Non-thermodynamic factors affect competition between thermophilic chemolithoautotrophs from deep-sea hydrothermal vents

ABSTRACT Various environmental factors, including H2 availability, metabolic tradeoffs, optimal growth temperature, stochasticity, and hydrology, were examined to determine if they affect microbial competition between three autotrophic thermophiles. The thiosulfate reducer Desulfurobacterium thermolithotrophum (Topt72°C) was grown in mono- and coculture separately with the methanogens Methanocaldococcus jannaschii (Topt82°C) at 72°C and Methanothermococcus thermolithotrophicus (Topt65°C) at 65°C at high and low H2 concentrations. Both methanogens showed a metabolic tradeoff shifting from high growth rate–low cell yield at high H2 concentrations to low growth rate–high cell yield at low H2 concentrations and when grown in coculture with the thiosulfate reducer. In 1:1 initial ratios, D. thermolithotrophum outcompeted both methanogens at high and low H2, no H2S was detected on low H2, and it grew with only CO2 as the electron acceptor indicating a similar metabolic tradeoff with low H2. When the initial methanogen-to-thiosulfate reducer ratio varied from 1:1 to 104:1 with high H2, D. thermolithotrophum always outcompeted M. jannaschii at 72°C. However, M. thermolithotrophicus outcompeted D. thermolithotrophum at 65°C when the ratio was 103:1. A reactive transport model that mixed pure hydrothermal fluid with cold seawater showed that hyperthermophilic methanogens dominated in systems where the residence time of the mixed fluid above 72°C was sufficiently high. With shorter residence times, thermophilic thiosulfate reducers dominated. If residence times increased with decreasing fluid temperature along the flow path, then thermophilic methanogens could dominate. Thermophilic methanogen dominance spread to previously thiosulfate-reducer-dominated conditions if the initial ratio of thermophilic methanogen-to-thiosulfate reducer increased. IMPORTANCE The deep subsurface is the largest reservoir of microbial biomass on Earth and serves as an analog for life on the early Earth and extraterrestrial environments. Methanogenesis and sulfur reduction are among the more common chemolithoautotrophic metabolisms found in hot anoxic hydrothermal vent environments. Competition between H2-oxidizing sulfur reducers and methanogens is primarily driven by the thermodynamic favorability of redox reactions with the former outcompeting methanogens. This study demonstrated that competition between the hydrothermal vent chemolithoautotrophs Methanocaldococcus jannaschii, Methanothermococcus thermolithotrophicus, and Desulfurobacterium thermolithotrophum is also influenced by other overlapping factors such as staggered optimal growth temperatures, stochasticity, and hydrology. By modeling all aspects of microbial competition coupled with field data, a better understanding is gained on how methanogens can outcompete thiosulfate reducers in hot anoxic environments and how the deep subsurface contributes to biogeochemical cycling.

T he rocky subseafloor harbors an estimated 40% of the overall bacterial and archaeal biomass on Earth (1,2), but the local and global biogeochemical impact of this life is largely unknown.Deep-sea hydrothermal vents provide a window into the fundamental microbial and biogeochemical processes that occur within the rocky subseafloor.At Axial Seamount, the chemistries of low-temperature (<50°C) hydrothermal fluids from two sites, Marker 113 and Marker 33, had similar pH and concentrations of H 2 S, ΣCO 2 , and CH 4 suggesting overall similarity in the source fluids (3,4) (Table S1).However, among the thermophiles collected from exiting low-temperature hydrothermal fluid, Marker 113 was mostly populated by the methanogens Methanocaldococcus and Methanothermo coccus, while Marker 33 was mostly the sulfur and thiosulfate reducer Desulfurobacterium based on metagenomic and culture-based analyses (4, 5) (Fig. S1).This suggests that these thermophilic, hydrogenotrophic autotrophs at these sites occupy similar ecological niches and likely compete for H 2 but with differing outcomes.The purpose of this study was to examine possible factors that lead to one group of organisms outcompeting the other in hydrothermal systems.
