The bacterial strains and plasmids used in this work are listed in Table 1. The highly In resistant strain Rhodanobacter sp. B2A1Ga4 (Caldeira et al., 2021) was isolated from Panasqueira mine, Portugal. The growth of Rhodanobacter strains was performed using Reasoner’s 2A broth medium (R2Ab), containing the following per liter: 0.5 g yeast extract, 0.5 g proteose peptone, 0.5 g casein, 0.5 g glucose, 0.5 g soluble starch, 0.3 g K₂HPO₄, 0.024 g MgSO₄, and 0.3 g sodium pyruvate. The growth of Escherichia coli, for plasmid construction, was performed in Luria-Bertani medium (LB), containing the following per liter: 10 g tryptone, 5 g yeast extract, and 5 g NaCl. Solid media were prepared with addition of 15 g per liter of agar to the liquid media. The growth of Rhodanobacter strains was evaluated by measuring the optical density at 600 nm (OD₆₀₀) after different incubation times at 25°C. Metal stock solutions were prepared at concentrations of 0.5 M indium(III) chloride (InCl₃) (Acros Organics), 0.2 M Ga(III) nitrate (GaN₃O₉) (Alfa Aesar), 0.2 M aluminum chloride (AlCl₃) (Acros Organics), 0.1 M nickel(II) chloride (NiCl₂) (Merck), 0.2 M neodymium(III) chloride (NdCl₃) (Alfa Aesar), 0.05 M yttrium(III) chloride (YCl₃) (Sigma), 0.05 M scandium(III) chloride (ScCl₃) (Alfa Aesar), 1 M sodium arsenate dibasic (Na₂AsO₄) (Sigma), 0.2 M lanthanum(III) chloride (LaCl₃) (Alfa Aesar), 1 M cooper(II) chloride (CuCl₂) (Merck), 1 M zinc chloride (ZnCl₂) (Sigma), 0.25 M sodium tellurite (Na₂TeO₃) (Sigma), and 1 M sodium chromate (Na₂CrO₄) (Sigma) and were sterilized by filtration.

An In susceptible B2A1Ga4 strain was obtained by random mutagenesis, by mobilization of the suicide plasmid pSUP5011 from the donor strain E. coli S17-1 to the recipient strain B2A1Ga4, using the filter plating methodology as described previously (Caldeira et al., 2021). The transconjugants grown on R2A plates supplemented with 100 μg/mL kanamycin and 50 μg/mL phosphomycin at 25°C were replicated to R2A media with 1 mM In. The clones unable to grow in the presence of this concentration of In for 5 days were recovered and subjected to further analyses.

The interrupted gene in mutant strain was analyzed by inverse polymerase chain reaction (IPCR) as described previously (Caldeira et al., 2021). Briefly, 1 μg of total bacterial DNA was digested with different restriction enzymes. Ligation reactions were prepared using 0.5 μg of digested DNA with 10 U of enzyme T4 ligase (ThermoFisher Scientific) in a final volume of 50 μL and incubated overnight at 16°C. After ligation and purification, PCR mixtures were prepared in a final volume of 50 μL containing 0.5 μg ligated DNA, 0.04 U/μL Platinum™ Taq DNA polymerase (Invitrogen, ThermoFisher Scientific), PCR buffer (1×), 1.5 mM MgCl₂, 2% KB extender, 0.3 mM nucleoside triphosphates (dNTPs), and 0.4 μM of each primer (forward tn5_1775f: CCT TGC GCA GCT GTG CTC GAC GTT GTC ACT and reverse IR2: CGG GAT CCT CAC ATG GAA GTC A). DNA amplification was performed for 4 min at 94°C, followed by 30 cycles of 45 s at 94°C, 45 s at 60°C and 6 min at 68°C. The IPCR products were visualized on 1% agarose gel; the DNA bands were removed, purified using the E.Z.N.A. ^® Gel Extraction Kit (Omega Bio-Tek), and sequenced (STAB VIDA, Portugal). The sequences were submitted to BLAST in order to identify the interrupted gene by the transposon insertion.

