Metallosphaera sedula (DSMZ 5348) cultures were grown aerobically in DSMZ88 Sulfolobus medium containing 1.3 g (NH₄)₂SO₄, 0.28 g KH₂PO₄, 0.25 g MgSO₄⋅7 H₂O, 0.07 g CaCl₂⋅2 H₂O and 0.02 g FeCl₃⋅6 H₂O dissolved in 1 L of water. After autoclaving, Allen’s trace elements solution was added to 1 L media resulting in 1.80 mg MnCl₂⋅4 H₂O, 4.50 mg Na₂B₄O₇⋅10 H₂O, 0.22 mg ZnSO₄⋅7 H₂O, 0.05 mg CuCl₂⋅2 H₂O, 0.03 mg Na₂MoO₄⋅2 H₂O, 0.03 mg VSO₄⋅2 H₂O, and 0.01 mg CoSO₄ final concentration. The pH was adjusted to 2.0 with 10 N H₂SO₄. Chemicals of high purity grade were used for media preparation.

In this study, four mineral mixtures of Mars regolith simulants (MRS) were used to examine whether these minerals could provide nutrients/energy sources necessary for lithoautotrophic growth of M. sedula (Tables 1, 2). The mineral mixtures of synthetic Martian regolith analogs were assembled in accordance to data on the structural and chemical composition of Martian minerals identified in meteorites (McSween, 1994) and by recent orbiter and rover missions (Bibring et al., 2005; Poulet et al., 2005; Chevrier and Mathé, 2007; Mustard et al., 2009; Morris et al., 2010) reflecting current knowledge of environmental changes on Mars (Boettger et al., 2012). JSC Mars-1A Martian Regolith Simulant is a palagonitic tephra (volcanic ash altered at low temperatures), produced by Orbital Technologies Corporation, Madison, WI, United States. The only phases detected by x-ray diffraction are plagioclase feldspar and minor magnetite. Iron Mossbauer spectroscopy also detected traces of hematite, olivine, pyroxene and/or glass (Morris et al., 1993). One of the Mars regolith simulants was phyllosilicate-rich (Phyllosilicate Mars Simulant, P-MRS), containing a high percentage of smectite clays like montmorillonite, kaolinite, and chamosite, as well as carbonates (siderite, hydromagnesite). The other mixture (Sulfatic Mars Simulant, S-MRS) was characterized by its high gypsum and goethite content. Both mineral mixtures consist also of pyroxene and plagioclase (gabbro), olivine, quartz, and the anhydrous ferric oxide hematite (Gooding, 1978; Boettger et al., 2012). The mineral composition of P-MRS is modeled based on the phyllosilicate-rich sites on Mars, which formed during an aqueous weathering regime with neutral to alkaline conditions in the Noachian epoch (>3.7 Ga), either on the surface or in the subsurface at hydrothermal areas (Tables 1, 2) (Bibring et al., 2006; Chevrier and Mathé, 2007; Halevy et al., 2007; Bishop et al., 2008; Ehlmann et al., 2011). S-MRS represents the Martian sediments with a high sulfate content that presumably correspond to the acidic conditions in the Hesperian epoch (3.7–3.0 Ga) (Bibring et al., 2006; Bullock and Moore, 2007; Chevrier and Mathé, 2007). The final inorganic mineral mixture investigated in this study was the Martian soil simulate MRS07/52. This sample was supplied by German Aerospace Center and was produced to have comparable constituents to that of the soil on Mars (Moeller et al., 2008) (Tables 1, 2).

