The Bacillus subtilis yqgC-sodA operon protects magnesium-dependent enzymes by supporting manganese efflux

Microbes encounter a myriad of stresses during their life cycle. Dysregulation of metal ion homeostasis is increasingly recognized as a key factor in host-microbe interactions. Bacterial metal ion homeostasis is tightly regulated by dedicated metalloregulators that control uptake, sequestration, trafficking, and efflux. Here, we demonstrate that deletion of the Bacillus subtilis yqgC-sodA (YS) complex operon, but not deletion of the individual genes, causes hypersensitivity to manganese (Mn). YqgC is an integral membrane protein of unknown function and SodA is a Mn-dependent superoxide dismutase (MnSOD). The YS strain has reduced expression of two Mn efflux proteins, MneP and MneS, consistent with the observed Mn sensitivity. The YS strain accumulated high levels of Mn, had increased reactive radical species (RRS), and had broad metabolic alterations that can be partially explained by the inhibition of Mg-dependent enzymes. Although the YS operon deletion strain and an efflux-deficient mneP mneS double mutant both accumulate Mn and have similar metabolic perturbations they also display phenotypic differences. Several mutations that suppressed Mn intoxication of the mneP mneS efflux mutant did not benefit the YS mutant. Further, Mn intoxication in the YS mutant, but not the mneP mneS strain, was alleviated by expression of Mg-dependent, chorismate-utilizing enzymes of the menaquinone, siderophore, and tryptophan (MST) family. Therefore, despite their phenotypic similarities, the Mn sensitivity in the mneP mneS and the yqgC-sodA deletion mutants results from distinct enzymatic vulnerabilities.


Abstract
Microbes encounter a myriad of stresses during their life cycle.Dysregulation of metal ion homeostasis is increasingly recognized as a key factor in host-microbe interactions.
Bacterial metal ion homeostasis is tightly regulated by dedicated metalloregulators that control uptake, sequestration, trafficking, and efflux.Here, we demonstrate that deletion of the Bacillus subtilis yqgC-sodA (YS) complex operon, but not deletion of the individual genes, causes hypersensitivity to manganese (Mn).YqgC is an integral membrane protein of unknown function and SodA is a Mn-dependent superoxide dismutase (MnSOD).The YS strain has reduced expression of two Mn efflux proteins, MneP and MneS, consistent with the observed Mn sensitivity.The YS strain accumulated high levels of Mn, had increased reactive radical species (RRS), and had broad metabolic alterations that can be partially explained by the inhibition of Mgdependent enzymes.Although the YS operon deletion strain and an efflux-deficient mneP mneS double mutant both accumulate Mn and have similar metabolic perturbations they also display phenotypic differences.Several mutations that suppressed Mn intoxication of the mneP mneS efflux mutant did not benefit the YS mutant.Further, Mn intoxication in the YS mutant, but not the mneP mneS strain, was alleviated by expression of Mg-dependent, chorismate-utilizing enzymes of the menaquinone, siderophore, and tryptophan (MST) family.Therefore, despite their phenotypic similarities, the Mn sensitivity in the mneP mneS and the yqgC-sodA deletion mutants results from distinct enzymatic vulnerabilities.

INTRODUCTION
Metal ion homeostasis relies on a careful balance between metal import, export, and storage mechanisms (1).Bacterial growth may be restricted due to either metal ion limitation or excess, and both mechanisms are important during host-microbe interactions (2,3).Nutritional immunity refers to the ability of the mammalian host to restrict access to essential nutrient metal ions such as iron (Fe), zinc (Zn), or manganese (Mn) (2).The host may also restrict the growth of intracellular pathogens through metal intoxication (3)(4)(5).In response, bacteria induce the expression of metal exporters that contribute to survival in the host (6,7).
Metal ion efflux systems are regulated in opposition to metal importers.In Bacillus subtilis, the MntR metalloregulatory protein binds to Mn to repress the expression of the MntH and MntABCD importers (8).As Mn levels rise further, MntR activates the transcription of the MneP and MneS cation diffusion facilitator (CDF) efflux proteins (9).Although B. subtilis normally tolerates up to 1 mM Mn ion, mutants deficient in Mn homeostasis are much more sensitive (9,10).
