The B. pseudomycoides sp., B. subtilis and E. coli strains, plasmids and sequences of oligonucleotides used in this study are available in Supplementary Tables 1–3. Plasmids were constructed using standard molecular biology methods (Ausubel et al., 2003) and amplified in E. coli MM294 or DH5α cells. PCR fragments were amplified from B. pseudomycoides NCr1a (Tamindžija et al., 2019) or B. subtilis PY79 chromosomal DNA. The nucleotide sequence of chrA cds and its upstream reagion was extracted from NCr1a genomic DNA using primers complementary to B. pseudomycoides DSM 12442 type strain genomic sequence (NCBI Reference Sequence: NZ_CM000745.1; locus tag “bpmyx0001_46360”; protein_id = “EEM14515.1”). The PCR fragment was sequenced (Eurofins Genomics LLC). The acquired sequence was translated into a protein using Vector NTI software (Invitrogen) and compared against the protein database using NCBI Standard Protein Blast tool, to confirm we extracted the sequence encoding the putative chromate transporter. The translated sequence was also compared to sequences of other previously characterized chromate transporters such as ChrA from B. cereus SJ1 (NCBI Reference Sequence: NZ_CP072774.1; protein_id = “WP_048538372.1”; “GeneID:67509833”), Lysinibacillus fusiformis ZC1 (NCBI reference Sequence NZ_ADJR01000017.1; locus_tag = “BFZC 1_RS06765”; protein_id = “WP_192807144.1”), Ochrobactrum triticii TnOtChr (GenBank: EF469735.1; protein_id = “ABO7 0325.1”), Shewanella oneidensis MR-1 (NCBI Reference Sequence: NC_004347.2; locus_tag = “SO_RS04615”; protein _id = “WP_011071255.1”), Shewanella sp. ANA-3 (NCBI Reference Sequence: NC_008577.1; locus_tag = “SHEWANA 3_RS17310”; protein_id = “WP_041412773.1”), (Pseudomonas aeruginosa pUM505 (NCBI Reference Sequence: NC_01613 8.1; locus_tag = “HS786_RS00455”; protein_id = “WP_0031 17270.1”).

To construct the B. subtilis strain P_chrA-chrA, a 1,414 bp region of NCr1a chromosomal DNA containing the chrA gene together with 317 bp of its upstream region was amplified using chrABP_Up_AccF and chrABP_STOP_BamR primers. This fragment was then cloned into Acc651 and BamHI sites of integrative vector pSG1729. The presence of the insert in the resulting plasmid was verified by restriction and sequence analysis. Resulting pSG17PchrA-chrA plasmid was used for transformation of B. subtilis PY79 as described previously (Feucht and Lewis, 2001). The integration of P_chrA-chrA into amyE locus was confirmed by PCR. To construct B. subtilis ΔywrBA strain, we used the vector pMAD designed for allelic replacement in multiple Gram positive hosts in a way that was described before (Arnaud et al., 2004). Shortly, we used chromosomal DNA of B. subtilis PY79 as a template as well as primers U1000_ywrB_BamF6 and U1000_ywrB_SalR2 to amplify region 936 bp upstream of ywrB gene and cloned the PCR fragment into BamHI, SalI restriction sites of pMAD. Similarly, primers D1000_ywrA_SalF2 and D1000_ywrA_EcoRIR were used to amplify region 948 bp downstream of ywrA gene and the PCR fragment was cloned into SalI, EcoRI restriction sites of pMAD. Primers Kan_new_SalF and Kan_new_SalR2 were used to amplify the sequence of the Kanamycin marker fom pUK19 plasmid template. The PCR fragment carrying promoter together with the coding sequence of kanamycin resistance gene was cloned into SalI restriction site. The final construct, pMAD Up_ywrB:kan:down_ywrA was verified by restriction and sequence analysis. To replace ywrBA operon in B. subtilis PY79 with the kanamycin resistance cassette, the construct pMAD Up_ywrB:kan:down_ywrA was introduced into B. subtilis cells by electroporation (conditions listed in section below). The transformation mixture was plated on Erm and X-gal containing plates and incubated overnight at 37°C. Transformants that appeared blue on X-gal media were the ones containing pMAD construct in the cell as a shuttle. These were pooled and re-streaked onto LB agar media containing X-gal, Erm and Kan, followed by incubation overnight. Liquid LB media containing Erm, Kan were inoculated by blue transformants and incubated overnight at 37^°C with shaking. The overnight culture was used for inoculation of fresh LB to starting OD600 0.4, which was followed by a 2-h incubation at 37°C with shaking. After that, the temperature was raised to a non-permissive 40°C to promote plasmid integration into the chromosome. The culture was incubated at 40°C for 6 h. Eventually, cultures were serially diluted and plated onto LB agar plates containing X gal and Kanamycin. Transformants that appeared white on Kan, X gal containing plates were tested for Erm sensitivity. Clones which were white on X-gal, resistant to kanamycin but sensitive for erythromycin were the ones that completed allelic replacement and lost the plasmid, leaving kanamycin cassette at the locus of former ywrBA operon. The lack of ywrBA operon as well as the presence of kanamycin cassette in the resulting strain was confirmed by PCR and sequence analysis.

