Geochemical bio-signatures in Martian analogue basaltic environments using laboratory experiments and thermochemical modelling

The identification of geochemical bio-signatures is important for assessing whether life existed on early Mars. In this paper, experimental microbiology and thermochemical modelling were combined to identify potential inorganic bio-signatures for life detection on early Mars. An analogue mixed microbial community from an analogue terrestrial fluvio-lacustrine environment similar to an ancient lacustrine system at Gale Crater was used to study microbial dissolution of a basalt regolith simulant and the formation of bio-signatures over a short time frame (1°month) at 14°C, 2 bar. Microbial growth influenced element dissolution (Mg, Fe, Mn, Ca and K) and the formation of morphologies and Fe-Si amorphous layers on mineral surfaces. Thermochemical models were performed at 14°C, 2 bar; the results were compared with experimental data to predict bio-signatures that would occur over geological timescales. The pH was varied to simulate abiotic and biotic experimental conditions. Model results suggest that, at water to rock ratios of 100 to 38, a less complex secondary mineral assemblage forms during biotic dissolution compared to abiotic weathering. Carbonates, quartz, pyrite and hydroxyapatite form under biotic conditions, whereas in the abiotic system magnetite and phyllosilicates would also precipitate. These results could be used to distinguish between abiotic and biotic basalt weathering processes, aiding the interpretation of data from Mars exploration missions.

Since rocks of basaltic composition dominate the Martian surface (Nyquist et al., 2001;Christensen et al., 2005;Filiberto, 2008;Treiman and Filiberto, 2015;Morris et al., 2016;Mangold et al., 2017), weathering processes and/or brine evaporation has evolved clay minerals, carbonates, sulfates and other alteration minerals Poulet et al., 2007;Ehlmann et al., 2008;Mustard et al., 2008;Ehlmann et al., 2009;Bridges et al., 2015;Stern et al., 2015;Bultel et al., 2019;Horgan et al., 2020). The presence of such secondary minerals, for example at Gale and Jezero Craters, indicates a variety of alteration conditions and water-rock reactions that may have controlled the local chemical conditions (e.g., chemistry of the fluids, redox state of elements, release of elements into the fluids) and, thus, the availability of bio-essential elements. Understanding the formation of secondary minerals is therefore critical to determining the potential habitability of aqueous environments (Schulte et al., 2006;Hand et al., 2007;Bridges and Schwenzer, 2012).
Bio-signatures are formed by microbial activity (e.g., Vandevivere et al., 1994;Banfield et al., 2001;Westall et al., 2015;Price et al., 2018;Tan et al., 2018) and can be divided in two main categories: 1) organic bio-signatures, which are biomolecules produced by organisms as part of their metabolic and reproductive machinery (Summons et al., 2011;Röling et al., 2015;Westall et al., 2015;Hays et al., 2017;Vago et al., 2017); and 2) inorganic bio-signatures, which include morphological fossils, sedimentary structures, isotope fractionation, and mineral alterations that are the result of microbial activity (Banfield et al., 2001;Cady et al., 2003;Westall et al., 2015;Vago et al., 2017;McMahon et al., 2018). Such bio-signatures can be preserved within the geological record and used as evidence of the presence of life in terrestrial and extraterrestrial environments, including early Martian aqueous systems. Owing to the detrimental effect of the conditions at the present-day surface of Mars, which can degrade organic molecules (Ten Kate et al., 2005;Dartnell, 2011;Hays et al., 2017;Vago et al., 2017), inorganic bio-signatures, such as secondary alteration minerals, may be more appropriate for assessing whether life existed on early Mars. For example, biotically produced secondary minerals may be preserved in Martian rocks and be detectable through in-situ measurements by on-going and up-coming rover missions (Vago et al., 2017).
To identify and use, unambiguously, specific secondary mineral assemblages as bio-signatures for life detection, it is fundamental to have a comprehensive comparison between biotic and abiotic weathering processes. Investigating analogue environments is one way of informing our understanding of potential bio-signatures. Unfortunately, the heterogeneity of the natural environment, the lack of control over changes during even a short sampling period, and the difficulty of finding relevant abiotic controls, make such work challenging. Moreover, although basalt weathering and secondary alteration mineral formation have been widely investigated in the field and in laboratory experiments under abiotic (Gislason and Eugster, 1987;Oelkers and Gislason, 2001;Wolff-Boenisch et al., 2006) and biotic conditions (Wu et al., 2007;Olsson-Francis et al., 2010;Olsson-Francis et al., 2012;Olsson-Francis et al., 2015;Olsson-Francis et al., 2016;Olsson-Francis et al., 2017;Price et al., 2018;Olsson-Francis et al., 2020) these studies have been only over short timeframes (months-years). This makes it difficult, if not impossible, to predict what would happen over years or even over geological time scales when the rock is fully dissolved or more likely, subject to the effects of incongruent dissolution. In these circumstances, the formation of amorphous and leached layers, and secondary mineral precipitation, may occur, influencing the availability of cations for use in biological metabolism and, thus, the formation of biosignatures (e.g. Welch and Ullman, 1993;Benzerara et al., 2004;Benzerara et al., 2005).
In order to overcome this problem, thermochemical modelling can be used to study, theoretically, alteration processes that can occur over geological timescales. Owing to its capability to predict secondary mineral assemblages and variations in fluid chemistries by assessing reaction pathways during water-rock interactions, thermochemical modelling has been widely applied to study alteration processes that happened, or may have happened, in terrestrial and Martian aqueous environments (Griffith and Shock, 1997;Kühn, 2004;Zolotov and Mironenko, 2007;Marion et al., 2008;Bridges and Schwenzer, 2012;Catalano, 2013;Filiberto and Schwenzer, 2013;Schwenzer and Kring, 2013;Bridges et al., 2015;Melwani Daswani et al., 2016;Zolotov and Mironenko, 2016;Ramkissoon et al., 2021). Only recently, Olsson-Francis et al. (2017) used thermochemical modelling in conjunction with laboratory-based experiments to investigate and compare the alteration minerals that may form during microbial basalt weathering over a range of different timescales. The study suggested that an aerobic chemoorganoheterotrophic bacterium (Burkholderia sp. strain B_33), during long-term weathering of a naturally occurring altered basalt, would produce a less chemically and mineralogically diverse secondary mineral assemblage consisting of Fe-hydroxide and kaolinite, than under abiotic conditions, where chlorite would also be formed (Olsson-Francis et al., 2017). That study also demonstrated the utility of combining laboratory experiments and thermochemical modelling to identify secondary minerals that could be used to distinguish, unambiguously, weathering processes that may have occurred on early Mars (Olsson-Francis et al., 2017).
In this paper, we investigate the impact of microbial activity on a simulated Rocknest environment and the formation of inorganic bio-signatures using a combination of laboratorybased experiments and thermochemical modelling. The Rocknest environment was simulated by preparing a Martian regolith simulant with a composition similar to that of the Rocknest basalts (global Martian basaltic soils, Blake et al., 2013;Schmidt et al., 2014) and combining this with a minimal medium. A buffer was added to the medium to compensate for the reactions that would occur between the basalt and a CO 2 rich headspace, with the intent to create a stable environment with circumneutral pH  similar to ancient Martian aqueous systems (Grotzinger et al., 2014;Vaniman et al., 2014). To study the fluid chemistry variations and secondary minerals formed under biotic conditions we use a terrestrial analogue mixed anaerobic community consisting of chemoorganotrophs and chemolithotrophs. The analogue microbial community was isolated from a Mars analogue environment, the anoxic intertidal zone of the River Dee (United Kingdom), and had previously been deemed a plausible analogue to study habitability of ancient lacustrine systems on early Mars (e.g. Gale Crater) using laboratory-based simulation experiments (Curtis-Harper et al., 2018).
Experimental and model results were compared in order to identify chemical and mineralogical signatures that are uniquely produced by microbial weathering and may be used as inorganic bio-signatures for life detection on Mars. Instrumentation onboard current Martian rovers could provide mineralogical and geochemical data that would assist in the identification of potential bio-signatures, e.g., the CheMin (XRD) instrument and the Laser Induced Breakdown Spectroscopy (LIBS, ChemCam) and the Alpha Particle X-Ray Spectrometer (APXS) instrument on board of the Mars Science Laboratory Rover Curiosity (Grotzinger et al., 2012) and the most recent SuperCam (a LIBS instrument) and Raman instruments on board of the Perseverance rover (Williford et al., 2008).

