Bioenergetic Constraints on Microbial Hydrogen Utilization in Precambrian Deep Crustal Fracture Fluids

Study site

A suite of five boreholes at the 1067-m level of Thompson Mine, Manitoba, were sampled in June 2006 (between two and 10 months after drilling) to investigate their aqueous and gaseous geochemistries and microbiologies. Thompson is a nickel mine located within 2.1 Ga metasedimentary, metavolcanic and ultramafic rocks that overlie Archean felsic and mafic gneisses, granulites and amphibolites of the Canadian Shield (Brooks and Theyer Citation1981). The boreholes were specifically chosen to span a wide range of conductivities. All the boreholes had gases naturally evolving from the fluids at the time of sampling. Drilling water (derived from surface lakes and circulated down to the mining operation levels) was also sampled.

Field measurements/sampling

Conductivity, pH and temperature were measured in situ with standard field probes. O₂ was measured in situ using a CHEMetrics colorimetric test kit (detection limit 0.1 ppm). All boreholes at the Thompson Mine were close to vertical, and completely full of water at the time of sampling, with water naturally discharging from the boreholes. The gas flow rate was measured in triplicate by temporarily sealing the borehole with a presterilized (autoclaved) rubber stopper with a tube inserted in it and by measuring the time taken for gas to displace a known volume of water. Gas samples were collected in preevacuated 125-ml acid washed borosilicate serum vials containing 50-µl saturated HgCl₂ solution. Water flow rates were measured by removing 2 l of water from the borehole and by measuring the time taken to refill.

Samples for aqueous geochemistry and microbiology were taken via a 50-cm long sterilized Tygon tube attached to a sterile 50-ml plastic syringe. Samples for ion analysis were filtered through 0.45-μm cellulose nitrate filters and were collected in 30-ml High-Density Polyethylene (HDPE) bottles. Water samples for the measurement of stable isotopic composition of water were sampled in 30-ml HDPE bottles with no headspace, and analyzed at the University of Waterloo. Samples for H₂-oxidation rate experiments and Most Probable Number (MPN) analyses were collected in the field in 125-ml acid washed, 6 × MQ rinsed, and baked (450°C, overnight) borosilicate serum vials (Product #223748, Wheaton Industries Inc., NJ, USA) sealed with 14-mm-thick butyl rubber stoppers (Product #2048–117800, Bellco Glass Inc., NJ, USA). Bottles for aerobic incubations initially contained ambient air, while bottles for anaerobic incubations had previously been flushed for 10 min with N₂, and a small amount of freshly synthesized FeS was added to ensure reducing conditions. Water (50 ml) was injected into the bottles through the stoppers using sterile 50-ml syringes leaving 110-ml headspace. The samples were maintained at 4°C during transport, and inoculations occurred within 36 h of sampling.

Laboratory chemistry measurements

Cation analyses were carried out at the University of Waterloo by coupled plasma atomic emission spectroscopy. Precision was ±5%. Reduced iron (Fe²⁺) was analyzed by the Ferrozine method (Stookey Citation1970), with a precision of 8%. Anions were analyzed by Dionex Ion Chromatography (IC). Acetate was analyzed by IC in a separate run after removing Cl⁻ with Alltech Maxi-Clean™ IC-Ag cartridges (Product #30258, Alltech, Kentucky, USA), using an AS11HC column and a gradient eluent program ramping from 50 mM NaOH to 1 mM NaOH. Dissolved Inorganic Carbon (DIC) was performed by the acidification of samples to ≤pH 2.5 and analysis of released CO₂ by infrared spectroscopy at the University of Arizona.

Compositional analyses of gas samples were performed at the Stable Isotope Laboratory at the University of Toronto. Concentrations of CH₄, C₂ H₆, C₃H₈, i-C₄H₁₀ and n-C₄H₁₀ were quantified on a Varian 3400 GC equipped with a flame ionization detector. The hydrocarbons were separated on a J&W Scientific GS-Q column (60 m × 0.32 mm ID) with a helium gas flow and the following temperature program: 32°C for 6 min, rising to 220°C at 20°C min⁻¹. Concentrations of H₂, He, O₂ and N₂ were analyzed on a Varian 3800 GC equipped with a micro-thermal conductivity detector. Gases were separated using a Varian Molecular Sieve 5A PLOT column (25 m × 0.53 mm ID) with an Ar gas flow and the following temperature program: 35°C for 6 min, then the temperature was increased to 220°C at 20°C min⁻¹. Reproducibility for triplicate analyses was better than ±5%.