Models are used to predict microbial competition and biogeochemical impacts using a combination of deterministic and stochastic variables.Competition between chemoli thoautotrophs at hydrothermal vents is often predicted using thermodynamic models based on the Gibbs energy available for given metabolisms and the fluid geochemistry of the environment (6)(7)(8).However, in resource-limited environments, microorganisms can undergo a metabolic tradeoff between microbial growth rate and cell yield (Y x/p , biomass produced/mole of metabolite produced) (9,10), which may provide a compet itive advantage for some organisms.Previous work showed that Methanocaldococcus jannaschii switched from fast growth rates and low growth yields in high H 2 conditions to slow growth rates and high growth yields in low H 2 conditions (11).However, it is unknown if a similar tradeoff occurs in thermophilic autotrophic sulfur/thiosulfate reducers and how this might impact competition in resource-limited environments.Also, staggered optimal growth temperatures of thermophilic organisms can create compet itive advantages depending on subseafloor geometry and hydrology, which can vary the residence time of hydrothermal fluids at different temperatures (12).Furthermore, stochastic factors, such as dispersal rates, drift, and population bottlenecks, can create niche exclusion and affect community composition (13).Some vent sites experience large shifts in microbial community composition following volcanic eruptions (14)(15)(16) possibly triggering stochastic recolonization events of new vent sites that could impact the outcome of competition.
In this study, competition between the thermophilic anaerobic chemolithoautotrophs Methanocaldococcus, Methanothermococcus, and Desulfurobacterium was examined to determine if the predominance of a type of organism or its metabolism was due primarily to the Gibbs energy for its metabolism or under what circumstances other factors, such as H 2 availability, metabolic tradeoffs, relative cell concentration, incubation temperature, and hydrology, might influence the outcome of competition.These data were used to parameterize a non-dimensional reactive transport model that included thermophilic and hyperthermophilic hydrogenotrophic methanogenesis and thermo philic hydrogenotrophic thiosulfate reduction and considered varying parameters, such as initial cell concentrations and residence times, of the fluids at different temperatures.

H 2 limitation and metabolic tradeoffs
H 2 availability for cell growth was assumed to become increasingly limited going from monoculture to coculture conditions and from high initial H 2 in the headspace to low initial H 2 (left-to-right for each organism in Fig. 1).Each set of high and low H 2 incuba tions ended when Desulfurobacterium thermolithotrophum in monoculture reached late logarithmic growth phase (~4-5 cell doublings).For each condition, D. thermolithotro phum grew to a higher cell concentration than Methanothermococcus thermolithotrophi cus at 65°C (Fig. 1A) and M. jannaschii at 72°C (Fig. 1B).At both growth temperatures, the maximum cell concentrations of D. thermolithotrophum were generally the same for high H 2 monocultures, low H 2 monocultures, and low H 2 cocultures but lower for high H 2 cocultures (Fig. 1A and B).Similarly, the specific growth rates of D. thermolithotrophum were the same for all four growth conditions at 65°C and in five of six pairwise compari sons at 72°C (Fig. S2A and B).
At both growth temperatures, the maximum cell concentration of the methanogens was lower in high H 2 coculture, low H 2 monoculture, and low H 2 coculture conditions relative to high H 2 monoculture conditions (Fig. 1A and B).At 65°C, the specific growth rate of M. thermolithotrophicus was unchanged in coculture relative to monoculture at both high and low H 2 concentrations but was lower at low H 2 concentrations relative to high H 2 concentrations (Fig. S2A and B).At 72°C, the specific growth rate of M. jannaschii was lower in coculture relative to monoculture at both high and low H 2 concentrations.
At both temperatures, each methanogen decreased the amount of CH 4 produced per cell in low H 2 coculture conditions relative to high H 2 monoculture conditions (Fig. 1C and D).As a result, the cell yield increased from 1.9 × 10 12 cells/mol CH 4 produced by both high H 2 monocultures to 1.4 × 10 13 cells/mol CH 4 produced by both low H 2 co-cultures indicating a rate-yield metabolic tradeoff with decreasing H 2 availability.In high H 2 conditions, D. thermolithotrophum produced 577 fmol H 2 S/cell, and cell yield was 1.7 × 10 12 cells/mol H 2 S produced at both temperatures in mono-and coculture (Fig. 1C and D).However, no H 2 S was detected when D. thermolithotrophum was grown in low H 2 conditions despite having the same maximum cell concentrations and specific growth rates as in high H 2 conditions (Fig. 1C and D; Fig. S2A and B).This suggests that D. thermolithotrophum may not require thiosulfate as a terminal electron acceptor, especially at low H 2 concentrations, and increased its cell yield on low H 2 also as part of a rate-yield metabolic tradeoff.
To test this, D. thermolithotrophum was grown at high and low H 2 concentrations without the addition of Na 2 S 2 O 3 or any other terminal electron acceptor other than CO 2 .At high H 2 concentration, the specific growth rate of D. thermolithotrophum with thiosulfate was 1.28 ± 0.19 h −1 (95% CI) (Fig. 2).With high H 2 but without thiosulfate, D. thermolithotrophum grew at a specific growth rate of 0.44 ± 0.07 h −1 .At low H 2 concentration, the specific growth rates of D. thermolithotrophum with and without thiosulfate were each the same as for high H 2 concentration without thiosulfate (Fig. 2).

