/106 cells, statistically indistinguishable from strain AAS1 grown on thiosulfate (P > 0.05). Heat- killed controls (105ºC, 30 mins) of strain R1-1 lost a greater amount of Hg(II) from the growth medium than uninoculated controls (P < 0.02), and the rate of Hg(II) loss is in the heat killed was slightly higher than in uninoculated controls (P = 0.054).

The results clearly show that growing cultures of Hydrogenivirga sp. R1-1 and

Hydrogenobaculum sp. AAS1 removed Hg(II) from their growth media. For strain

AAS1 this activity progressed at a higher rate when the energy source was S2O3, as

compared to H2.

Production of Hg(0) by growing cultures of AAS1 and R1-1

To determine if Hg(II) that was lost during growth was reduced to Hg(0) by strains AAS1 and R1-1, end point mass-balance experiments were performed (Table 2.5,

Figure 2.7). When grown to early stationary phase on S2O3, strains AAS1 and R1-1 produced 3.8 + 2.4, and 3.1 + 1.3 μmol Hg(0), respectively. When grown on H2, strain

AAS1 reduced 2.2 + 0.79 μmol Hg(II), statistically similar to when grown on S2O3.

Production of Hg(0) by heat killed controls was approximately 10 fold lower than by growing cultures for all treatments (Figure 2.7). Mass balance calculations showed

41 recoveries of 85 to 122% of the Hg(II) that was added at the beginning of the experiments

(Table 2.5).

10 Hg 203

µmol µmol 5

O 3 2 2 - H - GC - HK - GC - HK - GC - HK 0 - S T 2 2 O 3 O 3 O 3 O 3 T 0 2 2 2 2

AAS1 - H AAS1 - H AAS1 - S AAS1 - S R1-1 - S R1-1 - S

Figure 2.7: Reduction of Hg(II) to Hg(0) by strain AAS1 and R1-1. Means + 1 SD of triplicate cultures are shown, along with the e- donor provided. Abbreviations: GC – Growing Culture; HK – Heat Killed Control. Relevant statistics can be found in Table 2.5.

Table 2.5: Reduction of Hg(II) to Hg(0) by Aquificae cultures. 1 Strain e- donor Treatment2 Medium Hg Headspace Hg Recovery (%) (μM)3 (μM) GC 3.91 + 0.18a 2.19 + 0.79d 121.93 + 5.19a H 2 HK 4.32 + 0.80ab 0.21 + 0.06e 90.51 + 4.51bc AAS1 GC 5.02 + 1.47ab 3.80 + 2.38d 88.28 + 19.29ab S O 2 3 HK 10.87 + 2.60c 0.34 + 0.03f 112.20 + 13.19abc GC 6.69 + 0.48b 3.09 + 1.34d 97.92 + 9.18bc R1-1 S O 2 3 HK 8.13 + 0.48c 0.32 + 0.01f 84.66 + 4.83c 1Different letters indicate statistical significance (p<0.05), similar letters indicate no significant difference by Student’s t-test. 2GC, Growing Cultures; HK, Heat Killed. 3 - HgCl2 at t(0) was 5 μM and 10 μM, when H2 and S2O3 are used as e donors respectively.

MR activities by crude cell extracts of strains AAS1 and R1-1

To further examine if the merA homologs in strains AAS1 and R1-1 encoded for active MR, I measured specific rates of Hg dependent NAD(P)H oxidation in crude cell extracts of both strains. Preliminary experiments showed a preference (higher activity ratio in treatments with and without Hg) for NADH by R1-1 extract and for NADPH for

AAS1 extract (data not shown) and therefore all further experiments were performed with each strains’ preferred reductant. Extracts of strain H1 did not exhibit Hg-specific enzyme activities as compared to a no-Hg control when tested at the optimal growth temperature of this strain (0.19 + 0.22 mU mg-1, 70° C).

Given the developing understanding of the importance of mer in thermophilic microbes (8, 132, 157), I determined the effect of temperature on MR activities in crude cell extracts of strains AAS1 and R1-1. Results indicate that the optimal temperature for

Hg dependent NAD(P)H oxidation corresponded with each organism’s optimal growth temperature (Figure 2.8). Maximum apparent specific MR activity for AAS1 was 37.7 +

1.1 mU mg-1 at 50° C, near this strain’s optimal growth temperature of 55° C. Likewise,

43 optimal specific activity in strain R1-1 extracts were observed at this strain’s optimal growth temperature (70ºC) and were ~6 fold lower, at 6.4 + 0.2 mU mg protein-1, as compared to the apparent specific activities of strain AAS1. Results indicated that strain

AAS1’s MR was active at a temperature range of 30 to 70ºC, whereas extracts of strain

R1-1 were active from 60 up to 87ºC. Very low activities were recorded in both extracts at temperatures below 40ºC (Figure 2.8). These results suggest that strains AAS1 and

R1-1 produced active MR likely encoded by their merA homologs.

50 8 Act. (mU mg R1-1 Spec. ) -1 40 6

30 4 20

2 10 -1 ) AAS1 Spec. Act. (mU mg 0 0 0 20 40 60 80 100 Temperature (°C)

Figure 2.8: Effect of temperature on specific MR activity. Specific activities of cell-free extracts were determined at increasing temperatures for Hydrogenobaculum sp. Y04AAS1 () and Hydrogenivirga sp. 128-5-R1-1 (⚫). Average of three to five replicates + 1 standard deviation are represented. One activity unit = nmole NAD(P)H -1 - oxidized min . Cultures were prepared for the assay by growth using S2O3 as an e donor.

Induction of mer activities

mer operon expression is regulated in both Bacteria (147) and Archaea (131).

Bacterial mer operons are mostly regulated by MerR and in some cases by ArsR-like regulators (10). An ArsR-like regulator was described in the mer operon of Sulfolobus solfataricus P2 (132). tBLASTn searches with MerR of T. thermophilus HB27

(YP_004764) and ArsR of S. yellowstonense SS-5 (SULYE_0236) as queries did not detect ORFs recognized as those coding for either regulator in the proximity of the merA homologs in the genomes of either strain AAS1 or strain R1-1. In order to test if merA expression in these strains was induced by Hg(II), I determined levels of merA transcripts and of MR specific activities in cultures that were grown in presence or absence of

Hg(II). Induction of merA transcription was determined as the ratio of transcript levels, normalized to the levels of a constitutively expressed gyrA gene, obtained for cultures that were grown with or without 1 µM Hg(II). Primer sets RT-HBmerA(f/r), RT-

HBgyrA(f/r), and RT-HVmerA(f/r), RT-HVgyrA(f/r) were designed for this study

(Materials and Methods; Table 2.2). Results clearly showed that exposure to this sub toxic concentration of Hg(II) did not induce merA expression as indicated by ratios of 1 for the first 3 hours following exposure (Figure 2.9). In contrast, a similar approach clearly showed induction of merA in T. thermophilus HB27 and in an E. coli culture carrying a chromosomal insertion of Tn501 (157). These results, suggesting that merA expression in strains AAS1 and R1-1 was constitutively expressed, a finding that is supported by similar MR specific activities in crude cell extracts of cultures that were grown in presence or absence of 2 to 5 μM Hg(II) (Figure 2.10).

1.5

1.0

0.5 Induction Fold Induction A 0.0 0 50 100 150 200 Time (mins)

1.5

1.0

0.5 Induction Fold Induction B 0.0 0 50 100 150 200 Time (mins)

1.5

1.0

0.5 Induction Fold Induction C 0.0 0 50 100 150 200 Time (mins)

Figure 2.9: merA fold induction for strain AAS1 grown on H2 (A) and S2O3 (B), and strain R1-1 growing on S2O3 (C). The mean + 1 standard deviation of triplicate cultures are shown. Induction fold represents the relative induction fold of merA in cultures grown in presence of Hg(II) relative to those grown in its absence. Data were normalized to expression of a constitutively expressed control, gyrA.

30 )

-1 a a 20 a

10 Specific Activity (mU mg 0 AAS1 R1-1 Strain

Figure 2.10: Effect of Hg on MR activities in crude cell extracts of strains AAS1 and R1-1. Cell free extract activities were determined for cultures growth with (filled column) or without (clear column) HgCl2 in the media. Strains AAS1 and R1-1 were grown in 5 and 10 μM HgCl2, respectively. Similar letters above columns indicate no significant difference by Student’s t-test (P<0.05).

Discussion

This study is the first to demonstrate mer activities in the Aquificae phylum, and only the second description of mer activities in deep branching bacterial lineages (157).

With these studies, along with previous work characterizing mer in the Archeon

Sulfolobus solfataricus P2 by Schelert et al. (131, 132), we may begin to understand the natural history of mer by comparing mer from microorganisms inhabiting geothermal environments that belong to deep branching microbial lineages to the more thoroughly studied mer systems in heterotrophic, aerobic Proteobacteria and Gram-positive bacteria.

The Aquificae mer operons

Evolution of the mer operon has been previously addressed. Osborn et al. (106) hypothesize that mer evolved at a time when global Hg concentrations were higher than at present day due to greater volcanic activities. Furthermore, Helmann et al. (60) proposed that the evolution of mer preceded the divergence of Gram-negative and Gram- positive bacteria based on amino acid sequence conversation between the merR of sp. and Pseudomonas sp. These studies only included data of mer operons from recently diverged lineages, namely the Proteobacteria and Gram-positive bacteria, representing a relatively short time span in microbial evolution. A recent work on MR evolution used phylogenetic reconstructions to show that MR from thermophilic taxa were consistently found among early branching lineages while being scarcely distributed among more recently evolved lineages (8). Furthermore, distribution patterns among extant microorganisms combined with thermodynamic modeling of the effect of redox on ionic mercury speciation suggested evolution of MR only after the oxygenation of Earth, for sufficient bioavailable Hg(II) was most likely unavailable on early Earth (8). An

48 updated phylogenetic reconstruction of MerA from all prokaryotes showed that sequences of three Aquificae in a cluster with the sequence of the bacterium

Desulfovibrio desulfuricans SSM1, a strictly anaerobic heterotroph isolated from a deep- sea hydrothermal vent (150) branched sister to all archaeal sequences (Figure 2.1).

Together, this clade was sister to a cluster that included all other bacterial MerA. Thus the updated phylogeny slightly modifies conclusions from the previous reconstructions where Aquificae sequences branched basal to all Bacteria and Archaea (8). The new reconstruction clearly places the origin of MR in a thermophilic ancestor of the Bacteria and confirms the previously identified horizontal gene transfer from an Aquificae ancestor to the Archaea (10). Taken together, these results highlight an important evolutionary position for Aquificae merA, potentially representing an ancestral state of

MerA amongst all prokaryotes, or perhaps an evolutionary link of mer functions between the Archaea and the Bacteria.

The mer operons in strain AAS1 and R1-1 are structurally simpler than most other mer operons, only including homologs of merT and merA, and possibly merP homologs

(Figures 2.2, 2.3). Among thermophilic prokaryotes T. thermophilus HB27 and S. solfataricus P2, whose operonic structure includes only merA and merR (157) or arsR- like (131, 132) genes, the Aquificae mer operons are unique by the presence of transport, and absence of regulatory functions. Aquificae genomic BLAST searches with MerR from T. thermophilus HB27 (YP_004764) and ArsR from the aquificaeon S. yellowstonense SS-5 (EEP61239) reveal no ORF in the immediate gene neighborhood to merA encoding for either one of these two metallo-regulatory proteins. Homology in conserved regions of MerA, MerT, and MerP is observed in an alignment alongside that

49 of the well-characterized mer operon of transposon Tn501 (Figure 2.2). The MerT homologs in both strains R1-1 and AAS1, bear the three membrane spanning sequences and the two signature CC pairs, one in the first membrane spanning sequence and the other in the inner cytoplasmic loop (123), suggesting the possibility of functional transporters. This however is not the case for the MerP homologs. While the characteristic metal binding motif is present in both, the lack of a leader peptide in the

AAS1 homolog and the lack of sequence similarity to known leader peptides of the amino acid terminus in R1-1 question whether the gene products of these homologs can be translocated into the periplasm, the cellular location of MerP (10). As little is known about protein translocation in early bacterial lineages and without experimental evidence, one cannot predict functional periplasmic mer function in strains AAS1 and R1-1. It should be noted that the amino terminus of the MerP homolog of R1-1 includes a vicinal arginine pair followed by a 9 hydrophobic amino acids sequence, reminiscent of a Tat motif which is known to play a role in prokaryotic cross membrane transport (93).

Despite the low probability that this sequence is a Tat motif, as described in results, experimental evidence may be able to show whether or not this sequence directs transport of the MerP homolog to the periplasm in strain R1-1.

NmerA is a metallochaperone-like N terminal domain, which shares homology to metal chaperones whose primary function is intracellular trafficking of soft-metal ions

(10). Presence of NmerA may enhance the efficiency of Hg(II) removal by active MR

(73). Under conditions of depleted glutathione, decreased Hg(II) resistance was observed in an experimentally created E. coli carrying MR lacking an NmerA (77). Among emerging similarities between merA in previously characterized S. solfataricus and T.

50 thermophilus, with the two newly characterized Aquificae merA is the presence or absence of the N terminus domain, NmerA (8).

Clear distribution patterns among microbial taxa have been observed for full

MerA, those containing an NmerA domain, and core MerA, those lacking NmerA (8).

Later branching microbial lineages with full MerA include the Firmicutes, Beta- and non- marine Gammaproteobacteria. Core MerA dominates in the Alphaproteobacteria, the marine Gammaproteobacteria, the Actinobacteria, and the Archaea. It has been proposed that these patterns may be related to the types and levels of intracellular thiol agents whose role is to assure a reduced cytoplasm (8). The presence of NmerA among early branching thermophilic lineages is more variable, with full MerA in

Hydrogenivirga sp. R1-1, H. thermophilus TK-6, and D. desulfuricans SSM1, while core

MerA are found in T. thermophilus strains HB27 and HB28 and in Hydrogenobaculum sp. AAS1. Both S. solfataricus and T. thermophilus represent early microbial lineages and were isolated from high-temperature, high-sulfide geothermal environments with low oxygen solubility, it has been suggested that NmerA may not be needed for enhanced Hg resistance in these organisms (157). Strain AAS1 was isolated from Obsidian Pool, YNP, also a high temperature, sulfur rich environment (140), and is lacking NmerA.

Interestingly, NmerA is present in strain R1-1, isolated from hydrothermal vents in the

Lau Basin, South Pacific. Hydrothermal vents are characterized by rapid mixing of anoxic hydrothermal fluid with oxygenated seawater (154), where anoxia and high sulfide concentrations are transitory. It should be noted that the distribution patterns of

NmerA presence or absence suggest multiple acquisition and loss events, and

51 phylogenetic reconstructions showed that NmerA evolved at least twice, with independent events leading to NmerA in the Proteobacteria and the Firmicutes (8).

Mercury resistance in Hydrogenobaculum sp. Y04ASS1 and Hydrogenivirga sp. 128-

5-R1-1

Observed Hg(II) resistance of both AAS1 and R1-1 are significantly higher than observed resistance of the only other characterized mer system in thermophilic bacteria,

T. thermophilus HB27, where 4 μM HgCl2 completely inhibited growth (157). This was

notably lower than 60 and 10 μM HgCl2, which completely inhibited growth of AAS1 and R1-1, respectively, when grown on S2O3. HB27 was grown on rich medium, further highlighting its higher sensitivity, as Hg bioavailability is known to be decreased in the presence of organic ligands (7). The mer operon of HB27 lacks homologous sequences to transport proteins (MerT), whereas similar sequences are found in both AAS1 and R1-

1 (Figure 2.3). The presence of a MerT-like transport protein, resulting in efficient Hg transport to the cytosol, may confer higher resistance in AAS1 and R1-1 as compared to

HB27.

When grown on S2O3, strain AAS1 and R1-1 exhibited different Hg-resistance profiles (Figure 2.5). Strain R1-1 was more resistant at 5 μM Hg(II) and less resistant at

10 μM Hg(II) than AAS1. mer operons from both strains encode for homologs of MerA,

MerT, and possibly MerP. However, the putative mer operon of R1-1 encode for a full

MerA that includes the n-terminal domain NmerA as well as for a likely MerP homolog

(Figure 2.3). It is possible that at low Hg concentrations, NmerA and MerP-like proteins enable increased Hg resistance in R1-1 due to increasing efficiency of Hg movement in

52 the cell. However, at higher Hg concentrations, this system may be overwhelmed, resulting in increased Hg toxicity.

203 In the S2O3 supplemented medium, an abiotic loss of Hg from the growth medium was observed, whereas no such loss occurred in the H2 amended medium. This suggests that uncharged Hg(II) complexes are more stable at elevated temperatures than

the charged Hg-S2O3 complexes. High percent recovery and statistically indistinguishable recovery of headspace 203Hg for both cultures in the mass balance assays (Figure 2.7 and Table 2.5) suggest that the aforementioned observed loss of Hg(II) may be simply due to 203Hg evaporation from the growth medium.

Unfortunately, the observation of an abiotic loss of Hg(II) from the growth

medium in the presence of S2O3, with no such loss when H2 is included, renders any comparison of Hg speciation affecting its toxicity irrelevant in this study. At the commencement of growth, uninoculated controls of strain AAS1 and R1-1 had only 4.8 +

1.3 and 5.5 + 1.4 and μmol Hg(II) remaining in the medium, respectively, approximately half of the original starting concentration. No significant amount of Hg was lost

abiotically when H2 was included as the electron donor (P > 0.05). Thus, the possibility

remains, that the increased “Hg-resistance” of strain AAS1 in S2O3-relative to H2- supplemented media was simply the result of proportionally less Hg(II) in the aqueous fraction.