δ² H-H₂ analyses were performed at the University of Toronto using a Finnigan MAT Delta+XL isotope ratio mass spectrometer interfaced with an HP 6890 GC and a micropyrolysis furnace. Gases were separated using a 60-m J&W Scientific GS-Q column (60 m × 0.32 mm ID) with the following temperature program: 35°C increasing to 120°C at 5°C min⁻¹, rising to 220°C at 10°C min⁻¹, holding at 220°C for 10 min. Total error incorporating both accuracy and reproducibility was ±5‰ with respect to V-SMOW (after Ward et al. (Citation2000)). δ¹⁸O and δ²H of H₂O were analyzed at the University of Waterloo using offline preparatory methods and gas source stable isotope mass spectrometry. Precisions were 0.15 and 2.0‰ for δ¹⁸O and δ²H, respectively.

Laboratory microbiological measurements

Thermodynamic modeling

Estimating the impact of osmolyte synthesis on the % energy available for growth

The estimated percentage of ΔG^r required to combat osmosis during new cell growth in the fluids was calculated as follows. The concentration of osmolytes (M_o, with units of mmol cm⁻³) required within microbial cells at varying ionic strengths were calculated using Equation [2]. This equation is based on the regression equation of intracellular osmolyte concentrations measured in halophilic growth cultures at a range of ionic strengths (Brown Citation1990; Oren Citation1999) (Figure 1)[2]

Figure 1. Intracellular osmolyte concentrations of Dunaliella salina grown at different ionic strengths (data from experiments of Brown (Citation1990)).

Display full size

Next, the ATP required to synthesize M_O in 1 cm³ of cells (ATP_M) was calculated from Equation Equation[3][3] :[3] where a is the ATP_M molar equivalent required to autotrophically synthesize compatible solutes within a microbial cell (using a = 30, 54, 58, 54, 85 and 109 mmol of ATP mmol⁻¹ for glycerol, glycine betaine, ectoine, glucosyl glycerol, sucrose and trehalose, respectively, and a = 0.59 mmol of ATP mmol⁻¹ for the uptake of KCl by ion transporters, based on the theoretical calculations of Oren (Citation1999) and Stouthamer (Citation1973)). The relative percentage of ATP used for osmolyte synthesis or KCl uptake (% osmolyte) relative to normal cellular growth under nonsaline conditions (ATP_cell) was then calculated for the range of Thompson fluid ionic strengths using Equation Equation[4][4] :[4] where ATP_cell is assumed to be 30 mmol ATP cm⁻³ (Oren, Citation1999; Stouthamer, Citation1973).

Aqueous geochemistry

The temperature of the borehole fluids ranged from 21.9 to 22.7°C (Table 3). The conductivities of the borehole fluids ranged from 38.0 to 120 mS cm⁻¹, while the conductivity of drilling water was lower at 0.5 mS cm⁻¹ (Table 3). The calculated ionic strengths of the borehole fluids ranged from 0.6 to 6.4 M. The aqueous geochemistry of the borehole fluids was dominated by Na⁺ (up to 1370 mM), Cl⁻ (up to 3510 mM) and Ca²⁺ (up to 1090 mM) (Table 3). Nitrate was detected in one of the five borehole fluids (503 μM). Sulfate was detected in two of the five borehole fluids (105 and 409 μM). Dissolved O₂ was at or close to the detection limit in all the borehole fluids, but higher in the drilling water (0.16 mM) (Table 3). Total inorganic carbon (DIC) in the drilling water and two freshest boreholes ranged from 125 to 151 mM, with lower values (28–39 mM) in the three most saline boreholes. Acetate concentrations ranged from 53 to 496 μM in the borehole fluids, and 5 μM in the drilling water (Table 3).