Previous work investigated Mn intoxication using an mneP mneS (PS) effluxdeficient strain (11).When grown in a minimal medium (MM), the Mn sensitivity of the PS strain can be suppressed by loss of function mutations in qoxA (part of the major aerobic respiratory quinol oxidase, QoxABCD) or mhqR (repressor of the methylhydroquinone-induced genes) (11).MhqR-regulated proteins function to reduce or degrade quinones and other reactive species (12,13).Mn intoxication was thereby associated with production of reactive radical species (RRS) due to a dysfunction of the Qox (cytochrome aa3) quinol oxidase (11).
For reasons not yet understood, genetic suppression of Mn sensitivity is often dependent on the growth medium.For example, in MM with malate as a carbon source, but not with glucose, the Mn sensitivity of the PS mutant was suppressed by an insertion within the yqgC-sodA operon (iTn-sodA) (11).Conversely, MM with glucose as a carbon source, but not with malate, the Mn sensitivity of the PS mutant was suppressed by the inactivation of mpfA (11), which encodes a Mg efflux pump (14).
Inactivation of mpfA leads to increased intracellular Mg levels, suggesting that Mn intoxication can result from competition with Mg (14).Although Mg is the most abundant divalent cation in cells, it binds macromolecules less tightly than those in the Irving-William series: [Mn(II)<Fe(II)<Co(II)<Ni(II)<Cu(II)>Zn(II)].Thus, Mg is susceptible to replacement by metals with higher affinities towards protein ligands (including Fe, Mn, and Zn).
Here, we describe the importance of the yqgC-sodA complex operon in Mn homeostasis.A strain with a deletion of the yqgC-sodA operon (YS mutant) was defective in the expression of mneP and mneS and is as sensitive to Mn as a PS effluxdeficient strain.Both strains accumulated high levels of intracellular Mn, displayed high levels of RRS, and had similar alterations in their metabolic profiles.Despite these similarities, the basis for Mn intoxication appeared to be distinct.Several mutations that suppressed Mn sensitivity of the PS mutant had little benefit for the YS strain.
Conversely, the YS strain, but not the PS strain, was rescued by elevated expression of chorismate-utilizing, Mg-dependent enzymes of the menaquinone, siderophore, and tryptophan (MST) family.These findings reveal that Mn excess leads to specific metabolic disruptions, and the consequences of these disruptions can vary between phenotypically similar strains.

Mutants deleted for the yqgC-sodA operon (YS) are hypersensitive to Mn
Manganese-dependent superoxide dismutase (MnSOD) is one of the most abundant proteins in the Bacillus proteome (15) and is the major Mn-binding protein in the cell (16).The sodA gene is transcribed together with the upstream yqgC gene as part of a complex, two-gene operon, yqgC-sodA (Fig. 1a).YqgC encodes an unknown function (DUF456) membrane protein of 160 amino acids.Transcription initiates both upstream of yqgC and at two different promoters within the 178 bp yqgC-sodA intergenic region (17), with the upstream promoter associated with the production of a 128 nt long 5'-untranslated region designated ncr2103 (18) or S936 (17) (Fig. 1a).
Previously, we noted that Mn intoxication can lead to an accumulation of reactive radical species (RRS) (11), suggesting that MnSOD may be a protective protein under conditions of Mn intoxication.To explore the role of the yqgC-sodA operon in Mn homeostasis, we generated in-frame deletions of the yqgC and sodA coding regions using the BKE collection of mutant strains (19).In these strains, the coding region is replaced by an erythromycin cassette that can be removed using plasmid pDR244 expressing the Cre-Lox recombinase (19).In addition, we used a CRISPR-based mutagenesis approach (20) to delete just the S936 element, the S936-sodA region, or the yqgC-sodA operon in its entirety (DyqgC-S936-sodA; YS).The yqgC, sodA, and S936-sodA deletion strains all grew well in LB medium both with and without amendment with 150 µM Mn (Fig. 1b, 1c).However, the YS operon deletion was unable to grow in LB medium amended with 150 µM Mn (Fig. 1c).