To construct the B. subtilis P_chrA-chrA ΔywrBA strain, the aforementioned plasmid pSG17P_chrA-chrA was used to introduce chrA coding sequence under the control of its promoter into amyE locus of ΔywrBA strain. The resulting strain was resistant to kanamycin and chloramphenicol. The presence of amyE:P_chrA-chrA chloramphenicol cassette was verified by PCR using primers Up_chrABP_Prom_AccF, Up_chrABP_Prom_AccR.

To create B. subtilis strain P_chrA-gfp-chrA, the upstream region of chrA gene was amplified using primers Up_chrABP_Prom_AccF, Up_chrABP_Prom_AccR and chromosomal DNA of NCr1a as a template. The product of 315 bp was cloned into Acc651 restriction site of pSG1729, upstream of gfp gene present in the plasmid (Feucht and Lewis, 2001), creating the intermediate pSG17P_chrA-gfp, which was sequenced to confirm the correct orientation and sequence of the upstream chrA promoter containing region. Subsequently, chrA coding sequence was amplified using primers chrABP_GFP_EcoF, chrABP_GFP_EcoR and chromosomal DNA of NCr1a as a template. The product of 1,095 bp, chrA cds was cloned into EcoRI restriction site of pSG17PchrA-gfp intermediate, to create translation fusion with gfp positioned upstream. The orientation of the insert was verified by restriction analysis and resulting plasmid P_chrA-gfp-chrA was sequenced prior introduction into B. subtilis PY79 amyE locus.

The procedure of preparation of electro-competent cells of B. subtilis, as well as conditions for electroporation, was combined according to previously published works (Lu et al., 2012; Zhang et al., 2015). Briefly, electro-competent cells were prepared by incubation of B. subtilis culture in LSBP media (Zhang et al., 2015) at 37°C until OD600 0.85–1.0 was reached. Subsequently, the culture was cooled on ice for 10 min, centrifuged at 4°C and washed 4 times with the electroporation medium (Lu et al., 2012). The cells were resuspended in an appropriate amount of electroporation medium and stored at –70°C. For electroporation, 60 μl of the competent cells were mixed with 10–200 ng/μl plasmid in an ice-cold electroporation cuvette (1 mm gap, Bio-Rad, United States). After incubation for 5–10 min on ice, the mixture was pulsed using a pulser apparatus (Bio-Rad, United States) with parameters 25 μF, 200 Ω, 15 kv/cm. Subsequently, 1 ml of recovery medium was added to the transformation mixture and cells were incubated for 3 h at 37°C with shaking. Eventually, cells were gently pelleted by centrifugation at room temperature, the pellet was resuspended in 100 μl of the supernatant and suspension was plated on selective LB agar media, followed by overnight incubation at 37°C.

For B. subtilis cultivation, a fresh single colony from overnight LBA plate was spread-inoculated onto LB plate and incubated overnight at 30°C. B. pseudomycoides NCr1a cultivation was started by inoculation of a fresh single colony to a liquid LB followed by incubation overnight at 28°C with shaking (150 rpm). Then, the overnight cultures were harvested and re-suspended in 1X Phosphate Buffer Saline (PBS), pH 7.4 (0.137 M NaCl; 0.0027 M KCl; 0.01M Na₂HPO₄; 0.0018 M KH₂PO₄), washed 2 times, and the OD600 was set to 4.0 in 1X PBS. The dense cell suspension was used for inoculation of M9 media (Atlas, 2010), supplemented with L-Tryptophane to a final concentration of 20 μg⋅ml^–1; Thiamin to a final concentration of 1 mM; and Cas-amino acids to final concentration 0.01%, where necessary supplemented with K₂CrO₄ as Cr(VI). The starting OD600 was set to 0.15. Samples were cultivated at 37°C (B. subtilis) or 28°C (B. pseudomycoides NCr1a) with shaking (150 rpm) and their growth was monitored by measuring OD600 in 1 h intervals.