Competition at varying initial methanogen:thiosulfate reducer ratios
Each incubation ended separately when the cultures for each condition reached stationary growth phase.When D. thermolithotrophum was grown at 65°C in monocul ture and in coculture with M. thermolithotrophicus, the maximum cell concentrations of D. thermolithotrophum at the end of growth were the same for all monoculture conditions and for 10 6 :10 5 and 10 6 :10 4 initial ratios of M. thermolithotrophicus:D. thermolithotrophum (Fig. 3A).However, the maximum cell concentration of D. thermoli thotrophum at the end of growth decreased significantly when the initial cell ratio was 10 6 :10 3 of M. thermolithotrophicus:D. thermolithotrophum relative to the monocultures and for coculture ratios of 10 6 :10 4 relative to 10 6 :10 5 .The maximum cell concentration of M. thermolithotrophicus at the end of growth for cells grown at 65°C and 10 6 :10 4 of M. thermolithotrophicus:D. thermolithotrophum was lower than that of M. thermolithotrophi cus grown in monoculture, but the maximum cell concentration at the end of growth for the methanogen at 10 6 :10 3 initial cell ratio was the same as the monoculture (Fig. 3A).
At 72°C, the maximum cell concentrations of D. thermolithotrophum at the end of growth were the same for all mono-and M. jannaschii coculture conditions regardless of their initial cell concentration (Fig. 3B).The maximum cell concentration of M. jannaschii at the end of growth in coculture with D. thermolithotrophum was lower than that of M. jannaschii in monoculture for all initial cell ratios (Fig. 3B).

Reactive transport modeling
The outcome of competition between hyperthermophilic and thermophilic methano gens and thermophilic autotrophic thiosulfate reducers was predicted using a reactive transport model with varying hydrologic shape functions (x b ) and dilution rates (Q′ vt ) at varying temperatures (Fig. 4).High x b values represent pipe-like flow with little  4B).For all conditions, hyperthermophilic methanogens were dominant at low Q′ vt values (i.e., high residence times) (Fig. 4C through F).Thermophilic autotrophic thiosulfate reducers were dominant when Q′ vt values increased (i.e., relatively lower residence times) and when x b values decreased (i.e., residence times increase with decreasing temperature).Thermophilic methanogens similarly were dominant as Q′ vt values continued to increase and x b values decreased.
When initial thermophilic methanogen cell concentrations were increased to 10 2 -fold higher than the hyperthermophilic methanogens and thermophilic autotrophic thiosulfate reducers, the thermophilic methanogens were increasingly dominant at high Q′ vt -low x b values (Fig. 4D and F) relative to equal initial cell concentrations for the three organisms (Fig. 4C and E).Thermophilic methanogens were also more dominant when H 2 concentrations in pure hydrothermal vent fluid increased from 300 µM (Fig. 4C and  D) to 950 µM (Fig. 4E and F

DISCUSSION
Understanding the factors that influence competition between thermophilic chemoli thoautotrophs in hydrothermal environments provides insight into modern subseafloor processes as well as life on the early Earth and the search for extraterrestrial life.In this study, the thermophilic autotrophic thiosulfate reducer D. thermolithotrophum outcompeted thermophilic and hyperthermophilic methanogens at high and low H 2 concentrations when initially in equal cell concentrations with abundant thiosulfate present in the microcosms.This result was expected based on the standard Gibbs energy for each reaction with thiosulfate reduction producing more energy for growth than methanogenesis (8).However, differences in optimal growth temperatures and temperature growth ranges for the three groups of thermophiles influenced their predicted competition in a reactive transport model.Because hydrothermal fluid is the source of H 2 , hyperthermophilic methanogens have sole access to H 2 among the three groups of organisms until the hydrothermal fluid-seawater mixture cools to tempera tures in the range of the other two groups.Hyperthermophilic methanogens dominated when residence time was sufficiently high between 85°C and 72°C.Both microcosm experiments and the reactive transport model showed that thermophilic methanogens had an increasing predicted competitive advantage when the residence time of the fluid increased with decreasing temperature (i.e., low x b or plume-shaped geometry) and when they initially outnumber thermophilic thiosulfate reducers at least 10 2 -to 10 3 -fold.