Alignment of YqaA protein sequence of Rhodanobacter sp. B2A1Ga4 with representative sequences of the DedA protein family was generated by ClustalW (Thompson et al., 1994). The selection of the representative DedA proteins from the NCBI database was based on previous studies that gave some insights about the possible roles, or evolutionary relationships among the DedA domain containing proteins (Supplementary Table 1). Protein location and secondary structure predictions were obtained by using the server PSORTb v 3.0.3 (Yu et al., 2010) and both programs TMHMM server 2.0 and CCTOP (Constrained Consensus Topology Prediction) (Krogh et al., 2001; Dobson et al., 2015).

The minimum inhibitory concentrations (MICs) for In were determined using the standard broth microdilution method in R2Ab medium for native and mutant B2 strains, and in R2Ab medium supplemented with gentamicin for complemented strains, B2_p and B2_pyqaA. Bacterial growths were analyzed after 48 h of incubation at 25°C. MIC values were the lowest concentration of In that inhibited a visible growth of the microorganism.

The resistance of the mutant strain to In and to other metals from the same group of the periodic table (Ga and Al) was compared with the wild strain, through the evaluation of their bacterial growth curves in presence of these metals. Both strains were grown in R2Ab medium in four different conditions: control (without metal), 0.4 mM Ga, 0.4 mM Al, and 0.2 mM In. Growths were performed at 25°C, 140 revolutions/min (rpm), and evaluated measuring the OD₆₀₀ for 24 h.

The ability of B2A1Ga4 and mutant B2 strains to grow in presence of other metals, antibiotics, SDS, and at different temperatures was also evaluated to assess the potential effects of the gene mutation under different conditions. The resistance of both strains to metals and SDS at different concentrations was evaluated in R2Ab medium, in triplicate, measuring the OD₆₀₀ after 48-h incubation at 25°C. The antibiotic resistance profiles were evaluated using the standardized disc susceptibility testing method (Jorgensen and Turnidge, 2007), and results were obtained after 48 h of incubation at 25°C. Finally, the temperature assays were performed by growing both strains in R2A medium at different temperatures (20, 25, 30, 37, 40°C) for 5 days.

The complemented strains B2_pyqaA and B2_p were grown in R2Ab medium supplemented with gentamicin (15 μg/mL) at 25°C, 140 rpm, without or with different concentrations of In (0.1 and 0.2 mM In), and In resistance was evaluated measuring OD600 at two times of incubation, 16 and 24 h, which correspond to the end of exponential growth phase and to the stationary growth phase, respectively.

MTT assay (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide assay) was used to evaluate the cellular metabolic activity based on the protocol previously described (Wang et al., 2010). Briefly, samples from growths in the presence and absence of In were collected at two different times. Samples were centrifuged and washed twice, and the pellets were resuspended in 1 mL R2Ab medium. The cell suspensions were diluted to an OD₆₀₀ of 0.2 with the medium. Reactions were prepared with 200 μL of the diluted cell suspensions and 20 μL of MTT stock solution (5 g/mL) and were incubated with the cap open at 25°C for 1 h. After the incubation time, the mixtures were centrifuged at 10,000g for 2 min, and the supernatant was discarded. The pellets (crystals) were dissolved in 2.5 mL of dimethyl sulfoxide (DMSO), and 1 h later, the samples were quantified spectrophotometrically at 550 nm.