Scanning Transmission Electron Microscopy with energy dispersive spectroscopy (STEM-EDS) was used to obtain elemental composition maps and to compare elemental content in the two examined bacterial strains’ cells. Elemental content was reported in atomic percentages (at%), which is the percentage of one kind of atom relative to the total number of atoms. Data were acquired by using an X-ray EDS system (Esprit/Quantax Bruker) in STEM mode of a microscope FEI Tecnai Osiris [200 kV X-FEG field emission gun, X-ray detector (Super-X) with 4 mm × 30 mm windowless SDD diodes and 0.9 sr collection angle] at following conditions: 200 kV, 1.2 nA beam current as was described previously (Jamroskovic et al., 2016). Quantitative EDS analysis was carried out using the Cliff–Lorimer standard-less method with thickness correction using K-series. The physical Bremsstrahlung background was calculated based on the sample composition. Some elements such as Cu originating from the Cu grid were removed from quantification after the deconvolution procedure in the quantification process. Elemental concentrations in atomic% and net counts (signal above background) were derived from deconvoluted line intensities within a 95% confidence level. The processing time and acquisition rates were adapted to get the most accurate data for a specific element such as Ca, P, Cr, and Mn. The experimental spectra were collected with no pile-up artifacts. A correction for specimen drift was applied during acquisition to improve elemental mapping accuracy.

Strains were cultivated as described above, in M9 media supplemented with ascending concentrations of K₂CrO₄ to a final concentration of Cr(VI) of 0 mM, 0.1, 0.3, 0.5, and 1.0 mM. The growth of strains was monitored by measuring OD600 in 30 min intervals. During each experiment, each strain was grown in a triplicate for each concentration of Cr(VI) tested. The growth curve presented for each strain and concentration of Cr(VI) is a result of median calculated from OD600 values gained in four independent experiments.

Strains were cultivated as described above. After 6 h of exposure to Cr(VI), decimal serial dilutions of the culture grown without Cr as well as of cultures exposed to Cr(VI) were prepared. The selected dilutions were plated onto M9 agar plates in duplicates. The number of CFU for each concentration of Cr(VI) was assessed for each strain, and their viability was expressed as a percentage of CFU counts of the same strain cultivated without Cr. The experiment was performed 3 times in an independent manner.

In all experiments, the strains to be analyzed were grown as described above. For all strains grown in the presence of Cr(VI), cultures grown without Cr(VI) served as negative controls. After the staining with fluorescence dyes, 1 ml of sample was centrifuged at RT, 5,000 rpm, 3 min. Except for Propidium iodide staining, the samples were washed 3 times by 1 × PBS and subsequently were 100 × concentrated. 0.5 μl of the sample was put on poly-lysine-covered slide and subjected to fluorescence microscopy.

For CellROX™ Deep Red (Thermo Fisher Scientific) staining, strains NCr1a and B. subtilis PY79 were grown in M9 media supplemented with 1 mM K₂CrO₄ [Cr(VI)] for 6 h. Thirty minutes prior harvest, the CellROX™ Deep Red reagent was added to 5 ml of culture to reach a final concentration of 5 μM. Images were obtained using the Texas Red filter.

For Live/Dead staining, Rhodamin 123 and Propidium Iodide (Sigma Aldrich) were used. In the case of Rhodamin 123 staining, B. subtilis strains were grown in M9 media supplemented with 1 mM K₂CrO₄ [Cr(VI)] for 6 h. Twenty minutes prior the harvest, Rhodamin 123 was added to cell cultures to a final concentration 0.1μg⋅ml^–1. Images were obtained using the FITC filter. In the case of Propidium Iodide (Sigma Aldrich) staining B. subtilis strains were grown in M9 media supplemented with 1 mM K₂CrO₄ [Cr(VI)] for 6 h. Then PI was added to 1 ml of the sample to reach the final concentration of 1 μg⋅ml^–1 and was left to incubate for 5 min at the RT, followed by centrifugation as described above. No washing by PBS was involved, as this resulted in de-staining. Images were obtained using the DS Red Filter.

For mGFP-ChrA fluorescence microscopy, P_chrA-GFP-ChrA B. subtilis strain was grown as described above in M9 media supplemented with ascending concentrations of K₂CrO₄ [Cr(VI)] for 6 h. Images were obtained using the FITC filter.