H 2 availability is important for the ability of methanogens to compete with autotro phic thiosulfate reducers.Thermophilic methanogens were more predominant in the reactive transport model when the H 2 concentration in pure hydrothermal vent fluid increased from 300 to 950 µM and when the residence time of fluids increased with decreasing fluid temperatures (plume-shaped geometry).Monod kinetics showed that M. jannaschii had an H 2 K s of 67 µM and required at least 17-23 μM H 2 for growth in a bioreactor (3).For D. thermolithotrophum HR11, H 2 K s was 30 µM and required at least 3 µM H 2 for growth in a bioreactor (17).The maximum growth rate/H 2 -Ks ratio for D. thermolithotrophum HR11 is fourfold higher than that for M. jannaschii (17).Therefore, high-temperature methanogens appear to require highly elevated H 2 concentrations (or fluxes) to compete with Desulfurobacterium for H 2 .In nature, high-temperature methanogens are often more abundant than Aquificales, including Desulfurobacterium, at vent sites with higher H 2 concentrations (15,16,(18)(19)(20).Desulfurobacterium is often more abundant than high-temperature methanogens at vent sites with more moderate H 2 concentrations (4,(21)(22)(23).
The rate-yield metabolic tradeoff represents two divergent strategies that are important for microbial competition and biogeochemistry.The first is slow growth but efficient metabolism and high cell yields when resources are scarce; the second, fast growth but inefficient metabolism and low cell yields upon rich resources (9,10).This tradeoff was shown for M. jannaschii, which decreased its growth and CH 4 produc tion rates and increased its cell yield when grown with very low H 2 flux to the cells (11).This might have provided thermophilic methanogens with a growth advantage over Desulfurobacterium spp.This study demonstrated that a similar rate-yield meta bolic tradeoff occurs in D. thermolithotrophum when it is shifted from high to low H 2 concentrations.It also suggested that D. thermolithotrophum, which was believed to be an obligate sulfur or thiosulfate reducer, could use CO 2 as its sole added terminal electron acceptor through CO 2 fixation.At high initial H 2 concentrations, D. thermolitho trophum grew faster when thiosulfate was available as a terminal electron acceptor but grew slowly without the addition of thiosulfate.At low initial H 2 concentrations, the growth rates were the same, and no H 2 S was detected with or without the addition of thiosulfate.Therefore, the organism may only require sulfur or thiosulfate for extra energy generation when the environment has a high flux of H 2 .
D. thermolithotrophum BSA fixes CO 2 using the Arnon-Buchanan cycle (i.e., the reductive TCA cycle) and relies on ATP citrate lyase to split citrate into oxaloacetate and acetyl-CoA for biosynthesis (24,25).Based on its whole-genome sequence (26), D. thermolithotrophum HR11 has a Hyn-type (Group 1) hydrogenase in its membrane with a twin-arginine transport signal suggesting that its catalytic subunit is oriented toward the periplasm.It also encodes all the genes for the Arnon-Buchanan cycle including a membrane-bound fumarate reductase.This suggests that D. thermolithotro phum oxidizes H 2 in the periplasm and passes electrons to menaquinones, which D. thermolithotrophum BSA possesses (24), and then to fumarate reductase where the electrons enter the Arnon-Buchanan cycle for CO 2 reduction.A proton motive force for ATP generation via oxidative phosphorylation is generated by the oxidation of H 2 and menaquinones by hydrogenase and fumarate reductase, respectively.Using CO 2 as a terminal electron acceptor as well as for biosynthesis would provide D. thermolithotro phum with a growth advantage in hydrothermal environments that can often be limited in exogenous electron acceptors at high growth temperatures.
Subseafloor hydrology was previously shown to impact methanogen community composition at Axial Seamount.The plume-like (low x b ) hydrology of Marker 113 resulted in a shorter residence time for hotter fluid compared to cooler fluid, favoring Methanothermococcus, while the pipe-like (high x b ) geometry of Marker 33 had a longer residence time at hotter temperatures favoring Methanocaldococcus (12).In this study, the non-dimensional reactive transport model from Stewart et al. (12) was expanded to include thermophilic autotrophic growth of Desulfurobacterium spp., which are also present in this system (4).The new modeling results support the idea that Marker 113 has shorter residence times for hotter fluid, while Marker 33 has a longer residence time for hotter fluid.Using both CH 4 anomaly in hydrothermal fluids and measured propor tions of Methanocaldococcus, Methanothermococcus, and Desulfurobacterium based on metagenomics to determine the best fit, the modeling results suggest that the residence time of the fluids between 85°C and exiting the seafloor increased from what was reported by Stewart et al. (12).This increase in residence time is understandable given the increased demand for H 2 with the additional non-methanogenic hydrogenotroph.Furthermore, the model showed that stochastic effects, such as Methanothermococcus, initially outnumbering Desulfurobacterium 10 2 -fold can increase the Q′ vt −x b parameter space where Methanothermococcus can become the predominant thermophile in the system.Therefore, the stochastic colonization and establishment of an organism with relatively lower Gibbs energy for its metabolism may lead to that organism predominat ing a niche by excluding an organism with a higher Gibbs energy for its metabolism through a combination of relative abundances of organisms and relative residence times at varying temperatures in the system.
