After liquid aspiration, a piece of the basalt was aseptically removed from the tubes and analyzed with Scanning Electron Microscopy (see Section SEM). Next, 5 ml filtered mineral water was added and the solution was probe sonicated for 3 min at 20 kHz, 4 W at low-intensity to release the biofilm and intact cells from the biofilm from the basalt. The protocol for sonication was determined before the flight experiment based on previously described protocols (Kobayashi et al., 2009) and was found to be optimal at the conditions described above. After sonication, the solution was again analyzed as described in the post-flight analysis.

Samples were stained and analyzed with flow cytometry to analyze physiology and impact of space flight conditions on the CH34 cells. This was done according to the optimized procedures, described in Buysschaert et al. (2016), Van Nevel et al. and SLMB recommendation for characterizing drinking water communities (SLMB, 2012; Van Nevel et al., 2013; Buysschaert et al., 2016). Cell suspensions were diluted 10,000 times in 0.2 μm filtered Evian mineral water, as this gave the lowest background fluorescence. Next, the different dyes were added and cell suspensions were incubated at 35°C. The tested dyes include DiBAC₄(3) (Sigma Aldrich, U.S.A.), cFDA (Sigma Aldrich, U.S.A.), SYBR Green (Sigma Aldrich), and PI (Sigma Aldrich). The different dye concentrations and incubations times are shown in Table 3. Work solutions were all prepared in DMSO and kept at 4°C. These were prepared from stock solutions in DMSO kept at −20°C. For cFDA and DiBAC₄(3) cells were centrifuged and washed with Evian water before analyzing the samples with flow cytometry to eliminate background signals from the staining solution. For the other dyes, stained cell solutions were analyzed directly after incubation and the cells were not washed.

Stained bacterial suspensions were analyzed on an Accuri C6 (BD, Erembodegem) with a blue (488 nm, 20 mW) and red (640 nm, 14.7 mW) laser which was calibrated according to the manufacturer's recommendation. Standard optical filters were used and included FL-1 (530/30 nm), FL-2 (585/40 nm), and FL-3 (670 LP) for the blue laser and FL-4 (675/25 nm) for the red laser. The dyes DiBAC₄(3), cFDA, and SYBR Green were all detected with FL-1, PI was detected on FL-3. A quality control with 6 and 8 peaks fluorescent beads (by manufacturer BD, Erembodegem) and cleaning cycle was performed prior to experiments to assess both the accuracy (bead count and position) and the cleanliness of the machine. Samples were analyzed using the Accuri C6 software (version 1.0.264.21). The flow files were uploaded to FlowCyte database (FR-FCM-ZYZQ).

The different dyes are used to test for different physiological parameters and determine the effect of the flight conditions on CH34. Sybr Green (SG) was used to stain all cells as this stain can enter both intact and damaged cells independent from their physiological state. SG enters the cells in both permeabilizated and non-permeabilizated cells due to its positive charge which allows it to pass it through the membrane (Veal et al., 2000; Berney et al., 2007, 2008). DiBAC₄(3) staining allows differentiating between cells with an intact membrane potential and normal polarization and cells which have a depolarized membrane or lost membrane function as DiBAC₄(3) can only enter the latter due to its anionic structure. Once in the cell it binds to positively charged proteins or hydrophobic regions. No binding of DiBAC₄(3) with outer membrane structures is observed. Increased depolarization of the membrane also causes more influx of DiBAC₄(3) while hyperpolarization causes a decrease in influx and thus fluorescence (Berney et al., 2008; Muller and Nebe-Von-Caron, 2010; Sträuber and Müller, 2010). cFDA staining can passively enter cells and is metabolized by esterase enzymes in the cytoplasm to its fluorescent product cFluorescein (cF) that accumulates in the cytoplasm and is thus an indicator for enzymatic cell activity. Negative charges present in cF ensures that it is better retained in the cell, avoiding leakage of the product as well as reducing background signal by unspecific binding (Sträuber and Müller, 2010).