We next quantified Mn sensitivity on MM-malate agar plates (Fig. 1d).The YS strain exhibited a clearance zone equivalent to the Mn-sensitive mneP mneS (PS) efflux knockout strain (Fig. 1d).This zone of clearance was not observed in the yqgC and sodA single mutants.Although both the YS and PS mutants are Mn sensitive, a quadruple mutant (PSYS) was no more sensitive than the YS or PS double mutants (Fig. 1d).Thus, cells lacking the entire YS operon (yqgC-sodA) are highly Mn sensitive.This high sensitivity was also seen in cells grown in liquid cultures: the MIC for Mn in MM-malate revealed a 400-fold increase in sensitivity in the YS strain (MIC of 5 µM) compared with the wild-type (WT) or single deletions strains (MIC of 2 mM) (Fig. S1).
Thus, the presence of either yqgC or sodA is sufficient to confer resistance to elevated Mn.

Mn intoxication has similar consequences in YS and PS
Since the YS mutant was as Mn sensitive as the efflux-deficient PS mutant (Fig. 1d), we compared these two strains by monitoring Mn accumulation, production of RRS, and metabolome profiles.When grown in LB medium and challenged with 150 µM Mn, the intracellular level of Mn increased in all strains, but most dramatically in the YS and PS mutant strains (Table 1).After Mn challenge, the YS strain accumulated ~15x the level of Mn seen in the WT strain, and the YSPS strain accumulated ~7-fold more (Table 1).In contrast, Mn accumulation was unaffected by the deletion of either yqgC or sodA (Table 1).In general, elevated levels of Mn were correlated with a decrease in intracellular Mg, but there was little change in Zn levels (Table 2).
Previously, the PS efflux deficient strain was found to accumulate RRS when challenged with Mn (11).These studies monitored RRS using the fluorogenic reporter 2′,7′-dichlorodihydrofluorescein diacetate (DCFDA).Intracellular DCFDA is deacetylated to generate DCF, which is readily oxidized by one-electron oxidizing species, such as hydroxyl radical and other reactive radicals generated by the Fenton reaction (21,22).Like the PS strain, the YS mutant and the PSYS mutant also accumulated high levels of RRS after Mn challenge (Fig. 1e).We conclude that the YS strain, like the PS strain, accumulates Mn intracellularly and experiences increased stress from RRS.

Mn sensitivity in the YS strain is correlated with defects in Mn efflux
We hypothesized that the Mn sensitivity of the YS strain may result from a defect in the expression or function of the MneP and MneS efflux pumps.We therefore monitored the levels of mneP and mneS mRNA using real-time quantitative reverse transcription PCR (qRT-PCR) (Fig. 2a).In the WT background, the three single mutants (yqgC, S936, and sodA), and the S936-sodA mutant, both transcripts were detected and were, as expected, induced by exposure of cells to 150 µM Mn for 30 min.In contrast, in the YS deletion strain the mneP and mneS transcripts were greatly reduced in abundance and were no longer inducible by Mn (Fig. 2a).These observations suggest that the YS operon deletion precludes efficient expression of mneP and mneS under these conditions.
Transcriptional induction of mneP and mneS by excess Mn is dependent on the MntR transcription factor (9).We hypothesized that the lack of mRNA accumulation for the mneP and mneS genes might result from a failure of MntR to activate transcription.
We therefore tested whether the Mn sensitivity of the YS mutant could be rescued by overexpression of MneP from an IPTG-inducible promoter.Indeed, induction of MneP restored the ability of YS to grow in LB medium amended with 150 µM Mn (Fig. 2b).