In conclusion, this study demonstrated the utility of coupling metagenomics, field geochemistry, thermodynamic predictions, and biogeochemical modeling to formu late hypotheses regarding the distribution and metabolic potential of microbes (27).Thermodynamic predictions of Gibbs energy associated with various autotrophic metabolisms remain a valid predictor of which thermophilic microbes will predominate in hydrothermal environments.In most instances, the thermophilic thiosulfate reducers in this study outcompeted the thermophilic methanogens.However, this study also showed the importance of other factors, such as staggered optimal growth temper atures, hydrology, and stochastic effects, and how in certain circumstances methano gens might outcompete thiosulfate reducers.Future thermodynamic predictions and biogeochemical modeling may also need to account for a shift in metabolism toward increased CO 2 fixation and less metabolic product formation as a means for autotrophic growth.Other aspects of competition modeling that were not accounted for herein are the formation of biofilms in the subseafloor, possible spatial separation of meta bolic processes within those biofilms, and antibiotic production.Pressure likely has a minor impact on microbial competition as little effect was observed on the growth, community composition, and metatranscriptomes of a natural assemblage of thermo philic hydrogenotrophic autotrophs sampled from Marker 33 hydrothermal fluid when incubated at 55°C on the seafloor at in situ pressure relative to a parallel shipboard incubation at 0.1 MPa (28).

Growth medium and microorganisms used
Three thermophilic organisms were obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ, Braunschweig, Germany) and used for monoand coculture experiments: M. jannaschii DSM 2661 (T opt 82°C) (29), M. thermolithotrophi cus DSM 2095 (T opt 65°C) (30), and D. thermolithotrophum HR11 (T opt 72°C) (DSM 100454) (17,26).All chemical reagents were purchased from Sigma-Aldrich.The growth medium is based on DSM medium 282 (29) and is composed of the following per liter: 30 g of NaCl, 4.1 g of MgCl 2 •6H 2 O, 3.40 g of MgSO 4 •7H 2 O, 0.33 g of KCl, 0.25 g of NH 4 Cl, 0.14 g of CaCl 2 •2H 2 O, 0.14 g of K 2 HPO 4 , 10 mL of DSM medium 141 trace elements solution, 10 mL of DSM medium 141 vitamins solution, 0.1 mL of 0.1% (wt/vol) each of Na 2 WO 4 •2H 2 O and Na 2 SeO 4 solution, 1 g of NaHCO 3 , 1 g of Na 2 S 2 O 3 , and 50 µL of 0.5% (wt/vol) resazarin.The medium was pH balanced to 6.00 ± 0.05.The medium was reduced with 0.025% (wt/vol) cysteine-HCl.All cells were grown in 60-mL serum bottles sealed with butyl rubber stoppers containing 25 mL of growth medium.For high H 2 conditions, 1.6 atm of H 2 and 0.4 atm of CO 2 were added to the headspace of each bottle based on the standard growth medium (29).For low H 2 conditions, which was based on the minimum amount of H 2 needed to support monoculture growth of each organism, 0.1 atm of H 2 , 1.5 atm of N 2 , and 0.4 atm of CO 2 were added to the headspace of each bottle.All gases were provided by AirGas.Using the geochemical modeling program Geochemist's Workbench 10.0, the estimated aqueous H 2 concentrations were 1.2 mM and 85 µM for the high and low H 2 conditions, respectively, and were representative of the expected H 2 concentrations at Marker 113 and Marker 33 (3,12).