Data are conflicting when attempting to understand the difference in Hg resistance between strain R1-1 and AAS1 beyond those suggested by the amino acid

sequences of MerA and MerP (see above). When grown on S2O3, strain AAS1 and R1-1 are statistically indifferent in rate and mass of Hg(II) lost from growth medium (Figure

2.6), as well as mass of Hg(0) production (P > 0.05; Figure 2.7 and Table 2.5). The only measure by which the activities of the two strains differ were that the apparent specific

MR activities of AAS1 were 5 fold greater than those in extracts of R1-1 (P < 0.05;

Figure 2.8). Yet in other instances, MR activities in extracts of the two strains were similar (Figure 2.10). All together, a high variability in apparent specific activities was observed throughout the study. Given the inherent crude nature of assaying enzymes in crude extracts, this may not be surprising and a finer comparison between the activities of the aquificaeal MR may await assays of purified enzyme preparations. At this juncture, I can only suggest that there is no clear difference between the Hg-resistance capabilities of the two newly characterized strains.

merA transcription was not induced by the addition of subtoxic concentrations of

Hg(II) to mid-log phase cultures of both strains AAS1 and R1-1 under all conditions tested, suggesting a constitutive expression of mer functions (Figure 2.9). This observation was verified by MR analysis in crude cell extracts, showing no statistically significant differences in apparent specific rates between culture grown with (5-10 μM

Hg[II]) or without Hg(II) prior to harvesting (P > 0.05; Figure 2.10). This observation was not unexpected, given the absence of merR or arsR-like gene homologs in the genomic neighborhood surrounding mer genes in either culture. Previously, similar experiments showed a 1300 and 65 fold increase in merA transcription upon addition of 1

μM Hg(II) to cultures of E. coli containing Tn501 and of T. thermophilus HB27 (157). In light of the issues described above regarding an abiotic loss of Hg(II) in the presence of

S2O3, it is promising that merA transcription in strain AAS1 grown on H2 did not appear to be induced by an Hg(II) spike. With the addition of constitutively expressed mer

54 systems reported in this study, the story of evolution of mer becomes clearer; from a constitutively expressed to highly regulated system (Figure 2.11).

Taken together with previously characterized thermophilic mer in HB27, it is fascinating that among thermophiles, optimal temperature for MR activity corresponds directly with the optimal growth temperature of each organism while the optimal temperature for the Tn501’s enzyme was > 20 °C higher than the optimal growth temperature of P. aeruginosa, the mesophilic Hg resistant bacterium from which Tn501 was isolated (Figure 2.12) . It has been previously proposed that this discrepancy between the optimal

10,000 Tn501 HB27 1,000 AAS1 - H2 AAS1 - S2O3 100 R1-1 - S2O3

10 Induction Fold Induction 1

0.1 20 40 60 Time (mins)

Figure 2.11: merA fold-induction in strains AAS1 and R1-1 as compared with induction of the mer operons of HB27 and Tn501. The mean + 1 standard deviation of triplicate cultures are shown. Induction fold was calculated as described in the legend to Figure 2.9 except that merA expression in Tn501 and HB27 was normalized to expression of 16S rRNA genes rather than to gyrA.

50 ) 1 - g

m 40

U m (

30 y t i v i t

c 20 A

c i f i c 10 e p S 0 0 20 40 60 80 100 Temperature (°C)

Figure 2.12 Effect of temperature on MerA activities. Specific activities of cell-free extracts were determined as previously described for: Hydrogenobaculum sp. Y04AAS1 (); Hydrogenivirga sp. 128-5-R1-1 (x10) (⚫); T. thermophilus (157) (♢); Tn501(154) (△). Average of three to five replicates + 1 standard deviation are represented. One activity unit = nm NAD(P)H oxidized min-1. growth and enzyme activity temperatures could be interpreted as a relic of MR evolution at high temperature environments (157). My results, therefore, strengthen this conclusion and the identification of thermophilic organisms from high temperature environments as the origin and early evolution of the microbial Hg detoxification system (8, 154, 157).

This study adds to a growing body of literature characterizing mer function in deep branching Bacteria (157) and Archaea (131, 132) and will inevitably aid in our understanding of the evolution and characteristics of early prokaryotic mer systems.

Future studies may increase our understanding of mer in Aquificae through purification

of MR, or expression of merA in E. coli. Futhermore, Hg(II) inputs other than HgCl2 can be included to better understand the effect of Hg speciation on its toxicity. By further comparing mer systems from deep branching microbial lineages to those from organisms inhabiting metal contaminated environments, we hope to expand our comprehension of the evolution of mer-mediated microbial mercury detoxification.

Chapter 3: Diversity and Distribution of merA in Geothermal Environments and on a Global Scale: Novel Insights into the Ecological Structure of Mercury Resistant Communities.

Introduction:

Mercury (Hg) is a potent neurotoxic substance, affecting both humans and wildlife due to it’s global distribution (10) and the accumulation of methyl-mercury

(MeHg) in aquatic and terrestrial food webs (22). Due to its high affinity for sulfur and sulfhydryl groups, Hg represents the most toxic heavy metal to microorganisms (98). In

Bacteria (8, 10, 157) and Archaea (142), the presence of an elaborate system of Hg- detoxification, the mercury resistance (mer) system, facilitates survival at elevated Hg concentrations. The mer system includes a suite of enzymes that are involved in the reduction of Hg(II) to Hg(0), and sometimes the degradation of MeHg. If present, additional mer-encoded proteins are involved with Hg transport into the cell as well as regulating expression of the mer operon (10). The core function of mer is the reduction of Hg(II) to Hg(0) which may then be volatilized and removed from the local environment. This transformation is catalyzed by the enzyme mercuric reductase (MR), an NAD(P)H dependent flavin oxidoreductase, encoded by the merA gene (10).

Most of our knowledge of metal resistance is based on research with microbes from recently contaminated environments (8). Research addressing controls on the diversity and distribution of merA have been conducted in contaminated soils and sediments (26, 97, 105), natural waters (9), and biofilms (112). Features such as shallow

(146), and deep-sea vents (32), volcanoes (102), geysers (101), hot springs (22, 74), and fumaroles (38) are known geologic sources of Hg to the environment. Thermal features

57 in Yellowstone National Park (YNP) have been shown to reach concentrations of micrograms total Hg per liter (22, 74, 136), similar to concentrations found in environments with high input of anthropogenic Hg (96, 129). Mercury is released in the aqueous phase (74, 156), and accumulates across trophic levels in YNP (22). On a global scale, recent UN estimates show that approximately one-third to one-half of Hg emissions to the atmosphere come from geothermal sources (24), yet very little is known about the interactions of microbial communities with Hg in geothermal environments.

Culture independent techniques have been used to amplify Hg resistance genes from hot springs in YNP (156) and Coso Hot Springs, CA (142), and Hg resistant organisms have been previously isolated from YNP hot springs (49, 157), and other geothermal environments (28, 131, 154).

The tendency of a lineage to retain ancestral ecological traits and environmental distribution is known as niche conservatism (163), which can be realized by observable patterns in the distribution of phylotypes within assemblages (phylogenetic structure) along a physicochemical gradient (59, 160). Microbial communities have been shown to adapt to life under elevated Hg concentrations by enrichment of resistant organisms (6,

92, 115, 116), synthesis of mer gene products (96, 129), and horizontal gene transfer of mer genes (8, 28, 115). Patterns in the distribution of merA have been previously investigated (97, 156). However the degrees to which Hg-resistant communities are structured relative to physical and chemical gradients in the environment and to the total species pool in that environment, remain largely unknown.

Geothermal springs in YNP are geochemically diverse and exhibit physical and chemical gradients within (82) and among (100, 139) geothermal features. These

58 conditions provide an ideal environment in which the effect of physical and chemical parameters on the structure of Hg-resistance communities can be observed. Additionally, aforementioned physical and chemical gradients create strong selective pressures that may result in the presence of an assortment of ecological niches, with some locations supporting the growth of certain Hg-resistant microorganisms and limiting the growth of others.

This study was initiated to assess the effect of geochemical gradients on merA gene diversity and distribution. Hot springs were chosen for this study to emphasize the effect of physicochemical conditions previously suggested to have an effect on merA diversity and distribution, namely pH, DOC, and sulfide (156). Here, I report the distribution and phylogenetic diversity of merA in two geochemically diverse YNP hot springs, and the degree to which these are structured relative to the microbial community as a whole as represented by the diversity of 16S rRNA gene PCR amplification products. These data were then corroborated by a global meta-analysis of environmental merA gene sequences available on NCBI (http://www.ncbi.nlm.nih.gov/popset).

Phylogenetic analysis demonstrates evidence of ecological structure, realized by significant clustering of phylotypes within assemblages. Furthermore, cluster analyses revealed different clustering patterns in 16S rRNA and merA gene assemblages, suggesting unique physicochemical controls on 16S rRNA and merA gene communities.

Meta-analysis revealed similar structure on merA communities from 23 assemblages on a global scale. These results obtained begin to shed light on the dispersal limitations and niche conservatism of Hg resistant community, as deciphered by the diversity and distribution of merA, in both YNP geothermal environments and on a global scale.

Materials and Methods

Sample Collection and Analysis

Approximately 1 g samples of microbial mat and sediment were collected from two YNP hot springs with contrasting geochemical conditions that may influence both

16S rRNA and merA diversity and distribution. Bijah spring (44° 45.670’ N, 110°

43.857’ W) is located in the vicinity of Nymph Lake within the Gibbon River watershed in YNP. Succession Spring (44°43’75.7” N, 110°42’72.7” W) is located in the Hundred

Springs Plain areas of Norris Geyser Basin, YNP (Figure 3.1).

Figure 3.1: Selected geothermal springs in Yellowstone National Park. Hot springs selected to sample for this study are underlined. Adapted from Wang et al. (156).

Bijah Spring is characterized as a basic silicate-sinter-rich hot spring (74), whereas Succession Spring is an acid-sulfate-chloride spring (82, 156). Samples were collected on June 18-19th, 2007 near each spring source, at the first site where visual observation of microbial biomass was observed, and at additional sites according to visual observation of colored-precipitate, indicative of unique metabolic processes by the spring microbial community (Table 3.1). Water temperature and pH were recorded on-

60 site at the time of sampling using a model 59002-00 Cole-Parmer temperature compensated pH meter (Vernon Hills, IL, USA). All samples were taken aseptically with

EtOH sterilized spatulas and forceps, placed in a 1.5 mL cryovial (Fisher Scientific,

Waltham, MA) and immediately flash frozen on dry ice. Samples were kept on dry ice at all times during transport to the laboratory, where they were kept at -80° C until analysis.

DNA Extraction and PCR Amplification of 16S rRNA and merA genes

Approximately 0.1 g of microbial mat and sediment was subjected to DNA extraction using the PowerSoil DNA Kit (Mo Bio Laboratories, Solano Beach, CA) as previously described (97). DNA was quantified using a nanodrop spectrophotometer

(Thermo Scientific, Wilmington, DE). To ensure the presence of PCR-amplifiable DNA from each extract, 16S rRNA genes were amplified using Universal Bacterial primers

27F(5’-AGAGTTTGATCMTGGCTCAG-3’)/519R (5′-GWATTACCGCGGCKGCTG-

3′) (75). PCR products obtained were used in preparation of the 16S rRNA gene clone library. merA primer sets 1-4 were used to amplify merA DNA from major clades in the merA phylogeny; target sequences, reagent concentrations, and reaction conditions used have been previously established (156). Primer sets used to generate merA PCR products used for clone library construction are described in table 3.2.

Creation of 16S rRNA and merA clone libraries

16S rRNA and merA genes clone libraries were created as previously described using the pGem-T Easy cloning kit (Invitrogen, Carlsbad, CA) (156). PCR products were purified using the QIAquick gel extraction kit (Qiagen, Valencia, CA), cloned using pGEM®-T Easy (Promega Corp., Madison, WI), and transformed into competent MAX

Efficiency Escherichia coli DH10B (Invitrogen, Carlsbad, CA).

61 !

d d c b ( a 3.1: Table Succession 2 Succession 1 Bijah 2 Bijah 1 Site Measurements were taken sampled from precipitated elemental flocs sulfur Temperature andwere pH taken at the time sampling 18 of (June NA NA Sample taken at visualfloc. first observation sulfur of Numbers in parenthesis denote filtered Hg. 156

) – .

not available.

a 3.2: Table Geochemical parameters study sites in the of two YNP hotselected springs this study. for Information from Wang etInformation Wang from al. 4 3 2 1 set Primer d

6 3 45 12 from Source from

Distance

(m)

merA Target content GC

(mole %) - specific primer PCR sets High High Low Low 45.5 57 59 65 Temp (

o C)

2.50 2.56 8.10 7.76 pH ( 156

Sulfolobus acidophilum solfataricus, Thermoplasma thermophilus Proteobacteria, Firmicutes, Actinobacteria Proteobacteria, Firmicutes, Actinobacteria Proteobacteria, Targeted

) . 324 14.4 9.9

Sulfide Sulfide ( μ

g/L) c

a merA

NA 12.8(10.2) 14.8(12)

THg from microbesfrom belonging to: Firmicutes, Actinobacteria (ng/L) d (64.6)

water -

19, 2007); all other geochemical data areet Wang from al.,

NA 1.05 2.17 Gaseous Hg Hg Gaseous Dissolved (ng/L) d

( 156 b

) . NA 0.38 0.21 Gaseous Hg Hg Gaseous

Reactive (ng/L) d , Thermus , Thermus

0.96 0.579 0.652 (mg/L)

DOC DOC c Amplicon size

1246 1249 (bp) 662 310 8.99 2.31 0.96 THg

( μ

g/g)

solid

650 284 27 MeHg

(pg/g) c

solid

Transformants were selected by growth on X-gal (40 μg/mL, Fisher, Waltham, MA) plates including 100 μg/mL ampicillin. Plasmid DNA was isolated using the Plasmid

Mini-Kit (Qiagen, Valencia, CA). Insert presence was verified by digestion using

EcoR1, inserts of expected size were then sequenced by Genewiz Inc (South Plainfield,

NJ) using universal primers M13(-21) and M(13)R.

Determination of Mole Percent of G + C of merA PCR Products

The G + C % of sequenced amplicons obtained with primer sets 2 and 3 were calculated by uploading and analyzing the sequences in the Genomics percentage of G ∼

C Content Calculator (http:// www.sciencebuddies.org/science-fair-projects/projects_ ideas/Genom_GC_Calculator.shtml).

Phylogenetic Analysis of 16S rRNA gene and merA clones

Both merA and 16S rRNA gene Sequences were aligned using ClustalX (version

2.08) (76) using the IUB DNA weight matrix with default gap extension and opening penalties; alignments were manually adjusted if necessary. For merA amplicons, forward and reverse strands were sequenced and aligned for redundancy as a quality control measure. Sequence identity matrices were created in the DNADIST program within

PHYLIP (version 3.6) (40) using the Jukes-Cantor nucleotiode substitution model, and were used within the program MOTHUR (134) to define and group operational taxonomic units (OTU’s) and to perform rarefaction analysis. A 97% nucleotide sequence similarity was used as a cutoff point to define 1 OTU for 16S rRNA gene clones. Due to high sequence similarity of merA gene clones, a 99% cutoff to define an

OTU was used to provide a more resolved snapshot of community diversity.

To properly assess the phylogenetic position of the deduced MerA that were encoded by the merA gene clones, DNA sequences were translated using the translate tool on the ExPASy proteomics server (http://web.expasy.org/translate/). For phylogenetic reconstruction, a randomly selected representative MerA sequence for each

OTU (OTU defined as described above) from YNP microbial mat MerA sequences and

MerA reference sequences obtained from GenBank

(http://www.ncbi.nlm.nih.gov/genbank/) were aligned as described above using a Gonnet

250 protein weight matrix and edited in Seaview when necessary (52). The NmerA extension, which is not a universal feature of MerA (8), was not included in the phylogenetic analysis. Dihydrolipoamide dehydrogenase sequences, paralogs of MerA

(48), from Thermus thermophilus HB27 (YP_005669), Magnetospirillum magneticum

AMB1 (YP_423326), and Pseudomonas fluorescens Pf01 (YP_351398) were used as outgroups. 16S rRNA nucleotide sequences were aligned and edited as described above.

The 16S rRNA sequences of the archaeons Sulfolobus sulfataricus P2 (X03235) and

Pyrobaculum calidifontis JCM 11548 (AB078332) were used as outgroups.

MrBayes (version 3.1) (67) was used to determine the phylogenetic position of the 16S rRNA and MerA homologs. 16S rRNA gene phylogentic reconstructions were determined using the GTR substitution model with gamma shaped rate variation and a proportion of invariable sites as recommended by the ModelTest Server (ver. 3.8) (111) using a total of 100 bootstrap replicates. For MerA phylogenetic reconstructions, ProtTest

(54) was used to select the Whelston and Goldman (WAG) evolutionary model with fixed amino-acid frequencies and a proportion of invariable sites as recommended by the server. Topologies for both 16S rRNA gene and MerA phylogenies were sampled every

10 generations for 1 x 106 generations with a burnin value of 2 x 105 (standard deviation between split frequencies < 0.04). Consensus phylograms were viewed and edited in

TreeEdit (version 1.0a10) (114).

Nucleotide sequence Accession Numbers

All nucleotide sequences included in this study have been deposited to GenBank under the accession numbers JQ228628 to JQ228763 and JQ228764 to JQ228802 for 16S rRNA and merA clone libraries respectively.

Community Analysis

Putative merA and 16S rRNA gene sequences from each of the 4 environments were imported into MEGA4 (ver. 4.0.1) (151) and aligned with ClustalW using the IUB substitution matrix and default gap extension and opening penalties. The aligned nucleic acid sequences for both the 16S rRNA and merA genes were exported from MEGA4 and used for phylogenetic analyses.

Both 16S rRNA and merA alignments did not include reference sequences except sequences that were used as outgroups which served as a “root” for all further analyses.