Images were acquired using an Olympus BX63 microscope equipped with a sCMOS Zyla-4.2P camera (Andor, Oxford Instruments, Belfast, United Kingdom). Olympus CellP imaging software and ImageJ v1.53f software were used for image acquisition and analysis.

The 200 ml cultures were started as usual, in M9 media supplemented with ascending concentrations of K₂CrO₄, from 0 to 0.5 mM. Pellets were frost in liquid nitrogen and stored at –80°C. Prior use, pellets were concentrated in 2 ml of the Solubilization buffer [20 mM TrisHCl (pH 8), 150 mM NaCl, 1 mM AEBSF] to OD600 = 20. To each suspension, the lysozyme was added to a final concentration of 1 mg/ml and consequently, the cells were lysed by sonication (15.000 kHz, 20 times for 9 s with a 1-min break and constant cooling). 50 μl of the “TOTAL” fraction was collected for each sample, as were fractions “SOLUBLE” and “PELET” following centrifugation at 30,000 rpm; 6°C for 30 min. All samples were mixed with 2 × Sample Laemmli Buffer (4% SDS, 20% glycerol, 10% 2-mercaptoethanol, 0.004% bromphenol blue and 0.125 M Tris HCl, pH. 6.8) and boiled for 3 min before loading 20 μl of each sample to 12% SDS gel. The proteins were blotted onto nitrocellulose membrane by semi-dry transfer (Biometra Fast Blot, Analytik Jena, GmbH) at 205 mA, 10 W, 12 V, for 1 h. After washing the membrane with TBST buffer (20 mM Tris, 150 mM NaCl, 0.1% (w/v) Tween^® 20) and blocking with 5% (w/v) milk in TBST, the membrane was cut in half according to the molecular weight protein standard PageRuler™ Plus Prestained (Thermo Fisher Scientific) bands. The upper half of the membrane containing GFP-ChrA fusion protein (70 kDa) was incubated with the monoclonal mouse anti−GFP antibody [Anti-GFP antibody (9F9.F9) (ab1218); Abcam, Cambridge, CB2 0AX, United Kingdom] diluted to 1:1,000 in 5% (w/v) milk in TBST overnight, at 4°C. The other half of the membrane, containing major RNA-polymerase subunit SigA a member of Sig70 protein family (43 kDa) was incubated with the monoclonal mouse anti-Sig70 antibody [(ab12088) Anti-RNA polymerase sigma 70 antibody (2G10); Abcam, Cambridge, CB2 0AX, United Kingdom] diluted to 1:1,000 in 5% (w/v) milk in TBST for 1 h, RT, as described previously (Strickland et al., 1988; Jovanovich et al., 1989; Breyer et al., 1997). Subsequently, the membranes were incubated with anti-mouse HRP conjugated antibody (Promega) diluted 1:2,500 in 5% (w/v) milk in TBST for 1 h, RT. The Western blot analysis was performed with a modified ECL peroxidase detection kit (Bio-Rad). The signal was developed using RTG films, scanned and subjected to ImageJ analysis. For quantification of Western blots, the adaptation of the Luke Miller protocol was used.¹ The representative image shows the GFP and SigA signal coming from the same sample for each concentration of Cr(VI) tested. All samples were run on the same gel, and blotted to the same membrane. The GFP and SigA signal were developed separately. Altogether, in this study we present the results of three independent Western blots, where the average SigA signals were synchronized so that the GFP-ChrA signals could be compared and plotted together with standard errors. In this process, the average values of the percentage of peaks for the SigA signal in one Western blot were normalized to that of the second, and to that of the third Western blot. All profile plot percentage values were then recalculated with this normalization coefficient.