To obtain an idea about cell morphology and biofilm structure Scanning Electron Microscopy (SEM) was performed. Five microliters of the solution as well as a piece of basalt was taken and transferred to a 0.2 μm filter (Milipore) and fixed two times with fixation solution [3% gluteraldehyde (w/v) in 0.15 M cacodylate solution, pH 7.6] in filter holders. Between each fixation step, the membrane was left to dry for 20 min at room temperature. Afterwards the filter surface was washed three times with the wash solution (0.15 M cacodylate solution). Next, filter holders were wrapped with parafilm and stored overnight at 4°C to let the filters dry. The next day, the filter surface was rinsed with ethanol, in ascending concentrations (30, 50, 70, 90, 95, 100% v/v), once for each concentration. Between each step the filters were left to dry for 10 min. The final rinsing with 100% ethanol was done three times. To dry the surface completely, the ethanol solution was replaced with hexamethyldisilazane (HMDS), and the filters were again rinsed three times, with 10 min of incubation between each step. Next, the filters were air-dried in a desiccator for storage. For visualization, these filters were taped onto an aluminum stub using carbon tape and coated with gold particles using the ScanCoat machine (2 × 300 ms, 6–8 mbar argon; 50 mA plasma tension). They were directly thereafter visualized with the JEOL JSM6610LV SEM Microscope with a W filament.

To measure intracellular ATP levels the BIOTHEMA intracellular ATP kit HS was used and adapted from the manufacturer's protocol for smaller volumes. Cell suspensions were diluted 1/100 and 25 μl of this dilution was added in a cuvette. Twenty-five microliters of ATP eliminating agent was added and the cuvette was left to incubate for 10 min to degrade all the extracellular ATP. Twenty-five microliters of cell lysis “extractant BS” was added to the cell solution, vortexed and immediately thereafter 200 μl “ATP reagent HS” was added. Luminescence was measured immediately after (Kikkoman Lumitester c-100). Five microliters of a 100 nmol/l ATP standard was added as an internal control and light emission (luminescence) was measured again. From the overall amount of ATP (in pmol) measured in the samples, i.e., the value I, then the average intracellular ATP concentration per cell was calculated as follows: (IsampleIstandard+sample-Isample)total cell count (with SG in flow) in measured sample

To measure intracellular PHB levels, a Nile Red staining protocol was developed, based on the work of Degelau et al. (1995), for C. metallidurans CH34 and adapted to our test conditions. Nile red binds selectively to non-polar lipid droplets inside cells and can be used to detect the presence of storage lipids (PHA/PHB) via fluorescence spectrophotometry (Greenspan and Fowler, 1985; Johnson and Spence, 2010). A Nile red working solution (200 mg/l) was prepared from a 1 g/l Nile red stock solution in DMSO, stored at −20°C. Ten microliters of the cell solution was diluted in 100 μl sterile Evian mineral water in a micro titer plate (MTP). Five microliters of the work solution was added and the MTP was incubated for 30 min at 30°C. Water without cells but with dye was used as a negative control and water without cells and without dye was used for background correction (blanco). Afterwards fluorescence of the lipid-bound Nile red was determined with a Thermo Scientific™ Fluoroskan Ascent™ Microplate Fluorometer. Nile red solution has an excitation peak at 544 nm and an emission peak at 590 nm. The normalized amount of intracellular PHB per cell was then calculated with following formula: (FI590 nm, 10 μl sample-FI590 nm, negative control-FI590 nm, blanco)OD600 nmtotal cell count (with SG) in measured sample

At the end of the experiment, the in-organic element concentrations in the water were measured to compare changes in water chemistry induced by the rocks and cells during the experiment. Four milliliters of the supernatant from the vials without basalt fragments was taken from the tubes and centrifuged (10,000 × g; 15 min, 20°C) to pellet the cells. This supernatant was then filtered through a 0.22 μm filter to remove particles and 20 μl of 70% nitric acid was added before final inorganic element concentrations in the cell-free supernatant were determined by ICP-OES (Inductively Coupled Plasma Optical Emission Spectroscopy). Details on the wavelengths used for ICP-OES analysis are provided in Supplementary Table S1. For concentrations below the detection limit, values were changed to zero and analyzed in the subsequent statistical analysis as such.

To assess cell viability and cultivability after space flight, a serial dilution of the cell suspensions was plated on R2A and 284 MM agar (Figures 2A,B respectively). Cultivability significantly decreased (p < 0.0001) in the space flight experiment, compared to the initial cell number, irrespective of the presence of basalt, leading to 50% of the population remaining cultivable. Statistical analysis have been summarized in Table 4. A greater loss of cultivability was observed for the ground experiment with only 1 and 0.1% of the population still being cultivable, with and without basalt, respectively. These results indicated that cultivability decreased significantly less during space flight compared to the ground control and that the presence of basalt positively affected cultivability of the ground control.