Preparation and characterization of the mars regolith simulant
For this study, a regolith simulant was prepared that was compositionally similar to basaltic float rocks analysed by Frontiers in Astronomy and Space Sciences frontiersin.org 03 Curiosity at Rocknest, Gale Crater (Schmidt et al., 2014) ( Table 1). The simulant consisted of basalt from Le Cheix Puy de Dome (France) ( Table 1), which was selected to represent a typical unweathered basalt, whereby any macroscopically visible weathering was separated out of the sample during the sample preparation. The basalt was supplemented with aegirine, an iron rich pyroxene (FeO 29.21 wt%, Table 1) from Mount Malosa (Malawi). This was to ensure that the simulant contained an equivalent iron content to an average Martian basalt, which is higher than terrestrial basalts (Longhi et al., 1992), and to correct for Fe speciation. This is important when exploring Martian habitability since the Fe 2+ /Fe 3+ ratio is crucial for microbial metabolism (Nixon et al., 2013;Ramkissoon et al., 2021). Both the basalt and the aegirine were purchased from Richard Tayler Minerals (United Kingdom).
The major element composition of the aegirine (Table 1) was obtained from polished thin sections using a CAMECA SX100 Electron Microprobe (EMPA) operated with a spot size of 10 mm, accelerating voltage of 20 kV and a beam current of 20 nA. Analysis of the basalt's major element composition (Table 1) was carried out on powdered samples using an ARL 8,420 + dual goniometer wavelength-dispersive X-ray Fluorescence (XRF) spectrometer.
To produce the regolith simulant, the basalt and aegirine were separately crushed using a Tema swing mill, for 8 min. The crushed materials were sieved to a fraction size of between 100 μm and 250 μm, the fine particles were removed by ultrasonication in MilliQ water (Olsson-Francis et al., 2010;Olsson-Francis et al., 2017), and the products were dried for 24 h, at 80 C and at 1 bar. The 100-250 μm size fraction was selected to include materials that have a grain size comparable to dust-size particles of the Martian regolith observed by Curiosity's Mars Hand Lens Imager instrument in sedimentary deposits at various locations at Gale crater (Minitti et al., 2013;Weitz et al., 2018). Moreover, the selection of a such small grain size was intended to speed up the formation of secondary minerals mitigating for the short duration of the dissolution experiments. The aegirine was mixed to the basalt in a 1.75:1 ratio to achieve an overall iron-rich composition similar to that of Rocknest basalt (Schmidt et al., 2014), which has 6.65 wt% more Fe 2 O 3(Tot) than the Le Cheix Puy de Dome basalt (Table 1).