Images obtained by CellRox™ Deep Red fluorescence microscopy were saved with optimal settings as TIF stacked images by CellP imaging software directly during the experiment. For analysis in ImageJ,² stacked images were set to the same values of brightness, contrast and intensity, and split to a phase-contrast image and fluorescence signal image. All phase-contrast images were adjusted to the same values of threshold, rendered binary and processed with watershed function to discern cells in chains. Such image was used for counting of particles (cells) with particle size characteristics being 250− infinity pixel^2, circularity 0.00–0.8 to fit rod-like Bacillus sp. cells. By this approach total number of cells in each image was obtained. Phase-bright or translucent cells were considered to be dead and were counted manually in each image hence the number of live cells was estimated by subtraction of the number of dead cells from the number of all (total) cells. The image with the fluorescence signal was inverted, and after subtraction of background by Rolling Ball function was processed by Auto Local Thresholding. Up-coming measurements were set to analyze the area taken up by the signal, integrated density of the signal and minimal and maximal mean value of the signal. Selections, where these characteristics were to be analyzed, were defined by the outlines of the signal foci. These were obtained after particle analysis of a binary version of fluorescence image, where the particle was defined as being of size 50-infinity pixel^2, with 0.00–1.00 circularity, to fit the above-noise size and any shape of fluorescence signal. Outlines of fluorescence signal foci were saved as ROIs for each image. Particle analysis set for assessment of the signal area, integrated density and maximal and minimal mean value provided these data for each ROI. Per each type of sample [PY79_0 mM Cr(VI); PY79_1 mM Cr(VI); NCr1a_0 mM Cr(VI); NCr1a_1 mM Cr(VI)] six images were analyzed in total, together representing over 3,000 cells. Overlay of fluorescence signal derived ROIs with the phase-contrast image enabled exclusion of signals coming from dead cells or surrounding media. From these data, the average of integrated density of the fluorescence signal was estimated for each type of sample and was taken as the means to express the overall signal intensity. Similarly, the average of the signal integrated density in the live cells was estimated as well as the number of live cells displaying the fluorescence signal.

Bacteria employ several mechanisms to minimize the impact of Cr on their vital processes. After the limits of these defenses are met, the build-up of intracellular Cr species intracellularly is inevitable.

Using the STEM-EDS technique that allows quantitative mapping of chromium in individual cells, we compared the cellular accumulation of chromium for two species with differing Cr tolerance. The minimal inhibitory concentration of Cr has been previously estimated for both strains (Tamindžija et al., 2019).

The results show that, after exposure to 1 mM Cr(VI) for 4 h, NCr1a strain accumulates significantly less Cr than the strain PY79 of B. subtilis. The accumulation of Cr by NCr1a cells was consistently very low, with little variability between individual cells, specifically 0.03 ± 0.01 atomic% (Figures 1A,C). In contrast, B. subtilis cells showed greater variability in Cr content, making the average content of intracellular Cr significantly higher, specifically 0.27 ± 0.2 atomic% (Figures 1A,C). After 18 h of 1 mM Cr(VI) stress, the amount of Cr accumulated by NCr1a strain increased to 0.09 ± 0.03 atomic% (Figures 1A,C,D). The amount of Cr accumulated by the cells of B. subtilis strain after 18 h remained approximately at the same level, specifically 0.3 ± 0.23 (Figures 1A,C), suggesting that this strain could already have received a lethal dose of Cr after 4 h of exposure and prolonged incubation had no further effect. To assess whether the priming of a cell culture with sub-inhibitory concentration of Cr(VI) could induce a more effective Cr stress response, we treated the exponential-phase cultures of the strains with sub-inhibitory Cr(VI) concentration of 0.01 mM of for an additional hour. Subsequently, cells of both strains were exposed to 1 mM Cr(VI) for 1 h. Results show, that while the induction prior exposure to the high concentration had no effect on Cr accumulation by NCr1a cells (Figures 1B,C), B. subtilis PY79 cells clearly benefited from the pre-treatment (Figures 1B,C).

After Cr(VI) enters the cell, it is reduced by abundant non-specific reducing agents. The intracellular reduction of Cr(VI) is associated with turning molecular oxygen to superoxide and eventually to hydrogen peroxide or hydroxyl radicals (Shi and Dalal, 1990a,b). The degree of intracellular accumulation of Cr(VI) is thus proportional to the concentration of free radicals and the level of oxidative stress in the cell. To assess whether B. subtilis PY79 is subjected to greater oxidative stress than the Cr-resistant B. pseudomycoides NCr1a, we visualized free radicals within cells using the CellROX™ Deep Red Reagent, the fluorescent probe for detection of oxidative stress (see section “Materials and Methods”). Samples of both strains were subjected to fluorescence microscopy after 6 h of growth in the presence of 1 mM Cr(VI) and compared with the control samples grown without Cr. During this experiment, we did not see significant differences in percentage of live cells between Cr-treated or Cr-untreated samples of the same strain, or between the different strains respectively (Figure 2A), however, only translucent phase-bright cells were considered as dead (Pospíšil et al., 2020; Supplementary Table 4). One of the categories based on which we compared the strains was the CellROX™ Deep Red signal intensity (Figure 2A). The total signal intensity was calculated from the means of the values of integrated signal density recorded for each sample (see section “Materials and Methods”). This value includes the CellROX™ Deep Red signal coming from both, the live and the dead cells. In this category, the significant increase was observed only in cells of B. subtilis PY79 strain treated by 1 mM Cr(VI) (Figure 2A and Supplementary Table 4). The dead cells having already succumbed to Cr oxidative stress were naturally giving off substantial CellROX™ Deep Red signal. To get rid of the signal coming from already dead cells, we estimated the percentage of living cells displaying CellROX™ Deep Red signal per each sample; as well as the intensity of the CellROX Deep Red signal coming exclusively from these living cells (Figure 2A). B. subtilis PY79 treated by 1 mM Cr(VI) had a higher percentage of living cells displaying CellROX™ Deep Red signal than any other of the analyzed strains. In addition, also the intensity of the fluorescence signal displayed only by the living cells was the greatest in B. subtilis PY79 strain treated by 1 mM Cr(VI).