Gas and water flow rates, gas compositions and gas and water isotopic values

Gas flow rates in the Thompson Mine boreholes ranged from 0.2 to 45 ml min⁻¹ (Table 4). Water flow emanating from the highest ionic strength borehole (6.4 M) was between 0.17 and 2.0 ml min⁻¹. Water flow in the remainder of boreholes was not measured. The gas to water ratio in the 6.4-M borehole was between 0.15 and 1.8. δ¹⁸O-H₂O and δ²H-H₂O for the drilling water were −12.9 and −112‰, respectively, falling slightly below the meteoric water line (Figure 2), consistent with their source in a local lake. All borehole fluids were more depleted in δ¹⁸O-H₂O and more enriched in δ²H-H₂O than drilling water (Table 5), with overall trends of δ¹⁸O depletion and δ²H enrichment with increasing salinity, and values elevated above the meteoric water line (Figure 2).

Figure 2. The stable isotopic composition (δ¹⁸O-H₂O and δ²H-H₂O) of borehole fluids and drilling water at the Thompson Mine from the current study, compared to prior data from the same mine. The dotted line shows the meteoric water line (Craig Citation1961). Solid circles depict borehole fluids and drilling water from this study. Remaining open symbols depict Thompson Mine borehole fluids described in the work by Frape et al. (Citation1984). Open circles = fresh, open inverted triangles = brackish, open triangles = saline, and open squares = brines.

Display full size

The composition of the gases in borehole fluids was dominated by N₂ (49–62% v/v) and CH₄ (35–46% v/v). There were detectable concentrations of C2+ alkanes (C₂H₆, C₃H₈, i-C₄H₁₀ and n-C₄H₁₀), with a logarithmic decrease in concentration from CH₄ to C₄H₁₀ (Table 4). H₂ was detected in all borehole fluids (0.03–2.7% v/v). Helium was detected in all boreholes, with concentrations ranging from 2.4 to 3.3% v/v (Table 4). H₂ concentrations generally increased with Cl⁻ concentration, with a polynomial best fit rather than linear relationship (R² = 0.99; Figure 3). δ² H-H₂ of Thompson borehole fluids was within a narrow range of −753 to −762‰ (Table 5) consistent with values typically observed in crystalline brines in Precambrian rocks (Sherwood Lollar et al. Citation2007).

Figure 3. Chloride versus H₂ gas for the suite of borehole fluids sampled at the Thompson Mine. The nonlinear relationship suggests that H₂ concentrations are not controlled purely by the mixing of a H₂-poor freshwater end member with a H₂-rich saline end member, but that there may be microbial utilization of H₂ in fresher waters.

Display full size

Microbial H₂-oxidation microcosm experiments

Microcosm experiments demonstrated H₂ consumption in the two freshest borehole fluids (0.6 and 1.9 M) (Figure 4). Maximum H₂ consumption rates (the steepest gradient between two time points) in the freshest borehole were 0.66 μmol H₂ l⁻¹ day⁻¹ under aerobic conditions and 0.45 μmol H₂ l⁻¹ day⁻¹ under anaerobic conditions. Maximum H₂ consumption rates in the second freshest borehole were 0.24 μmol H₂ l⁻¹ day⁻¹ under aerobic conditions and 0.12 μmol H₂ l⁻¹ day⁻¹ under anaerobic conditions (Figure 5). Maximum rates of H₂ consumption in the experiments conducted with the drilling fluid were 0.23 μmol H₂ l⁻¹ day⁻¹ and 0.29 μmol H₂ l⁻¹ day⁻¹ under aerobic and anaerobic conditions, respectively (Figure 5). There was no detectable H₂ consumption either in the three most saline boreholes (2.6, 2.8 and 6.4 M) (Figure 4) or in killed controls (not shown) in incubations that lasted 78 days.

Figure 4. Aerobic and anaerobic H₂ consumption microcosm experiments from Thompson Mine fracture fluids. There was a significant H₂ consumption in the drilling water and two lowest ionic strength borehole fluids (0.6 and 1.9 M) under both aerobic (a) and (b) anaerobic conditions. There was no detectable H₂ consumption in killed (autoclaved) controls (not shown). Solid square = 6.4 M, solid circle = 2.8 M, solid inverted triangles = 2.6 M, open triangle = 1.9 M, open diamonds = 0.6 M, and open circles = 0.002 M (drilling water).

Display full size

Figure 5. Evidence of microbial utilization of H₂ in Thompson Mine borehole fluids and drilling water. (a) Numbers of culturable H₂-utilizing microorganisms, based on Most Probable Number analysis. Solid circles = aerobic oxidizers, open hexagons = Fe(III)-reducers, open circle = SRBs, open squares = methanogens, open triangles = putative acetogens (acetogens are classed as presumptive, as confirmatory acetate production was not measured). (b) Maximum rates of microbial H₂ utilization based on H₂ oxidation experiments. Solid circle = aerobic H₂ oxidation, open circle = anaerobic H₂ oxidation.