Similarly, expression of MntE, a Staphylococcus aureus CDF (23) that is a homolog of B. subtilis MneS (41% identity) and MneP (26% identity), also restored Mn resistance (Fig. 2b).In contrast, expression of the Escherichia coli FieF CDF protein (22% identity to MneP, 21% identity to MneS), implicated in Zn, Fe, and Mn efflux (24,25), did not restore Mn resistance to the B. subtilis YS mutant (Fig. 2b).Thus, the Mn sensitivity in the YS mutant strain is due to a failure to properly express Mn efflux pumps (MneP and MneS), rather than an inability of efflux pumps to function properly.
To further test the hypothesis that MntR activation of mneP and mneS transcription is deficient in the YS strain we monitored the ability of PmneP-lacZ and PmneS-lacZ promoter-lacZ fusions to be induced by Mn. Surprisingly, amendment with 10 to 20 µM Mn efficiently induced these promoters in the YS strain (Fig. 2c).Expression was similar to that seen in PS strains (Fig. 2c), indicating that activation by MntR is unimpaired in the YS strain.Similarly, the ability of MntR to function as a Mn-dependent repressor, as monitored using a PmntH-cat-lacZ reporter fusion was unimpaired (data not shown).The lack of mRNA accumulation in the YS strain, as seen in LB medium both with and without Mn shock (Fig. 2a), may therefore result from an effect of the YS operon deletion on mRNA synthesis or stability at a step after the MntR-dependent initiation of transcription.The nature of this effect is presently unclear.

Genetic suppression of Mn sensitivity in the YS mutant
Since the YS operon deletion leads to a reduction in expression of the mneP and mneS genes, we anticipated that YS and PS mutants might have very similar effects on cell physiology.To test this hypothesis, we asked whether suppressor mutations previously shown to rescue Mn sensitivity of the PS strain (11) also rescue the YS strain (Fig. 3a).Unexpectedly, none of the mutations that improved fitness of the PS mutant strain on MM-glucose (mhqR, pyk, mpfA, ytfP-opuD) were able to improve the growth of the YS strain on this medium (Fig. 3b).The mutation that most strongly rescued growth of the PS strain on MM-malate (qoxA) had a modest beneficial effect in the YS strain (Fig. 3c).A small beneficial effect was also noted for the pyk mutation in YS (Fig. 3c), even though this mutation was not beneficial to the PS strain on this medium (and was selected on MM-glucose (11)).While the basis for these genetic interactions is poorly understood, these differences imply that, despite their phenotypic similarities, the processes that limit growth in the PS and YS strains may be different.
To better understand the effects of Mn on the YS strain we next selected spontaneous mutants that formed colonies in the zone of clearance around Mn disks in zone of inhibition assays.Results from whole genome-resequencing (Supplementary file S1C) revealed that these isolates often carried missense mutations affecting either of two TerC homologs (MeeF, MeeY) implicated in Mn export and exoprotein metalation (26,27).As expected based on prior work (26), null mutations in meeF or meeY did not increase Mn resistance in the YS background.Therefore, these missense mutations were recreated by CRISPR-based genome editing to test whether their introduction was sufficient for Mn resistance.In each case, CRISPR-generated strains with only the altered function meeF* or meeY* alleles were Mn resistant.Since TerC homologs are implicated in Mn efflux (26,27), we used ICP-ES to measure the impact of meeF* or meeY* alleles on Mn levels.Our ICP-ES results (Table 1) suggest that the meeF* allele (Phe225Val) decreased Mn accumulation in the YS cells, likely accounting for the increased Mn resistance.In contrast, meeY* (Trp71Arg) and meeF* (Ile206Thr) led to only modest reductions in Mn accumulation (Table 1).
Next, we used mariner transposon (mTn) mutagenesis to isolate mutations that could suppress Mn sensitivity of the YS strain (Fig. 3d).We recovered 8 different transposants with insertions in a variety of loci (Supplementary Table S1D), all of which were genetically linked to the observed Mn resistance.We here focused on strain HBYL1110 with an mTn insertion after position 87484 in the Bacillus reference genome (NC_000964; ( 28)) within yazB.The yazB gene is part of a complex operon including genes for folate biosynthesis (folB-folK-yazB-yacF-lysS) (Fig. 3d).YazB is predicted to be a small (69 amino acid) DNA-binding protein, suggestive of a possible regulatory function.Prior studies of a yazB mutant did not reveal changes in the expression of the upstream folate genes (29).We confirmed that an in-frame yazB deletion mutant also suppressed Mn sensitivity of the YS strain (Fig 3e).