Microbial community
A microbial community from the sub-surface intertidal anoxic zone of the River Dee, United Kingdom (53˚21′15.40 N, 3˚10′24.95 W) was used for the biotic experiments. This mixed community was selected because it contains chemolithoautotrophic and chemoorganoheterotrophic microbes that are analogues for potential microbial life that may have existed in the fluviolacustrine environment at Gale Crater (Amils et al., 2007;Grotzinger et al., 2014;Curtis-Harper et al., 2018). The sample was collected as part of a previous study (Curtis-Harper et al., 2018). MiSeq sequencing demonstrated that the bacterial community was dominated by the families Hyphomicrobiaceae (28%), Flavobacteriaceae (23%), Alteromonadaceae (16%), whilst over 90% of the archaeal community was dominated by Thaumarchaeaota phylum (Nitrosopumilus genus) (Curtis-Harper et al., 2018) ( Table 2).
The microbial enrichment was prepared by adding 5 g of dry sediment to an anaerobic enrichment medium containing per litre: 2 g of trypticase peptone (Sigma Aldrich), 2 g of yeast extract (Sigma), 0.3 g of KCl, 1 g of NH 4 Cl, 3 g of Na 2 SO 4 , 23 g of NaCl, 0.5 g of Na-Lactate, 2 g of MgCl 2 , 0.35 g of K 2 HPO 4 , 0.1 g of Na-thioglycollate (C 2 H 3 NaO 2 S) and 0.1 g of ascorbic acid (C 6 H 8 O 6 ), and was prepared anaerobically, as previously described (Hungate, 1966). Vitamin solution (Balch et al., 1979) and SL10 trace element solution (Widdel et al., 1983) were added post-autoclaving (121°C, 15 psi, 15 min). 100 ml of medium was aseptically transferred (under anaerobic conditions) to sterilised Wheaton bottles (Speers et al., 2009), and after inoculation, the headspace was pressured to 2 bar with 80% CO 2 /20% H 2 . The cultures were incubated at 14°C, for 700 h (29 days), and subsequently transferred twice to eliminate any crossover from the sediment.

Dissolution experiments
The dissolution experiments were conducted with a (W/R) E of 100/33 (we denote W/R E as the experimental water to rock ratio and denotes the amount of rock and the amount of water) with a nutrient-limited anaerobic minimal medium, which contained the following (per litre): 1 g of NH 4 Cl, 2 g of Nalactate (C 3 H 5 NaO 3 ), 1 g of Na-thioglycollate (C 2 H 3 NaO 2 S), 1 g of ascorbic acid (C 6 H 8 O 6 ), 37 g of NaCl, 13.25 g of Na 2 CO 3 . To prepare the regolith simulant, 8.09 g of basalt and 5.91 g of aegirine were dispensed into acid-washed Wheaton bottles, which had been purged with N 2 to remove O 2. The regolith was autoclaved at 121 C for 15 min and after cooling, 68 ml of medium was added, and the pH was adjusted to pH 7.0 with filtered sterilised 10 mM NaOH. For the inoculum, exponential phase cells were harvested by centrifugation (4,000 × g, for 5 min), washed, and resuspended in the minimal medium to give a final cell density of 10 7 to 10 8 cell mL −1 . A 1% inoculum was used for the biotic experiments (designated V1, V2, V3, V4 and V5) and abiotic controls were prepared in parallel (designated VA1, VA2, VA3, VA4 and VA5). The headspace of all of the Wheaton bottles were overpressured to 2 bar with 80% CO 2 /20% H 2 , prior to incubating at 15 C for 700 h. Microbial growth was measured every 24-48 h over the course of the 29 days experiment (closed system simulation) using the acid-binding Sybr, as previously described by Olsson-Francis et al. (2017). The duration of the dissolution experiment was also dictated by the experimental design (closed system simulation) that limited the availability of nutrients for microbes and by the fact that, interrupting the experiments after 700 h (point at which the microbial community reached the stationary phase), it was avoided a reduction in abundance and diversity of microbes relating to the selection of specific metabolic groups. To monitor microbial growth in absence of the regolith simulant, experiment controls were also performed combining the microbial community and the minimal medium only. In such cases, no microbial growth was observed; for this reason, this aspect will not be further discussed.

Chemical analyses of the medium
An Orion 3-Star Thermo Scientific bench top meter with an uncertainty of 0.01 pH unit was used to measure the pH of the media. A Prodigy High Dispersion Inductively Coupled Plasma-Atomic Emission Spectrometer (ICP-AES) was used to measure the total elemental concentration of dissolved elements in the medium. In triplicates, 1 ml aliquots were removed aseptically from each Wheaton bottle after 700 h. The aliquots were centrifuged at 400 × g, for 2 min to eliminate any residual regolith simulant from the analytical sample, and then acidified using nitric acid (10%). Detection limits for the ICP-AES were defined as 3 times above blank level. The ICP-AES data were corrected for the loss of elemental mass during sampling and for the decrease in fluid volume, as described in Wu et al. (2007).