The dihydrolipoamide dehydrogenase nucleotide sequences from Bacillus cereus

03BB102 (CP001407) and Alkaliphilus metalliredigens QYMF (CP000724) served as outgroups for merA phylogenetic reconstruction. Likewise 16S rRNA genes from

Acidilobus sulfurireducens str. 18D70 (EF057391) and Caldisphaera draconis str. 18U65

(EF057392) served as outgroups for bacterial 16S rRNA gene phylogenetic reconstruction when necessary.

The phylogenetic position of bacterial 16S rRNA and merA genes was assessed using PhyML (ver. 3.0) (54) using the GTR substitution model with gamma-shaped rate

65 variation and a proportion of invariable sites as recommended by ModelTest server (ver.

3.8) (111). A total of 100 bootstrap replicates were performed for each phylogenetic reconstruction. The merA and bacterial 16S rRNA gene consensus phylograms were rate- smoothed using the data driven penalized likelihood approach (125) as implemented in the Ape (ver. 2.5) (107) package within R (ver. 2.10.1) using an alpha smoothing parameter of 0.50. The resulting rate-smoothed phylograms for 16S rRNA, and merA gene nucleotide sequences were used as inputs for the following analysis.

Community ecological and statistical analyses

Both merA and bacterial 16S rRNA gene rate-smoothed phylograms were used to

calculate Rao’s quadratic entropy (Dp), net relatedness index (NRI), and the nearest taxon index (NTI) using Phylocom (ver. 4.0.1) (161). I tested if these metrics differed significantly from a null model based on a randomly assembled community using a two- tailed significance test based on 1, 000 independent permutations. Significance was assumed when the random null model was supported in > 975 of the permutations

(P<0.05). Importantly, Dp represents an abundance weighted (e.g., frequency by which a terminal is detected in a clone library) measure of the total branch length associated with an assemblage relative to the total sequence pool. Assemblages with a relatively higher

DP exhibit a greater phylogenetic diversity relative to the total sequence pool. Phylocom was also used to construct a community phylogenetic distance matrix using Rao phylogenetic distances. The Rao phylogenetic distance matrix was used in a hierarchical cluster analysis, using 1000 bootstrap replicates within the R package pvclust

(http://www.is.titech.ac.jp/~shimo/prog/pvclust/). Hierarchical clustering was based on

Ward’s agglomerative correlation method (159).

Rooted 16S rRNA gene and merA gene rate-smoothed phylograms were also used to calculate a distance matrix using the UniFrac metric (80). UniFrac differentiates environments based on the sum of branch lengths that are unique to each environment.

The abundance weighted normalized UniFrac distance matrix was visualized by random start Nonmetric Multidimensional Scaling using MetaMDS within the R package vegan as in Rousk et al. (124).

Global merA Gene Community Analysis

In addition to merA sequences obtained in this study, environmental merA sequences were recovered from the NCBI database into a local database. merA community data was available for a total of 7 environments; anaerobic sediment enrichments from Berry’s Creek, Meadowlands, NJ (Accession numbers: DQ132515–

DQ132604) (97), soils from Sunday Lake, Adirondack Park, NY (Accession numbers:

EU292169-EU292209) (Yu, Unpublished), biofilms associated with marine algae in the vicinity of Cornwallis Island, Nunavut, Canada (Accession numbers: DQ408728-

DQ408744) (112), soils from East Fork Poplar Creek, Oak Ridge, TN (Accession numbers: EF460128-EF460310) (105), microbial mats from hot springs, YNP, WY

(Accession numbers: EU259721-EU259733; FJ639661-FJ639666) (156) and two sediment samples collected from medium-size fish farms in the Turku, Finland

Archipelago and two small-to-medium scale farms in the Stockholm, Sweden

Archipelago, Northern Baltic Sea (110). Primer sets used in the creation of clone libraries from each location are shown in table 3.3. Environmental putative merA gene sequences from each of the 21 bacterial and 2 archaeal assemblages were imported into

MEGA4 (ver. 4.0.1) (151) and aligned with ClustalW using the IUB substitution matrix

67 and default gap extension and opening penalties. Alignments were trimmed to include only regions of homology shared by all sequences. Dihydrolipamide dehydrogenase genes from Alkaliphilus metalliredigens QYMF (Amet_1254) and Bacillus cereus

NC7401 (AP007209) were included as outgroups. The resulting 240 bp alignment of merA nucleic acid sequences were exported from MEGA4 and used for phylogenetic analyses. Rooted, rate-smoothed phylograms were used to calculate a diversity and distance matrices using Phylocom and UniFrac as described in “community ecological and statistical analysis”. Ward’s agglomerative hierarchical clustering was not performed with the global merA assemblages.

68 Sources: a 3.3: Table ! β National Park Hot Springs, Yellowstone Archipeligo, Baltic Sea, Fish Farm Turku and NY, USA Sunday Lake, Adirondak Park, Oak Ridge, TN Cornwallis Island, Nunavut, CA NJ, USA Berry’s Creek, Meadowlands,

–

Betaroteobacteria;

Ni Chadhain

merA Site Study

primerPCR sets used Stockholm

et al. γ

–

Gammaproteobacteria; f

(

97 ) merA4r merA4f merA3r merA3f merA2r merA2f merA1r merA1f merAr merAf merA1r merA1f A5 A7s A5 A1s A5 A1s Primer ;

Name 2 - - - Schaefer n.R n.R n.R - - -

n.F n.F n.F

to obtain 5’ 5’ 5’ 5’ 5’ 5’ 5’ 5’ 5’ 5’ 5’ 5’ 5’ 5’ 5’ 5’ 5’ 5’

------

CGCYGCRAGCTTYAAYCYYTCRRCCATYGT ATCCGCAAGTNGCVACBGTNGG GGTGGAAGTCCAGTAGGGTGA CGCYGCRAGCTTYAAYCYYTCRRCCATYGT ATCCGCAAGTNGCVACBGTNGG TCCGCAAGTNGCVACBGTNGG TCCGCAAGTNGCVACBGTNGG et al. CTTCCGGTTGCWACSAGWATYTCATCTACCTC GGMGGAACWTGYGTTAAYGTWGGYTGYGTCCC CGAGCYTKARSSCYTCGGMCAKSGTCAGGTAGG CGTSAACGTSGGSTGCGTGCCSTCCAAG CGCYGCRAGCTTYAAYCYYTCRRCCATYGT CCATCGGCGGCACYTGCGTYAA GTGCCGTCCAAGATCATG ACCATCGTCAGRTARGGRAAVA CGATCCGCAAGTGGCIACBGT ACCATCGTCAGRTARGGRAAVA ACCATCGTCAGRTARGGRAAVA

sequences included in

( 129

–

) Firmicutes. ; 3

Oregaard

Sequence

et al.

- 3’

the

( 105 - - - - 5’ 3’ 3’ 3’ global meta ------

3’ 3’

3’ 3’ 3’ 3’

) ;

4 Pitkanen - 3’

- - - - 3’ 3’ analysis. 3’

et al.

- 3’ -

- 3’

3’

( 110 Amplicon

1246 bp 1249 bp 662 bp 310 bp 360 bp 310 bp 288 bp 280 bp 280 bp Size ) ; 5

Wang

et al Phyla/class High G+C Low G+C Gram ( Archaea Target β β

β β β

/ / ( γ γ / / / γ γ γ 156 /f /f

- )

)

a .

Chadhain Shaeffer Source Wang Wang Virta Ni 4

5 5

2 1

Results

16S rRNA and merA Gene Clone Libraries

Geothermal springs selected for this study represent two sites of contrasting geochemistry, as well as varying temperatures and pH that are commonly encountered in

YNP hot springs (Table 3.1). Bijah Spring is characterized as a high temperature, low sulfide spring, which becomes slightly more alkaline downstream, whereas Succession

Spring is characterized as high sulfide, low pH, and high temperature at the source

(76ºC), with quick cooling as the spring water flow downstream through a shallow stream

(69). Water in both sites contains elevated Hg concentrations (Table 3.1), making them suitable locations for the study of the impact of Hg on microbial communities.

Microbial mats were collected from two locations within each site representing increasing distance from the source and DNA extracted from the mats was successfully amplified by PCR with bacterial universal 16S rRNA gene primers (27f/519r). Extracted

DNA fractions from each sample were used in PCR with the four merA-specific primer sets that had been designed to target the known diversity of merA that was known in

May, 2006 (Table 3.2); a low degree of sequence homology prevented construction of a single set of degenerative primers (156). Products of PCR were obtained with at least 2 bacterial merA primer sets from each of the samples (Table 3.4).

Table 3.4: PCR Amplification of merA from YNP mat DNA extracts Primer Seta Site 1 2 3 4 Bijah 1 + + - - Bijah 2 + - + - Succession 1 - + +* - Succession 2 + + +* - a+: PCR product detected, -: PCR product was not detected, * indicates low reaction yield as indicated by weak bands after gel electrophoresis

Failure to obtain products with primer set 4, which targeted archaeal merA, suggested an absence or limited distribution of archaeal merA in the sampled sites.

Primer set 3, designed to target the widest range of merA diversity, amplified DNA from

Bijah 2, with weak amplification observed with DNA from Succession 1 and 2. Primer set 1 amplified DNA from all but Succession 1, whereas primer set 2 amplified DNA from all but Bijah 2. Taken together, the data indicate a wide distribution of bacterial merA in the springs included in this study.

Phylogenetic Analysis of 16S rRNA and merA Clone Libraries

All merA clones obtained through PCR products generated with primer 2 using template DNA from Bijah 1 (9 clones), Succession 1 (14 clones), and Succession 2 (14 clones) or primer 3 using DNA from Bijah 2 (13 clones), when translated, were homologous to known MerA amino acid sequences. Although, primers 2 and 3 were designed to target merA with low (primer 2) or high (primer 3) GC mole % (156), there was no significant difference in GC mole % of clones amplified in this study (Table 3.5).

Table 3.5: merA clone library composition Site Primer Set Used Number of Clones GC mole % a Bijah 1 2 9 65.09 + 0.23 Bijah 2 3 13 64.95 + 0.52 Succession 1 2 14 65.15 + 0.31 Succession 2 2 14 65.28 + 0.31 aGC mole % of clones screened represents the average value for all clones sequenced + 1 standard deviation. GC mole % for all groups was statistically similar (P < 0.05).

Rarefaction analysis using nucleotide sequences was used to determine the i) relative species richness of each community, and ii) quality of the sampling effort for each site tested for both 16S rRNA gene and merA clone libraries. Within Bijah Spring, the 16S rRNA gene clone library from the downstream site exhibited a greater diversity at the 97% sequence similarity OTU cutoff level than the upstream site (Figure 3.2).

Similarly, the downstream Succession Spring site exhibited greater 16S rRNA gene based species diversity than the site closest to the source. This was not surprising, given the less extreme conditions of the downstream, relative to the upstream, locations of the springs (Table 3.1). Succession 1 was the only sample tested to reach an asymptote in the rarefaction analysis, suggesting the clone library captured the diversity of the bacterial community. Similar trends can be noted for the merA-specific community, where Bijah 2 had greater species richness than Bijah 1. The rarefaction curves for Succession Spring merA sequences appear nearly identical and both seem less diverse than the collection of sequences from Bijah Spring sites (Figure 3.2).

15 16S rRNA Bij1 Bij2 SS1 10 SS2

5 Number of OTU's (97% Cutoff)

0 0 20 40 60 Number of Sequences 15 merA

5 Number of OTU's (99% Cutoff)

0 0 5 10 15 Number of Sequences

Figure 3.2: Rarefaction analysis of 16S rRNA gene and merA clone libraries. 1 OTU was defined at the 97%, and 99% nucleotide sequence similarity cutoff for 16S rRNA gene and merA sequences, respectively. Succession Spring 1 and 2 are indistinguishable in the merA panel.

Phylogenetic reconstructions using sequences of cloned 16S rRNA gene PCR products show that distinct communities inhabiting Bijah and Succession springs (Figure

3.3). Succession 1 was dominated (n = 81/89) by clones with >97% homology to members of the Aquificae. Interestingly, clones obtained from Succession 2 template

DNA showed >97% homology to chloroplast DNA from Cyanidioschyzon merolae (n=

41/47), a unicellular red algae known to inhabit highly acidic environments (29).

Amplified bacterial sequences from Succession 2 shared homology with Acidisphaera spp. (n=2/47), Thiomonas spp. (n=1/47), and uncultured bacterial clones (n=3/47).

Clones from Bijah 1 were dominated by photosynthetic bacteria; namely Synecococcus spp. (n=23/68), Chloroflexus spp. (n=11/68), and Roseiflexus spp. (n=8/68). Green non- sulfur bacteria (n=7/68) as well as clones sharing <97% homology with known sequences in the NCBI database (n=15/68) were also well represented. Bijah 2 was noted by the same dominance of photosynthetic bacteria Synecococcus spp., Chloroflexus spp., and

Cyanobacterium spp. (n=17/37). Taxa normally noted for their cosmopolitan distribution

(12), the Firmicutes (n=5/37) and Bacteriodetes (n=7/37) were also amplified in Bijah 2.

Perhaps as artifacts of animal dung, E. coli (n=1/37) and Shigella sp. (n=1/37) sequences were obtained from Bijah 2.

0.2 Archaea Alphaproteobacteria Beta/Gammaproteobacteria Deltaproteobacteria Chloroflexi Firmicutes Actinobacteria Deinococcus/Thermus Bacteroidetes Aquificae Cyanobacteria Figure 3.3: Bayesian phylogram of 16S rRNA gene sequences representing the microbial communities of Bijah Spring 1 (red), and 2 (orange) and Succession Spring 1 (Yellow), and 2 (Green) are shown for each unique OTU (97% similarity). When more than one clone in represented by a sequence, the number of represented clones is provided in parenthesis. Nodes labeled with black, gray, green, yellow circles and no circles indicating > 90, > 80, > 70 and > 60 and < 70 % posterior probability values, respectively. The side bar on the right shows taxa affiliation of corresponding clades in the phylogram according to the key below. Archaeons Pyrobaculum caldifontis, and Sulfolobus solfataricus sequences were used as an outgroup and the bar indicates 2 substations per 10 positions.

Translated merA clones from the 4 libraries were dominated by sequences most similar to MerA of the Gammaproteobacteria; all 9 sequences from Bijah 1 and 5 of 13 sequences from Bijah 2 belonged to this group. Similarly, 12 of 14 and 11 of 14 clones from Succession 1 and 2, respectively, were affiliated with the Gammaproteobacteria.

Bayesian phylogenetic reconstruction reveals MerA sequences from Bijah 1, Succession

1, and Succession 2 all cluster in the beta- and gammaproteobacterial clade (Figure 3.4).

These sequences were most closely affiliated with the MerA of Cupriavidus metallidurans CH34, a mesophilic, facultative chemolithotrophic betaproteobacterium known to thrive in environments with toxic concentrations of metals (90). More variable

MerA sequences were noted in Bijah 2 as compared to Bijah 1, as indicated by sequences most similar to those of the Bacteriodetes (n=7/13), and Betaproteobacteria (n=1/13). All

7 Bacteriodetes-like MerA clustered with Rhodothermus marinus DSM 5452, a deep branching, thermophilic, bacterium isolated from an Icelandic hot spring (3).

Betaproteobacterial-like MerA comprised the remaining sequences from Succession 1

(n=2/14) and Succession 2 (n=3/14).

Figure 3.4: Bayesian phylogram of MerA clones. Representative clones for sites Bijah Spring 1 (red) and 2 (orange) and Succession Spring 1 (Yellow) and 2 (Green) representing each unique OTU (99% sequence similarity). The total number of clones represented by each sequence is indicated in parenthesis. Bar indicates 3 substitutions per 10 positions. Nodes labeled with black, gray, green, yellow circles and no circles indicating > 90, > 80, > 70 and > 60 and < 70 % posterior probability values, respectively. The side bar on the right shows taxa affiliation of corresponding clades in the phylogram according to the key below. Dihydrolipoamide dehydrogenase from Magnetospirillum magneticum, Pseudomonas fluorescens, and Thermus thermophilus were used as outgroups.

Community Ecological Analysis

Phylogenetic reconstruction using nucleotide sequences from a total of 201 16S rRNA and 51 merA gene amplicons recovered from both sites tested were used to generate well-supported phylogenies for use in community phylogenetic analysis.

Results generated using both merA nucleotide and MerA amino acid sequences were congruent and therefore gene sequences were used in community analysis in order to capture a similar resolution of phylogenetic diversity for comparison with 16S rRNA gene sequences. Sequences similar to chloroplast DNA of the Eukaryote C. merolae were omitted from all subsequent ecological analysis. Rao’s phylogenetic diversity index

(Dp), an abundance weighted, β-diversity index comparing the weighted branch lengths associated with a particular assemblage relative to the total sequence pool (Rao, 1982), was computed for the four 16S rRNA and merA gene assemblages, normalized to taxa

abundance to control for unequal sampling effort. A higher Dp metric indicates a greater phylogenetic diversity relative to that of the total sequence pool. In both springs tested,

Dp increased in the downstream sample site relative to the site closest to the source for both the 16S rRNA gene and merA sequence assemblages (Table 3.6).

Table 3.6: Phylogenetic diversity metrics for gene assemblages recovered from mat microbial communities in Bijah (B) and Succession (S) springs, YNP. Phylogenetic Diversity Metricsa Gene Sampling Diversity Clustering assemblage Site b b (Dp) NRI P-value NTI P-value B1 0.334 15.957 < 0.01 1.674 0.03 B2 0.408 8.911 < 0.01 1.109 0.13 16S rRNA S1 0.291 29.139 < 0.01 1.157 0.13 S2 0.472 1.297 0.10 0.703 0.24 B1 0.213 4.071 < 0.01 2.505 < 0.01 B2 0.436 1.897 0.03 2.924 < 0.01 merA S1 0.495 1.305 0.09 0.687 0.25 S2 0.504 0.952 0.17 0.719 0.24 a Dp, Rao’s Quadratic Entropy; NRI, net relatedness index; NTI, nearest taxon index. bP-values were determined by comparison with randomly generated phylogenies.