Since YS strains are defective in expression of mneP and mneS, we monitored the effect of the yazB on these two genes.Remarkably, Mn-dependent induction of mneS was restored in the YS yazB strain, but expression of mneP expression was still non-responsive to Mn stress (Fig 3f).This suggests that the yazB suppressor works by restoring the expression of mneS, consistent with the conclusion above that YS fails to properly express the MneP or MneS efflux proteins.We attempted to test this hypothesis by constructing a YS yazB mneS quadruple mutant, but were repeatedly unable to obtain this strain.

Mn intoxication elicits similar changes in the metabolome of YS and PS strains
Next, we used metabolomics to monitor the global changes in metabolism upon challenge of WT, yqgC, sodA, and YS strains grown in LB with and without 150 µM Mn and compared these changes to those seen in the efflux-deficient PS mutant.The yqgC and sodA single mutants had relatively modest metabolic perturbations after Mn challenge.In contrast, the Mn-sensitive YS and PS strains had similar and wide-ranging changes in cellular metabolism (Fig. S2).
These metabolic perturbations led us to hypothesize that Mn might inhibit enzymes of the menaquinone, siderophore, and tryptophan (MST) family (Fig. 4a).B. subtilis encodes several MST-family enzymes that are involved in the synthesis of the bacillibactin siderophore (DhbC), folate (PabB), and menaquinone (MenF).MST enzymes bind chorismate, with the carboxylate group coordinated to divalent magnesium (Mg 2+ ), and catalyze a variety of addition reactions (with either water or ammonia as nucleophile) that may or may not be coupled to the elimination of pyruvate (30)(31)(32).Other Mg-dependent enzymes may also be inhibited under these conditions, including Mg-dependent amidotransferases involved in aromatic amino acid (Trp, Phe, Tyr) biosynthesis.In addition, asparagine levels are greatly reduced in the Mnchallenged YS and PS mutants (Fig. S2), and asparagine synthase is a Mg-dependent amidotransferase (33).Similarly, inhibition of the HisC/AroJ amidotransferase may account for the observed increase in phenylpruvate and decrease in phenylalanine (Fig. S2).
Prior work demonstrates that ferrous Fe can displace Mg and thereby serve as a nanomolar inhibitor for multiple MST family enzymes, including isochorismate synthases (Pseudomonas aeruginosa PchA and E. coli EntC) and salicylate synthase (Yersinia enterocolitica Irp9) (32).We therefore hypothesized that Mn, previously shown to be inactive as a cofactor for isochorismate synthase (30), might serve as a general inhibitor of MST family enzymes in B. subtilis.To test this hypothesis, we asked whether supplemental Mg could suppress Mn sensitivity.Indeed, Mg reversed the Mndependent growth inhibition of the YS strain in both liquid culture (Fig 4c) and on plates (Fig 4d).Elevated levels of Mg also reversed the Mn-dependent growth inhibition of the PS and PSYS strains on LB medium (Fig. 4d).Thus, Mn-dependent inhibition of Mgdependent enzymes is correlated with growth inhibition.

Induction of MST family enzymes reverses Mn intoxication of YS but not PS
To test the hypothesis that inhibition of Mg-dependent MST enzymes is limiting for growth, we engineered strains with inducible expression of several of these enzymes.Surprisingly, induction of any of the three MST proteins suppressed Mn intoxication of the YS strain (Fig. 4e, f) but not the PS strain in LB medium (Fig. S3).
Induction of these chorismate-utilizing enzymes had no effect on growth in the absence of Mn challenge (Fig. S4).These results suggest that Mn intoxication inhibits a similar set of enzymes in the PS and YS strains, but the specific metabolic deficiencies that underlie growth inhibition differ.