Interestingly, the ROS signal seemed qualitatively different in B. subtilis cells compared to NCr1a strain. In NCr1a cells, the CellROX™ Deep Red signals appeared to be exclusively localized in spots or patches enclosed by a membrane-like border and remained in this arrangement also in dead NCr1a cells (Figure 2B). NCr1a strain belongs to B. cereus sensu lato group and as such tends to form polyhydroxyalkanoate (PHA) granules during unfavorable environment conditions (Kihara et al., 2017; Ray and Kalia, 2017; Obruca et al., 2018). PHAs are highly reduced carbon-rich storage biopolymers presumably surrounded by a membrane (Obruca et al., 2018). ROS signal in NCr1a seems to localize to these intracellular structures and hints to their underestimated potential in oxidative stress alleviation.

Altogether, we have observed no significant differences between samples of NCr1a strains before or after exposure to 1 mM Cr(VI), in all the categories analyzed. On the other hand, the PY79 strain of B. subtilis showed major differences between treated and untreated samples in all other categories (Supplementary Table 4 and Figure 2A). Based on these results, we assume that B. subtilis PY79 suffers more from oxidative stress upon Cr(VI) exposure than NCr1a strain. This was in accordance with the STEM-EDS results (Figure 1C) indicating that PY79 accumulates more Cr intracellularly than the environmental B. pseudomycoides NCr1a strain.

Resistance to Cr can be ensured by multiple layers of protection (Nies, 2003). To assess the contribution of Cr efflux to the overall Cr resistance of B. pseudomycoides NCr1a strain, we were engaged in futile efforts to knock out its chrA locus. To be able to study the ChrA efflux from NCr1a, although in heterologous system, we cloned the chrA gene, together with its promoter into the amyE locus of the model organism, B. subtilis PY79. B. subtilis is itself moderately Cr-tolerant as it tolerates time-limited exposure to Cr(VI) in sub-millimolar concentrations and can resume the growth after the concentration of the toxic metal decreases (Garbisu et al., 1998). B. subtilis harbors an innate Cr efflux system in its genome, consisting of two SCHR chromate transporters ywrB (chr3N) and ywrA (chr3C), expression of which is regulated by a Cr-responsive Lrp-type regulator ChrS (Díaz-Magaña et al., 2009; Aguilar-Barajas et al., 2013). In contrast, the Cr-resistant NCr1a strain likely harbors only chromosomally encoded putative LCHR chromate transporter ChrA with no regulator sequences detected in the close vicinity of its promoter region. The putative chrA sequence together with its operator and promoter region was PCR extracted from the NCr1a genomic DNA and sequenced (see section “Material and Methods”). To confirm that the extracted sequence corresponds to the putative chromate transporter, the translation of this sequence was compared against the protein database as well as to the amino acid sequences of previously characterized LCHR chromate transporters (see section “Material and Methods”) from Bacillus cereus SJ1 (He et al., 2010); Lysinibacillus. fusiformis ZC-1 (He et al., 2011), Ochrobactrum triticii TnOtChr transposon (Branco et al., 2008), Pseudomonas aeruginosa pUM505 plasmid (Alvarez et al., 1999), Shewanella oneidensis MR-1 (Gang et al., 2019), and Shewanella sp. ANA-3 strain (Aguilar-Barajas et al., 2008; Supplementary Figure 1). The comparison shows the greatest homology of ChrA from NCr1a strain to a putative ChrA from B. pseudomycoides DSM 12442 and to the previously characterized chromate transporters of B. cereus SJ1 and L. fusiformis ZC1 with 90 and 70% of conserved residues, then to both Shewanella sp. with 59% of conserved residues. The lesser homology is shown to chromate tranporters from P. aeruginosa and O. triticii transposon, with conserved residues homology of 47 and 42% (Supplementary Figure 1 and Supplementary Table 5).