The NRI is a measure of phylogenetic clustering of sequences across the phylogram, whereas the NTI metric is a measure of the phylogenetic distance to the nearest taxon for each taxon within a community, which reflects the extent of terminal

(branch tip) clustering (160). Both metrics were normalized to taxa abundance to control for unequal sampling effort. Positive NRI and NTI metrics indicate that co-occurring species in each community are more phylogenetically related than expected by chance

(phylogenetic clustering), whereas negative NRI and NTI scores indicate a lesser phylogenetic relationship (phylogenetic over-dispersion). That is, positive NRI and NTI indicate lower than expected phylogenetic diversity in the assemblage given the species richness of that assemblage. All four sites within both the 16S rRNA gene and merA specific community exhibited positive NRI and NTI values with 8 of the 16 total assemblages being very well supported (P≤0.03) as compared the null model. In all

cases, assemblages with a relatively higher Dp also exhibited a lower NRI or NTI, again suggesting relative phylogenetic over-dispersion in the downstream sites, and greater clustering in communities near the source within each hot spring tested. Overall, all NRI

79 and NTI metrics were positive, indicating overall phylogenetic clustering as compared to the total sequence pool (NRI), and within each assemblage (NTI), highlighting the uniqueness of each assemblage relative to the total sampled community.

In order to further understand the degree of phylogenetic structure in microbial mat 16S rRNA and merA gene communities, agglomerative hierarchical clustering and

UniFrac analysis were performed. Agglomerative hierarchical cluster analysis shows unique assemblage clustering for the 16S rRNA gene and merA specific communities

(Figure 3.5; A, B). Approximately unbiased P-values (AU values) are shown at each node, indicating the significance of each cluster as determined by multi-scale bootstrap resampling. The 16S rRNA gene dendrogram shows well supported assemblage clustering based on hot spring, with S1 and S2 forming a well supported clade, and B1 and B2 forming a distinct clade, although with less statistical support. B1 along with S1 and S2 form a well-supported cluster in the merA dendrogram, with B2 branching distinctly to the larger clade.

These data are corroborated by nonmetric multidimensional scaling plot ordination of the pairwise UniFrac distances, which illustrates the influence of sampling location on the retrieved 16S rRNA- and merA-gene assamblages (Figure 5; C, D). B1 and B2 cluster together with a weaker relationship between S1 and S2 in the 16S rRNA plot. Again, different patterns are observed in the merA plot, showing the assemblages from B1 and S1 in a tight cluster. A weaker relationship is observed between S2 and B2.

16S rRNA merA

3.0 2.5

2.5 Root Root 2.0

2.0

1.5

96 1.0 Height

Height 1.0

0.5

0.5 98 100

92 100 A 79 B 0.0 0.0 B1 B1 B2 S1 S2 S1 S2

Stress: > 0.01! 0.10 Stress: > 0.01! B1 S2 C D S1 0.05 B2 0.00 NMDS2 -0.05

B1 S1

B2 -0.10 S2 -0.15

-0.20 -0.15 -0.10 -0.05 0.00 0.05 NMDS1 Figure 3.5: Phylogenetic distance based cluster analysis of 16S rRNA (A, C) and merA (B, D) gene sequences using hierarchical cluster analysis (A, B) and UniFrac (C, D) for Bijah (B1, B2) and Succession (S1, S2) springs. Hierarchical clustering was based on Ward’s agglomerative correlation used within pvclust in R. AU values are shown for each node. Non-metric multidimensional scaling (NMDS) plots (C, D) derived from pairwise abundance weighted UniFrac distance with symbols color coded by spring.

merA Global Meta-Analysis

741 total nucleotide sequences were compiled for most available merA sequences within the PopSet function in the NCBI database (http://www.ncbi.nlm.nih.gov/popset/) to determine how the observed trends in the two sites included in this study relate to merA diversity on a meta-scale (Table 3.3, 3.7). Again, all sites exhibited positive NRI and NTI metrics, indicating overall phylogenetic clustering as compared to the total

81 sequence pool (NRI), and within each assemblage (NTI). All 21 Bacterial assemblages yielded well-supported (P<0.05) NRI results, while 16 of 21 NTI results were statistically well supported (P<0.05). This observation again highlights the uniqueness of each assemblage relative to the total sampled community. The two archaeal assemblages were

not included in Dp, NTI and NRI diversity analysis.

Rate-smoothed merA phylograms were also used to create a distance matrix using the UniFrac metric, visualized by NMDS (Figure 3.6). Results show that the assemblages cluster in three main groups. One group contains two assemblages of archaeal sequences from YNP (156). Another includes sequences from Frying Pan West

Pool in YNP along with samples from pristine and polluted fish farms in the Turku and

Stockholm Archipelago in Finland and Sweden, respectively (110). The third and largest group encompasses assemblages from Bijah and Succession Springs from two studies in

YNP (49, 156), Berry’s Creek, Meadowlands, NJ (97), East Fork Poplar Creek, Oak

Ridge, TN (105), Cornwallis Island, Nunavut, Canada (112), and Sunday Lake,

Adirondack Park, NY (Yu, personal communication). Within this larger group, assemblages form tight clusters based on sampling location.

Table 3.7: Phylogenetic diversity metrics for merA gene assemblages recovered from microbial communities from sites listed in table 3.3. Number Phylogenetic Diversity Metricsa Sampling Site of Diversity Clustering Clones (Dp) NRI P-valueb NTI P-valueb Arctic 17 0.44 3.01 < 0.01 1.48 0.06 BC_AnOx 71 0.53 4.19 < 0.01 1.34 0.08 BC_Ox 19 0.45 2.73 < 0.01 1.27 0.10 F1_Fm 69 0.36 2.55 < 0.01 1.54 0.01 F1_Pr 85 0.03 3.57 < 0.01 1.01 0.01 F2_Fm 66 0.02 4.72 < 0.01 1.43 0.02 OR_B 63 0.07 4.76 < 0.01 1.50 0.01 OR_B_Act 19 0.34 5.93 < 0.01 1.98 < 0.01 OR_C 49 0.01 3.10 < 0.01 0.42 0.39 OR_D 51 0.48 3.84 < 0.01 1.48 0.01 S_Fm 65 0.40 2.72 < 0.01 1.61 0.02 S_Pr 74 0.03 3.75 < 0.01 1.40 < 0.01 SunLa_Soil 24 0.01 5.67 < 0.01 2.15 < 0.01 SunLa_SURN 17 0.01 4.14 < 0.01 1.83 < 0.01 YNP_Bij1 3 0.01 3.44 < 0.01 2.33 < 0.01 YNP_FPWP_3 3 0.20 2.20 0.04 2.20 0.04 YNP_FPWP_2 6 0.36 1.83 0.05 1.48 0.07 YNP_ZF_Bij1 5 0.01 2.23 0.02 2.23 0.02 YNP_ZF_Bij2 10 0.12 3.43 < 0.01 2.26 < 0.01 YNP_ZF_SS1 14 0.01 4.83 < 0.01 2.33 < 0.01 YNP_ZF_SS2 11 0.02 3.81 < 0.01 2.22 < 0.01 aAbbreviations: Arctic - Cornwallis Island, Nunavut, CA (112); BC - Berry’s Creek, Meadowlands, NJ (AnOX – Anaerobic enrichment, Ox - Aerobic Enrichment) (97); F1, F2 – fish farm, Turku Archipelago, Sweden (110); OR - lower East Fork Poplar Creek floodplain, Oak Ridge, TN (105); S – Fish Farm, Stockholm Archipelago, Sweden (110); SunLake – Sunday Lake, Adirondak Park, NY (Soil – upland forest soil, SURN- sediment sample from lake bank) (164); YNP – Yellowstone National Park (Bij – Bijah Spring; FPB – Frying Pan Spring Bowl; FPWP – Frying Pan West Pool (merA# primer set used included after underscore) ; SS – Succession Spring) (49, 156). b Dp Rao’s Quadratic Entropy; NRI, net relatedness index; NTI, nearest taxon index. cP-values were determined by comparison with randomly generated phylogenies.

83 0.6 0.4 0.2

3 3 3 2 NMDS2 2 2

0.0 2

3 -0.2 -0.4

-0.4 -0.2 0.0 0.2 0.4 0.6

NMDS1 Fish Farms Meadowlands Succession Spring Sunday Lake Nunavut Frying Pan West Pool Oak Ridge Bijah Spring Archaea

Figure 3.6: Non-metric multidimensional scaling (NMDS) plot of merA gene assemblages listed in table 3.7. Filled circles and squares represent bacterial and archaeal merA assemblages, respectively. Color Code: black, Meadowlands; blue, Arctic; green, Sunday Lake; red, Yellowstone National Park (boxed circles, Frying Pan West Pool; light circles, Succession Spring; dark circles, Bijah Spring); white, Oak Ridge; yellow, fish farms in Finland/Sweden. Numbers inside YNP assemblage circles indicate merA primer set number used (156). Dotted lines have been added to illustrate clustering patterns.

Discussion

The diversity and distribution of protein-coding genes provide insight to the factors, both biotic and abiotic, that have controlled the evolution of these genes in a given environment (21). Recent studies begin to shed light on how microbial community structure responds to physicochemical controls (42, 63, 81), however our knowledge of factors affecting controls on functional guilds is even less developed (21, 23, 59, 156). In this study, merA served as a proxy for examining the factors that have influenced the ecology and evolution of microbes that detoxify Hg by transforming Hg(II) to Hg(0) in

Hg-containing hot springs in YNP. These microbes may partake in Hg cycling in the geothermal spring environment.

Previous studies documented the occurrence of various forms of Hg in geothermal environments (5, 38, 56, 94, 95, 100), as well as its presence and accumulation in geothermal food webs (22, 61, 74). Recent work investigates the effect of various abiotic factors which may affect merA distribution in YNP hot springs (156), but my study represents the first to investigate ecological controls on phylogenetic structuring of the merA-specific community. Wang et al. (156) previously suggested a possible effect of pH, dissolved organic carbon, dissolved total Hg, and sulfide on constraining distribution of merA. However, models failed to show statistically well-supported trends for the effect of these environmental variables on the distribution of merA lineages (156). The two hot springs chosen for this study have contrasting environmental factors such as temperature (45.5 - 65° C), pH (2.50-8.10), total Hg in filtered spring water (10.2-64.6 ng/L), and sulfide (9.9 – 324 μg/L) (Table 3.1); they were chosen to emphasize how

85 varying ecological niches may affect the diversity and distribution of merA in geothermal environments (156).

Bacterial universal primers were used to generate 16S rRNA gene clone libraries whereas one of two merA-specific PCR primer sets (Table 3.2) were used to generate community composition data. These primer sets were designed to amplify all merA genes as revealed by the current gene phylogeny at the time of primer design in May,

2006 (156). It was not possible to generate clone libraries from the four samples using the same merA primer set (Table 3.4). I, therefore, chose to generate clone libraries from amplicons obtained with primer set 2, designed to amplify low GC% merA from the

Proteobacteria, Firmicutes, and Actinobacteria for B1, S2, and S3, and with primer 3, intended to amplify high GC mole % merA from a wide range of bacterial phyla (Table

3.2) for B2. There is an extensive overlap in the merA loci which are amplified by these two primer sets (156). Furthermore, despite the target GC mole % for each primer set, the clone libraries of merA amplicons across all 4 sites had statistically indistinguishable average GC mole % (Table 3.4).

Aligned nucleic acid sequences were edited whereby regions of the alignment used for phylogenetic reconstruction only included areas of overlapping homology, resulting in 1220 bp region of the ~1240 bp amplicon being used for the analysis.

However, observations from the MerA phylogram (Figure 3.4) show that clones derived from merA primers 2 and 3 clearly cluster in different regions of the tree. This, together with the fact that PCR products were obtained with only primer set 2 or 3 from the two

Bijah sites (Table 3.4) clearly suggest that the two primers target different merA loci in environmental extracts. The question then arises whether the patterns observed in both

86 the hierarchical and UniFrac cluster analysis included in this chapter could be due to true phylogenetic differences in merA assemblages, or may be an artifact introduced by the different PCR primer sets.

Six different primer sets were used to generate the bacterial assemblages included in the global merA analysis (Table 3.3). Alignments used for meta-analysis were similarly edited to the local YNP alignments, although using a 240 bp region of homology within the 280-1249 bp amplicon sizes of environmental merA sequences. It is promising that merA meta-analysis reveals tight clustering based on sampling location

(Figure 3.6) rather than on the primer set used to retrieve the sequenced PCR products.

For example, products of primer set 2 were obtained from 3 different YNP springs, two of which cluster in the large cluster in the NMDS plot while the third, originated in

FPWP clustered with the loci from fish farms from Northern Europe. Conversely, the

FPWP cluster consisted of primer sets 2 and 3 products (Fig. 3.6). These observations may suggest that the physical and chemical parameters of the sampled location affects ecological structure more than the bias introduced by using different PCR primer sets.

Next-generation sequencing technology, facilitating the retrieval of environmental sequences without the use of PCR and therefore avoiding this issue may best serve future studies on merA diversity and distribution.

merA amplicons were detected in each of the four sites included in this study, indicating a wide distribution of mer function among a variety of geothermal ecological niches (Table 3.3). Conservation of merA presence across wide gradients suggest Hg may be toxic to organisms inhabiting these environments, as well as highlights the importance of mer as a Hg detoxification mechanism in geothermal environments. It was

87 expected that archaeal merA genes (detected by primer set 4) would be detected in

Succession Spring, as archaeal merA has been shown to be constrained to acidic, high sulfide springs (156). Furthermore, sequenced archaeal genomes containing merA homologs are largely from acidophilic strains (8). In my study, however, archaeal merA genes were not observed in any sample, as indicated by lack of DNA amplification by archaeal-specific PCR primer sets, demonstrating either the lack of amplifiable archaeal

DNA by the included primer sets or dominance of bacterial merA even in low pH high sulfide sites.

Amplification products derived from individual PCR primer sets designed to amplify the merA homologs in the genome of the Aquificae Hydrogenivirga sp. 128-5-

R1-1 and Hydrogenobaculum sp. Y04AAS1 (119) were observed in B1 and B2, whereas amplification products were only observed with primers designed to target the merA of strain AAS1 in samples from Succession Spring (data not shown). These data suggest that primer sets used to construct the merA clone libraries may have lacked the specificity to amplify putative merA from Aquificae and therefore failed to show the presence of merA from early bacterial lineages in YNP hot spring microbial mats.

Previous work has proposed that mer systems of microbes that inhabit hot springs may be the best model for studying the evolution of the broadly distributed Hg-resistance system (8, 156), as recent phylogenetic reconstructions suggest merA originated after oxygenation of the biosphere in a thermophilic bacterium most related to Aquificae (8).

This is supported by the recovery of 7 merA sequences from a high temperature (59° C)

Bijah 2 site representing deep branching thermophilic lineages, namely in the

Cytophaga–Flavobacteria–Bacteriodetes (CFB) phylum (Rhodothermus marinus DSM

4252; 87% identity) (Figure 3.4). The most updated MerA phylogenetic reconstructions show the CFB cluster in a basal position to the bacterial clade (Figure 2.1). It has been proposed that this group may serve as an evolutionary link of mer function between early-evolving bacterial lineages from high temperature environments to later evolving lineages inhabiting more temperate habitats (156).

The distribution of merA lineages may be correlated with the distribution of microbial taxa in the Hg-containing springs of YNP. Similar observations were made with regard to the distribution of other functional genes. For example, the distribution of aroA-like genes, encoding arsenite oxidase, in YNP, showed a correlation with

Aquificae-like 16S rRNA gene sequences in environmental clone libraries (57).

Furthermore, the amplification of archaeal merA from acid-sulfide hot springs (156) is consistent with previous observations of archaeal dominance in low pH, high sulfide thermal features in YNP (70). Recent phylogenetic reconstructions of all merA genes from complete microbial genomes shows a topology that mirrors that of the 16S rRNA gene phylogeny (8). In my study, similar trends were observed in 16S rRNA and merA gene sequences retrieved from B2. 8/37 16S rRNA clones clustered with representatives from the CFB, similarly 7 merA sequences clustered with CFB. Likewise, the finding of amplification of merA from Hydrogenobaculum sp. specific primers (not shown) along with a dominance of 16S rRNA sequences clustering in a clade with Aquificae in S1 further support the hypothesis that merA distribution is related to the distribution of microbial taxa in a given environment.

It has been previously demonstrated that despite the supposed cosmopolitan distributions of microorganisms, geographically isolated microbial communities exist in

89 geothermal hot springs. Specifically, Takacs-Vesbach et al. (148) suggest that the ancient volcanic caldera in YNP delineate microbial assemblages. Recent work showing a deep level of phylogenetic clustering (NRI > NTI) of the nitrogenase gene (nifH), supports the hypothesis that ancient geographic regions (calderas), the most recent of which formed nearly 600,000 years ago, impose geographic constraints which may structure communities (59). Similar deep phylogenetic clustering of both 16S rRNA and merA gene sequence assemblages were observed in this study (Table 3.6). 3 of 4 16S rRNA and merA sequence assemblages had NTI values lower that NRI, suggesting phylogenetically deep rather than shallow level of clustering. The only exception (B2), was marked by the presence of more recent branching Cyanobacteria in the 16S rRNA gene phylogeny (Figure 3.3). This finding is also consistent with a growing body of literature suggesting a thermophilic origin, occurring after the oxygenation of Earth’s biosphere, of the prokaryotic Hg resistance system (8, 49, 154, 157).