Of all the products of pathways dependent on MST enzymes, the simplest to assay is the siderophore bacillibactin and its precursors.In B. subtilis 168 strains this pathway is not fully functional due to the presence of the sfp 0 mutation that inactivates the Sfp 4-phosphopantetheinyl transferase necessary for the activity of non-ribosomal peptide synthases (34,35).As a result, 168 strains fail to properly modify the DhbB and DhbF enzymes with the phosphopantetheine carrier and produce the bacillibactin precursors 2,3-dihydroxybenzoic acid (DHBA) and 2,3-dihydroxybenzoylglycine (DHBG), known collectively as DHBA(G) (36).These precursors are easily assayed in cell supernatants of iron-starved cells using spectrophotometry.We monitored DHBA(G) production using fur mutants that derepress the expression of the entire dhb operon for bacillibactin production (37).As expected, DHBA(G) production was elevated in cells grown in MM-malate in the fur mutant background.However, DHBA(G) production was not elevated if the medium was amended with Mn, or if the fur mutant was present in the YS mutant background (Fig. 4g, Fig. S5).Further, DHBA(G) production was increased even above the levels seen in the fur mutant in strains overexpressing the DhbC isochorismate synthase, and synthesis was no longer inhibited by Mn (Fig. 4g).DHBA(G) synthesis was also increased by expression of a different isochorismate synthase, MenF.These observations are consistent with the hypothesis that Mn is a generalized inhibitor of MST enzymes, and this inhibition can be overcome by Mg supplementation or by enzyme overproduction.Unexpectedly, the induction of pabB also increased the expression of DHBA(G) (Fig. 4g, Fig. S5).
Although PabB is an MST-family enzyme, it is not known to have isochorismate synthase activity.Thus, the mechanism of this effect may be indirect, perhaps through the binding of Mn and protection of DhbC from Mn inhibition.

DISCUSSION
Metalloenzymes are ubiquitous in biology, but their activity is contingent on metalation with the proper metal cofactor.Many metalloproteins obtain the cognate metal ion from a buffered cytosolic pool dependent on metal availability and the metal affinity of the protein active site (38).According to the Irving-Williams series, divalent transition metals typically bind with affinities in the order: Mn(II) < Fe(II) < Co(II) < Ni(II) < Cu(II) > Zn(II).These thermodynamic affinities often determine the hierarchy for metal occupancy within a specific metalloenzyme.The metalation of enzymes that require high-affinity metals, such as Cu or Zn, typically involves ligand exchange reactions with cellular metabolites or protein metallochaperones.For lower affinity metals, such as Mn and Fe, the buffered metal pools are more loosely held, and metal acquisition may occur unaided from pools of rapidly exchanging ions.In both cases, metalation by more tightly held and slower exchanging metals (from higher in the Irving-Williams series) can lead to enzyme mismetalation and inactivation.For example, Zn intoxication has often been linked to the mismetalation of proteins that normally require Mn or Fe for their activity (39)(40)(41).
Even though Mn and Fe are relatively weak-binding metals (at the lower affinity end of the Irving-Williams series), cells can still experience metal intoxication when their homeostasis systems are disrupted.In B. subtilis, Fe intoxication is apparent in strains lacking the PfeT Fe efflux ATPase (42,43), and Mn intoxication is apparent in effluxdeficient PS strains (9,11) and, as shown here, in the absence of the yqgC-sodA operon (YS).The molecular basis for intoxication by these more weakly interacting metal ions is not well understood, but previous results have suggested that excess Mn may interfere with Mg-dependent processes (14,44).For example, mutations in a B. subtilis Mg-efflux system (MpfA) can increase tolerance to high Mn levels (14).
However, the Mg-dependent pathways that are inhibited by Mn have not been defined.
Prior work has shown that Mn intoxication of the PS mutant derives, in part, from a dysfunction associated with the QoxABCD terminal oxidase and is correlated with production of RRS (11).