Mono-and coculture incubations
To determine how H 2 concentration affects cell yield and growth competition, growth medium in serum bottles were inoculated separately with D. thermolithotrophum, M. jannaschii, and M. thermolithotrophicus each in monoculture or in coculture between D. thermolithotrophum and M. jannaschii or D. thermolithotrophum and M. thermolithotrophi cus (Fig. S3).A logarithmic growth phase culture was used to inoculate each bottle such that the initial concentration of each organism in each bottle was 10 6 cells/mL based on the detection limit of the Petroff-Hausser counting chamber.D. thermolithotrophum and M. jannaschii, in mono-and coculture, were incubated at 72°C (the optimum growth temperature of D. thermolithotrophum), while D. thermolithotrophum and M. thermoli thotrophicus, in mono-and coculture, were incubated at 65°C (the optimum growth temperature of M. thermolithotrophicus).Each combination of cells and temperatures was incubated separately using high and low H 2 concentrations as described above.
The concentration of total cells in each bottle was determined at various time points during growth using light microscopy (Nikon Eclipse 55i) and a Petroff-Hausser counting chamber.The proportions of methanogens and D. thermolithotrophum in each cocul ture sample were determined using fluorescence microscopy (Nikon Eclipse 55i with a CoolLED pE-300 light source) and the autofluorescence and coccoid shape of the methanogens (30) relative to the non-fluorescent rod-shaped D. thermolithotrophum (17) (Fig. S4).All sample bottles were removed from the incubator when one of the organ isms in monoculture (usually D. thermolithotrophum) reached late logarithmic-to-early stationary growth phase.The specific growth rate (per h) of each organism in mono-and coculture was determined by fitting an exponential curve to cell concentration versus time.NaOH (0.1 M final concentration) was added to each bottle.The amount of CH 4 in each bottle was determined by measuring the volume of gas in the bottle and using a gas chromatograph fitted with a flame ionization detector (SRI 8610C) and a HayeSep D packed column (Supelco, 6′×⅛″ stainless steel).The amount of sulfide in each bottle was determined using a spectrophotometer (Thermo Spectronic Genesys 10) and methylene blue (31).The product yield per cell was estimated from the total amount of H 2 S or CH 4 produced per bottle divided by the total number of thiosulfate reducers or methanogens per bottle, respectively.Data are presented in Table S2.
To determine how initial methanogen:thiosulfate reducer cell concentrations (per mL) ranging from 10 6 :10 5 to 10 6 :10 2 affect growth competition, growth medium in serum bottles was inoculated as described above except that the initial cell concentration of D. thermolithotrophum varied at 10 5 , 10 4 , 10 3 , and 10 2 cells/mL relative to 10 6 cells/mL for each methanogen (Fig. S5).All samples were incubated at 72°C or 65°C as described above with the high H 2 concentration only.Maximum cell concentrations and the total amount of methane and sulfide produced at the end of the incubation were determined as described above.The cells in each growth condition were allowed to reach late logarithmic-to-early stationary growth phase before removal from the incubator.Data are presented in Tables S3 and S4.
A two-way analysis of variance with a Tukey honestly significant difference post-hoc analysis was used to compare one dependent variable (maximum cell concentration, growth rate, growth yield) based on two independent variables as well as any potential interaction between the two (H 2 concentration and mono-versus coculture conditions).Comparisons were not made across temperatures or organisms except when that was the only variable.A two-sample statistical test (either Student's t-test or Wilcoxon rank sum) was used to compare the maximum cell concentrations of the thiosulfate reducer and methanogens in coculture to determine which organisms outcompete the other.All statistical analyses were performed in RStudio using R statistical software.

Reactive transport model
The reactive transport model used was a modification of that described previously by Stewart et al. (12) with the addition of hydrogenotrophic, thermophilic thiosul fate reduction to the model.The model described the growth of hydrogenotrophic methanogens and thiosulfate reducers in a mixture of pure high-temperature hydrother mal fluid and 2°C seawater with unlimited thiosulfate that was transported along a one-dimensional flow path consisting of a series of n boxes.High-temperature hydro thermal fluid, which naturally lacks Mg 2+ , entered the first box with the fluid composition of pure high-temperature hydrothermal fluid, then flowed from box-to-box, and at each box was progressively diluted with an equal amount of 2°C seawater (composition in Table S5) until it exits the last box with a fluid composition like that observed for low-temperature hydrothermal vents on the seafloor.Because Mg 2+ is biologically conservative, it acts as a tracer for the dilution of hydrothermal fluid with seawater, which contains Mg 2+ , and is used to constrain the reactive transport model with field data.Because the true length of the flow path is unknown, the model was non-dimensional ized with respect to space such that the sum of all the box volumes was equal to one and represents the total volume of the subseafloor mixing zone feeding the vent outflow.