The membrane potential was measured according to a described protocol (Sikdar et al., 2013). Strains were grown in presence of 0.2 mM In and without metal for 10 h at 25°C and 140 rpm. Samples of 2 mL of growths with OD₆₀₀ of approximately 0.8 were collected (the sample volumes of growths with lower OD₆₀₀ were suitably adjusted). The pellets were resuspended with the working solution JC-1: 16 μL DMSO, 4 μL JC-1 stock solution (5 mg/mL), and 1 mL of permeabilization buffer (10 mM Tris, pH 7.5, 1 mM EDTA, 10 mM glucose). After 20-min incubation at 30°C (in the dark), the fluorescence was read: green fluorescence (λ_em = 530 nm and λ_ex = 485 nm) and red fluorescence (λ_em = 595 nm and λ_ex = 485 nm). Control experiments were performed with cells previously treated with 100 μM (3-chlorophenyl) hydrazonomalononitrile (CCCP) for 30 min. The results were obtained by the ratio fluorescence (red/green).

To quantify the total of In accumulated by cells, strains were grown in R2Ab medium supplemented with 0.1 mM In at 25°C with 140 rpm, and cells were collected at different growth times. Samples were centrifuged at 4,000 rpm for 20 min at 4°C, the cellular pellets were washed twice with cold phosphate-buffered saline solution (containing the following per liter: 8 g NaCl, 0.2 g KCl, 1.44 g Na₂HPO₄, 0.24 g KH₂PO₄, pH 7.4), and the final cellular pellets were lysed by acid treatment (5% HNO₃), heated at 50°C for 1 h, and then centrifuged at 13,000 rpm for 10 min. The In amount in the intracellular supernatants was quantified by inductively coupled plasma mass spectrometry (ICP-MS). Pellets were neutralized with NaOH 0.5 M and then used to quantify the total protein concentration by Bradford method (Bradford, 1976).

In addition, in specific assays, In was also quantified to differentiate the intracellular and extracellular In accumulation. Native and mutant strains were grown in presence of 0.1 mM In; samples were collected, centrifuged, and washed as previously described. The cellular pellets were resuspended with a solution of 20 mM EDTA and incubated for 15 min with 100-rpm agitation (Hudek et al., 2009). The samples were centrifuged, and the supernatants were used for quantification of In by ICP-MS, which correspond to the cellular surface metal binding fraction (biosorption fraction). The pellet was lysed by acid treatment (5% HNO₃), heated at 50°C for 1 h, and then centrifuged at 13,000 rpm for 10 min. These supernatants were quantified by ICP-MS that correspond to intracellular metal fraction.

The mutant B2 was complemented with the empty vector pBBR1MCS-5, which served as a control, and with this same vector carrying the yqaA gene of B2A1Ga4 strain.

The yqaA gene was amplified from strain B2A1Ga4. PCR reaction consisted of a mixture of 2 ng B2A1Ga4 DNA, 0.04 U/μL Platinum™ Taq DNA polymerase (Invitrogen, ThermoFisher Scientific), PCR buffer (1×), 1.5 mM MgCl₂, 0.3 mM dNTPs, 0.4 μM of specific forward and reverse primers comprising the enzyme restriction sequences for HindIII and XbaI, respectively, and sterile ultrapure water until the final volume of 50 μL. The PCR program was 4 min at 94°C, followed by 30 cycles of 1 min at 94°C, 1 min at 60°C, and 1 min at 72°C.

To clone the amplified gene into the plasmid pBBR1MCS-5, the PCR product and the plasmid were digested with the restriction enzymes HindIII and XbaI (Takara) for 2 h at 37°C. After purification of the restriction reactions, the gene was ligated to the digested vector through T4 ligase enzyme (ThermoFisher) to obtain pyqaA plasmid. The ligation reaction was purified with the E.Z.N.A. ^® Gel Extraction Kit (Omega Bio-Tek) and transformed into competent E. coli S17-1 cells. A clone from plates of LB agar with 100 μg/mL ampicillin and 15 μg/mL gentamicin was selected to colony-PCR reaction, and the PCR product was used to confirm the correct construction by sequencing (STAB VIDA, Portugal).