Each assemblage carries significantly different bacterial communities clustering with high support (5 of 6 nodes > 90% AU) in an agglomerative hierarchical cluster analysis (Figure 3.5, A, B). Both hierarchical clustering and pairwise UniFrac distance as visualized by NMDS (stress > 0.01) indicates different clustering patterns between all 4 sites included in this study when using the 16S rRNA gene and merA phylograms (Figure

3.5, C, D).

Taken together, these data may indicate the different selective forces which act on the species, as revealed by the 16S rRNA gene, and on merA evolution. It should be noted however, that the merA phylogram for the Bijah 2 assemblage was prepared using

PCR amplicons using a different primer set (Table 3.4). This was possibly due to

90 inherent community differences as indicated by the lack of amplification with a single primer set across all 4 sites included in this study (Table 3.3). However, the possibility cannot be ruled out that the basal position of the B2 merA assemblage by hierarchical cluster analysis is an artifact of the use of a different primer set than was used for the other 3 assemblages as previously discussed. Even disregarding the B2 assemblage, the merA assemblages from primer set 2 (B1, S1, and S2) yield different clustering patterns by UniFrac, adding support to the hypothesis that the physicochemical controls act differently to shape 16S rRNA and merA assemblages in YNP hot springs.

Meta-analysis of 21 bacterial merA assemblages incorporating 741 environmental merA gene sequences revealed overall positive NRI and NTI metrics, indicating that co- occurring species in each community are more phylogenetically related than expected by chance, or phylogenetic clustering. Six of the first 7 most diverse assemblages by Rao’s

Dp were sampled from more recently, and highly contaminated environments such as

Berry’s Creek, NJ and Oak Ridge, TN (97, 105). It should also be noted that 5 of the 10 least diverse merA assemblages originated from samples from YNP hot springs, and 6 of

7 samples from YNP yielded relatively low Dp values (<0.12). This observation may reveal a lesser diversity of merA assemblages in geothermal environments, where communities may have evolved in the presence of Hg for their entire life history (8, 28,

156). Analyses of 16S rRNA gene diversity revealed communities of greater diversity in soil (133, 152), oceans (143), and salt marshes (155) than hot springs (64, 113), highlighting the “extreme” nature of the hot spring environment. Likely, merA gene diversity mirrors 16S rRNA relative gene diversity in the sampled environments. The possibility cannot be ruled out that the specificity of merA PCR primer sets used did not

91 capture the full diversity of this locus in these environments. This is supported by the failure of primer sets 1-4 to amplify merA from pure cultures of strains

Hydrogenobaculum sp. AAS1 and Hydrogenivirga sp. R1-1 (not shown), possibly underestimating merA diversity of these environments.

Nineteen of the 21 bacterial assemblages had NTI values lower than NRI, supporting the findings from the Bijah and Succession study (see above) of a phylogenetically deep level of clustering rather than just at the terminals. Although this finding may be explained by the dramatic differences expected in the physical and chemical parameters of the environments from which merA sequences were obtained and included in this meta-analysis (Table 3.3). UniFrac analysis of 21 bacterial and 2 archaeal merA assemblages, comprising 782 sequences, reveal 3 assemblage clusters based on sampling location (Figure 3.6). Two assemblages of archaeal sequences from

YNP form a single cluster, and assemblages from Frying Pan West Pool, YNP, and 2 fish farms together with their associated pristine site controls in Finland and Sweden formed a second well-defined cluster. A third cluster was comprised of the remaining assemblages from YNP, Nunavut (High Arctic), the Adirondacks, Oak Ridge, and the Meadowlands.

Tight clustering of assemblages from fish farm and control sediments highlight the uniqueness of the environments sampled. Assemblages from Frying Pan West Pool

(FPWP), YNP form a cluster with fish farm sediment, apart from other YNP bacterial sequence assemblages. Phylogenetic reconstruction of clones generated from YNP samples in Wang et al. (156) reveal clustering of FPWP clones in the proteobacterial clade, unrelated to clones from other sampled hot springs. Although, proteobacterial-like merA sequences were well represented in all but Bijah Spring assemblages from Wang et

92 al. (156) which were dominated by Bacteriodetes-like merA. Percent similarity to known merA was lower in fish farm (57-100%) and FPWP (66-94%) assemblages than in other

YNP assemblages (93-97%). Other merA assemblages from the large cluster including

Adirondack, Arctic, Meadowlands, Oak Ridge, and YNP assemblages did not include clones that had a low similarity by BLAST search to known merA (105, 110, 112, 156).

Thus, the presence of merA sharing low homology to known merA could explain the distant clustering of FPWP and fish farm sediment assemblages. Taken together, these data may support the previous observations of merA communities structured by the physical and chemical parameters of the site from which they are located.

In summary, the phylogenetic diversity of 16S rRNA and merA gene assemblages appear to be constrained by physicochemical conditions characterizing the ecological niche of the sampled community. The data further suggest both 16S rRNA and merA- specific gene assemblages are dispersal limited across all sites, and furthermore, share a deep phylogenetic relationship indicating that these communities have evolved over geological timescales. Global meta-analysis revealed tight clustering of assemblages based on sampling site, supporting the hypothesis that merA diversity and distribution patterns are structured by the physical and chemical parameters of their environment.

These data bring to light an interesting question; is the observed clustering of merA assemblages based on ecological structure, or selection based on PCR primer set used?

Further analysis must be performed to address this issue. The need for greater degeneracy in PCR primer sets designed to amplify merA from deep branching microbial lineages, including the recently MerA from the recently characterized Aquificae phylum

(49) is apparent. This study also brings in to question the use of PCR based methods for

93 study of environmental merA communities. It is clear that the use of next-gen sequencing techniques in future studies regarding merA diversity and distribution, which avoid the need for PCR primers altogether, will be beneficial in reducing procedural bias.

Geothermal features of YNP provide a variety of Hg-rich environments, populated by early evolving microbial lineages. Gene sequences recovered from such an environment aid in our understanding of biogeographical controls on merA diversity and distribution, the evolution of MerA, and thus Hg resistance and detoxification.

Chapter 4 – Summary and Conclusions

Geothermal environments such as YNP hot springs are hypothesized to be the habitat in which microbial resistance to Hg evolved, for microbial life has possibly been exposed to toxic concentrations of Hg since the beginning of life on Earth (8, 122, 154).

Features such as shallow (146), and deep-sea vents (32), volcanoes (102), geysers (101), hot springs (22, 74), and fumaroles (38) are known geologic sources of Hg to the environment. Thermal features in YNP have been shown to be highly enriched in Hg, reaching concentrations of micrograms per liter total Hg (22, 74, 136), similar to concentrations found in environments with high input of anthropogenic Hg (96, 129).

Mercury is released in the aqueous phase (74, 156), and accumulates across trophic levels in YNP (22).

Culture independent techniques have been used to amplify Hg resistance genes from hot springs in YNP (156) and Coso Hot Springs (142), and Hg resistant organisms have been previously isolated from geothermal environments (28, 131, 154). The large scale sequencing of microbial genomes has resulted in an increased availability of merA sequences and allowed for a robust analysis of gene evolution, further supporting an origin and early evolution of Hg resistance among thermophilic microbes from geothermal environments (8).

The main objective of this thesis dissertation was to gain a better understanding of microbially-mediated Hg-resistance in geothermal environments. Specifically, I was interested in i) identifying organisms within the deep branching phylum Aquificae, dominant primary producers in many geothermal environments, which are capable of reducing toxic Hg(II) to Hg(0) through the activities of mercuric reductase (MR), and ii)

95 investigating the degree to which Hg-resistant communities are structured relative to physical and chemical gradients in YNP hot springs and on a global scale.

I obtained two cultures within the phylum Aquificae, whose genome sequencing revealed homologs to merA, the gene that codes for MerA, the subunit of MR, in the genomes of Hydrogenobaculum sp. Y04AAS1 (AAS1) and Hydrogenivirga sp. 128-5-

R1-1 (R1-1) (119). An updated phylogeny that included all MerA homologs in sequenced microbial genomes confirmed the basal position of the Aquificales sequences in the MerA phylogeny (Figure 2.1). The two MerA homologs of Hydrogenobaculum sp.

AAS1 and Hydrogenivirga sp. (R1-1) clustered together with a third Aquificae merA homolog from the recently sequenced genome of Hydrogenobacter thermophilus TK-6

(4) and a merA homolog of Deferribacter desulfuricans SSM1 in a basal position to all

Archaeal sequences with accompanying posterior probability values of 1 (Figure 2.1).

This cluster shared a common, likely bacterial, ancestor with a large cluster consisting of the remaining bacterial MerA sequences.

The amino acid sequences of both Aquificae loci contained signature motives known to be required for MerA activity, including the redox-active site, the vicinal CC pair at the carboxy-terminus, and two tyrosine residues; one approximately 60 residues downstream of the redox active site, and one approximately 20 residues upstream of the vicinal CC pair (Figure 2.2) (10).

In both Aquificae strains, ORF’s upstream of the putative merA code for proteins that bear homology to mer transport functions. Directly upstream of the putative MerA in both AAS1 and R1-1 are loci that share homology to MerP, including the signature metal binding motif sequence. Although homologous sequences are found in both

96 strains, the ~19 residue sec-type signal sequence that is present in proteobacterial MerP is not found in either Aquificae strains; the MerP homolog in strain R1-1, though, has an amino terminus that bear some similarity to documented signal leader peptides (Figure

2.2). This sequence is missing in the locus of strain AAS1, questioning the transport of its product to the periplasmic space and thus its functioning as a MerP. ORF’s encoding for MerT homologs are found upstream of the putative merP genes in both genomes.

TMpred (62) determined the positioning of three hydrophobic inner membrane spanning sequences. Further homology includes the vicinal CC pair in the first membrane spanning sequence, and a Cys pair located in the cytoplasmic loop between the second and the third membrane embedded sequences (123). No ORF in the proximity of these mer operons appears to contain signature sequences associated with mer function. Thus, it seems the mer operons in the two Aquificae strains consists of merT-merP-merA-like genes (Figure 2.3).

In order to further understand the putative Hg-resistance in the two Aquificae strains, I tested the ability of strain AAS1 and R1-1 to grow in elevated Hg concentrations relative to a third Aquificae strain that did not have a merA homolog in its genome, Persephonella marina str. H1 (H1) (119). Mercury is known to form complexes with media components, which control its bioavailability and therefore toxicity (39).

This study utilized defined growth media that included either S2O3 or H2 as a sole energy source. Because of the high affinity of Hg to sulfur (98) it was important to determine the exact chemical speciation of Hg(II) in the growth media. I, therefore, modeled the

speciation of Hg in both types of medium when S2O3 or H2 are included. Interestingly, it

-1 was found that, in the presence of 5-60 μmol L HgCl2, all Hg(II) speciated as negatively

97 charged Hg-thiosulfate complexes, with 84% of Hg[II] as the -2 dithiosulfate complex and the remaining 16% as the -4 trithiosulfate complex. In the absence of thiosulfate and

presence of H2, Hg speciated mostly as uncharged HgCl2 and HgCl(OH) complexes

(Table 2.2).

Hg resistance profiles (Fig 2.5) reveal the ability of both strains AAS1 and R1-1 to grow in elevated concentrations of Hg(II). Traditionally, it is thought that Hg(II) concentrations over 10 μM select for Hg-resistant organisms (28). Strain AAS1 was able

to grow using both S2O3 and H2 as the sole electron donor. It was found that the IC50 for

strain AAS1 was reduced from 13.5 μM to 2.8 μM Hg(II) when grown with S2O3 as compared to H2 as the electron donor. Resistance patters differed in strain AAS1 and R1-

1 when grown using S2O3 as the sole electron donor. When grown on S2O3, the IC50 was

13.5 μM Hg(II), and 7.6 μM Hg(II) for strains AAS1 and R1-1, respectively. Expectedly, the lowest tested Hg(II) concentration inhibited growth of strain H1. Taken together, these findings suggest that merA homologs in AAS1 and R1-1 may specify Hg-resistance by encoding for an active MR.

In order to better understand the Hg resistance of both Aquificae strains, I performed a series of tests to determine if Hg(II) was lost from the growth medium in

203 growing cultures of AAS1 and R1-1 (Figure 2.6). Radiolabelled HgCl2 was removed from growth media by both cultures prior to the commencement of growth in all tested conditions. Both strains AAS1 and R1-1 removed Hg(II) from the growth medium at a greater rates than uninoculated controls (P < 0.05), and there was no difference in rate of

Hg(II) removed between the two cultures (P > 0.05). My findings clearly indicate that growing cultures of R1-1 and AAS1 removed Hg(II) from their growth media. In the

presence of S2O3, an abiotic loss of Hg(II) was observed. At the commencement of growth in live cultures, the uninoculated controls of strain AAS1 and R1-1 had only 4.81

+ 1.30 and 5.45 + 1.36 and μmol Hg(II) remaining in the media, respectively, approximately half of the starting concentration. Unfortunately, the observation of an

abiotic loss of Hg(II) from the growth medium in the presence of S2O3, with no such loss when H2 is included, renders any comparison of Hg speciation affecting its toxicity impossible in this study.

203 To verify that the product of the aforementioned loss of HgCl2 in growing cultures of AAS1 and R1-1 was in fact Hg(0), I performed end-point mass balance

experiments using an acidified KMnO4 trapping solution to capture any Hg(0) that may have been produced by Hg(II) reduction and volatilized from the medium to the headspace (Table 2.4, Figure 2.7). Both strains AAS1 and R1-1 produced ~10 fold greater amount of Hg(0) than did heat-killed controls (P < 0.05), and there was no

difference in mass of Hg(II) reduced between both cultures grown on S2O3 (P > 0.05).

Mass balance calculations showed recoveries of 85 to 122% of Hg(II) that was added at the beginning of the experiments (Table 2.3). These results are consistent with the possibility of mer mediated Hg-resistance in AAS1 and R1-1.

Mercury reduction is mediated by the activity of the enzyme MR. In order to determine if Hg(II) reduction by strains AAS1 and R1-1 was mediated by MR, the apparent specific enzymatic activity rates in crude cell extracts of both strains were determined (Figure 2.8). Maximum apparent specific activity for AAS1 was 37.7 + 1.1 mU mg protein-1, whereas rates observed in extracts of strain R1-1 were ~6 fold lower, at

6.4 + 0.2 mU mg protein-1. Interestingly, the optimal temperature of MR activity

99 corresponded with each strain’s optimal growth temperature, 55 and 70° C for strain

AAS1 and R1-1, respectively. No difference was noted in activities of AAS1 extracts

when cultures were grown with S2O3 or H2 as the electron donor (P > 0.05). Taken together, these results strongly suggest mer-mediated Hg resistance in

Hydrogenobaculum sp. Y04AAS1 and Hydrogenivirga sp. 128-5-R1-1.

The genome sequences of strains AAS1 and R1-1 showed no ORF’s with homology to the known regulators of mer operon, merR and ArsR. To determine if expression of mer functions are induced by the presence of Hg(II), I tested if sub-toxic

Hg(II) concentrations induce merA transcription using reverse transcription-qPCR. I

observed that up to 3 hours after the addition of HgCl2 to exponentially growing cultures, levels of merA transcripts, normalized to levels of the housekeeping gene gyrA

transcripts, were similar to those of cultures grown without HgCl2 (Figure 2.10).

Previous work using a similar approach clearly showed induction of merA in T. thermophilus HB27 and in an E. coli culture carrying a chromosomal insertion of Tn501

(157). Thus, expression of merA in the two Aquificae strains may be constitutive, a conclusion supported by similar MR specific activities in crude cell extracts of cultures that were grown in presence or absence of 2 to 5 μM Hg(II) (Fig. 2.11). These results clearly show merA expression in strains AAS1 and R1-1 to be constitutively expressed.

Taken together, data presented in Chapter 2 clearly suggest a mer-mediated Hg resistance in Hydrogenobaculum sp. Y04AAS1 and Hydrogenivirga sp. 128-5-R1-1, representing the earliest bacterial lineage to have this system. This observation is the first documentation of mer-mediated Hg resistance in the Aquificae phylum, and the second, following the discovery of this system in Thermus thermophilus HB27 (157), to

100 demonstrate Hg resistance in deep branching bacterial lineages from geothermal environments.

Physicochemical and chemical conditions have previously been suggested to have an effect on merA diversity and distribution (156). To expend on these observations, phylogenetic analyses were initiated using nucleotide sequences to assess the degree to which environmental gradients control merA gene diversity and distribution. Hot springs were chosen for this study to emphasize environmental factors that have been suggested to affect merA diversity and distribution, namely pH and sulfide (156).

Genomic DNA was extracted from microbial mat biomass from two geochemically diverse YNP hot springs. Bijah spring (44° 45.670’ N, 110° 43.857’ W) is located in the vicinity of Nymph Lake within the Gibbon River watershed in YNP.

Succession Spring (44° 43’75.7” N, 110° 42’72.7” W) is located in the Hundred Springs

Plain areas of Norris Geyser Basin, YNP (Figure 3.1). Bijah Spring is characterized as a neutral/basic silicate-sinter-rich hot spring, whereas Succession Spring is an acid-sulfate- chloride spring (156). These springs have several contrasting physical chemical characteristics (Table 3.1).

Clone libraries were generated with PCR amplicons from primer sets created to amplify bacterial 16S rRNA and merA genes. Nucleotide sequences obtained were first used to organize sequences by operational taxonomic unit (OTU). 1 OTU was defined at the 97% sequence similarity for 16S rRNA gene clone libraries and at the 99% sequence similarity level for merA clone libraries due to the high degree of similarity in sequenced merA clones. A randomly selected representative sequence for each OTU was included in Bayesian phylogenetic reconstructions. Nucleotide sequences were used as input for

101

16S rRNA gene reconstructions (Figure 3.2) whereas deduced amino acid sequences were used as input for MerA reconstructions (Figure 3.3).

