Over the last several decades, investigations of Earth's subsurface and other extremely low-biomass systems have refined our understanding of the environmental limits of life, driven by methodological advances that permit agnostic life detection of biology and their respective physical biosignatures and chemical biomarkers. These advances enable mission concepts centered on microbiological processes that facilitate identification of both active life and preserved biosignatures through measurements of metabolism and associated biochemical markers that, on Mars, are more likely to be retained below the surface. Terrestrially, although biological processes can exert a significant influence on Earth's crust, the presence of habitable conditions does not necessarily imply the existence of cellular life. The Viking missions constituted the first direct life-detection experiments on Mars but produced equivocal outcomes, prompting subsequent exploration strategies to emphasize surface habitability rather than direct biological testing. Leveraging progress in subsurface microbiology and planetary exploration, we contend that Mars missions are now poised to shift toward direct tests for extant microbial activity in the subsurface, with metabolic processes serving as a broadly applicable indicator of life.

Following the Viking experiments in 1976, many of the original inferences of biological metabolism have been replicated via abiotic mechanisms we now know are plausible on Mars’ surface. While in many cases subsequent experiments have cast doubt on whether Viking truly detected life, numerous other studies since Viking have greatly expanded our knowledge of life’s limits and microbial metabolism. In particular, increased characterization of Earth’s subsurface has revealed the astounding complexity and adaptability of life, highlighting chemically based metabolisms as potentially strong targets for future life detection missions. Over the same time frame, we have gained knowledge of putatively more habitable regions in Mars’ subsurface, relative to the original Viking lander surface sites, that could host similar organisms. In this review, we discuss the wealth of knowledge concerning the habitability of zones across Mars’ surface/subsurface, and we suggest specific microbial metabolisms that should be targeted in future life detection missions based on laboratory and field studies under analogous conditions on Earth and with consideration of recommendations from the larger Astrobiology community. The ability to leverage these advancements in subsurface research toward the incorporation of increased specificity in future life detection efforts is additionally discussed in the context of current Mars subsurface mission progress and planetary protection and defense concerns.

Serpentinization produces hydrogen and methane through abiotic water-rock interactions, potentially supporting chemotrophic life in planetary subsurface environments. Serpentine deposits in the Martian Noachian landscapes of Nili Fossae and the Southern Highlands have been considered as potential paleo-habitable zones. However, the geochemical and physical conditions of Martian serpentinization fluids are poorly constrained because of limited data on serpentinite composition and formation environment. Furthermore, the co-occurrence of magnesite and magnesium-enriched serpentines on Mars remains enigmatic. To address such gaps, we investigated antigorite-magnesite paleo-serpentine bodies along the Highland-Vijayan suture of Sri Lanka as Martian analog sites, using thermodynamic batch reaction models to constrain alteration fluids and the production of primary (H2) and secondary (CH4) serpentinization products. Geochemist's Workbench models combined field X-Ray Fluorescence (XRF) observations with varied protolith compositions (ultramafic or gneissic), precursor fluids (seawater or freshwater), temperatures (100-600 degrees C), volumetric water-to-rock ratios (1-100,000), and CO2 partial pressures (0.01-1 bar). Models successfully reproduced the co-occurrence of antigorite and magnesite observed on Mars, with both minerals forming at 100 degrees C across water-to-rock ratios. Despite their Mg-enriched composition, ultramafic protoliths produced H2 yields (up to 229 mmol/kg at W/R 1 at 100 degrees C), supporting chemotrophic populations up to 107.5 cells/mL. Geochemical models indicate Mg enrichment from ultramafic mineralogy and Fe contribution from regional gneisses. Our thermodynamic equilibrium results show that Mg-rich serpentine systems with sufficient ferrous iron can produce biologically significant H2, establishing Sri Lankan serpentinites as valuable analogs for Noachian Mars habitability.