The pyqaA construction and the empty plasmid were mobilized into the B2 mutant by biparental separated conjugations (del Campo et al., 2012). The conjugated clones were obtained on plates of R2A supplemented with 100 μg/mL kanamycin and 15 μg/mL gentamicin incubated at 25°C for 3 days. A clone designated B2_p and a clone designated B2_pyqaA were selected for further assays.

A scanning electron microscopy (SEM) equipment, Vega3 from Tescan GmbH, was used to observe possible morphological surface cell changes in strains when exposed to In. Native and mutant cells were grown in R2Ab medium in presence of 0.1 mM In. After 24 h of incubation, the OD600 was measured, and bacterial suspensions were prepared to a final OD600 of approximately 0.3; 10 μL of these bacterial suspensions was placed on sterilized surfaces, fixated with 2.5% (vol/vol) of glutaraldehyde, and dehydrated by a grade of ethanol incubations [70, 80, 90, 95, and 100% (vol/vol)]. The bacterial suspension was shaken for 20 min in the glutaraldehyde solution and in each ethanol solution (Kaláb et al., 2008). Prior to observations, the samples were sputtered coated with a thin film of 10 nm of gold to eliminate the charging effect. Each sample was observed in three different areas with an accelerating voltage of 20 keV, in the secondary electron mode.

Each result is indicated as the mean value of two or three independent experiments (number of independent experiments is indicated in the caption of each figure) ± the standard derivation. The statistical analysis of all results was performed using GraphPad Prism version 5.00 for Windows (GraphPad Software, 2007), using two-way analysis of variance followed by Tukey multiple-comparisons test.

Approximately 1,500 random mutants of strain B2A1Ga4, generated by Tn5 transposon mutagenesis, were tested for their ability to grow on R2A plates containing 1 mM In. One clone unable to grow in solid medium with 1 mM In, designated B2, was selected to confirm its In sensitivity in liquid medium comparatively and for detailed genetic analysis. Inverse PCR using transposon-specific primers followed by DNA sequencing of the amplified fragments identified the Tn5 insertion site in the mutant B2. Tn5 was located into an open reading frame encoding for a protein with a high degree of homology to a DedA family protein (uncharacterized membrane protein YqaA) [National Center Biotechnology Information (NCBI) Reference Sequence: WP_108473013.1], which is a highly conserved protein family poorly characterized (Doerrler et al., 2013).

DNA sequencing of the inverse PCR products and analysis of the genome of Rhodanobacter sp. B2A1Ga4 (DDBJ/ENA/GenBank under the accession JADBJR000000000.1) showed that the mutated gene is placed between genes encoding for a peptidoglycan DD-metalloendopeptidase family protein and a protein-L-isoaspartate(D-aspartate) O-methyltransferase, as shown in Figure 1.

The Rhodanobacter sp. B2A1Ga4 genome, besides the YqaA protein (MBQ4853883) with 202 amino acids, encodes two other members of the DedA family protein, with 206 (MBQ4854938) and 170 amino acids (MBQ4855745), respectively. These two members of the DedA protein family do not have significant amino acid similarity with the YqaA protein; however, they share 25% of amino acid identity with each other.

Alignment of YqaA protein of Rhodanobacter sp. B2A1Ga4 with homolog sequences of the DedA protein family showed the presence of two short motifs separated in length by approximately 50 amino acids (Figure 2), previously identified in DedA protein family (Keller and Schneider, 2013; Tábara et al., 2019). These two motifs have the consensus sequences of [F/Y]XXX[R/K] (motif 1) and GXXX[V/I/L/M]XXXX[F/Y] (motif 2), respectively, and are conserved in most of the proteins that belong to the DedA family. In YqaA protein of Rhodanobacter sp. B2A1Ga4 motif 1 is slightly different, having a histidine (H) and tryptophan (W) in the last two positions of motif 1. However, motif 2 is fully conserved in YqaA protein, including the glycine in the first position of motif 2, the most conserved residue in all DedA protein family. The secondary structure prediction of YqaA protein suggested that this protein is located in the cytoplasmatic membrane and contains five predicted transmembrane domains with a cytoplasmic N-terminal and an extracellular C-terminal, respectively.