Display full size

H₂-utilizing MPN counts

Cultivable H₂-utilizing microorganisms were detected in the 0.6-M borehole fluid, the 1.9-M borehole fluid and drilling water, but not in the 2.5-M, 2.7-M or 6.4-M borehole fluids (Table 6 and Figure 5). The largest mean estimated cultivable cell concentrations were in the 0.6-M borehole fluid, with an estimated ∼50,000 cells ml⁻¹ of SRBs and putative acetogens, nearly 9800 cells ml⁻¹ of aerobic H₂ oxidizers, and smaller numbers (6 cells ml⁻¹) of Fe(III) reducers (Table 6). All these functional microbial groups were present in lower numbers in the 1.9-M borehole fracture fluid and drilling water (Table 6).

Cultivable aerobic H₂ oxidizers and Fe(III) reducers were only detected in brackish media (0.06 M in Table 6). SRBs were documented in both brackish and saline media (0.06 and 0.8 M in Table 5, respectively) in the 0.6-M fluid, although with higher cell numbers in brackish media. Putative acetogens grew in brackish media in the drilling fluid and 0.6-M borehole, whereas low (6.5 cells ml⁻¹) but equal numbers of putative acetogens grew in media of all three salinities in the 1.9-M borehole (Table 6). Methanogens were only detected in one of the borehole fluids (1.9 M), and this was in supersaline media only (2.5 M in Table 6).

The only cultivable counts present in borehole fracture fluids but not in drilling fluids were saline SRBs, saline and supersaline putative acetogens, and supersaline methanogens (Table 6). Note that the ionic strength of the MPN media (2.5 M) was less than half of the ionic strength of the most saline borehole fluid (6.4 M), but within 11 and 4% of the ionic strengths of the next two most saline borehole fluids (Table 6).

Thermodynamic modeling results and comparison to viable numbers of microbes

One atmosphere pressure scenario

Theoretical mixing of the Thompson saline end member (6.4 M ionic strength) with the freshwater end member (drilling water) indicated that there was sufficient energy (ΔG^r) for H₂-based catabolic reactions (>20 kJ mol⁻¹) at all ionic strengths (Table 7) using the stoichiometry in Table 1. 20 kJ mol⁻¹ is a commonly cited minimum energy threshold required to drive ATP synthesis (Scholten and Conrad Citation2000). Aerobic H₂ oxidation gave the highest free energy yield (up to 972 ± 11 kJ mol⁻¹ H₂) (Table 7). There was a gradual decrease in the free energy for aerobic H₂ oxidation with increasing ionic strength (Table 7 and Figure 6a), and opposing trends of gradually increasing free energy with greater ionic strength for various anaerobic H₂-utilization reactions (Table 7 and Figure 6b).

Table 7. ΔG^r, Gibbs Free Energy (kJ mol⁻¹) of H₂-based catabolic reactions based on the theoretical mixing of drilling fluid with the highest ionic strength borehole fluid. First number shows mean ΔG^r assuming 1 atm pressure; number in brackets assumes all gases are in dissolved phase in pressurized fractures.

Download CSVDisplay Table

Figure 6. Calculated Gibbs free energies of H₂-based catabolic reaction (G^r), based on the theoretical mixing of the most saline Thompson borehole fluid (6.4 M) with freshwater drilling fluid. We assumed that the oxygen was derived from the palaeo-meteoric end member, and that H₂ was derived from the fracture fluids. (a) G^r, all reactions. (b) G^r, showing anaerobic reactions only. Closed circles = aerobic oxidation, open circles = sulfate reduction, open squares = methanogenesis (CO₂), closed squares = methanogenesis (HCO₃₋), open triangles = acetogenesis (CO₂), closed triangles = acetogenesis (HCO₃₋).

Display full size

High-pressure scenario (all gases dissolved)

Under the scenario that all gases were dissolved, approximating high-pressure conditions within closed fracture systems (Sherwood Lollar et al. Citation1993), the free energy available for all H₂-based redox couples increased (Table 7). The increase in ΔG^r ranged from a minimum of −23.5 kJ mol⁻¹ to a maximum of −50.2 kJ mol⁻¹ depending on redox couple and ionic strength (Table 7).