Despite their ubiquity, viruses remain understudied in many of Earth’s environments. The deep subsurface contains a large portion of Earth’s biomass and can be expected to hold a rich store of viruses as well. We investigated the viral community in a 1.2-kilometer-deep dolomitic aquifer accessed in the Moab Khotsong gold mine in South Africa by microscopically quantifying virus-like particles (VLP) and cells and conducting a metagenomic survey of viruses and their associations with the microbial community. VLP:cell ratios were consistently lower (∼1:1) than those found in most shallower aquifers. Viral sequences were recovered from both the <0.2 µm fraction, presumably representing free virions, and the >0.2 µm fraction that also included microbial cells. Viruses were diverse and 78% were novel, and the viral communities were relatively unchanged at four sampling dates spanning five years. Host prediction indicated viral infection of the dominant microbes, including Thiomicrospiraceae and Rhodocyclaceae , with temperate phages. Viral genomes include auxiliary metabolic genes (AMGs), with most of them being Type II, which favor host persistence. Taken together, VLP:cell ratios, type II AMGs, and temporally stable viromes point to dominance of lysogeny and piggyback-the-persistent viral-host interactions in the deep aquifer. The many novel viruses observed here suggest that the terrestrial subsurface is an untapped reservoir of viral diversity. ### Competing Interest Statement The authors have declared no competing interest. U.S. National Science Foundation, https://ror.org/021nxhr62, EAR2026858

The ICDP DSeis project accomplished full-core drilling and borehole-logging of the seismogenic zone of the 2014 M5.5 Orkney earthquake, South Africa. Three NQ-holes (total 1.6 km in length), drilled from 2.9 km depth at the Moab Khotsong gold mine, penetrated mostly intact hard rock, including 2.9 Ga meta-sedimentary and altered andesite (Crown) formations dipping ~20°SE. Subparallel altered gabbroic diorite sills intrude the formations. After the borehole penetrated the Crown Formation and approached the steeply-dipping planar cluster of earthquake aftershocks, it encountered a potassic lamprophyre dyke several meters thick. The lamprophyre was intact close to the dyke contact, with mineral assemblages of augite, actinolite, and biotite. The talc and calcite content and magnetic susceptibility increased towards the centre of the dyke, while the augite and actinolite content decreased. The lamprophyre rock mass then became brecciated, with a substantial fraction of gouge. The Crown Formation adjacent to the dyke contact was also brecciated. Friction tests made on lamprophyre gouge (which contains about 20 wt% talc) yielded very low friction coefficients, similar to the results of previous wet friction experiments (Yabe et al. EGU 2024). The DSeis drilling also intersected a non-potassic dyke rich in actinolite and chamosite about 300 m east of the potassic lamprophyre dyke. Whilst this dyke hosted no indications for aftershocks, the extracted brine was more hypersaline and older than any brine previously sampled from deep South African gold mines (Nisson et al., 2023). The hypersaline brine was non-meteoric in composition, with dissolved organic carbon concentrations sufficient to support deep life. Both dykes show significant spatial variation in composition which we attributed to contamination/assimilation and metamorphism, depending on which formations (~20°SE dip) the dykes cut. We postulate that the localization of aftershocks in ‘streaks’ subparallel to the strata is a result of this compositional heterogeneity. DSeis has successfully penetrated, sampled, and studied the aftershock sequence on the upper edge of the Orkney earthquake rupture. However, important questions regarding the nucleation and rupture of the earthquake that will only be solved by studying the strong motion source of the mainshock. The proposed PROTEA scientific drilling project aims to probe the Orkney earthquake's strong motion sources (the heart). The existing DSeis hole, the new PROTEA hole, and the connecting horizontal tunnels at 2.9 km depth will allow us to deploy a 3D distributed acoustic sensing (DAS) network with a vertical span of several hundreds of meters, and a horizontal span of about 1 km. Using both active and passive seismic sources, we expect to image the 3D structure of the reflectors precisely. Moab Khotsong has offered the team access to borehole cores that have sampled numerous dykes and sills; as well as access to the database of lithology and geological structure mapped on the mining horizons at 2-3 km depth. These data cover a much broader volume than the DSeis and PROTEA projects, and will significantly extend and enhance the interpretation. Acknowledgements: South African gold mines, related firms, SATREPS, Kakenhi (21224012), ICDP, JSPS Core-to-Core Program, MEXT Kochi Core Center, NSF, DFG, NRF, Ritsumeikan and Kyoto Univs.