Chemolithoautotrophic cultivation of M. sedula was performed in DSMZ88 Sulfolobus medium described above in 1 L glassblower modified Schott-bottle bioreactors (Duran DWK Life Sciences GmbH, Wertheim/Main, Germany), equipped with a thermocouple connected to a heating and magnetic stirring plate (IKA RCT Standard/IKA C-MAG HS10, Lab Logistics Group GmbH, Meckenheim, Germany) for temperature and agitation control. Each bioreactor was equipped with three 10 mL graduated glass pipettes, permitting carbon dioxide and air gassing (with the gas flow of 9 mL min^-1, adjusted to five bubbles s^-1 by using 8 mm valves (Serto, Frauenfeld, Switzerland)) and sampling of culture, respectively (Supplementary Figure 1). The graduated pipettes used for gassing were connected by silicon tubing to sterile 0.2 μm filters (Millex-FG Vent filter unit, Millipore, Billerica, MA, United States). The graduated pipettes used for sampling were equipped with a Luer-lock system in order to permit sampling with sterile syringes (Soft-Ject, Henke Sass Wolf, Tuttlingen, Germany). The offgas was forced to exit via a water-cooled condenser (Ochs GmbH, Bovenden, Germany). For the cultivations of M. sedula at 73°C the temperature inside the bioreactors was controlled by electronic thermocouple via the heating and magnetic stirring plates. M. sedula inocula were obtained by resuspending a chemolithoautotrophically grown cell pellet formed by centrifugation at 6000 × g for 15 min in DSMZ88 media without organic carbon source and inorganic metals/metalloids as energy sources. For chemolithoautotrophic growth cultures with initial pH of 2.0 were supplemented with 1 g/liter Martian regolith simulants, no further pH adjustments were introduced during the cultivation. The minerals were temperature sterilized at 180°C in a heating chamber for a minimum of 24 h prior to autoclaving at 121°C for 20 min. Abiotic controls consisting of uninoculated culture media supplemented with MRSs were included in all the experiments. Growth of cells was monitored by phase contrast/epifluorescence microscopy and metal release. Precise cell enumeration was found to be difficult due to the interference with mineral particles of similar size and round shaped morphology and shielding of M. sedula by iron mineral precipitates, especially at later stages of the growth. To visualize wiggling cells on solid particles, a modified “DAPI” (4′-6′- Diamidino-2-phenylindole) staining was used (Huber et al., 1985); afterward the cells were observed and recorded with ProgRes^® MF cool camera (Jenoptik) mounted on Nikon eclipse 50i microscope, equipped with F36-500 Bandpass Filterset (ex, 377/50 nm; em, 447/60 nm).

Cultures of M. sedula grown on Martian Regolith Simulants (Supplementary Figure 2) were fixed in 2% paraformaldehyde (PFA) at room temperature for 1 h and subsequently washed three times in 1x PBS (phosphate buffer saline) with centrifugation steps of 3000 × g for 5 min between each exchange. Cells were extracted from sediment after Braun et al. (2016) as follows: 1 ml sediment from each sample was centrifuged at 5000 × g for 5 min, the supernatant discarded, re-suspended in 1.5 ml Mili-Q that included 0.2 ml methanol and 0.2 ml detergent mix (100 mM EDTA, 100 mM sodium pyrophosphate decahydrate and 1% v:v Tween 80) and shaken at 750 rpm for 60 min. To separate cells from sediment particles samples were sonicated at 30% power three times for 15 s using an ultrasonic probe. Finally, a gradient centrifugation was applied consisting of three 2 ml Nycodenz layers of 30, 50, and 80% on top of a 2 ml sodium polytungstate solution with a 2.23 g per ml density and centrifuged at 5000 × g for 2 h at 4°C. The microbial fraction contained within the supernatant above the sodium polytungstate was extracted with a glass pipette. To further clean the samples from fine sediment particles the gradient centrifugation was repeated a second time in the same manner. M. sedula 16S rRNA phylotype specific probe was designed with the software package ARB (Ludwig et al., 2004) and labeled with 4x Atto488 via Click chemistry (biomers.net GmbH, Ulm, Germany) (see Table 3). Cells were mounted on 10 well Diagnostica glass slides (Thermo Fisher Scientific Inc. Waltham, MA, United States) and MiL-FISH conducted directly on slides with 30% formamide and a 3 h hybridization time (Schimak et al., 2015). Positive control for the specificity of the phylotype specific probe M.sedula_174 was provided by use of the same M. sedula DSM5348 culture published in Schimak et al. (2015). After hybridization, slides were washed for 15 min at 48°C according to Manz et al. (1992) (14 to 900 mM NaCl, 20 mM Tris-HCl [pH 8], 5 mM EDTA [pH 8], and 0.01% SDS) at a stringency adjusted to the formamide concentration used (Manz et al., 1992). DNA staining with DAPI (4 = ,6-diamidino-2-phenylindole) followed (10 mg/ml) for 10 min after which slides were rinsed in distilled water three times. Vectashield (Vector Laboratories, Burlingame, CA, United States) mounting medium was applied and slides closed with a coverslip. Fluorescence images were taken with an AxioCam Mrm camera mounted on an Nikon Eclipse 50i illuminated by a Nikon Intensilight C-HGFI light source and equipped with a F36-525 Alexa 488 (ex, 472/30 nm; em, 520/35 nm) filter cube. Images were recorded with the Windows based AxioVision (release 4.6.3 SP1) imaging software and any image-level adjustments made either therein or using the Mac OS X based Adobe Photoshop version 12.0.4. The figure table depicted was composed using Mac OS X based Adobe Illustrator version 15.0.2 and images cropped by use of clipping masks.