When H₂ stoichiometries of the reactions in Table 2 are normalized to unity to better compare relative reaction yields, aerobic oxidation showed the greatest ΔG^r (Figure 6a). Sulfate reduction showed the highest ΔG^r energy yields of the anaerobic H₂ oxidizing reactions (Figure 6b), followed by methanogenesis (Figure 6b). There was no apparent similarity between ΔG^r and measured numbers of cultivable H₂-utilizing microorganisms or rates of microbial H₂ oxidation (Figures 5 and 6).

The total energy available for microbial catabolic reactions of H₂ with O₂, SO₄²⁻, HCO₃⁻ and CO₂ with theoretical mixing of the end-member brine with the freshwater end member is illustrated in Figure 7. The largest total free energy yield was due to aerobic H₂ oxidation, with the largest potential energy yields at ∼1-M ionic strength (∼100 J L⁻¹) (Figure 7). Smaller total energies were available from methanogenesis, acetogenesis and sulfate reduction, all with maximum yields at ≤2-M ionic strength (Figure 7).

Figure 7. Total energy available from the mixing of the highest ionic strength Thompson Mine borehole fluid (6.4 M) with freshwater drilling water. See the main text for detailed calculations. Closed circles = aerobic oxidation, open circles = sulfate reduction, open squares = methanogenesis (CO₂), closed squares = methanogenesis (HCO₃₋), open triangles = acetogenesis (CO₂), closed triangles = acetogenesis (HCO₃₋).

Display full size

Relative energetic cost of osmolyte synthesis relative to cell building during growth

The proportion of energy required for potentially combating osmosis via the uptake of KCl did not exceed that required for normal cell growth over the range of borehole ionic strengths (0–6.4 M), with a maximum of 30% at an ionic strength of 6.4 M (Figure 8). The energy costs for potentially combating osmosis through the production of organic osmolytes within cells were far higher than those for KCl (Figure 8). At an ionic strength of 2.5 M, organic osmolyte synthesis required an estimated 81–96% of the combined energetic cost of cell building and osmotic regulation (Figure 8). At an ionic strength of 6.4 M, these percentages rose from 91 to 98% of the combined energetic cost (Figure 8).

Figure 8. Modeling the impact of organic osmolyte synthesis and KCl uptake on the % anabolic energy available for cell growth over the range of ionic strengths of the Thompson Mine borehole fluids. See main text for calculations. Solid circles = KCl, open circles = glycerol, inverted triangles = glycine betaine, open triangle = ectoine, solid square = glucosylglycerol, open square = sucrose, closed diamond = trehalose.

Display full size

H₂ from ancient Precambrian fluids supports the growth of microorganisms

The trend of δ¹⁸O-H₂O and δ² H-H₂O values above the meteoric water line (Figure 2) suggests mixing between a meteoric-water-derived end member and a saline end member whose δ¹⁸O and δ²H values have been modified by water–rock interactions over geological time (Frape et al. Citation1984; Ward et al. Citation2004). The drilling water falls to the right of the meteoric water line, consistent with isotopic enrichment by evaporation in the surface lake from which the drilling water is derived. δ²H-H₂ of Thompson borehole fluids was within a narrow range of −753 to −762‰, within the range of continental abiogenic H₂ described from a variety of Precambrian Shield terrains (Sherwood Lollar et al. Citation2007). These values are typical of abiogenic H₂ in ancient Precambrian Shield fracture fluids that have isotopically re-equilibrated with water over 10³–10⁵ years or more (Sherwood Lollar et al. Citation2007).

The nonlinear relationship between the concentration of H₂ and Cl⁻ in the suite of borehole fluids from the Thompson Mine (Figure 3) cannot be explained by conservative mixing of a H₂-poor meteoric end member and a H₂-rich saline end member. The simplest explanation for the apparent preferential loss of H₂ in lower ionic strength borehole fluids (≤1.9 M) is microbial utilization of H₂ as proposed by Sherwood Lollar et al. (Citation2006, Citation2007). This hypothesis is consistent with measured rates of potential microbial H₂ consumption with the utilization of H₂ under both aerobic and anaerobic conditions detected only in the two lowest ionic strength borehole fluids (0.6 and 1.9 M; Figure 4). Preferential microbial H₂ utilization in lower ionic strength borehole fluids was also consistent with the detection of cultivable H₂-utilizing microbes in the 0.6-M and 1.9-M borehole fluids, with none detected at ionic strengths ≥2.5 M (Figure 5).