The determination of the MIC values for both strains (the native strain and the mutant B2) was performed in liquid assays using In concentration ranging between 1 and 0.01 mM. The In MIC values were 1 and 0.5 mM for native and mutant strains, respectively. These results confirmed that gene mutation led to a loss of In resistance.

The growth of the native and mutant B2 strains in medium supplemented with the selected metals from the 13th group of the periodic table, In, Ga, and Al, was evaluated along with incubation time (Figure 3). The growth curves of strain B2A1Ga4 in all conditions (with or without metals) exhibited a similar profile. However, in the case of mutant B2, the bacterial growth curve was clearly different in presence of In comparatively to the control situation. This difference is particularly visible at late exponential phase (12 h of incubation). The other tested metals (Ga and Al) did not affect the growth profile of the mutant B2.

As bacterial growth might be impaired by gene mutation, native and mutant strains were tested under different conditions of temperature, presence of several antibiotics and metals, and presence of SDS. The strains exhibited similar results in the tested conditions (Table 2). Both strains showed susceptibility to chloramphenicol, gentamicin, erythromycin, nalidixic acid, rifampicin, and tetracycline and resistance to amoxicillin, ampicillin, colistin, and polymyxin B. The two strains grew at 20 and 25°C but did not grow at higher temperatures. Moreover, both strains showed the same resistance/susceptibility profile for all tested metals and SDS concentrations.

The importance of the DedA family protein YqaA in conferring resistance to In in Rhodanobacter sp. B2A1Ga4 strain was confirmed by comparing the In MIC values and growths of the mutant B2 complemented with a functional DedA family protein (B2_pyqaA) with the B2 mutant control carrying the empty vector (B2_p). The In MICs for both strains, B2_p and B2_pyqaA, were 0.25 and 0.5 mM, respectively. These values are not the same as those obtained for native or mutant B2 strains, but this can be explained because the complemented mutant strains carry an additional plasmid with a gentamicin resistance gene that affected the bacterial behavior.

Comparing the OD₆₀₀ measurements of the two complemented B2 strains, the growth of the control strain B2_p was significantly affected by In exposure; however, the metal did not affect the growth of strain B2_pyqaA (Figure 4). At 16 h of incubation time, only 0.2 mM of In had a negative effect in the growth of control cells that showed half OD₆₀₀ of the value measured in the control condition (without In) or for strain B2_pyqaA. The highest differences were obtained at 24 h of incubation, showing significant differences between OD₆₀₀ of growth of B2_p and B2_pyqaA strains in presence of both In concentrations tested (0.1 and 0.2 mM).

MTT assay was used to compare the cellular metabolic activity of mutant with native strain and the yqaA complemented mutant (B2_pyqaA) with the control strain (B2_p) when exposed to In, as an indicator of cell viability or cytotoxicity (Figure 5). The native strain in presence of In showed a decrease of activity of 1.2- and 1.7-fold at 10 and 24 h of incubation, respectively, when compared to the control condition. In the case of mutant strain, In exposure resulted in a very stronger impact on cellular activity, showing a decrease of 2.4- and 12.1-fold for 10 and 24 h of incubation, respectively (Figure 5A). The results of MTT assays also showed that In exposure affected drastically the cellular activity of control strain B2_p, but had only a slight effect in strain B2_pyqaA (Figure 5B). The presence of 0.1 mM In reduced the cellular activity of strain B2_p to approximately 2-fold of the control situation at 16 and 24 h. The highest concentration of In tested (0.2 mM In) showed a much stronger impact in cellular activity of strain B2_p, with reductions of 4.3- and 22.7-fold at 16 and 24 h, respectively. B2_pyqaA strain showed only a minor reduction of 1.3-fold of cellular activity with 0.2 mM In at 24 h of growth. Therefore, significant differences between the metabolic activities of B2_p and B2_pyqaA were observed for both evaluated conditions, different In concentration and incubation time.