The putative Cr-transporter of B. pseudomycoides and Cr-transporter of B. subtilis belong to different evolutionary lines within CHR super family (Díaz-Pérez et al., 2007), which led us to question whether one could offer better protection than the other.

First, we studied the effect of ChrA expression on Cr tolerance of the resulting B. subtilis strain, referred to as the P_chrA-chrA strain. To assess whether ChrA expression would compensate for the loss of YwrBA, the innate Cr efflux system of B. subtilis during chromate stress, we prepared the strain expressing ChrA in a ΔywrBA background, referred to as strain P_chrA-chrA ΔywrBA. As control strains, we used a parental PY79 strain and a deletion ΔywrBA strain. These strains were subjected to chromate stress represented by varying concentrations of Cr(VI) in the liquid culture. The growth of each strain was recorded as OD600 values over time and was taken as an indicator of tolerance for, or susceptibility to Cr(VI). A mild increase of B. subtilis Cr(VI) tolerance in the presence of 1 mM Cr(VI) was observed when ChrA was expressed in the wild-type background (Figures 3A,B). The ΔywrBA strain displays Cr sensitivity, albeit it still exhibits some growth in the presence of 0.1 mM Cr(VI) (Figure 3C). The inability of ΔywrBA to adapt to other sub-millimolar concentrations of Cr(VI) indicates that chromate efflux transporters might be the key determinants of survival under sub-millimolar Cr(VI) stress for B. subtilis (Figures 3A,C). On the other hand, wildtype PY79 and the ΔywrBA strain display similar growth in the presence of 1.0 mM Cr(VI), suggesting that, for PY79 wild-type strain, there is no effective mechanism to fend off Cr stress at this Cr(VI) concentration. Expression of ChrA in the ΔywrBA background results in elevated Cr tolerance across the spectrum of all tested Cr(VI) concentrations, especially at 1 mM Cr(VI) (Figures 3B–D). Hence, ChrA from the environmental B. pseudomycoides strain NCr1a is not only able to compensate for the absence of the chromate transporter operon ywrBA, but seems to increase Cr tolerance at concentrations of Cr(VI) otherwise inhibitory to the parental PY79 strain (Figures 3A,D). The doubling time (Td) was estimated from the exponential phase of growth of each strain for each level of Cr stress tested (see section “Materials and Methods”) (Figure 3E and Supplementary Figure 2). The shortest doubling time across all concentrations of Cr(VI) is associated with the P_chrA-chrA ΔywrBA strain, whereas the doubling time of P_chrA-chrA strain is shorter at 1 mM Cr(VI) than that of PY79 strain (Figure 3E and Supplementary Table 6). The differences between doubling times of strains were found to be significant by statistical testing using Two-way Anova without replication (Supplementary Table 7). From all strains tested only P_chrA-chrA ΔywrBA displays statistically significant differences in T_d across the concentrations of Cr(VI) used when compared to the parental ΔywrBA strains as well as PY79 strain (Supplementary Table 8). This approach, however, only tells how fast the mass is doubled in the fastest phase of growth, and doesn’t consider other phases, such as the length of the lag phase, the stationary phase, or the onset of decline.

It is well established that Cr stress leads to damage of DNA, lipids and proteins (Roundhill and Koch, 2002), thus the defects in cell division in susceptible strains are inevitable. The method based on measuring OD600 over time, however, partially covers up these defects since both, accumulation of debris as well as elongation of cells (instead of division) add up to overall optical density of the culture.

Thus we tested the colony outgrowth of strains subjected to different levels of Cr(VI) stress, to discern the successful division events. The method used is based on the count of viable colonies formed after cultures were exposed to Cr(VI) for 6 h, serially diluted and plated onto minimal agar media without Cr (see section “Materials and Methods”). Strains grown without Cr(VI) and treated afterward in the same manner, served as a reference. The ability of cells to successfully divide and their progeny to grow is referred to as the “fitness” of strains in this study and was expressed as the percent CFU to that of the reference strain (Figure 4). Results show that even the lowest concentration of Cr(VI) (0.1 mM) affects the ability of PY79 strain to divide and grow. With increasing concentrations of Cr(VI), the number of colonies decreases dramatically, which is likely attributable to intracellular accumulation of Cr(III) and/or overall damage linked to oxidative stress. On the other hand, we see that strains expressing ChrA fare better when exposed to 0.1 mM Cr(VI) than parental strains PY79 and ΔywrBA. These results suggest that the expression of chrA under the control of its native promoter increases Cr tolerance in cells exposed to sub-inhibitory concentrations of Cr at least during the first 6 h from the start of exposure (Figure 4). After exposure of strains to the higher Cr(VI) concentration (0.3 mM), the ChrA expression in P_chrA-chrA strain improves outgrowth of colonies. However, this is not the case for the P_chrA–chrAΔywrBA strain, which displays zero outgrowth at this level of Cr stress but shows some outgrowth at 0.5 mM (Figure 4).