“Palaeopickling” of hydrogenotrophic deep crustal ecosystems?

End-member brines in isolated fractures can mix with meteoric-derived water through mining activity or natural tectonic fracturing, providing potential blooms of subsurface microbial growth (Sherwood Lollar et al. Citation2007; Sleep and Zoback Citation2007). Our results show that when the resulting mixed fluids have an ionic strength ≤1.9 M, the utilization of abiogenic H₂ with available electron acceptors can potentially be rapid under both aerobic and anaerobic conditions, resulting in very brief (weeks and months) periods of subsurface growth (Figure 4), followed by resource depletion. Following these subsurface blooms, much lower rates of microbial growth (below the detection limits of our methods) could potentially be sustained by the continual production of H₂ via radiolysis or low-temperature serpentinization (Lin et al. Citation2005, Citation2006a; Sherwood Lollar et al. Citation2014).

In borehole fluids >1.9 M, the numbers of cultivable H₂-utilizing microorganisms were below the detection limit (Table 6). This by itself cannot be taken as conclusive evidence for the lack of hydrogenotrophs, due to the well-documented difficulties of growing more than a small subsection of environmental microorganisms in the laboratory (Amann et al. Citation1995). For example, we cannot rule out the possibility that microbial growth in the MPN vials was limited by some essential trace nutrient. It is further possible that some hydrogenotrophic microorganisms at ionic strengths >1.9 M were present and were potentially cultivable, but due to slow growth rates, they would take longer periods of incubation to detect. Further studies could complement our culture-based approach with genetic culture-independent methods targeting both molecular phylogeny and functional genes within the borehole fluids, including functional genes for CO₂ reduction and hydrogenase activity, alongside more general measurements of microbial presence and activity including total cell counts, ATP concentration (Karl and Holm-Hansen Citation1978) and hydrogenase activity assays (Soffientino et al. Citation2006). In addition, more sensitive culture-based approaches such as concentrating microbial cells from borehole fluids prior to inoculation could be used.

The inability to culture hydrogenotrophs in higher ionic strength fluids was, however, consistent with the results of the H₂-amended microcosm experiments. The lack of any detectable microbial H₂ utilization in H₂-oxidation microcosm experiments after 78 days (Figure 4) indicates that rates of microbial H₂ consumption in borehole fluids >1.9 M were, if present, substantially slower than those at weaker ionic strengths. This is seemingly at odds with the potential ΔG^r yields, which predicted greater energy availability for anaerobic hydrogen utilizing reactions at higher ionic strengths (Figure 7). While we cannot rule out that in situ hydrogenotrophic growth was inhibited by some essential trace nutrient or cofactor, our thermodynamic calculations (Figure 7) suggest that bioenergetic factors alone can provide a potential explanation for this apparent paradox (Head et al. Citation2014; Oren Citation1999). Recently, it has been suggested that the high bioenergetic cost of forming osmolytes within microbial cells may retard the biodegradation of low-temperature subsurface petroleum reservoirs, in a process termed “palaeopickling” (Head et al. Citation2014). Our theoretical calculations for the Thompson Mine fracture fluids suggest that “pickling” may also be an important factor in preventing the consumption of abiogenic H₂ gas in Precambrian Shield fractures. At ionic strengths >1.9 M, the energetic cost of synthesizing organic osmolytes (Figure 7) may become so high that microbial cell division in hydrogenotrophic microorganisms dramatically slows or stops (Figure 5 and Table 6). Few microbial groups are able to use the energetically much more favorable KCl as an osmolyte as it requires the adaption of all intracellular processes to high salt concentrations (Oren Citation1999).