The total residence time of fluid in the system is set by the spatially non-dimensional ized fluid flux exiting the seafloor (Q′ vt ), which has units of time −1 and can be thought of as a dilution factor that defines the timescale of hydrothermal fluid circulation.The residence time fluid spends at different temperatures along the flow path is controlled by both Q′ vt and the volume of each individual box along the flow path.To simplify the specification of box volumes, the volume of each box is given by the following formula: (Eqn. 1) where V is the box volume, n is the total number of boxes, x is a non-dimensional variable describing the position along the flow path and varies from 0 at the high-tem perature endmember to 1 at the point where fluid is venting into the deep ocean, and x b is a shape parameter describing the geometry of the subsurface mixing zone.The shape parameter allowed the model to transition between two different mixing regimes.flow path spread out as it rose, approximating an expanding plume percolating through the ocean crust.The fluid flux, Q′ vt , and the shape parameter, x b , were treated as tuning parameters that were adjusted to understand how flow characteristics influence the chemical concentrations and the microbial populations present in venting fluids.While Q′ vt sets the timescale of fluid flow, the shape parameter, x b , determined the relative amount of time spent at various temperatures along the flow path.All model parameters and boundary conditions are provided in Table S5.
The model state variables were concentration of CH 4 , [CH 4 ] (μmol/kg), concentration of H 2 , [H 2 ] (μmol/kg), concentration of a thermophilic methanogen with M. thermo lithotrophicus growth kinetics, [M the ] (cells/L), concentration of a hyperthermophilic methanogen with M. jannaschii growth kinetics, [M jan ] (cells/L), concentration of a thermophilic thiosulfate reducer with D. thermolithotrophum growth kinetics, [D the ] (cells/L), concentration of Mg 2+ , [Mg 2+ ] (mmol/kg), and temperature, T (K).Temperature was calculated assuming pure mixing and was given by the following equation, where f ht is the fraction of high-temperature fluid in the box determined from the Mg 2+ content of the venting fluid assuming zero Mg 2+ in the hydrothermal endmember, T ht is the temperature of the hydrothermal endmember fluid, T sw is the temperature of seawater, and C p,ht and C p,sw are the heat capacities of hydrothermal endmember fluid and seawater, respectively.During hydrothermal circulation, Mg 2+ is removed from solution within hours at high temperatures, and the Mg 2+ content of diffuse fluid indicates how much hot, zero-Mg 2+ endmember is in the fluid (32).
The rest of the variables were described by the following system of differential equations for each box i, which were solved using the method-of-lines.The model was implemented in the R software environment using the package ReacTran (33).
where Q is fluid flux.Methane (R CH4 ) and sulfide (R H2S ) production are calculated by, (Eqn.9) where, for CH 4 production, v max is the maximum rate of cell-specific CH 4 production, K H2 is the half-saturation constant for cell-specific CH 4 production, and T max is the optimum growth temperature of the methanogen or sulfur reducer (12).A conversion factor of 10 −9 is used to convert v max , which in Table S5 is expressed in terms of fmol/ cell/h to μmol/cell/h.For H 2 S production, Y x/p is the growth yield (this study) and μ max is the maximum growth rate (17).
The growth rates for methanogens (R M ) and thiosulfate reducers (R S ) are given by, where A is the Arrhenius constant, E a is the activation energy.These parameters were drawn from Stewart et al. (12,17).

Field constraints on the reactive transport model
Field measurements of Methanocaldococcus species, Methanothermococcus species, Desulfurobacterium species, CH 4 , and H 2 concentrations in exiting fluids at individual diffuse vents were used to constrain the modeled subseafloor methanogen and sulfur reducer abundances, CH 4 production, and the shape function of fluid mixing for each vent as in Stewart et al. (12).The top five best fits of the model to field data were determined by calculating the sum of squared variable costs using the modCost function from the R package FEM (34).Dissolved inorganic carbon was also measured in the fluids to ensure that it was not growth limiting.Thirty-seven diffuse hydrothermal fluid samples were collected from Marker 113 and Marker 33 in 2013, 2014, and 2015 using the deep-sea remotely operated vehicles Jason II and ROPOS (Table S1) as previously described (4,5).The concentration of Methanocaldococcus, Methanothermococcus, and Desulfurobacterium cells in the diffuse hydrothermal fluids was estimated from the product of the proportion of these organisms in annotated metagenome sequence read counts (4) and the total cell counts (5) (Table S1).Functionally annotated reads from metagenomes collected in 2013, 2014, and 2015 from Marker 113 and Marker 33 were taxonomically classified and proportioned relative to the whole metagenome for each sample (4).The proportions of annotated reads for Methanothermococcus, Methanocaldo coccus, and Desulfurobacterium were used to estimate their concentrations relative to the total number of cells in the same samples.