Overall, during both the growth and the outgrowth experiments we observed the positive effect of ChrA expression on the survival of B. subtilis strains exposed to Cr(VI) stress.

To visualize the effect of Cr(VI) on vital processes and at the same time assess cell survival, we employed Live/Dead fluorescence staining of bacterial cultures exposed to Cr stress (see section “Materials and Methods”) and fluorescence microscopy analysis. This approach provides us with direct insight into the cell-specific response to toxic stress, focusing especially on morphological changes and the effect of ChrA expression on the Live/Dead cells ratio. The principle of the staining is the diverse affinity of fluorescent stains to lipid membranes. Lipophilic Rhodamin 123 stains cells with intact membranes, but water-soluble Propidium Iodide is able to enter cells only when the cytoplasmic membrane is breached (Jepras et al., 1995). We compared samples of the parental PY79 and the P_chrA-chrA strains. Since the differences were the most obvious after exposure to extreme Cr stress, the strains were cultivated in media with 1 mM Cr(VI) for 6 h. As a result of Cr stress, both strains showed elongated and otherwise misshapen cells, indicating impairment of cell division (Figure 5A). Live/Dead staining revealed that even some phase-dark cells were stained by PI, indicating that they were dead although appeared alive. According to the statistics from Live/Dead fluorescence microscopy imaging, the parental PY79 strain displays more than 50% dead cells after cultivation with 1 mM Cr(VI). On the other hand, P_chrA-chrA strain shows only 13% of cells as dead, after the same treatment (Figure 5B).

To assess whether ChrA is stably expressed in B. subtilis PY79 and to monitor its expression during increasing Cr(VI) concentrations, we prepared N-terminal in-frame fusion of mgfp to the chrA coding sequence. The fusion was placed under the control of chrA upstream region containing the putative P_chrA promoter and introduced into the B. subtilis amyE locus. The expression of the GFP-ChrA in the resulting P_chrA-GFP-ChrA strain was monitored using fluorescence microscopy and anti-GFP IgG Western blotting (see section “Materials and Methods”). GFP fluorescence microscopy revealed that with increasing concentration of Cr(VI) the length of the cells also increases, indicating the inhibition of cell division. More importantly, the intensity of the GFP signal grows with the increasing concentration of Cr(VI) as well (Figure 5A). To be able to quantify the amount of GFP-ChrA expressed for each concentration of Cr(VI), P_chrA-GFP-ChrA lysates were subjected to anti-GFP Western blotting. To obtain a reference standard for each sample, the concentration of which does not vary with changing conditions, we chose to simultaneously visualize the housekeeping protein SigA of B. subtilis using anti-Sig70 antibody (see section “Materials and Methods”). This set-up allowed quantification of the relative amounts of SigA in each sample and normalization to the amount of GFP-ChrA for each sample. The ImageJ software was used for quantification of protein bands (see section “Material and Methods”). Western blot analysis of P_chrA-GFP-ChrA using anti-Sig70/anti-GFP IgG suggests that there is only a minimal amount of GFP-ChrA expressed in cells grown in the absence of Cr (Figures 6B,C). For concentrations of Cr(VI) from 0 to 0.3 mM it seems that with increasing concentration of Cr(VI), the GFP signal is also intensified (Figures 6A–C), indicating the inducibility of the P_chrA promoter by the presence of Cr(VI). Considering that ChrA from B. pseudomycoides and YwrBA from B. subtilis are chromate transporters belonging to different families, the mechanism of Cr efflux activation by the presence of Cr itself is likely conserved in these species. According to the GFP signal, however, the concentration of the fusion protein in cells doesn’t seem to increase any further at 0.5 mM Cr(VI). Either the expression of the fusion is stalled or even declines or the protein starts to be degraded at this concentration of Cr(VI).