The impact of space flight conditions and the presence of basalt on physiological changes within the cells were analyzed with flow cytometry by staining with specific functional dyes. The total cell numbers, measured with SYBR Green (SG), decreased slightly but significantly in samples without basalt (p < 0.0002) compared to the initial cell number as well as in the ground sample with basalt (p < 0.0336). When ground and flight samples were compared, flight samples with basalt significantly differed from the ground samples (p < 0.003, Figure 2C). When intact cell numbers were compared (by measuring SG positive cells when stained with both SG and PI, Figure 2D) to the initial setup (with 99% of the population intact in the initial setup) only the flight sample without basalt differs significantly, containing 60% intact cells (p < 0.0001). In addition, these samples also were significantly different from the flight samples with basalt (86%) (p < 0.0001) and ground with and without basalt (90%; p < 0.0001). Permeabilizated fractions (measuring the PI positive cells when stained with both SG and PI, Figure 2E) increased significantly in all test conditions (p < 0.0001) compared to the initial (contained 1% permeabilizated cells). Both flight conditions with and without basalt and ground control without basalt contained more permeabilizated cells then the ground with basalt, which only contained 5% permeabilizated cells. The number of cells stained with DIBAC₄(3) (Figure 2F), related to cells that lost their membrane potential, significantly increased (p < 0.0014) in all conditions compared to the initial amount in the initial culture (3%). Less membrane potential-defected cells were observed in the flight samples with basalt (16%) significantly differing (p < 0.0078) from the ground samples and flight samples without basalt, indicating that the presence of basalt and flight conditions reduced the number of cells that lost their membrane potential. Cell activity measurements, with cFDA (Figure 2G) showed that the number of active cells in the flight experiment with basalt did not significantly differ (p > 0.1647) from the initial setup (35%), in contrast to the other samples (p < 0.0095) with decreased numbers of active cells. Basalt also increased the number of active cells in the flight sample with basalt (26%) compared to other conditions.

Intracellular ATP was measured to determine the energy status of the cells (Figure 2H). When ATP levels were compared with the initial culture (p < < 0.0001), ATP levels decreased 3-fold in the stored cells in water, in all conditions. Flight had a significant effect (p < 0.0439) on ATP levels with flight samples containing more ATP/cell then the ground samples. Basalt did not impact ATP levels, neither for ground nor flight samples. In addition, the energy stock of the cells was estimated by measuring PHB content (Figure 2I). The PHB content per cell significantly increased after the 3 months of storage of the cells in water (p < 0.005), in all conditions, compared to the initial culture and on average doubling PHB concentration. No significant difference (p > 0.0792) was seen between ground and flight samples except for the ground samples without basalt (p < 0.0376) which contained more PHB then the ground and flight samples with basalt.

SEM analysis showed biofilm formation on basalt under flight conditions and in the ground control experiment, although the biofilm formed during flight conditions was less developed (Figure 3). On the ground the basalt surface is completely covered with biofilm cells, while for the flight samples some basalt surface can still be seen (Figure 3, flight). The level of biofilm formation in the ground control samples is comparable to that observed in preflight preparation experiments with the same setup (showed in Supplementary Figure S1).

When biofilm cells were cultured on R2A and 284 MM agar (Figures 4A,B respectively), the viable count for flight was significantly higher (p < 0.0058) than the ground samples. The number of cultivable cells in the biofilm fraction was also 1–1.4 log lower than in the planktonic cell fraction, indicating a significant impact (p < 0.0003) of the biofilm mode of growth on cultivability.

The total biofilm cell number was significantly lower (p < 0.0004; Figure 4C) in the flight samples compared to the ground. Flight samples also contained significantly more permeabilizated cells (p > 0.004; Figure 4E) and less intact cells (p > 0.0001; Figure 4D). No significant differences were observed for activity (Figure 4G) and membrane potential (Figure 4F) between flight and ground conditions.

No significant difference was observed between the intracellular ATP levels nor the PHB content of biofilm cells from the flight and ground experiment (Figure 4H). In contrast, the ATP content of the biofilm cells is 10x lower than the ATP content (p < 0.0006) of the planktonic cells while the intracellular PHB content does not significantly differ from the planktonic cells in suspension (Figure 4I).