There was no high-temperature hydrothermal venting within 0.5 km of Marker 113 or Marker 33 that could provide hydrothermal endmember H 2 concentrations.There fore, high-temperature endmember H 2 concentrations for these sites were estimated based on a trend of endmember H 2 and endmember Cl − for the closest high-tempera ture vents at Axial Seamount (12).The endmember Cl − concentration of Marker 113 is ~100 mmol/kg and that of Marker 33 is ~400 mmol/kg (from extrapolation of diffuse fluid Cl − to zero Mg 2+ ).Temperature and chlorinity for the high-temperature endmem bers were determined from the relationship of Mg 2+ and temperature (or chlorinity) and extrapolated to zero Mg 2+ concentration.The nearest high-temperature vents with similar Cl − endmembers are Diva vent (endmember Cl − 200 mmol/kg, endmember H 2 400-970 μmol/kg) and El Guapo vent (endmember Cl − 400 mmol/kg and endmember H 2 120-470 μmol/kg).Endmember H 2 concentrations at Marker 113 and Marker 33 were assigned values of 950 and 300 µmol/kg, respectively (12).These endmember H 2 values for the diffuse vent sites reflect the fact that Marker 113 has a vapor-dominated source with higher gas content than the source for Marker 33.

FIG 1
FIG 1 Growth of D. thermolithotrophum and M. thermolithotrophicus at 65°C (A, C) and D. thermolithotrophum and M. jannaschii at 72°C (B, D).The maximum cell concentrations (A, B) and the growth product yields (C, D) are shown.D. thermolithotrophum at high initial H 2 concentration is shown in blue shades, methanogens at high initial H 2 concentration are shown in red shades, and D. thermolithotrophum and methanogens at low initial H 2 concentrations are both shown in gray shades.Monocultures are the darker shades of all colors, while cocultures are the lighter shades.The statistical relevance (P < 0.05) of the data is shown separately for each organism (non-prime versus prime).Abbreviations: MC, monoculture; CC, coculture in 1:1 initial cell ratio; H, high initial H 2 concentration; L, low initial H 2 concentration.

FIG 2
FIG 2 Specific growth rate of D. thermolithotrophum in monoculture at 72°C at high (blue shade) and low (gray shade) initial H 2 concentrations with and without the addition of sodium thiosulfate as a terminal electron acceptor.The statistical relevance of the data is P < 0.05.

FIG 3
FIG 3 Maximum cell concentrations of D. thermolithotrophum and M. thermolithotrophicus at 65°C (A) and D. thermolithotrophum and M. jannaschii at 72°C (B).D. thermolithotrophum at high initial H 2 concentration is shown in blue shades, methanogens at high initial H 2 concentration are shown in red shades, and D. thermolithotrophum and methanogens at low initial H 2 concentrations are both shown in gray shades.Monocultures are the darker shades of all colors, while cocultures are the lighter shades.The statistical relevance (P < 0.05) of the data is shown separately for each organism (non-prime versus prime).The initial cell concentrations and initial ratios of methanogens to D. thermolithotrophum cells is shown along the x-axis for each condition.
).When compared with cell proportions and CH 4 anomalies measured in hydrothermal fluids from Marker 33 and Marker 113 at Axial Seamount (4, 5), the modeling results suggest that the hydrothermal fluid residence times from 85°C to exiting the seafloor at these sites are 63-100 and 40-77 h, respectively (Fig. 4C through F).Marker 33 showed more pipe-like fluid flow, while Marker 113 showed more plume-like fluid flow.

FIG 4
FIG 4 Reactive transport modeling results for (A) pipe-like (high x b ) and (B) plume-like (low x b ) fluid flow.The heat maps (C-F) show the predominant thermophile (M.jannaschii in red, D. thermolithotrophum in yellow, and M. thermolithotrophicus in blue) at varying x b values and subseafloor residence times below 85°C.H 2 concentrations in and temperature of pure hydrothermal vent fluids were 300 µM and 214°C (C, D) and 950 µM and 330°C (E, F) representing Marker 33 and Marker 113, respectively, at Axial Seamount.The initial M. jannaschii:D.thermolithotrophum:M. thermolithotrophicus ratios were either 1:1:1 (C, E) or 1:1:100 (D, F).The conditions that best fit the field results for Marker 33 and Marker 113 are indicated with an F symbol in the heat map.
If x b >> 1, then the flow path resembled a linear crack or straight pipe.If x b << 1, then the