The mineral precipitates (Supplementary Figure 3) were examined with a Zeiss Supra 55 VP scanning electron microscope (SEM), equipped with a spectroscope of dispersive energy (EDS), which was used for imaging and elemental analysis of precipitates. The samples were coated with a thin Au/Pd layer (Laurell WS-650-23 spincoater). The acceleration voltage applied was 5 kV and the EDS analyses were performed with a 120 μm aperture and a counting time of 50 s. In order to control the beam parameters, cobalt was used as a standard. Conventional ZAF matrix correction was used to calculate the final composition from the measured X-ray intensities. All the sample spots investigated by EDS were chosen randomly and each spot was measured three times. Table 4 represents the chemical composition of aluminum/chlorine containing microspheroids, which was taken as the average of the measurements from 20 randomly chosen spots.

To determine the extracellular concentrations of metal ions mobilized from the Martian regolith simulants, culture samples were clarified by centrifugation. Samples of the resulting supernatants were filtered (0.44 μm pore size) and analyzed by inductively coupled plasma-optical emission spectrometer (ICP-OES) Perkin Elmer Optima 5300 DV. All reported values are averages from duplicate samples.

The Electron Paramagnetic Resonance (EPR) spectra were recorded on an X-Band Bruker Elexsys-II E500 CW-EPR spectrometer (Bruker Biospin GmbH, Rheinstetten, Germany) at 90 ± 1 and 293 ± 1 K using a high sensitivity cavity (SHQE1119). Solid state EPR measurements were performed setting microwave frequency to 9 GHz, modulation frequency to 100 kHz, center field to 6000 G, sweep width to 12000 G, sweep time to 335.5 s, modulation amplitude to 20.37 G, microwave power to 15 mW, conversion time to 81.92 ms and resolution to 4096 points. The samples were put in EPR quartz tubes (Wilmad-LabGlass, Vineland, NJ, United States) and scanned three times, of which the average was used for analysis. The spectrum of an empty control tube was subtracted from all sample spectra. All spectra were analyzed with the Bruker Xepr software.

For the statistical analysis and graphical representation of the data the Excel 2016 (version 7.0) and Sigma plot (version 13.0) software packages were used.

Metallosphaera sedula was grown under chemolithoautotrophic conditions on MRSs (Table 5), and interaction of cells with the mineral particles was examined. After 21 days of CO₂-supplemented cultivation on MRSs as the sole energy source, phylogenetic identification and visualization of M. sedula cells was achieved with Multi-Labeled fluorescence in situ hybridization (MiL-FISH). For this purpose a newly designed phylotype specific probe targeting the 16S rRNA (Table 3) was used. The cultures of M. sedula grown on all four investigated in this study MRSs (JSA 1A, P-MRS, S-MRS, and MRS07/52) resulted in positive fluorescent signal with Atto488 labeled probes at a 30% formamide concentration (Figure 1). Only cells that show both DAPI and fluorescent signal were considered as positive hybridization with the target organism. Additionally, it was noted that cells of S-MRS and MRS07/52 cultures occur in mucus bound aggregates which can be inferred as extracellular polysaccharide substances (EPS) known for M. sedula (Auernik et al., 2008) (Figures 1G–L).

Inductively coupled plasma-optical emission spectrometer analysis of composition of major elements mobilized from all tested Martian regolith simulants showed the elevated levels of released S, K, Ca, Mg, Si, and Na in the growth medium (leachate solution) (Figure 2). A further change in trace elements (higher released Mn) occurred in cultures of M. sedula grown on JSA 1A, P-MRS, and S-MRS. Additionally, in case with MRS07/52 leachate solution was characterized by elevated levels of Fe, Ni, and to lesser extent Al in comparison to abiotic control (Figure 2D). The increased level of released Ni ions was also detected in S-MRS grown cultures. The elevated levels of Sr ions were measured in culture supernatants of M. sedula grown on P-MRS, S-MRS, and MRS07/52, while increased soluble Zn was represented in JSC 1A and MRS07/52 grown cultures. The drop in P concentrations was detected in leachate solutions of all the tested MRS. The decrease of total iron observed in M. sedula cultures grown on JSC 1A, P-MRS, and S-MRS could be possibly attributed to the formation of insoluble iron oxyhydroxides which form precipitated mineral phase and therefore are not included in the samples of leachate solution used for the ICP-OES analysis. This phenomenon of ‘lost iron’ has been already previously reported in case of microbial cultivation on extraterrestrial material (Gronstal et al., 2009).