Phylogenetic reconstructions using sequences of cloned 16S rRNA gene PCR products show distinct communities inhabiting Bijah and Succession springs. The upstream Succession site was dominated by clones that shared >97% homology (n =

81/89), and clustered with members of the Aquificae. Amplified bacterial sequences from the downstream Succession site clustering with the alpha- and delta-Proteobacteria, and shared homology with Acidisphaera spp. (n=2/47), Thiomonas spp. (n=1/47), a cluster was observed within the Actinobacteria and Deinococcus/Thermus group, sharing homology with uncultured bacterial clones (n=3/47). Clones from Bijah 1 were dominated by photosynthetic bacteria; namely Synecococcus spp. (n=23/68),

Chloroflexus spp. (n=11/68), and Roseiflexus spp. (n=8/68). Green non-sulfur bacteria

(n=7/68) as well as clones sharing <97% homology with known sequences in the NCBI database (n=15/68) were also well represented. Phylogenetic reconstructions mirror this observation, with clones from Bijah 1 clustering mainly within the Chloroflexi and

Cyanobacteria. Libraries from Bijah 2 were noted by the same dominance of clones clustering and sharing homology with photosynthetic bacteria, namely Synecococcus sp.,

Chloroflexus sp., and Cyanobacterium sp. (n=17/37), additionally, clones clustering and sharing homology with the Firmicutes (n=5/37) and Bacteriodetes (n=7/37) were also represented in Bijah 2.

Bacterial universal primers were used to generate 16S rRNA gene clone libraries whereas one of two merA-specific PCR primer sets were used to generate merA community composition data, and 6 different primer sets were used to generate bacterial

102 assemblages included in the global merA analysis (Table 3.3). Aligned nucleic acid sequences were edited whereby regions of the alignment used for phylogenetic reconstruction only included areas of overlapping homology, resulting in 1220 bp region of the ~1240 bp amplicons being used for the local YNP analysis, and a 240 bp region of the 280-1249 bp amplicon sizes of environmental merA sequences obtained from the

NCBI database for the global meta-analysis. Consequently, patterns observed in phylogenetic analysis included in this chapter may be due phylogenetic differences in merA assemblages, or may be an artifact of selection introduced by different PCR primer sets.

When clusters are delineated based on primer set, primer sets merA2 and merA3 form very loose clusters that encompassed all Bacterial assemblages (data not shown), which suggest the physical and chemical parameters of the sampled location affects ecological structure more than bias introduced by using primer sets merA2 and merA3.

All other primer sets used for meta-analysis form tight clusters of assemblages, thus it cannot be determined if the observed grouping is based on environmental differences or

PCR primer bias.

Bayesian phylogenetic reconstruction of deduced MerA sequences reveal clustering within the beta- and gammaproteobacterial clade for all clones recovered from

Bijah 1, Succession 1, and Succession 2 (Figure 3.3). These sequences were most closely affiliated with MerA of Cupriavidus metallidurans CH34, a mesophilic, facultative chemolithotrophic betaproteobacterium isolated from an environment with toxic concentrations of metals (90). More variable MerA sequences were noted in Bijah 2 as

103 compared to Bijah 1, as indicated by sequences most similar to those of the

Betaproteobacteria (n=1/13), and the deep branching Bacteriodetes phylum (n=7/13).

Taken together, Bayesian phylogenetic reconstructions indicate little overlap in the 16S rRNA gene and MerA communities. This is not unexpected, given the small sample size and the depth of sampling of each population within the community.

Detection of members of Aquificae, Deinococcus/Thermus, and CFB in the 16S rRNA gene phylogeny, and of representatives clustering with the CFB in MerA phylograms, reveal the presence of deep branching bacterial lineages. This is expected given the proposed similarity of geothermal environments to habitats present on early Earth (8, 88,

127).

With the goal of investigating the degree to which 16S rRNA and merA gene communities are structured by physical and chemical gradients in YNP hot springs, community ecological analyses were performed using phylogenetic reconstructions generated with nucleotide sequences from 16s rRNA and merA gene amplicons recovered

from both sites. Various among site (β) diversity metrics (Dp, NTI, NRI) were calculated to quantify the diversity and distribution of merA in the two sampled hot springs.

Dp increased in the downstream sample site relative to the site closest to the source for both the 16S rRNA gene and merA sequence assemblages in both springs tested, although with weak support for Succession Spring site 2 (Table 3.5). These data indicate a greater phylogenetic diversity in downstream sites relative to that of the total sequence pool. All four sites within both the 16S rRNA and merA specific communities exhibited positive NRI and NTI values, indicating overall phylogenetic clustering as compared to the total sequence pool (NRI), and within each assemblage (NTI); 8/16

104 metrics were statistically well supported in the complete analysis (Table 3.6).

Assemblages with a relatively higher Dp also exhibited a lower NRI or NTI in all cases, suggesting relative phylogenetic over-dispersion in the downstream sites, and greater clustering in communities near the source within each hot spring tested compared to the randomly generated null model. Overall, these observations highlight the uniqueness of each assemblage relative to the total sampled community.

Further phylogenetic cluster analysis of sampled 16S rRNA and merA gene assemblages revealed the impact of physical and chemical parameters in shaping merA

communities. A distance matrix derived from Rao’s Dp metric for each assemblage was used in an agglomerative cluster analysis. UniFrac was also used to compare the assemblages based on unique shared branch length in the input phylogeny. Both analysis reveal similar trends, with different observed clustering patterns generated from the 16S rRNA and merA gene phylogenies (Figure 3.5).

Similar methods were used to determine if the trends observed in the merA analysis described above would be corroborated by existing environmental sequence data from the NCBI database (Table 3.7, Figure 3.6). Similar trends were observed, with all assemblages exhibiting positive NRI and NTI metrics, indicating overall phylogenetic clustering as compared to the total sequence pool (NRI), and within each assemblage

(NTI). However, data were better resolved on the entire sequence pool scale (Table 3.7) than with the 4 Bijah and Succession assemblages (Table 3.6), with all 21 bacterial assemblages exhibiting well supported (P < 0.05) NRI results, while 15 of 20 NTI results were statistically well supported (P < 0.05). UniFrac analysis of 21 bacterial and 2 archaeal merA assemblages included in this meta-analysis reveal tight clustering of

105 assemblages based on sampling location (Figure 3.6). UniFrac results reveal 3 main assemblage clusters. One group containing 2 assemblages of archaeal sequences from

YNP (156). Another including sequences from Frying Pan West Pool in YNP along with samples from pristine and polluted fish farms in Turku (Finland) and the Stockholm

(Sweden) Archipelago (110). The third and largest group encompasses assemblages from

Bijah and Succession Springs from 2 studies in YNP (49, 156), Berry’s Creek,

Meadowlands, NJ (97), East Fork Poplar Creek, Oak Ridge, TN (105), Cornwallis Island,

Nunavut, Canada (112), and Sunday Lake, Adirondack Park, NY (Yu, personal communication). Within this larger group, assemblages form tight clusters based on sampling location. These data add support to previous observations of physical and chemical constraints on the community structure of hydA (21) and nifH (59) gene assemblages, encoding for [Fe-Fe]-hydrogenase and nitrogenase-iron protein, respectively.

The possibility cannot be ruled out that the specificity of PCR primer sets used did not amplify the full diversity of merA in the sampled environments, given the observation that merA primer sets 1-4 failed to amplify merA from pure cultures of

Hydrogenobaculum sp. AAS1 and Hydrogenivirga sp. R1-1, the Aquificae that are the topic of chapter 2, potentially underestimating the diversity of these environments. It is promising that assemblages created using merA primer sets 2 and 3 formed very loose clusters that encompassed all bacterial assemblages, suggesting the physical and chemical parameters of the sampled location effects ecological structure more than a possible bias introduced by using different primer sets to retrieve gene sequences from environmental

DNA extracts.

106

Data presented in chapter 3 begin to shed light on the degree to which merA communities are structured by physical and chemical parameters in YNP hot springs and on a global scale. Various phylogenetically derived diversity indices indicate i) greater diversity in sites further from the spring source, ii) greater clustering in communities near the source within each hot spring tested compared the randomly generated null model, and iii) possible controls on merA diversity and distribution based on the physical and chemical parameters of the tested environment. Global meta-analysis of most available environmental merA sequences may add support to the hypothesis of ecological control on merA communities, given the result of assemblages generated using merA primer sets

2 and 3. The possibility cannot be ruled out that the differences observed in aforementioned phylogenetic cluster analysis of merA gene assemblages are not simply due to PCR bias, which weakens the conclusions that are able to drawn from this study and warrant further study.

Taken together, the data from the entirety of this study provide further insight into our growing understanding of microbially mediated Hg-resistance in geothermal environments. The data presented in this thesis provide further insight into the ecological importance of the phylum Aquificae, and specifically into their probable primary role in the geothermal biotic Hg-cycle. My work suggests that microbes known to be highly abundant in geothermal environments (144), may play an important role in modulating

Hg toxicity in an environment that is constantly bathed in Hg. Furthermore, it appears that physical and chemical gradients structure Hg-resistant communities in YNP hot springs. These data also add further support to the hypothesis that mer-mediated Hg resistance originated in organisms which have most likely been exposed to toxic metals

107 since the oxygenation of the biosphere, most likely in an environment that is highly similar to a modern day hot spring.

Further characterization of Hg resistance in deeply branching bacterial lineages, as well as ecological studies to understand the environmental forces that shape Hg- resistant communities will help clarify differences in Hg resistance capabilities, and to provide increased resolution in ecological studies. Through methods such as enrichment and genome sequencing, our understanding of the ecological and evolutionary importance of the Aquificae in Hg(II) reduction in geothermal environments will be better understood. As I practiced in this study (chapter 2), it is common in studies on specific phenotype to include a negative control that lacks genes necessary to perform the function in question (58, 86). The common use of mutants defective in the studied function is currently not possible with the Aquificae as a genetic manipulation system is not yet available for this phylum, although lipid transferase (WaaA) from Aquifiex aeolicus, the first genome to be sequenced among the Aquificae (34), have been expressed in E. coli (85). The reproducible genetic manipulations of Aquificae appears to be a major challenge in understanding the metabolic capabilities and ecology of this phylum which according to most is the earliest bacterial lineage and represent life on early Earth.

The application of next-gen methods such as metagenomics and transcriptomics has potential to reveal previously undiscovered diversity and depth of coverage of sampled communities. From results described in this thesis, it would be beneficial for future studies regarding the ecological structure of Hg-resistant communities to focus on next-gen “omics” methods given the difficulty to obtain a single primer set to cover the

108 known diversity of merA. New frontiers in molecular microbial ecology methods will lead to exciting refinement of our current understanding of the ecological controls on the evolution and distribution of functional genes in microbial communities.

109

References

1. Abbott, M., J. Einerson, and R. Lee. 2004. Presented at the 7th International Conference on Mercury as a Global Pollutant, Ljubljana, Slovenia.

2. Aguiar, P., T. J. Beveridge, and A.-L. Reysenbach. 2004. Sulfurihydrogenibium azorense, sp. nov., a thermophilic hydrogen-oxidizing microaerophile from terrestrial hot springs in the Azores. Internation Journal of Systematic and Evolutionary Microbiology 54:33-39.

3. Alfredsson, G. A., J. K. Kristjansson, S. Hjorleifsdottir, and K. O. Stetter. 1988. Rhodothermus marinus, gen. nov., sp. nov., a Thermophibic, Halophilic Bacterium from Submarine Hot Springs in Iceland. Journal of General Microbiology 134:299-306.

4. Arai, H., H. Kanbe, M. Ishii, and Y. Igarashi. 2010. Complete genome sequence of the thermophilic, obligately chemolithoautotrophic hydrogen- oxidizing bacterium Hydrogenobacter thermophilus TK-6. J Bacteriol 192:2651- 2652.

5. Ball, J. W., R. B. McCleskey, D. K. Nordstrom, and J. M. Holloway. 2007. Water-chemistry data for selected springs, geysers, and streams in Yellowstone National Park, Wyoming, 2003–2005. , p. 183. In U. S. D. o. t. Interior (ed.), Washington D. C.

6. Barkay, T. 1987. Adaptation of aquatic microbial communities to Hg stress. Appl Environ Microbiol 53:2725-2732.

7. Barkay, T., M. Gillman, and R. R. Turner. 1997. Effects of dissolved organic carbon and salinity on bioavailability of mercury. Appl Environ Microbiol 63:4267-4271.

8. Barkay, T., K. Kritee, E. Boyd, and G. G. Geesey. 2010. A thermophilic bacterial origin and subsequent constraints by redox, light and salinity on the evolution of the microbial mercuric reductase. Environmental Microbiology.

9. Barkay, T., C. Liebert, and M. Gillman. 1989. Hybridization of DNA probes with whole-community genome for detection of genes that encode microbial responses to pollutants: mer genes and Hg(II) resistance. Appl Environ Microbiol 55:1574-1577.

10. Barkay, T., S. M. Miller, and A. O. Summers. 2003. Bacterial mercury resistance from atoms to ecosystems. FEMS Microbiology Reviews 27:355-384.

11. Barkay, T., and I. Wagner-Dobler. 2005. Microbial transformations of mercury: potentials, challenges, and achievements in controlling mercury toxicity in the environment. Adv Appl Microbiol 57:1-52.

110

12. Battistuzzi, F. U., A. Feijao, and S. B. Hedges. 2004. A genomic timescale of prokaryote evolution: insights into the origin of methanogenesis, phototrophy, and the colonization of land. BMC Evol Biol 4:44.

13. Ben-David, E. A., P. J. Holden, D. J. Stone, B. D. Harch, and L. J. Foster. 2004. The use of phospholipid fatty acid analysis to measure impact of acid rock drainage on microbial communities in sediments. Microb Ecol 48:300-315.

14. Bendtsen, J. D., H. Nielsen, D. Widdick, T. Palmer, and S. Brunak. 2005. Prediction of twin-arginine signal peptides. BMC Bioinformatics 6:167.

15. Benoit, J. M., C. C. Gilmour, and R. P. Mason. 2001. The influence of sulfide on solid-phase mercury bioavailability for methylation by pure cultures of Desulfobulbus propionicus (1pr3). Environ Sci Technol 35:127-132.

16. Bhaya, D., A. R. Grossman, A.-S. Steunou, N. Khuri, F. M. Cohan, N. Hamamura, M. C. Melendrez, M. M. Bateson, D. M. Ward, and J. F. Heidelberg. 2007. Population level functional diversity in a microbial community revealed by comparative genomic and metagenomic analyses. The ISME Journal 0:1-11.

17. Blank, C. E., S. L. Cady, and N. R. Pace. 2002. Microbial Composition of Near- Boiling Silica-Depositing Thermal Springs throughout Yellowstone National Park. Applied and Environmental Microbiology 68:5123-5135.

18. Bocchetta, M., S. Gribaldo, A. Sanangelantoni, and P. Cammarano. 2000. Phylogenetic Depth of the Bacterial Genera Aquifex and Thermotoga Inferred from Analysis of Ribosomal Protein, Elongation Factor, and RNA Polymerase Subunit Sequences. Journal of Molecular Evolution 50:366-380.

19. Boone, D. R., R. L. Johnson, and Y. Liu. 1989. Diffusion of the Interspecies

Electron Carriers H2 and Formate in Methanogenic Ecosystems and Its Implications in the Measurement of Km for H2 or Formate Uptake. Appl Environ Microbiol 55:1735-1741.

20. Botero, L. M., K. B. Brown, S. Brumefield, M. Burr, R. W. Castenholz, M. Young, and T. R. McDermott. 2004. Thermobaculum terrenum gen. nov., sp. nov.: a non-phototrophic gram-positive thermophile representing an environmental clone group related to the Chloroflexi (green non-sulfur bacteria) and Thermomicrobia. Arch Microbiol 181:269-277.

21. Boyd, E. S., T. L. Hamilton, J. R. Spear, M. Lavin, and J. W. Peters. 2010. [FeFe]-hydrogenase in Yellowstone National Park: evidence for dispersal limitation and phylogenetic niche conservatism. The ISME Journal:1-11.

22. Boyd, E. S., S. King, J. K. Tomberlin, D. K. Nordstrom, D. P. Krabbenhoft, T. Barkay, and G. G. Geesey. 2009. Methylmercury enters an aquatic food web

111

through acidophilic microbial mats in Yellowstone National Park, WY. Environmental Microbiology 11:950-959.

23. Boyd, E. S., J. R. Spear, and J. W. Peters. 2009. [FeFe] hydrogenase genetic diversity provides insight into molecular adaptation in a saline microbial mat community. Appl Environ Microbiol 75:4620-4623.

24. Branch, U. C. 2009. The Global Atmospheric Mercury Assessment: Sources, Emissions and Transport. In UNEP-Chemicals (ed.), Geneva, Switzerland.

25. Brock, T. D., and H. Freeze. 1969. Thermus aquaticus gen. n. and sp. n., a Non- sporulating Extreme Thermophile. Journal of Bacteriology:289-297.

26. Bruce, K. D., A. M. Osborn, A. J. Pearson, P. Strike, and D. A. Ritchie. 1995. Genetic diversity within mer genes directly amplified from communities of noncultivated soil and sediment bacteria. Mol Ecol 4:605-612.

27. Caldwell, S. L., Y. Liu, I. Ferrera, T. J. Beveridge, and A.-L. Reysenbach. In Press. Thermocrinis minervae sp. nov., a hydrogen and sulfur oxidizing, thermophilic member of the Aquificales from a Costa Rican terrestrial hot spring. Internation Journal of Systematic and Evolutionary Microbiology.

28. Chatziefthimiou, A. D., M. Crespo-Medina, Y. Wang, C. Vetriani, and T. Barkay. 2007. The isolation and initial characterization of mercury resistant chemolithotrophic thermophilic bacteria from mercury rich geothermal springs. Extremophiles 11:469-479.

29. Ciniglia, C., H. S. Yoon, A. Pollio, G. Pinto, and D. Bhattacharya. 2004. Hidden biodiversity of the extremophilic Cyanidiales red algae. Mol Ecol 13:1827-1838.