Cellular elemental uptake
To measure the intracellular elemental concentration, 1 ml aliquots were centrifugated (4,000 × g for 2 min) and the supernatant collected (to remove any excess regolith). The cells were harvested and washed three times in sterilised 0.5% HNO 3 by centrifugation (13,000 × g for 10 min) to remove excess media. The cells were acidified by resuspending the cells in 10% nitric acid. This allowed the elemental content of the cells to be measured using a Prodigy High Dispersion Inductively Coupled Plasma Atomic Emission Spectrometer (ICP-AES). The values obtained were multiplied by the number of cells measured at Frontiers in Astronomy and Space Sciences frontiersin.org 05 700 h and were corrected for the decrease in fluid volume during sampling, as previously described (Wu et al., 2007;Olsson-Francis et al., 2012).

Characterization of the simulant after dissolution experiments
After 700 h, the regolith simulant was removed from the Wheaton bottle, air-dried, and carbon coated (15-20 μm thickness) on aluminium stubs. The chemical composition and the morphological characteristics of the simulant were determined using a ZEISS Supra 55-VP Field Emission Gun -Secondary Emission Microscope (FEG-SEM) with an Energy Dispersive Spectroscopy (EDS) detector, which was operated at an accelerating voltage of 2-15 kV and a 7-10 mm working distance. Secondary electron images, backscatter electron images and EDS spectra obtained from FEG-SEM analysis were compared and used to identify the presence of secondary minerals that may be used as bio-signatures.

Thermochemical modelling
Thermochemical models were executed to identify reaction pathways, shifts in fluid chemistry and mineral formation that would occur over geological timescales under biotic and abiotic conditions in the considered fluid-rock systems. The code CHIM-XPT (previously CHILLER, Reed and Spycher, 2006;Reed et al., 2010) was used to perform stepwise titration simulations during which a specific amount of rock was titrated into a constant amount (1 L) of the starting fluid. The model assumes complete rock dissolution and each titration step is in equilibrium, and therefore can be interpreted independently. The reaction progress is expressed as water to rock ratio (W/R), which represents the amount of rock reacted with the fluid. By modelling at different (W/R) M (water to rock ratio used in the models), the software simulates chemical reactions that may happen under different environmental conditions (e.g. Reed, 1982;Filiberto and Schwenzer, 2013;Bridges et al., 2015). At very high W/R (~1,000,000), the model simulates an environment where a limited amount of rock react with a large mass of water (e.g., freshwater inflow, fluid percolating in a fracture, a rock surface exposed to regular precipitation), while at low W/R ratios (<1,000), the model represents a scenario where a large volume of rock reacts with a limited volume of water (e.g., stagnant water table like in lacustrine-sediment settings). More information on the code and method and its application and limitations can be found in Palandri and Reed (2004), Reed and Spycher (2006) and Reed et al. (2010).
The input data for the thermochemical modelling included the Mars regolith simulant composition (Table 1) and the chemical composition of the minimal medium (Table 3). Elements in the fluid summarised in Table 3 are given as component species and in the model calculation partitioned between several species as relevant to the pH and Eh conditions during the modeling. The code applies mass balance and mass action equations to calculate the equilibrium of secondary alteration mineral assemblages and fluid composition at different water/rock ratios. The modelling was carried out at 2 bar and 14°C, which was used to simulate the pressure and temperature conditions of the growth experiment. The models were run over a range of (W/R) M ratios between 10 5 and 1. Three models were conducted: in the first, the pH was allowed to vary and was treated as a free parameter to simulate rock weathering in an abiotic environment. In the other two models, the pH was set at 7.0 and 7.4 to reproduce the minimum and maximum pH conditions observed during microbial growth experiments simulating weathering progression in presence of the considered bacterial community. Full data tables, minerals suppressed in the modelling because not stable at low P and T conditions (2 bar, 14 C) or not forming in the simulated Martian environment and relative references are presented in Supplementary Material.

Microbial growth
The microbial community grew in the minimal medium with the Mars regolith simulant as the sole source of bio-essential elements. Exponential growth occurred between 24 h (1-5×10 5 cells mL −1 ) and 300 h (4-7 × 10 6 cells mL −1 ) after inoculation (Figure 1). After 300 h, the cultures reached stationary phase and cells numbers remained relatively constant between 5 and 6 × 10 6 cells mL −1 (Figure 1). No growth was detected in the abiotic controls.
In the biotic controls, the mean pH increased from 7.07 ± 0.08 to 7.36 ± 0.06 in the first 170 h and then stabilised to 7.29 ± 0.02 until the end of the experiments ( Figure 1A, Supplementary Material). For the abiotic controls, the pH increased from 6.98 ± 0.08 to 7.31 ± 0.08 ( Figure 1A, Supplementary Material).  Figure 1G, with a minimum and maximum difference of~30 μmolL −1 and~170 μmolL −1 , respectively, between the biotic and abiotic controls. The microbial growth phase influenced the dissolution of certain elements, e.g., after 300 h (exponential growth), K was higher under biotic conditions, and the concentrations of Fe, Ca and Mn were higher in the biotic experiments than the abiotic experiments only during the late stationary phase (>400-500 h).

Secondary alteration minerals
FEG-SEM with EDS was used to investigate secondary alteration products (secondary minerals, amorphous phases) formation and morphological changes on the surface of the simulant. Analysis of the abiotic samples demonstrated that the mineral surfaces showed little or no evidence of physical or chemical weathering (Figures  2A-C). Analysis of the regolith simulant from the biotic experiments showed some microbial attachment and evidence of chemical and physical changes on the mineral surfaces ( Figures 2D-F). This included morphological steps observed along mineral cleavage plains and amorphous layers with smooth, rounded and undulate surfaces observed on top of some weathered grains ( Figure 2D). SEM-EDS analysis demonstrated that the amorphous layers were mainly composed of Fe, Si, O and Na with lower amounts of Cl, Ca and Al (Figures 2E,F).  Table 4 shows all the minerals that precipitate in the abiotic and biotic models at different (W/R) M ratios. Variation in element concentration in the fluid that are not directly comparable with data from dissolution experiments but that are relevant for secondary minerals formation predicted by the models are reported in Supplementary Material.