To further study the interactions of M. sedula with the Martian regolith simulants, the surface of these minerals was examined for possible alterations upon M. sedula growth. Inspection by SEM of solid mineral phase in cultures of M. sedula grown on synthetic Martian regolith revealed the presence of sphere-like particles of variable size (0.3–3 μm). These microhemispheroids were represented in cultures of M. sedula grown on JSC 1A, P-MRS, and S-MRS and absent in the corresponding abiotic controls (Figure 3). The results of SEM-EDS analysis indicate that these hemispheroidal morphologies are mainly composed of oxygen, aluminum and chlorine and with no or low carbon content (Table 4 and Supplementary Figure 4). The solid mineral phase withdrawn from S-MRS grown cultures of M. sedula was especially enriched with aluminum/chlorine containing microspheroids, while both biotransformed JSC 1A and P-MRS had a minor occurrence of these particles. Figure 3C and Supplementary Figure 4 show budding microspheroids represented on biotransformed surface of S-MRS. SEM assisted investigations of a solid phase of biotransformed MRS07/52 showed the biofilm layer distributed over the mineral surface (Figure 3G). Such a deposited biofilm layer was absent in the corresponding abiotic control of MRS07/52 incubated in the growth medium at 73°C but without M. sedula (Figure 3H).

Electron Paramagnetic Resonance measurements were performed to (1) identify paramagnetic species of manganese and iron in the different MRSs samples and to (2) investigate the impact of M. sedula on MRSs with a possible effect on the oxidative state of manganese and iron. The four different MRSs mineral mixtures were incubated in cultivation medium in the presence (biotic) or absence (abiotic) of M. sedula. The spectra of untreated (raw) MRSs were investigated as well. The abiotic samples of JSC 1A and P-MRS contained Mn²⁺ as evident by a prominent signal at g = 2.4 with a broad linewidth of >1200 G (Figure 4 and Supplementary Tables 1, 2) typical for abiotic Mn²⁺ samples (Kim et al., 2011; Ivarsson et al., 2015). For JSC 1A, the biotic sample revealed an almost complete decline of the Mn²⁺ signal, while for P-MRS an alteration of the g = 2.4 signal in the biotic sample resulted to the appearance of a broad signal with g = 3.3, indicating dipolar interactions of mixed ionic states (Mn³⁺ and Mn⁴⁺) (Ivarsson et al., 2015). In biotic and abiotic samples of JSC 1A and P-MRS, Fe³⁺ could be identified by the characteristic g = 4.3 and g = 9 signals. A high spin d5 configuration of the Fe³⁺ can be inferred from the positions of these resonance signals. Figure 4B shows that the amplitude of high spin Fe³⁺ g = 4.3 signal is slightly lower in biotic P-MRS than in the corresponding abiotic sample, while the signals g = 11.5 and g = 9.9 measured at 90 and 293 K, respectively, are clearly diminished after M. sedula cultivation. Independent of the incubation conditions, the EPR spectra of all four MRSs samples recorded at 90 K revealed small signals with g-values in the range of 2.0 and 2.7, which were assigned to Fe³⁺ with low spin electron configuration. No significant changes were detected between the biotic and abiotic samples measured at 90 K with regard to the Fe³⁺ in the low spin state. In opposite, a continuously narrowing resonance signal g = 2 was recorded at 293 K in JSC 1A samples after cultivation with M. sedula. The linewidth (ΔH) remained unaltered for raw and abiotic JSC 1A (732 G) and decreased to 600 G after the cultivation with M. sedula (Figure 4 and Supplementary Table 2).

The EPR spectrum of the abiotic S-MRS sample exhibits a prominent resonance at g = 3.2 with a broad linewidth of 2061 G, which is absent in the EPR spectra of all the other MRSs samples measured at 90 K. This specific resonance is obtained from samples containing tetrahedral Mn²⁺ species (Xu et al., 1999). A pronounced decrease of the g = 3.2 resonance signal could be demonstrated for the biotic sample of S-MRS compared to the corresponding abiotic sample. In MRS07/52, Fe³⁺ species in the low and high spin state were identified based on the characteristic EPR signals as mentioned above. No manganese could be detected in MRS07/52. The minor difference in biotic and abiotic MRS07/52 samples occurred at the area of low spin Fe³⁺, which is even more detectable in spectra recorded at 293 K (Figure 4D). The slightly increased amplitude of high spin Fe³⁺ signal along with the shift of g-value from 8.7 for abiotic to 9 for biotic samples was also observed at 90 K. The strong resonance line at g = 4.1 which can be assigned to Fe³⁺ located in orthorhombic positions (Polikreti and Maniatis, 2002) is not altered in biotic and abiotic MRS07/52 samples (Figure 4D).