30. Corliss, J. B., J. Dymond, L. I. Gordon, J. M. Edmond, R. P. von Herzen, R. D. Ballard, K. Green, D. Williams, A. Bainbridge, K. Crane, and T. H. van Andel. 1979. Submarine thermal sprirngs on the galapagos rift. Science 203:1073-1083.

31. Craw, D. 2005. Potential anthropogenic mobilisation of mercury and arsenic from soils on mineralised rocks, Northland, New Zealand. J Environ Manage 74:283-292.

32. Crespo-Medina, M., A. Chatziefthimiou, N. S. Bloom, G. W. Luther, D. D. Wright, J. R. Reinfelder, C. Vetriani, and T. Barkay. 2009. Adaptation of chemosynthetic microorganisms to elevated mercury concentrations in deep-sea hydrothermal vents. Limnol Oceanogr 54:41-49.

33. Cuebas, M., D. Sannino, and E. Bini. 2011. Isolation and characterization of arsenic resistant Geobacillus kaustophilus strain from geothermal soils. J Basic Microbiol 51:364-371.

112

34. Deckert, G., P. V. Warren, T. Gaasterland, W. G. Young, A. L. Lenox, D. E. Graham, R. Overbeek, M. A. Snead, M. Keller, M. Aujay, R. Huber, R. A. Feldman, J. M. Short, G. J. Olsen, and R. V. Swanson. 1998. The complete genome of the hyperthermophilic bacterium Aquifex aeolicus. Nature 392:353- 358.

35. Eder, W., and R. Huber. 2002. New isolates and physiological properties of the Aquificales and a description of Thermocrinis albus sp. nov. Extremophiles 6:309-318.

36. Edmond, J. M., C. Measures, R. E. McDuff, L. H. Chan, R. Collier, B. Grant, L. I. Gordon, and J. B. Corliss. 1979. Ridge Crest Hydrothermal Activity and the Balances of the Major and Minor Elements in the Ocean: The Galapagos Data. Earth Planet Sc Lett 46:1-18.

37. Ekstrom, E. B., F. M. Morel, and J. M. Benoit. 2003. Mercury methylation independent of the acetyl-coenzyme A pathway in sulfate-reducing bacteria. Appl Environ Microbiol 69:5414-5422.

38. Engle, M., M. Gustin, F. Goff, D. Counce, C. Janik, D. Bergfeld, and J. Rytuba. 2006. Atmospheric mercury emissions from substrates and fumaroles associated with three hydrothermal systems in the western United States. . J Geophys Res-Atmosphere 11:D17304.

39. Farrell, R. E., J. J. Germida, and P. M. Huang. 1990. Biotoxicity of mercury as influenced by mercury(II) speciation. Appl Environ Microbiol 56:3006-3016.

40. Felsenstein, J. 2005. PHYLIP (Phylogeny Inference Package), p. Distributed by the Author, 3.6 ed. Department of Genome Sciences, University of Washington, Seattle, Seattle, WA.

41. Ferguson, T. J., and R. A. Mah. 1983. Isolation and characterization of an H(2)- oxidizing thermophilic methanogen. Appl Environ Microbiol 45:265-274.

42. Fierer, N., M. Hamady, C. L. Lauber, and R. Knight. 2008. The influence of sex, handedness, and washing on the diversity of hand surface bacteria. Proc Natl Acad Sci U S A 105:17994-17999.

43. Flores, G. E., Y. Liu, I. Ferrera, T. J. Beveridge, and A.-L. Reysenbach. 2008. Sulfurihydrogenibium kristjanssonii sp. nov., a hydrogen and sulfur oxidizing thermophile isolated from a terrestrial Icelandic hot spring. Internation Journal of Systematic and Evolutionary Microbiology 58:1153-1158.

44. Fouquet, Y., U. Von Stackelberg, L. Charlou, J. Donval, J. Erzinger, J. Foucher, P. Herzig, R. Mu ̈ he, S. Soakai, M. Wiedicke, and H. Whitechurch. 1991. Hydrothermal activity and metallogenesis in the Lau Back-Arc Basin. Natue 349:778-780.

113

45. Fournier, R. O. 1994. A field-trip guide to Yellowstone National Park, Wyoming, Montana, and Idaho--volcanic, hydrothermal, and glacial activity in the region. U.S. G.P.O., Denver, CO.

46. Fournier, R. O. 2005. Geochemistry and Dynamics of the Yellowstone National Park Hydrothermal System, p. 3-30. In W. P. Inskeep and T. R. McDermott (ed.), Geothermal Biology and Geochemistry of Yellowstone National Park. Montana State University Publications, Bozeman, MT.

47. Fox, B., and C. T. Walsh. 1982. Mercuric Reductase: Purification and Characterization of a Transposon-Encoded Flavoprotein Containing an Oxidation- Reduction-Active Disulfide. The Journal of Biological Chemistry 257:2498-2503.

48. Fox, B. S., and C. T. Walsh. 1983. Mercuric reductase: homology to glutathione reductase and lipoamide dehydrogenase. Iodoacetamide alkylation and sequence of the active site peptide. Biochemistry 22:4082-4088.

49. Freedman, Z., and T. Barkay. In Prep. Initial Characterization of Mercury Resistance in Aquificae; a promising candidate for the ancestral mer system?

50. Gotz, D., A. Banta, T. J. Beveridge, A. I. Rushdi, B. R. Simoneit, and A. L. Reysenbach. 2002. Persephonella marina gen. nov., sp. nov. and Persephonella guaymasensis sp. nov., two novel, thermophilic, hydrogen-oxidizing microaerophiles from deep-sea hydrothermal vents. Int J Syst Evol Microbiol 52:1349-1359.

51. Gotz, D., A. B. Banta, T. J. Beveridge, A. I. Rushdi, B. R. T. Simoneit, and A.-L. Reysenbach. 2002. Persephonella marina gen nov., sp. nov. and Persephonella guaymasensis sp. nov., two novel, thermophilic, hydrogen- oxidizing microaerophiles from deep-sea hydrothermal vents. Internation Journal of Systematic and Evolutionary Microbiology 52:1349-1359.

52. Gouy, M., S. Guindon, and O. Gascuel. 2010. SeaView version 4: A multiplatform graphical user interface for sequence alignment and phylogenetic tree building. Mol Biol Evol 27:221-224.

53. Griffiths, E., and R. S. Gupta. 2004. Signature sequences in diverse proteins provide evidence for the late divergance of the Order Aquificales. International Microbiology 7:41-52.

54. Guindon, S., and O. Gascuel. 2003. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Systematic biology 52:696- 704.

55. Guo, H. B., A. Johs, J. M. Parks, L. Olliff, S. M. Miller, A. O. Summers, L. Liang, and J. C. Smith. 2010. Structure and conformational dynamics of the metalloregulator MerR upon binding of Hg(II). J Mol Biol 398:555-568.

114

56. Hall, B. D., M. L. Olson, A. P. Rutter, R. R. Frontiera, D. P. Krabbenhoft, D. S. Gross, M. Yuen, T. M. Rudolph, and J. J. Schauer. 2006. Atmospheric mercury speciation in Yellowstone National Park. Sci Total Environ 367:354- 366.

57. Hamamura, N., R. E. Macur, S. Korf, G. Ackerman, W. P. Taylor, M. Kozubal, A.-L. Reysenbach, and W. P. Inskeep. 2009. Linking microbial oxidation of arsenic with detection and phylogenetic analysis of arenite oxidase genes in diverse geothermal environments. Environmental Microbiology 11:421- 431.

58. Hamamura, N., R. E. Macur, Y. Liu, W. P. Inskeep, and A.-L. Reysenbach. 2010. Distribution of Aerobic Arsenite Oxidase Genes within the Aquificales, p. 47-55, Interdisciplinary Studies on Environmental Chemistry - Biological Responses to Contaminants. Terrapub.

59. Hamilton, T. L., E. S. Boyd, and J. W. Peters. 2011. Environmental Constraints Underpin the Distribution and Phylogenetic Diversity of nifH in the Yellowstone Geothermal Complex. Microbial Ecol 61:860-870.

60. Helmann, J. D., Y. Wang, I. Mahler, and C. T. Walsh. 1989. Homologous metalloregulatory proteins from both gram-positive and gram-negative bacteria control transcription of mercury resistance operons. J Bacteriol 171:222-229.

61. Hirner, A. V., R. E. Krupp, A. R. Gainsford, and H. Staerk. 1970. Metal- organic associations in sediments. II. Algal mats in contact with geothermal waters. Appl Geochem 5::507–513.

62. Hofmann, K., and W. Stoffel. 1993. TMbase - A database of membrane spanning proteins segments. Biol. Chem. Hoppe-Seyler 374:166.

63. Horner-Devine, M. C., and B. J. Bohannan. 2006. Phylogenetic clustering and overdispersion in bacterial communities. Ecology 87:S100-108.

64. Huang, Q., C. Z. Dong, R. M. Dong, H. Jiang, S. Wang, G. Wang, B. Fang, X. Ding, L. Niu, X. Li, C. Zhang, and H. Dong. 2011. Archaeal and bacterial diversity in hot springs on the Tibetan Plateau, China. Extremophiles 15:549-563.

65. Huber, R., and W. Eder. 2006. Aquificales, p. 925-938. In M. Dworkin and S. Falkow (ed.), The Prokaryotes, vol. 7. Springer Reference, New York, NY.

66. Huber, R., W. Eder, S. Heldwein, G. Wanner, H. Huber, R. Rachel, and K. O. Stetter. 1998. Thermocrinis ruber gen. nov., sp. nov., a Pink-Filament- Forming Hyperthermophilic Bacterium Isolated from Yellowstone National Park. Applied and Environmental Microbiology 64:3576-3583.

67. Huelsenbeck, J. P., and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754-755.

115

68. Hugenholtz, P., C. Pitulle, K. L. Hershberger, and N. R. Pace. 1998. Novel Division Level Bacterial Diversity in a Yellowstone Hot Spring. Journal of Bacteriology 180:366-376.

69. Inskeep, W. P., and T. R. McDermott. 2005. Geothermal biology and geochemistry in Yellowstone National Park : proceeding of the Thermal Biology Institute workshop, Yellowstone National Park, WY, October 2003, p. xii, 352 p. In W. P. Inskeep and T. R. McDermott (ed.), Geothermal biology and geochemistry in Yellowstone National Park : proceeding of the Thermal Biology Institute workshop, Yellowstone National Park, WY, October 2003, 1st ed. Montana State University Publications, Bozeman, MT.

70. Inskeep, W. P., D. B. Rusch, Z. J. Jay, M. J. Herrgard, M. A. Kozubal, T. H. Richardson, R. E. Macur, N. Hamamura, R. Jennings, B. W. Fouke, A. L. Reysenbach, F. Roberto, M. Young, A. Schwartz, E. S. Boyd, J. H. Badger, E. J. Mathur, A. C. Ortmann, M. Bateson, G. Geesey, and M. Frazier. 2010. Metagenomes from high-temperature chemotrophic systems reveal geochemical controls on microbial community structure and function. PLoS One 5:e9773.

71. Jackson, W. J., and A. O. Summers. 1982. Biochemical characterization of

HgCl2-inducible polypeptides encoded by the mer operon of plasmid R100. J Bacteriol 151:962-970.

72. Jannasch, H. W., and M. J. Mottl. 1985. Geomicrobiology of deep-sea hydrothermal vents. Science 229:717-725.

73. Johs, A., I. M. Harwood, J. M. Parks, R. E. Nauss, J. C. Smith, L. Liang, and S. M. Miller. 2011. Structural Characterization of Intramolecular Hg(II) Transfer between Flexibly Linked Domains of Mercuric Ion Reductase. J Mol Biol 413:639-656.

74. King, S. A., S. Behnke, K. Slack, D. P. Krabbenhoft, D. K. Nordstrom, M. Burr, and R. G. Striegl. 2006. Mercury in water and biomass of microbial communities in hot springs of Yellowstone National Park, USA. Applied Geochemistry 21:1868-1879.

75. Lane, D. J., B. Pace, G. J. Olsen, D. A. Stahl, M. L. Sogin, and N. R. Pace. 1985. Rapid determination of 16S ribosomal RNA sequences for phylogenetic analyses. Proc Natl Acad Sci U S A 82:6955-6959.

76. Larkin, M. A., G. Blackshields, N. P. Brown, R. Chenna, P. A. McGettigan, H. McWilliam, F. Valentin, I. M. Wallace, A. Wilm, R. Lopez, J. D. Thompson, T. J. Gibson, and D. G. Higgins. 2007. Clustal W and Clustal X version 2.0. Bioinformatics 23:2947-2948.

77. Ledwidge, R., B. Patel, A. Dong, D. Fiedler, M. Falkowski, J. Zelikova, A. O. Summers, E. F. Pai, and S. M. Miller. 2005. NmerA, the metal binding domain

116

of mercuric ion reductase, removes Hg(II) from proteins, delivers it to the catalytic core, and protects cells under glutathione-depleted conditions. Biochemistry 44:11402-11416.

78. Lin, C. C., N. Yee, and T. Barkay. In Print. Microbial transformation in the mercury cycle. In G. Liu, Y. Cai, and N. O’Driscoll (ed.), Environmental Chemistry and Toxicology of Mercury. . John Wiley and Sons, Inc., Hoboken, NJ.

79. Lonsdale, P. 1977. Marine geoscience. Science 196:519.

80. Lozupone, C., and R. Knight. 2005. UniFrac: a New Phylogenetic Method for Comparing Microbial Communities. Applied and Environmental Microbiology 71:8228-8235.

81. Lozupone, C. A., and R. Knight. 2008. Species divergence and the measurement of microbial diversity. FEMS Microbiol Rev 32:557-578.

82. Macur, R. E., H. W. Langner, B. D. Kocar, and W. P. Inskeep. 2004. Linking geochemical processes with microbial community analysis: successional dynamics in an arsenic-rich, acid-sulfate-chloride geothermal spring. Geobiology 2:163-177.

83. Madigan, M. T., J. M. Martinko, D. A. Stahl, and D. P. Clark. 2010. Brock Biology of Microorganisms, 13th ed. Pearson/Benjamin Cummings, San Francisco, CA.

84. Magos, L., and T. W. Clarkson. 2006. Overview of the clinical toxicity of mercury. Annals of clinical biochemistry 43:257-268.

85. Mamat, U., H. Schmidt, E. Munoz, B. Lindner, K. Fukase, A. Hanuszkiewicz, J. Wu, T. C. Meredith, R. W. Woodard, R. Hilgenfeld, J. R. Mesters, and O. Holst. 2009. WaaA of the hyperthermophilic bacterium Aquifex aeolicus is a monofunctional 3-deoxy-D-manno-oct-2-ulosonic acid transferase involved in lipopolysaccharide biosynthesis. J Biol Chem 284:22248-22262.

86. Marszalkowski, M., J. H. Teune, G. Steger, R. K. Hartmann, and D. K. Willkomm. 2006. Thermostable RNase P RNAs lacking P18 identified in the Aquificales. RNA 12:1915-1921.

87. Martell, A. E., and R. M. Smith. 1974. Critical stability constants. Plenum Press, New York,.

88. Mathur, J., R. W. Bizzoco, D. G. Ellis, D. A. Lipson, A. W. Poole, R. Levine, and S. T. Kelley. 2007. Effects of Abiotic Factors on the Physiological Diversity of Bacterial Communities in Acidic Thermal Springs. Applied and Environmental Microbiology 73:2612-2623.

117

89. McCollom, T. M., and E. L. Shock. 1997. Geochemical constraints on chemolithoautotrophic metabolism by microorganisms in seafloor hydrothermal systems. Geochim Cosmochim Acta 61:4375-4391.

90. Mergeay, M., S. Monchy, T. Vallaeys, V. Auquier, A. Benotmane, P. Bertin, S. Taghavi, J. Dunn, D. van der Lelie, and R. Wattiez. 2003. Ralstonia metallidurans, a bacterium specifically adapted to toxic metals: towards a catalogue of metal-responsive genes. FEMS Microbiol Rev 27:385-410.

91. Meyer-Dombard, D. R., E. L. Shock, and J. P. Amend. 2005. Archaeal and bacterial communities in geochemically diverse hot springs of Yellowstone National Park, USA. Geobiology 3:211-227.

92. Muller, A. K., L. D. Rasmussen, and S. J. Sorensen. 2001. Adaptation of the bacterial community to mercury contamination. FEMS Microbiol Lett 204:49-53.

93. Muller, M. 2005. Twin-arginine-specific protein export in Escherichia coli. Research in Microbiology 156:131-136.

94. Nakagawa, R. 1995. Concentration of mercury in hair of diseased people in Japan. Chemosphere 30:135-140.

95. Nakagawa, R., and M. Hiromoto. 1997. Geographical distribution and background levels of total mercury in air in Japan and neighbouring countries. Chemosphere 34:801-806.

96. Nazaret, S., W. H. Jeffrey, E. Saouter, R. Von Haven, and T. Barkay. 1994. merA gene expression in aquatic environments measured by mRNA production and Hg(II) volatilization. Appl Environ Microbiol 60:4059-4065.

97. Ni Chadhain, S. M., J. Schaefer, S. Crane, G. J. Zylstra, and T. Barkay. 2006. Analysis of mercuric reductase (merA) gene diversity in an anaerobic mercury- contaminated sediment enrichment. . Environmental Microbiology 8:1746-1752.

98. Nies, D. H. 2003. Efflux-mediated heavy metal resistance in prokaryotes. FEMS Microbiol Rev 27:313-339.

99. Nies, D. H. 1999. Microbial heavy-metal resistance. Applied Microbiology and Biotechnology 51:730-750.

100. Nordstrom, D. K., J. W. Ball, and R. B. McClesky. 2005. Ground Water to Surface Water: Chemistry of Thermal Outflows in Yellowstone National Park, p. 73-94. In W. P. Inskeep and T. R. McDermott (ed.), Geothermal Biology and Geochemistry in Yellowstone National Park. Montana State University Publications, Bozeman, MT.