Abiotic model
The abiotic model ( Figure 3A) forms the baseline from which to compare the biotic models. From high to low (W/ R) M ratios the secondary mineral assemblage became more complex. Pyrite formed at (W/R) M ratio of 10 5 , which was stable over the considered (W/R) M ratio range, and progressively decreased from 100 wt% to <1 wt%. This reduction in pyrite coincided with other Fe-rich minerals forming and the (W/R) M ratio decreasing. Talc formed below a (W/R) M ratio of 100,000, it reached its maximum abundance (~21 wt%) at a (W/R) M ratio of 3,330 when quartz started to form. The values remained constant between 12 wt % and 16 wt% up to a (W/R) M ratio of 1. Between (W/R) M

Biotic models
When the pH was set at 7.0 and 7.4 (simulating the small change in the buffered medium), the models ( Figures 3C,D) gave similar results in terms of secondary mineral assemblages and fluid chemistries. Pyrite was stable over the considered (W/R) M ratio range with its abundance progressively decreasing from 100 wt% to <1 wt% when quartz started to form ((W/R) M ratio of~100,000). Below this value, quartz increased and was the most abundant mineral phase at a (W/R) M ratio lower than 25,000 (a maximum value of 67 wt% was measured at end of the titration range).
FIGURE 3 (Continued) quartz, carbonate, talc, magnetite, goethite in different amounts and traces of chlorite, nontronite, apatite and kaolinite (not shown in the diagram since its abundance is <<1 wt%). For more details on the variability of the minor phases (chlorite, nontronite, apatite and kaolinite) see supplementary materials A, B, C. The (W/R) D is represented by the shadowed area in between the two blue dotted vertical lines and it was calculated based on the element potassium because it is not incorporated into the microbes and into any of the expected mineral precipitates.
Frontiers in Astronomy and Space Sciences frontiersin.org When the pH was fixed at 7.4, quartz reached its maximum abundance of 68.8 wt% at a (W/R) M ratio of~3. Quartz and pyrite were the major phases that formed between (W/R) M ratios of 100,000 and 500. Siderite precipitated only below a (W/R) M ratio of 500, together with quartz (~62-65 wt%) and pyrite (~35-< 1 wt%). Siderite increased up to~34 wt% at a (W/R) M ratio of 11, and then started to decline. Below a (W/R) M ratio of 10, quartz continued to be stable with a maximum abundance of 69 wt%, siderite decreased from 34 wt% to <3 wt%, and goethite reached its maximum concentration of around 22 wt% at (W/ R) M ratio of 1. Below (W/R) M of 10, magnetite formed (<4 wt%) when the pH was set at 7.0, while it reached 7 wt% when the pH was set at 7.4. Huntite formed in traces (<1 wt%) only when the pH was set at 7.4 and between (W/R) M ratios of 50 and 24; whilst talc (<7.5 wt%) formed only when the pH was set at 7.4 below a (W/R) M of 3. Trace amounts (<1-3 wt%) of chlorite (clinochlore, daphnite), carbonates (rhodochrosite), kaolinite, nontronite and hydroxyapatite were also formed at different (W/R) M ratios.
Most of the dissolved elements, here given as component species, varied in concentration by, at least, one order of magnitude with the exception of Cl − , H 4 SiO 4 , NH 4 + and Na + that were relatively stable ( Figures 4C,D