Long-isolated subsurface brine environments (Ma-Ga residence times) may be habitable if they sustainably provide substrates, e.g. through water-rock reactions, that support microbial catabolic energy yields exceeding maintenance costs. The relative inaccessibility and low biomass of such systems has led to limited understanding of microbial taxonomic distribution, metabolism, and survival under abiotic stress exposure in these extreme environments. In this study, taxonomic and metabolic annotations of 95 single-cell amplified genomes were obtained for one low biomass (103-104 cells/ml), hypersaline (246 g/L), and radiolytically enriched brine obtained from 3.1 km depth in South Africa's Moab Khotsong mine. The majority of single-cell amplified genomes belonged to three halophilic families (Halomondaceae (58%), Microbacteriaceae (24%), and Idiomarinaceae (8%)) and did not overlap with any family-level identifications from service water or a less saline dolomite aquifer sampled in the same mine. Functional annotation revealed complete metabolic modules for aerobic heterotrophy (organic acids and xenobiotic oxidation), fermentation, denitrification, and thiosulfate oxidation, suggesting metabolic support in a microoxic environment. Single-cell amplified genomes also contained complete modules for degradation of complex organics, amino acid and nucleotide synthesis, and motility. This work highlights a long-isolated subsurface fluid system with microbial metabolism fueled by radiolytically generated substrates, including O2, and suggests subsurface brines with high radionuclide concentrations as putatively habitable and redox-sustainable environments over long (ka-Ga) timescales.

Investigations of abiotic and biotic contributions to dissolved organic carbon (DOC) are required to constrain microbial habitability in continental subsurface fluids. Here we investigate a large (101–283 mg C/L) DOC pool in an ancient (>1Ga), high temperature (45–55 °C), low biomass (10 2 −10 4 cells/mL), and deep (3.2 km) brine from an uranium-enriched South African gold mine. Excitation-emission matrices (EEMs), negative electrospray ionization (–ESI) 21 tesla Fourier-transform ion cyclotron resonance mass spectrometry (FT-ICR MS), and amino acid analyses suggest the brine DOC is primarily radiolytically oxidized kerogen-rich shales or reefs, methane and ethane, with trace amounts of C 3 –C 6 hydrocarbons and organic sulfides. δ 2 H and δ 13 C of C 1 –C 3 hydrocarbons are consistent with abiotic origins. These findings suggest water-rock processes control redox and C cycling, helping support a meagre, slow biosphere over geologic time. A radiolytic-driven, habitable brine may signal similar settings are good targets in the search for life beyond Earth.