101. Nriagu, J. 1989. A global assessment of natural sources of atmospheric trace metals Nature 338:47-49.

118

102. Nriagu, J., and C. Becker. 2003. Volcanic emissions of mercury to the atmosphere: global and regional inventories. Sci Total Environ 304:3-12.

103. Nunoura, T., M. Miyazaki, Y. Suzuki, K. Takai, and K. Horoshi. 2008. Hydrogenivirga okinawensis sp. nov., a thermophilic sulfur-oidizing chemolithoautrotroph isolated from a deep-sea hydrothermal field, Southern Okinawa Trough. Internation Journal of Systematic and Evolutionary Microbiology 58:676-681.

104. O'Neill, A. H., Y. Liu, I. Ferrera, T. J. Beveridge, and A.-L. Reysenbach. 2008. Sulfurihydrogenibium rodmanii sp. nov., a sulfur-oxidizing chemolithotroph from the Uzon Caldera, Kamchatka Peninsula, Russia, and emended description of the genus Sulfurihydrogenibium. Internation Journal of Systematic and Evolutionary Microbiology 58:1147-1152.

105. Oregaard, G., and S. J. Sorensen. 2007. High diversity of bacterial mercuric reductase genes from surface and sub-surface floodplain soil (Oak Ridge, USA). Isme J 1:453-467.

106. Osborn, A. M., K. D. Bruce, P. Strike, and D. J. Ritchie. 1997. Distribution, diversity and evolution of the bacterial mercury resistance (mer) operon. FEMS Microbiology Reviews 19:239-262.

107. Paradis, E., J. Claude, and K. Strimmer. 2004. APE: Analyses of Phylogenetics and Evolution in R language. Bioinformatics 20:289-290.

108. Pfaffl, M. W. 2001. A new mathematical model for relative quantification in real- time RT-PCR. Nucleic Acids Res 29:e45.

109. Phelps, D., and P. R. Buseck. 1980. Distribution of Soil Mercury and the Development of Soild Mercury Anomalies in the Yellowstone Geothermal Area, Wyoming. Economic Geology 75:730-741.

110. Pitkanen, L. K., M. Tamminen, A. Hynninen, A. Karkman, J. Corander, A. Kotilainen, and M. Virta. 2011. Fish Farming Affects the Abundance and Diversity of the Mercury Resistance Gene merA in Marine Sediments. Microbes Environ.

111. Posada, D. 2006. ModelTest Server: a web-based tool for the statistical selection of models of nucleotide substitution online. Nucleic Acids Res 34:W700-703.

112. Poulain, A. J., S. M. Ni Chadhain, P. A. Ariya, M. Amyot, E. Garcia, P. G. Campbell, G. J. Zylstra, and T. Barkay. 2007. Potential for mercury reduction by microbes in the high arctic. Appl Environ Microbiol 73:2230-2238.

113. Purcell, D., U. Sompong, L. C. Yim, T. G. Barraclough, Y. Peerapornpisal, and S. B. Pointing. 2007. The effects of temperature, pH and sulfide on the

119

community structure of hyperthermophilic streamers in hot springs of northern Thailand. FEMS Microbiol Ecol 60:456-466.

114. Rambaut, A., D. L. Robertson, O. G. Pybus, M. Peeters, and E. C. Holmes. 2001. Human immunodeficiency virus. Phylogeny and the origin of HIV-1. Nature 410:1047-1048.

115. Rasmussen, L. D., and S. J. Sorensen. 1998. The effect of longterm exposure to mercury on the bacterial community in marine sediment. Curr Microbiol 36:291- 297.

116. Rasmussen, L. D., and S. J. Sorensen. 2001. Effects of mercury contamination on the culturable heterotrophic, functional and genetic diversity of the bacterial community in soil. FEMS Microbiol Ecol 36:1-9.

117. Reysenbach, A.-L. 2011, posting date. The Aquificales Data Warehouse. [Online.]

118. Reysenbach, A.-L., A. Banta, S. Civello, J. Daly, K. Mitchell, S. Lalonde, K. Konhauser, A. Rodman, K. Rusterholz, and C. Takacs-Vesbach. 2005. Aquificales in Yellowstone National Park, p. 129-142. In W. P. Inskeep and T. R. McDermott (ed.), Geothermal Biology and Geochemistry in Yellowstone National Park. Montana State University Publications, Bozeman, MT.

119. Reysenbach, A.-L., N. Hamamura, M. Podar, E. Griffiths, S. Ferreira, R. Hochstein, J. F. Heidelberg, J. Johnson, D. Mead, A. Pohorille, M. Sarmiento, K. Schweighofer, R. Seshadri, and M. A. Voytek. 2009. Complete and Draft Genome Sequences of Six Members of the Aquificales. Journal of Bacteriology 191:1992-1993.

120. Robinson, N. J., and D. R. Winge. 2010. Copper metallochaperones. Annual review of biochemistry 79:537-562.

121. Rona, P. A., G. Klinkhammer, T. A. Nelsen, J. H. Trefry, and H. Elderfield. 1986. Black Smokers, Massive Sulphides and Vent Biota at the Mid-Atlantic Ridge. Nature 321:33-37.

122. Rosen, B. P. 2002. Transport and detoxification systems for transition metals, heavy metals and metalloids in eukaryotic and prokaryotic microbes. Comp Biochem Physiol A Mol Integr Physiol 133:689-693.

123. Rossy, E., O. Seneque, D. Lascoux, D. Lemaire, S. Crouzy, P. Delangle, and J. Coves. 2004. Is the cytoplasmic loop of MerT, the mercuric ion transport protein, involved in mercury transfer to the mercuric reductase? FEBS Lett 575:86-90.

120

124. Rousk, J., E. Baath, P. C. Brookes, C. L. Lauber, C. Lozupone, J. G. Caporaso, R. Knight, and N. Fierer. 2010. Soil bacterial and fungal communities across a pH gradient in an arable soil. Isme J 4:1340-1351.

125. Sanderson, M. J. 2002. Estimating Absolute Rates of Molecular Evolution and Divergence Times: A Penalized Likelihood Approach. Mol Biol Evol 19:101- 109.

126. Sargent, F., B. C. Berks, and T. Palmer. 2006. Pathfinders and trailblazers: a prokaryotic targeting system for transport of folded proteins. FEMS Microbiol Lett 254:198-207.

127. Sayeh, R., J. L. Birrien, K. Alain, G. Barbier, M. Hamdi, and D. Prieur. 2010. Microbial diversity in Tunisian geothermal springs as detected by molecular and culture-based approaches. Extremophiles 14:501-514.

128. Schaefer, J., J. Letowski, and T. Barkay. 2002. mer-Mediated Resistance and Volatilization of Hg(II) Under Anaerobic Conditions. Geomicrobiology Journal 19:87-102.

129. Schaefer, J. K., J. Yagi, J. R. Reinfelder, T. Cardona, K. M. Ellickson, S. Tel- Or, and T. Barkay. 2004. Role of the bacterial organomercury lyase (MerB) in controlling methylmercury accumulation in mercury-contaminated natural waters. Environ Sci Technol 38:4304-4311.

130. Schecher, W. D., and D. McAvoy. 1994. MINEQL+: A Chemical Equilibrium Program for Personal Computers. Environmental Research Software, Howell, ME.

131. Schelert, J., V. Dixit, V. Hoang, J. Simbahan, M. Drozda, and P. Blum. 2004. Occurrence and Characterization of Mercury Resistance in the Hyperthermophilic Archaeon Sulfolobus sulfataricus by Use of Gene Disruption. Journal of Bacteriology 186:427-437.

132. Schelert, J., M. Drozda, V. Dixit, A. Dillman, and P. Blum. 2006. Regulation of mercury resistance in the crenarchaeote Sulfolobus solfataricus. J Bacteriol 188:7141-7150.

133. Schloss, P. D., and J. Handelsman. 2006. Toward a census of bacteria in soil. PLoS computational biology 2:e92.

134. Schloss, P. D., S. L. Westcott, T. Ryabin, J. R. Hall, M. Hartmann, E. B. Hollister, R. A. Lesniewski, B. B. Oakley, D. H. Parks, C. J. Robinson, J. W. Sahl, B. Stres, G. G. Thallinger, D. J. Van Horn, and C. F. Weber. 2009. Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol 75:7537-7541.

121

135. Service, U. S. N. P. 2009, posting date. Geothermal Resources. [Online.]

136. Sherman, L. S., P. Blum, D. K. Nordstrom, R. B. McCleskey, T. Barkay, and C. Vetriani. 2009. Mercury isotopic composition of hydrothermal systems in the Yellowstone Plateau volcanic field and Guaymas Basin sea-floor rift. Earth and Planetary Science Letters 279:86-96.

137. Shima, S., and K. I. Suzuki. 1993. Hydrogenobacter acidophilus sp. nov., a Thermoacidophilic, Aerobic, Hydrogen-Oxidizing Bacterium Requiring Elemental Sulfur for Growth. International Journal of Systematic Bacteriology 43:703-708.

138. Shock, E., and P. Canovas. 2010. The potential for abiotic organic synthesis and biosynthesis at seafloor hydrothermal systems. Geofluds 10:161-192.

139. Shock, E., Holland M., Meyer-Dombard D.R., J. P. Amend, Osburn G.R., and T. P. Fischer. 2010. Quantifying inorganic sources of geochemical energy in hydrothermal ecosystems, Yellowstone National Park, USA. Geochim Cosmochim Acta 74:4005-4043.

140. Shock, E., H. M., M.-D. D.R., and J. P. Amend. 2005. Geochemical Sources of Energy for Microbial Metabolism in Hydrothermal Ecosystems: Obsidian Pool, Yellowstone National Park, vol. 1. Montana State University, Bozeman, MT.

141. Silver, S., and T. Phung le. 2005. A bacterial view of the periodic table: genes and proteins for toxic inorganic ions. J Ind Microbiol Biotechnol 32:587-605.

142. Simbahan, J., E. Kurth, J. Schelert, A. Dillman, E. Moriyama, S. Jovanovich, and P. Blum. 2005. Community Analysis of a Mercury Hot Spring Supports Occurrence of Domain-Specific Forms of Mercuric Reductase. Applied and Environmental Microbiology 71:8836-8845.

143. Sogin, M. L., H. G. Morrison, J. A. Huber, D. Mark Welch, S. M. Huse, P. R. Neal, J. M. Arrieta, and G. J. Herndl. 2006. Microbial diversity in the deep sea and the underexplored "rare biosphere". Proc Natl Acad Sci U S A 103:12115- 12120.

144. Spear, J. R., J. J. Walker, T. M. McCollom, and N. R. Pace. 2005. Hydrogen and bioenergetics in the Yellowstone geothermal system. Proceedings of the National Academy of Science 102:2555-2560.

145. Spiess, F. N., K. C. Macdonald, T. Atwater, R. Ballard, A. Carranza, D. Cordoba, C. Cox, V. M. Garcia, J. Francheteau, J. Guerrero, J. Hawkins, R. Haymon, R. Hessler, T. Juteau, M. Kastner, R. Larson, B. Luyendyk, J. D. Macdougall, S. Miller, W. Normark, J. Orcutt, and C. Rangin. 1980. East pacific rise: hot springs and geophysical experiments. Science 207:1421-1433.

122

146. Stoffers, P., M. Hannington, I. Wright, P. Herzig, and E. dr Ronde. 1999. Elemental mercury at submarine hydrothermal vents in the Bay of Plenty, Taupo volcanic zone, New Zealand. . Geology 27:931-934.

147. Summers, A. O. 1986. Organization, expression, and evolution of genes for mercury resistance. Annual Review of Microbiology 40:604-634.

148. Takacs-Vesbach, C., K. Mitchell, O. Jackson-Weaver, and A.-L. Reysenbach. 2008. Volcanic calderas delineate biogeographic provinces among Yellowstone thermophiles. Environmental Microbiology 10:1681-1689.

149. Takai, K., H. Hirayama, Y. Sakihama, F. Inagaki, Y. Yamato, and K. Horikoshi. 2002. Isolation and Metabolic Characteristics of Previously Uncultured Members of the Order Aquificales in a Subsurface Gold Mine. Applied and Environmental Microbiology 68:3046-3054.

150. Takai, K., H. Kobayashi, K. H. Nealson, and K. Horikoshi. 2003. Deferribacter desulfuricans sp nov., a novel sulfur-, nitrate- and arsenate- reducing thermophile isolated from a deep-sea hydrothermal vent. International Journal of Systematic and Evolutionary Microbiology 53:839-846.

151. Tamura, K., J. Dudley, M. Nei, and S. Kumar. 2007. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol Biol Evol 24:1596-1599.

152. Torsvik, V., and L. Ovreas. 2002. Microbial diversity and function in soil: from genes to ecosystems. Curr Opin Microbiol 5:240-245.

153. Varekamp, J. C., and P. R. Buseck. 1984. The Speciation of Mercury in Hydrothermal Systems, with Applications to Ore Deposition. Geochim Cosmochim Ac 48:177-185.

154. Vetriani, C., Y. S. Chew, S. M. Miller, J. Yagi, J. Coombs, R. A. Lutz, and T. Barkay. 2005. Mercury Adaptation among Bacteria from a Deep-Sea Hydrothermal Vent. Applied and Environmental Microbiology 71:220-226.

155. Wang, M., J.-K. Chen, and B. Li. 2007. Characterization of Bacterial Community Structure and Diversity in Rhizosphere Soils of Three Plants in Rapidly Changing Salt Marshes Using 16S rDNA. Pedosphere 17:545-556.

156. Wang, Y., E. Boyd, S. Crane, P. Lu-Irving, D. Krabbenhoft, S. King, J. Dighton, G. Geesey, and T. Barkay. 2011. Environmental Conditions Constrain the Distribution and Diversity of Archaeal merA in Yellowstone National Park, Wyoming, U.S.A. Microb Ecol.

157. Wang, Y., Z. Freedman, P. Lu-Irving, R. Kaletsky, and T. Barkay. 2009. An initial characterization of the mercury resistance (mer) system of the thermophilic

123

bacterium Thermus thermophilus HB 27. FEMS Microbiology Ecology 67:118- 129.

158. Ward, D. M., M. J. Ferris, S. C. Nold, and M. M. Bateson. 1998. A Natural View of Microbial Biodiversity within Hot Spring Cyanobacterial Mat Communities. Microbiology and Molecular Biology Reviews 62:1353-1370.

159. Ward, J. H. 1963. Hierarchical grouping to optimize an objective function. . J Amer Statist Assoc 58:236–244.

160. Webb, C. O. 2000. Exploring the Phylogenetic Structure of Ecological Communities: An Example for Rain Forest Trees. The American naturalist 156:145-155.

161. Webb, C. O., D. D. Ackerly, and S. W. Kembel. 2008. Phylocom: software for the analysis of phylogenetic community structure and trait evolution. Bioinformatics 24:2098-2100.

162. Werner, C., S. L. Brantley, and K. Boomer. 2000. CO2 emissions related to the Yellowstone volcanic system 2. Statistical sampling, total degassing, and transport mechanisms. Journal of Geophysical Research 105:10831-10846.

163. Wiens, J. J. 2004. Speciation and ecology revisited: phylogenetic niche conservatism and the origin of species. Evolution; international journal of organic evolution 58:193-197.

164. Yu, R. Q., I. Adatto, M. R. Montesdeoca, C. T. Driscoll, M. E. Hines, and T. Barkay. 2010. Mercury methylation in Sphagnum moss mats and its association with sulfate-reducing bacteria in an acidic Adirondack forest lake wetland. FEMS Microbiology Ecology 74:655-668.

165. Zeytun, A., J. Sikorski, M. Nolan, A. Lapidus, S. Lucas, J. Han, H. Tice, J. F. Cheng, R. Tapia, L. Goodwin, S. Pitluck, K. Liolios, N. Ivanova, K. Mavromatis, N. Mikhailova, G. Ovchinnikova, A. Pati, A. Chen, K. Palaniappan, O. D. Ngatchou-Djao, M. Land, L. Hauser, C. D. Jeffries, C. Han, J. C. Detter, S. Ubler, M. Rohde, B. J. Tindall, M. Goker, R. Wirth, T. Woyke, J. Bristow, J. A. Eisen, V. Markowitz, P. Hugenholtz, H. P. Klenk, and N. C. Kyrpides. 2011. Complete genome sequence of Hydrogenobacter thermophilus type strain (TK-6). Standards in genomic sciences 4:131-143.

124

Curriculum Vitae

Zachary Freedman

Education

2005-2012 Doctoral Candidate, Graduate Program in Ecology and Evolution, Department of Biochemistry and Microbiology, Rutgers, The State University of New Jersey, New Brusnwick, NJ 08901

2001-2005 B.S. Biology Fairfield University, Fairfield, CT 06824.

Publications

Lefcort, H., Z. Freedman, S. House, and M. Pendleton. (2008) Hormetic Effects of Heavy Metals in Aquatic Snails: Is a Little Bit of Pollution Good? EcoHealth. 5: 10-17

Wang, Y., Z. Freedman, P. Lu-Irving, R. Kaletsky, and T. Barkay. (2009) An Initial Characterization of the Mercury Resistance (mer) System of the Thermophilic Bacterium Thermus thermophilus HB27. FEMS Microbiol. Ecol. 67: 118-129

Freedman, Z., E. S. Boyd, and T. Barkay. The Effect of Evolutionary Pressures on Community Structure and merA Diversity and Distribution in Yellowstone National Park Hot Springs. In preparation.

Freedman, Z., and T. Barkay. Characterization of the Mercury Resistance (mer) System in the Phylum Aquificae, a Promising Candidate for the Ancestral mer System. In preparation.