Experimental regolith simulant dissolution
In this study, variations in fluid chemistry, as well as mineralogical and morphological changes to a Martian regolith simulant, were investigated to identify inorganic biosignatures produced by chemolithotrophic and chemoorganoheterotrophic microbes, which might be detected in the Martian rock-record and which might be analogue to those that may have lived in the fluvio-lacustrine system at Gale Crater on early Mars (e.g. Curtis-Harper et al., 2018). Frontiers in Astronomy and Space Sciences frontiersin.org ICP-AES analysis detected specific changes in the fluid chemistries that differed between the biotic and abiotic test groups, e.g., microbial mediated dissolution increased the dissolved concentrations of Mg, Fe, Mn, Ca and K. Previous studies demonstrated that excess protons and low molecular weight organic acids, produced as a by-product of microbial metabolism, can alter the pH of the fluid promoting the extraction of bio-essential elements from rocks Liermann et al., 2000;Huston and Logan, 2004;Wu et al., 2007) and the formation of secondary alteration minerals (Banfield et al., 2001;Berg et al., 2020). In our study, changes in element concentration could not be solely ascribed to the pH differences between the abiotic and biotic experiments. Given that the pH of the media remained nearly neutral during biotic experiments because of the buffering capacity of the simulant, it is possible that organic acids may have enhanced mineral dissolution locally, e.g. in pore spaces and close to mineral surfaces, without changing the pH of the whole system. Micro-reaction zones have previously been identified at microbial binding sites and are associated with high concentrations of organic acids (Hiebert and Bennett, 1992;Thorseth et al., 1992;Vandevivere et al., 1994). Although there is little evidence of direct microbial attachment to the mineral surface it is still possible that elemental dissolution in the biotic experiments has occurred through direct contact of the microbes with the simulant, and also indirectly (Hiebert and Bennett, 1992;Vandevivere et al., 1994;Banfield et al., 2001;Welch et al., 2002;Konhauser, 2007;Lian et al., 2008;Uroz et al., 2009;Gadd, 2010). Differences between the amount of element dissolved in the fluid in this study and previous studies (Wu et al., 2007;Olsson-Francis et al., 2017) could be explained by the inclusion of a buffer in the system that may have prevented elemental release as a result of high pH changes, or the use of a more complex microbial community instead of a single microbe type (Burkholderia sp. strain B_33-Olsson-Francis et al., 2017; Burkholderia fungorum - Wu et al., 2007). Heterotrophic microorganisms, such as Burkholderia fungorum, use organic compounds as source of energy and chemical equilibrium is more easily reached and dissolution is at minimum (Wu et al., 2007). In a mixed community, that contains also chemolithotroph microorganisms which require chemical compounds and elements (sourced from the fluid and the rock) to produce energy, dissolution rates increase due to the pH being far from equilibrium (Ramkissoon et al., 2021).
Morphological steps observed along mineral cleavage plains and Fe-Si rich amorphous deposits ( Figure 2D) identified by SEM on the surfaces of some aegirine grains in the biotic experiments are, respectively, evidence of enhanced mineral weathering and secondary mineral deposition following microbial action (Brantley and Chen, 1995;Benzerara et al., 2004;Benzerara et al., 2005;Wu et al., 2007). According to the elemental distribution in "spectrum 5" (Figure 2E), the amorphous deposit may represent partially dissolved aegirine and the formation of non-crystalline Fe-rich secondary clay at the mineral surface, characteristic of an early stage of pyroxene weathering (Nahon and Colin, 1982). We interpret this as evidence for mineralogical changes specific to the biotic experiments as nothing comparable was observed in the abiotic experiments. Peaks detected for Ca, Na and Cl are interpreted as CaCl 2 and NaCl deposits precipitated on top of the Fe-Si rich amorphous layer during sample drying before SEM analysis. Previous studies that used a naturally occurring, already altered, basalt as a Martian analogue observed similarities between secondary minerals formed during biotic laboratory experiments and those predicted by thermochemical models (Olsson-Francis et al., 2017). In our study, mineralogical features observed in the biotic samples and alteration phases predicted by biotic models are not directly comparable, as detailed analyses of secondary mineralisation formed experimentally were limited by the rarity of potential secondary minerals at detectable abundances. The use of an unweathered fresh basalt, of a small grain size (<250 μm) and the short duration of the experiments could have reduced the likelihood of alteration phase formation in our study, impeding a comparison with model results. Longer experimental run times in a simulated open-system environment would be needed to observe the development of a more complex alteration assemblage, the formation of larger quantities of secondary minerals and more pronounced alteration features that would be more easily detectable, and thus comparable with secondary minerals predicted by the biotic and abiotic models.

Thermochemical models of basalt dissolution
Thermochemical modelling was used to assist in the identification of possible inorganic bio-signatures, such as secondary alteration minerals and fluid chemistry variations, that form during microbial weathering of the Martian simulant over geological timescales in an aqueous system. To compare the experimental and model results, we focus on a specific (W/R) M range, between 100 and 38 (Figures 3, 4), that is equivalent to the actual (W/R) D ratio interval of the growth experiments. The (W/R) D ratio is defined as the amount of regolith dissolved during the experiments; it was assessed by focusing on the most soluble elements (e.g. K + ) in the fluids resulting from the experiments and calculated as outlined in Olsson-Francis et al. (2017). K + was specifically selected because it is not incorporated into microbial cells or (in large amounts) into any of the minerals predicted to precipitate by the models (see Results section). The (W/R) M ratio range considered here (between 100 and 38) is also equivalent to a lacustrinesedimentary setting where a limited amount of water interacts with a large volume of rock. As described in detail earlier, similar environments have been identified by rover missions at Gale and Jezero Craters (Grotzinger et al., 2014;Vaniman et al., 2014;Mangold et al., 2021) and are proposed as sites that could have supported microbial life in early Martian's history .
Under lacustrine-sedimentary conditions, a less complex secondary mineral assemblage forms during biotic dissolution compared to abiotic, where more mineral species precipitate (Figure 3). The models show that under biologically-mediated conditions, high amounts of quartz and carbonates (mainly siderite, but also traces of rhodochrosite and huntite) coexist with minor amounts of Fe-sulfides (pyrite) and trace amounts of phosphates (hydroxyapatite), whereas the abiotic system would also precipitate Fe-oxides (magnetite) and various amounts of phyllosilicate minerals (talc, chlorites, nontronite). Less chemically and mineralogically complex secondary mineral assemblages seem to be common features of basalt-fluid interactions under biotic conditions at 1 bar and temperature ≤25 C (Olsson-Francis et al., 2017), well below the threshold at which life can exist (<121°C, e.g., Conrad, 2014;Cockell et al., 2016).
The differences in the fluid compositions between an abiotic and a biotic environment are more pronounced in the models, where rock dissolution is simulated for longer periods of time (e.g., geological timescales), than in laboratory experiments which investigate rock alteration only for a few weeks. Within the considered (W/R) M ratio range between 100 and 38, that simulates lacustrine-sedimentary conditions, K, Si and Mn were observed in similar concentrations (10 −3 mol, 10 −4°m ol and 10 −4°m ol, respectively) in the biotic models as in the growth experiments (Figures 1, 4, Supplementary Material), suggesting that these elements are used by the microbes only in minimal amounts and that their behaviour is similar during short-and long-term biotic weathering. This is also supported by the fact that the cellular elemental uptake measured after biotic experiments was below the detection limits (Supplementary Material). Mg remains lower in the experimental fluid compared to model results. This could be related to microbial activity that impedes elemental dissolution by producing amorphous layers on top of some minerals (e.g., pyroxene) (Benzerara et al., 2004;Benzerara et al., 2005), promoting adsorption of polysaccharides onto mineral surfaces (Welch et al., 1999), producing high-molecular-weight polymers  and preventing the formation of etch-pits (Lüttge and Conrad, 2004). According to model results, Mg would not precipitate in large amounts in secondary minerals predicted to form under biotic conditions, and it would accumulate in the fluid during long-term regolith alteration. Fe, Ca and Al contents were higher in the experimental fluids than in the modelled fluids, suggesting that biological activity is a limiting factor for precipitation of these elements during short-term weathering.
Inhibition of precipitation could be related to different processes associated to specific biological functions. Microbes take up specific bio-essential elements (oxygen or phosphate) to form inorganic or organic compounds, including proteins and siderophores, that are used for internal vital processes; the production of polysaccharides and other extracellular polymeric substances is another way of how microbes can attract and use bio-essential elements for their external functions; finally, microbes also use various chemicals to produce and release organic compounds such as low Frontiers in Astronomy and Space Sciences frontiersin.org molecular weight organic acids (Welch and Ullman, 1993;Vandevivere et al., 1994;Kalinowski et al., 2000;Bennet et al., 2001;Wu et al., 2007;Uroz et al., 2009;Gadd, 2010;Olsson-Francis et al., 2015). These processes would prevent solubility limits being reached, impeding the precipitation of secondary minerals that could form in the considered water-rock system during long-term dissolution. As predicted by the models, these phases could be pyrite (FeS 2 ), siderite (FeCO 3 ), hydroxyapatite (Ca 5 (PO 4 ) 3 OH) and Al-rich minerals (kaolinite, Al 2 SiO 5 (OH) 4 ).