Geochemical and isotopic fluid signatures from a 2.9-3.2 km deep, 45-55 & DEG;C temperature, hypersaline brine from Moab Khotsong gold and uranium mine in the Witwatersrand Basin of South Africa were combined with radiolytic and water-rock isotopic exchange models to delineate brine evolution over geologic time, and to explore brine conditions for habitability. The Moab Khotsong brines were hypersaline (CaNa-Cl) with 215-246 g/L TDS, and Cl- concentrations up to 4 mol/L suggesting their position as a hypersaline end-member significantly more saline than any previously sampled Witwatersrand Basin fluids. The brines revealed low DIC (-0.266-�1.07 mmol/L) with high (-8.49-�23.6 mmol/L) DOC pools, and several reduced gaseous species (up to 46 % by volume H2) despite microoxic conditions (Eh = 135-16 1 mV). Alpha particle radiolysis of water to H2, H2O2, and O2 along with anhydrous-silicate-to-clay alteration reactions predicted 4 mol/L Cl- brine concentration and deuterium enrichment in the fracture waters over a period > 1.00 Ga, consistent with previously reported 40Ar noble gas-derived residence times of 1.20 Ga for this system. In addition, radiolytic production of 7-26 nmol/(L x yr) H2, 3-11 nmo l/(L x yr) O2, and 1-8 nmol/(L x yr) H2O2 was predicted for 1-100l g/g 238U dosage scenarios, supporting radiolysis as a significant source of H2 and oxidant species to deep brines over time that are available to a low biomass system (102-103 cells/mL). The host rock lithology was predominately Archaean quartzite, with minerals exposed on fracture surfaces that included calcite, pyrite, and chlorite. Signatures of 618Ocalcite, 613Ccalcite, D33Spyrite, 634Spyrite and 87Sr/86Sr obtained from secondary ion mass spectrometry (SIMS) microanalyses suggest several discrete fluid events as the basin cooled from peak greenschist conditions to equilibrium with present-day brine temperatures. The brine physiochemistry, geochemistry, and cellular abundances were significantly different from those of a younger, shallower, low salinity dolomitic fluid in the same mine, and both were different from the mine service water. These results indicate the discovery of one of few long-isolated systems that supports subsurface brine formation via extended water-rock interaction, and an example of a subsurface brine system where abiotic geochemistry may support a low biomass microbial community. & COPY; 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Deep within the Precambrian basement rocks of the Earth, groundwaters can sustain subsurface microbial communities, and are targets of investigation both for geologic storage of carbon and/or nuclear waste, and for new reservoirs of rapidly depleting resources of helium. Noble gas-derived residence times have revealed deep hydrological settings where groundwaters are preserved on millions to billion-year timescales. Here we report groundwaters enriched in the highest concentrations of radiogenic products yet discovered in fluids, with an associated 86 Kr excess in the free fluid, and residence times >1 billion years. This brine, from a South African gold mine 3 km below surface, demonstrates that ancient groundwaters preserved in the deep continental crust on billion-year geologic timescales may be more widespread than previously understood. The findings have implications beyond Earth, where on rocky planets such as Mars, subsurface water may persist on long timescales despite surface conditions that no longer provide a habitable zone.

Ca-Na-Cl fluids with high concentrations of dissolved reduced gases reside within fractures in crystalline Precambrian rocks around the world, and have been most intensively studied within South Africa, Fennoscandia and the Canadian Shield. In contrast to surface waters, shallow groundwaters, sedimentary basin brines and metamorphic fluids, the δ18O and δ2H values for these Ca-Na-Cl fluids typically plot to the left/above of the Global Meteoric Water Line (GMWL). To date, most interpretive frameworks for these fracture fluids have focused on their production via water-rock alteration reactions that affect both δ18O and δ2H values, resulting in co-variation of water isotope values above the GMWL. Such alteration processes include silicate hydration coupled with formation of secondary minerals, radiolytic H2 formation, isotopic exchange with a H2-rich gas, or isotope exchange with O and/or H-bearing minerals. This study presents the first compiled global isotopic dataset for these types of fluids, integrating a large amount of unpublished data with the previously published literature in order to investigate these fracture fluid systems on a global scale. Importantly this global perspective allows differentiation between fluids impacted by late-stage mixing with meteoric waters, from fluids that reflect the most saline end-members stored in the host rocks in hydrogeologic isolation from the surface hydrologic cycle. The most saline fluids are shown to occupy a more restricted range of δ18O-δ2H space than previously recognised, with end-member fluids from all of the Precambrian rock settings investigated occupying a range of δ2H δ18O isotope space within which there is no co-variation in these values. These findings suggest a set of common processes may define the isotopic signatures of these most saline end-member fluids in Precambrian settings around the world – creating common signatures identifiable in these fluids, despite differences in geologic setting. This study identifies the important role of oxygen isotopic exchange between primary fluids (associated with hydrothermal/metamorphic activity) and the host rocks, taking place under low temperature, low-volume, water-rock ratios over long (Ma) geologic timescales. This process results in progressive 18O depletion in the fluids over time, while δ2H values remain less affected. For each site the specific isotopic signature of the fracture fluid end-member depends on initial hydrothermal/metamorphic fluid composition, rates of isotopic exchange, water-to-rock ratios, and in-situ residence times. We suggest the often-observed co-variation in both δ18O-δ2H above the GMWL primarily results from late stage mixing of the fracture fluid end-members with (paleo)-meteoric water, resulting in isotopic regression back towards the GMWL, with end-points defined by the local meteoric-climatic conditions for each site.