The In accumulation, quantified by ICP-MS, showed that native and mutant strains have different profiles and levels of In accumulation (Figure 6A). The native strain showed the highest In accumulation values (2.80 ± 0.02 μg In/mg protein) at the exponential growth phase (8 h). For further incubation times, native cells showed lower In levels. The mutant B2 showed an increase of In amounts in the cells over the period of incubation with 4.67 ± 0.98, 12.53 ± 2.70, 14.89 ± 1.82, and 19.60 ± 1.14 μg In/mg protein at 8, 24, 48, and 96 h, respectively. These results showed that the mutant strain was able to accumulate up to 3.8-fold more In than the native strain.

The assay with EDTA was used to determine the main location of In in cells, if it was intracellular or complexed on the cellular surface. The fraction not treated with EDTA represents the total In amount in cells (intracellular and surface complexed In). After EDTA treatment, two fractions were obtained, a cellular fraction (intracellular In) and a supernatant fraction (surface complexed In). Figure 6B shows the values of In measured in these three fractions for samples collected at 48 h of growth. While the native strain showed 2.80 ± 0.02 μg In/mg protein in the total fraction, the mutant B2 showed 16.96 ± 0.82 μg In/mg protein. Mutant B2 accumulated significantly more In intracellularly than the native strain as the highest amount of In in the mutant cellular pellet corresponds to intracellular In (11.63 ± 1.67 and 1.52 ± 0.20 μg In/mg protein from mutant B2 and native strain, respectively). In the case of native cells, the highest amount of metal measured corresponds to In binding to the cellular surface (2.00 ± 0.23 μg In/mg protein). These results show that mutant strain was able to accumulate into the cells the largest fraction of In, with approximately 69% of the total In measured in cells.

In accumulation by the complemented mutant strains was evaluated to confirm the relevance of this studied DedA family protein (YqaA) in the process of In extrusion. Strain B2_pyqaA accumulated significantly lower amounts of In with 9.48 and 8.66 μg In/mg protein at 16 and 24 h, respectively, when compared to the control strain B2_p that showed 17.12 and 15.88 μg In/mg protein for the same incubation periods (Figure 6C). These results mean that B2_pyqaA cells accumulated approximately half of the In amount accumulated by the control strain.

Impaired DedA protein function could affect the membrane potential (Δψ). Thus, Δψ was measured using the probe JC-1 red/green dye. JC-1 is a membrane permeable dye that as a monomer exhibits green fluorescence at 530 nm, and in the presence of membrane potential, it aggregates showing red fluorescence at 595 nm. Therefore, membrane potential can be expressed as the ratio between the red (595 nm) and green (530 nm) fluorescence value ratio of red to green fluorescence (Jovanovic et al., 2006; Engl et al., 2011).

For both strains, native and mutant, there was no statistically significant difference in the membrane potential of cells grown in presence or absence of In (Figure 7). Moreover, fluorescence ratio values of mutant strain were similar to native strain indicating that compromised YqaA did not affect the membrane potential. The samples treated with CCCP served as a control of the experiment and showed similar fluorescence ratio.

The membrane potential was tested for both strains (B2_p and B2_pyqaA) with the protocol previously described (JC-1 probe). Once more, there was no significant difference in the membrane potential observed with and without the exposure to In, and the results obtained were similar for both strains (Supplementary Figure 1).

The Figure 8 shows the SEM images of the different strains (B2A1Ga4, mutant B2 and B2_pyqaA) grown with and without 0.1 mM In. The presence of In does not seem to interfere with the cellular morphology of the native strain. In case of the mutant B2, the In exposure resulted in visible cellular changes, in which the cells are shorter and wider than the native strain. Moreover, the gene complementation of mutant strain reverted the cellular morphology to the original characteristics.