Alteration minerals and life detection on Mars
Fe-phyllosilicates (nontronite and chlorite) and Fe-oxide (magnetite), predicted by thermochemical modelling to form only under abiotic conditions, have been found at several sites on Mars (Ehlmann et al., 2009;Carter et al., 2013;Ehlmann and Edwards, 2014;. Chlorite is an uncommon alteration mineral at very low-temperature since it forms preferentially in diagenetic and low-temperature metamorphic environments. However, it is thought to be authigenic in weathered sedimentary deltaic sandstones at 20-40 C (Grigsby, 2001) and has been found widespread on Mars (Ehlmann et al., 2011). Talc is also predicted by the abiotic model to form but its presence on Mars is still uncertain (Bristow et al., 2021). Talc is usually associated to higher temperature regimes (hydrothermal alteration) and metamorphic conditions; however, talc can be also authigenic in sedimentary deposit (Tosca et al., 2011 and references therein) or, as demonstrated in laboratory test performed at ambient conditions, it can precipitate as a secondary product in low-temperature aqueous systems as a precursor of Mg-rich carbonate deposits (Bricker et al., 1973;Tosca et al., 2011). In all these cases, talc precipitation seems strongly controlled by the pH and favoured under alkaline conditions (Tosca et al., 2011). Carbonate minerals (e.g., siderite and rhodochrosite) characteristic of both biotic and abiotic models have been observed in Martian meteorites (e.g., nakhlites, Changela and Bridges 2011;Melwani Daswani et al., 2016;Bridges et al., 2019) and by orbiting spacecraft and rovers on the Martian surface at various locations (e.g., Gale and Jezero Craters, Bultel et al., 2019;Bridges et al., 2019;Thorpe et al., 2020;Archer et al., 2020;Horgan et al., 2020). Other alteration minerals predicted to form under biotic and abiotic conditions such as quartz, pyrite and hydroxyapatite have also been discovered (Vaniman et al., 2014;. The carbonates huntite, which precipitates in trace amounts only in the biotic model at pH of 7.4, has not been directly observed on Mars, although it has been inferred to occur in Nili Fossae region (Palomba et al., 2009). However, since it can form at surface temperatures and pressures in fluvial-lacustrine evaporitic environments or as an alteration mineral in basalts weathered by cold solutions rich in Mg, Ca and carbonic acid (Kinsman 1967;Cole and Lancucki, 1975;Stanger and Neal, 1994;Akbulut and Kadir, 2003), it is likely to be present in ancient Martian systems similar to Gale and Jezero Craters. Huntite, could be particularly difficult to detect on Mars (especially if present in low amounts) since it is metastable at surface temperature (Garrels et al., 1960;Kinsman, 1967) and it is replaced with time by magnesite (Kinsman, 1967;Spotl and Burns, 1994;Marini, 2007), a more stable Mg-carbonate that is present on the Martian ground together with siderite, calcite, rhodochrosite and other Mg-and Fe-bearing alteration minerals Niles et al., 2013;Goudge et al., 2015). Finally, alteration phases predicted to form in the biotic system (siderite, rhodochrosite, quartz, pyrite, hydroxyapatite) are all minerals that naturally can precipitate in near-neutral pH aqueous environments at ambient temperature following the action of microbes (Lee et al., 2007;Ehrlich and Newman, 2009;Sun et al., 2014;Duverger et al., 2020).
The results of this study provide indications of which mineralogical and geochemical features need to be investigated when searching for inorganic bio-signature on Mars. Alteration effects and mineral deposits of such small scale (<15 μm) would be difficult, if not impossible, to be observed in-situ using the instruments on board of Martian rovers (Mars Hand Lens Imager instrument aboard Curiosity rover has a resolution up to 13.9 microns/pixel) (e.g. Edgett et al., 2012;Allwood et al., 2020). However, such information and observations are relevant for searching for the most likely sites to find putative life. The chemical changes associated with the observations in this study, while not resolved in the detail required to analyse them, might still be indicated through gradients and deviations from baseline mineral observations in spatially resolved techniques such as the Planetary Instrument for X-Ray Lithochemistry (PIXL) on Perseverance (Allwood et al., 2020). Moreover, a complex mineralogical assemblage and secondary minerals large enough to be detected by rovers' instruments are likely to have been formed in a natural Martian environment where the interaction between the substrate and the microbes, if ever been present, would have been prolonged over longer periods. Therefore, the results of this study are required for potentially informing the current sampling activities operated by the Perseverance rover, for planning future life detection missions and for interpreting the results of the analysis of Martian samples recovered by future sample return missions.