This paper reports on the outcomes of the ICDP drilling into seismogenic zones of M2.0-5.5 earthquakes in South African (SA) gold mines (DSeis; 2017-2018), the follow-up work in 2019, and planned post-drilling activity from 2020 onwards. In deep SA gold mines, seismogenic zones evolve ahead of thin tabular excavations. Normal faulting prevails because mining enhances the vertical maximum principal stress. At 1km depth at the Cooke 4 mine, we elucidated the evolution of the seismogenic zone with a dense acoustic emission network. In 2017, we successfully recovered both the metasedimentary host rock (mainly quartzite ~2.8 Ga) and samples of the seismogenic zone with well-preserved fracture systems using a triple-tube (BQ 1.5m-long). Subsequent laboratory work investigated critical characteristics of rock-rock friction. In 2014, an M5.5 earthquake, the largest in deep South African gold mining districts, took place. Dense seismic networks, both on the Earth’s surface and at 2-3 km depth, showed that this event was atypical because it was a sinistral event on an unknown geological structure below the mining horizon in West Rand Group strata (~2.9 Ga). Inversion and back-projection of the ground motion showed complicated but unilateral rupture propagation. The densest population of aftershocks shows a sharp upper cut-off and streaks, both dipping to the south. Its centroid lies outside the significant main rupture zone. In 2017, we commenced drilling at a site at 2.9km depth in a tension quadrant of the sinistral faulting, several hundreds of meters above the upper fringe of the M5.5 aftershock plane. During 2017-2018, we drilled holes, of a total length of 1.6 km. With a 1.5m NQ triple-tube for the critical section, we could recover the fault materials and the host rock with the seismic fracture system well preserved. Borehole logging and core curation in SA and laboratory work at international organizations, including Kochi Core Center Japan (KCC), followed during 2017-2019. With the geology data mapped on the mining horizons and the legacy seismic reflection data as additional information, the following picture is emerging: (a) transition of the stress regime from normal-faulting to sinistral-faulting; (b) stress localization; (c) heterogeneity in the aftershock distribution as well as the segregation between the main rupture and aftershocks, potentially correlated with significant heterogeneity in mechanical properties; (d) a role of an altered lamprophyre dike; (e) hypersaline brine with salinity even higher than measurements at other deep gold mines, potentially as old as brine found at Kidd Creek mine, Canada; and (f) abiogenic gas and organic carbon. These data sets allow us to address questions in earthquake and deep-life sciences raised in the ICDP Science Plan (2014-2019). In 2019, the ICDP Executive Committee described DSeis as a ‘successful’ project. To integrate and discuss the outcomes in greater depth and plan additional follow-up work, we are planning a post-drilling workshop in November 2020 or January 2021 at KCC before we return the imported critical section of the core to South Africa.

Many ecosystems are experiencing an increase in drought conditions as a consequence of climate warming and changing precipitation patterns. The stress imposed by these environmental changes can affect ecosystem processes such as the extracellular enzymatic degradation of carbon-containing leaf litter by soil microbial communities. However, the magnitude of these impacts may depend on the composition and metabolism of the microbial community. Based on the hypothesis of local adaptation, microbial communities native to warm-dry ecosystems should display a greater capacity to degrade leaf litter polymers with extracellular enzymes following exposure to warm-dry conditions. To test this hypothesis, we performed a microcosm study in which we monitored extracellular enzyme activity and respiration of microbial communities from five ecosystems along a southern California climate gradient, ranging from warmer, drier desert to wetter, cooler subalpine forest. To simulate drought and rewetting, we subjected microcosms to periods of high temperature and low moisture followed by a water pulse. We found that enzyme activity of wet-cool communities generally exceeded that of warm-dry communities across enzyme types for the five sites we considered. Additionally, we observed a significant decrease in respiration for all communities after longer durations of drought exposure. Although these findings did not align with our expectations of local adaptation, they suggest litter-inhabiting microbial communities are able to retain metabolic functioning in environmental conditions different from those of their native ecosystems. These results may imply that factors such as litter chemistry impose greater constraints than climate on community metabolic function. Overall, despite differences in local climates, microbial communities from semiarid regions may be metabolically adapted to maintain functioning in the face of drought.