The “pickling” of hydrogenotrophic microorganisms within some Precambrian Shield rocks may have implications for evaluating this type of geologic setting for proposed nuclear waste repository sites. Deep (500–1000 m) shafts driven into stable plutons within Canadian Shield rocks were once proposed as sites for the long-term disposal of nuclear fuel waste (Stroes-Gascoyne and West Citation1997). The excavation of such repositories could potentially mix ancient saline fluids with modern meteoric water, as documented in our Thompson Mine study. Within such settings, the radiolytic splitting of water by gamma radiation emitted from the high-level radioactive waste can provide additional sources of both H₂ (Gales et al. Citation2004) and oxidants (Lin et al. Citation1996a; Li et al. Citation2016). Additional H₂ could be provided by the corrosion of metals used in some waste containers (Libert et al. Citation2011). In the short term (years to hundreds of years) after radioactive waste burial, high levels of gamma radiation and increased temperatures could inhibit microbial growth in zones up to 20 cm adjacent to the waste containers (Stroes-Gascoyne and West Citation1997). Over time (hundreds to many thousands of years), microbial colonization may become possible, depending on a range of additional factors including available space for colonization and growth, hydrogeological connectivity, and nutrient supply (Stroes-Gascoyne and West Citation1997; Sherwood Lollar, Citation2011). Our Thompson Mine data suggest that if the ionic strength of fluids in fracture systems adjacent to radioactive waste is <1.9 M and radiation levels are sufficiently noninhibitory, the hydrogenotrophic microbiological activity in the vicinity of stored nuclear waste may be promoted by H₂ and oxidants produced by continued radiolysis. Over time, it has been suggested that the ongoing hydrogenotrophic activity could affect pH and Eh around waste containers, potentially altering the speciation of any radionuclides released from waste into adjacent fracture systems (Libert et al. Citation2011; Sherwood Lollar, Citation2011). One of the principal groups of microorganisms enumerated in our study using a culture-based approach were sulfate-reducing bacteria, with up to an estimated 49,900 cells ml⁻¹ grown in 0.6-M media, along with smaller populations of cultivable Fe(III)-reducers (Table 6). Importantly, both these physiological groups can potentially accelerate the corrosion of various types of metal containers. The stimulation of sulfate-reducing bacteria by H₂ in nuclear waste repository settings in crystalline rock has been identified as something that could potentially enhance rates of biocorrosion of many types of metal containers via the production of corrosive H₂S (El Hajj et al. Citation2010). Fe(III)-reducing bacteria can potentially increase corrosion via the reduction and alteration of passive surface films containing ferric iron (Libert et al. Citation2011; Potekhina et al. Citation1999). One way of potentially limiting the growth of these and other hydrogenotrophic microorganisms around radioactive waste sites might be to ensure that water in fracture systems immediately adjacent to future radioactive waste depositories is highly saline (>1.9 M). For example, the use of high ionic strength drilling fluid could potentially help to ensure that no pockets of fresher more habitable fluids are in close proximity to stored waste (Slater et al. Citation2013). However, most waste repositories rely on the use of bentonite clays as a barrier, and their swelling capacity, and hence effectiveness as a barrier decreases with increasing ionic strength (Studds et al. Citation1998). The use of high ionic strength fluids is therefore not a viable option to limit rates of microbial corrosion.

Another key driver in terrestrial deep crustal ecosystem research is to gain insight into potential analogous crustal ecosystems in the subsurface of other planets such as Mars, and to aid with future life detection planetary missions (McMahon et al. Citation2016; Michalski et al. Citation2013). It has been speculated that isolated crustal ecosystems could be present in the subsurface of Mars, isolated for the billions of years since Mars' surface became inhospitable to life. Subsurface crustal fluids on Mars are likely high-ionic-strength Ca-rich brines (Burt and Knauth Citation2003), comparable to those of our saline end-member Thompson Mine fracture fluid and similar ancient fluids isolated within Precambrian Shield terranes (Holland et al. Citation2013). Our study at the Thompson Mine suggests that hydrogenotrophic microorganisms are unlikely to be present in similar highly saline brines in the subsurface of Mars unless they base their osmotic regulation around the accumulation of inorganic (e.g., K⁺) rather than organic osmolytes (Figure 7). This supports a recent assertion that ionic strength is an important deciding factor in the habitability of Martian brines (Fox-Powell et al. Citation2016).

Further research into the bioenergetics of osmolyte regulation in halophiles in Precambrian Shield fracture systems is required to better constrain the habitability zones of terrestrial and other planetary body crustal environments. Future studies could expand on our study by extending the focus to the ionic strength constraints on heterotrophic microorganisms within crystalline Precambrian Shield rocks, as well as using a greater variety of more sensitive methods such as culture and culture-independent methods to target hydrogenotrophic activity in fracture fluids with ionic strengths >1.9 M.