Conclusion
In this study, laboratory-based experiments and thermochemical modelling were combined to investigate inorganic bio-signatures formed by an analogue microbial community from an anoxic inter-tidal zone of a Martian analogue fluvio-lacustrine system and that could be used as Frontiers in Astronomy and Space Sciences frontiersin.org evidence in the search for life in early Martian lacustrinesedimentary systems. Analogue microorganisms used in dissolution experiments were able to grow under environment conditions that were similar to that of the Gale Crater's aqueous environment at Yellowknife Bay. The analysis of simulated Martian water-rock systems after dissolution experiments identified mineralogical, microscopic and geochemical changes characteristic to the biotic test group that could be interpreted as potential bio-signatures. These changes are likely to be caused by the action of anaerobic microbes, analogues to those that could have existed in early Gale Crater aqueous system and would have enhanced mineral weathering and promoted secondary mineral formation. Thermochemical modelling has highlighted more significant differences between fluid chemistries and secondary alteration minerals that would form in reducing biotic and abiotic systems over geological time scales. Under biomediated lacustrinesedimentary conditions, a 'simpler' mineral assemblage is predicted to precipitate during long-term weathering. Quartz, carbonates and Fe-sulfides are the main secondary minerals to form under biotic conditions, whereas in the abiotic system Feoxides and phyllosilicate, in addition to the aforementioned minerals, also precipitate. Since Gale and Jezero Craters are the targets of current and future in-situ Mars exploration missions, identification and characterisation of geochemical bio-signatures produced by analogue microbial communities from terrestrial fluvio-lacustrine systems is important to determine the potential habitability of similar environments on early Mars. The results of this study reinforce the necessity to use complex analogue microbial communities and to combine laboratory experiments with thermochemical modelling when investigating the formation of inorganic bio-signatures that form over geological timescales, particularly where those studies may inform the search for habitable environments. The results obtained here, alone, may be not sufficient as unambiguous bio-signatures for life detection on Mars. However, they could be used in conjunction with other geological, geochemical and biological evidence to assess the presence of life in ancient Martian environments. More studies are required in order to build a more comprehensive body of evidence that can be used to identify, unambiguously, inorganic bio-signatures on Mars. Considering the outcomes of this and previous studies (Olsson-Francis et al., 2017), we want to empathise and tress that only using a more holistic approach that combines experimental microbiology, analytical geochemistry and thermochemical modelling, similarly to what was applied in this study, it will be possible to fully understand the evolution of fluid and rock chemistries under biotic conditions and geochemical bio-signatures formation in low-temperature aqueous systems on early Mars. In this context, thermochemical modelling represents an essential tool that allows to investigate reaction pathways and secondary minerals even at low-temperature when slow reaction rates make their study difficult under laboratory conditions. Future work, employing a manufactured, multicomponent simulant with a mineralogical and chemical composition more similar to the Rocknest basalt at Gale Crater (OUCM-1- Ramkissoon et al., 2019) and a microbial community that more likely to represent life on early Mars (as discussed in Macey et al., in review), could increase the fidelity of the type of inorganic bio-signatures (e.g., the secondary mineral assemblage) that a microbial community could develop under simulated Martian conditions. Also, the use a wider range of analytical techniques (e.g., NIR, FTIR, XRD, XPS, Raman Spectroscopy) relevant to either current (NASA's Mars Science Laboratory and Mars Perseverance) and future (ESA's Rosalind Franklin) Mars missions could help to better characterize the chemical, mineralogical and physical changes that occur on the simulant surface after dissolution experiments. This will inform the possible recognition of inorganic bio-signatures during ongoing exploration missions and the analysis of samples recovered by a future sample return mission.

Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.
Author contributions SC carried out the thermochemical modelling and interpreted the results and led the writing of the manuscript; EW carried out the microbiology experiments and analyses, the FEG-SEM work and the geochemical analyses; NR carried out preliminary thermochemical modelling and took part in manuscript revision; KO-F conceptualized and supervised the study, took part in microbial data analysis and manuscript writing; SS assisted in conceptualizing and supervising the study, took part in the interpretation of the thermochemical model results; VP assisted in supervising the study and took part in manuscript writing.

Funding
This work was supported by Research England Expanding Excellence in England (E3) fund (Grant code 124.18), as well as an STFC funded studentship awarded to EW, and United Kingdom Space Agency grant ST/ S001522/1 to SS.
Frontiers in Astronomy and Space Sciences frontiersin.org