Here, we focused on understanding the role of the yqgC-sodA operon in Mn homeostasis.The YS operon deletion mutant is phenotypically similar to the PS strain characterized previously (11).Both the YS and PS mutants display similar Mn sensitivity (Fig. 1d), accumulation of intracellular Mn when excess Mn is added (Table 1), Mninduced production of RRS (Fig. 1e), and changes to their metabolomes (Fig. 4b, S2).
In addition, the Mn sensitivity of both strains can be suppressed by high Mg levels (Fig. 4d).This similarity is supported by a lack of additivity between the PS and YS mutations (Fig. 1d), which is consistent with the observation that the YS strain is defective for expression of mneP and mneS (Fig. 2a), encoding two efflux proteins lacking in the PS strain.The mechanism underlying the reduced expression of mneP and mneS, and the lack of Mn induction (Fig. 2a), in the YS mutant is unclear but does not appear to be due to an inability of MntR to activate transcription (Fig. 2c).
Although the YS and PS strains are phenotypically similar, we were intrigued to observe that the YSPS strain accumulated less Mn than the YS strain after Mn shock (Table 1).One possibility is that the MneP and MneS CDF proteins contribute to Mn influx in the YS mutant strain under conditions of Mn excess.This is reminiscent of our previous observation that PS mutant strains lacking MntH, a proton-coupled Mn importer, accumulated more (rather than less) Mn when grown in Mn excess conditions (11).These findings lead to the hypothesis that these proton-coupled transporters may function in either direction, depending upon the strength of the relevant proton and metal ion gradients.
The phenotypes reported here are most evident in strains lacking both yqgC and sodA, but not in the single mutant strains.Superoxide dismutase is a broadly conserved protein found in all domains of life (45).In B. subtilis, MnSOD is the major Mn-binding protein in the cell (16) and protects cells against the reactive superoxide radical.The cotranscribed yqgC gene encodes an unknown function (DUF456) membrane protein.
YqgC and its homologs (COG2839) are present in >1000 sequenced genomes, including Deinococci, Mycobacteroides abscessus, and Salmonella enterica.Although both sodA and yqgC are broadly conserved in many bacteria, gene neighborhood analysis reveals that the yqgC-sodA genomic organization is predominantly seen in the Bacilli class (Fig. S6).
It is not obvious why the deletion of both genes is required to reveal a high level of Mn sensitivity.However, this sensitivity is correlated with RRS accumulation, as reported previously (11).Our prior work linked RRS production to a dysfunction of the major quinol oxidase (QoxA)-dependent electron transport chain (ETC).One possibility is that the YqgC membrane protein interacts with proteins of the ETC or with menaquinone itself to help prevent RRS production, and in its absence MnSOD helps prevent RRS accumulation.
Both the YS and PS strains had similar metabolic profiles when challenged with Mn (Fig. S2).We noted that several of the over-represented metabolites are substrates of Mg-dependent enzymes, consistent with the finding that elevated Mg suppresses intoxication (Fig. 4d).However, the proximal cause of growth inhibition in these two strains appears to be different.Overexpression of MST family enzymes rescued the YS strain (Fig. 4e), but not the PS strain (Fig. S3).We therefore conclude that Mg-dependent MST family enzymes are inhibited in these strains, but other Mg-dependent processes also fail in the PS strain, precluding growth rescue by simply overexpressing MST enzymes.Expression of any of several MST enzymes can rescue growth (Fig. 4e), even though they do not all catalyze the same reaction (Fig. 4a).In some cases, the expressed enzyme may be directly acting to accelerate the limiting reaction, and in others, it perhaps acts as a decoy to help buffer toxic Mn ions in the cell.A similar effect may account for the ability of induction of PabB to help increase the expression of DHBA(G) (Fig. 4g).
Our findings here extend our view of Mn homeostasis in B. subtilis (Fig. 5).In addition to the core MntR-regulated import (mntH, mntABC) and efflux (mneP, mneS) genes, Mn is also trafficked in the cell by two TerC family efflux proteins in support of exoenzyme metalation (27).We here demonstrate that YqgC and MnSOD are two additional factors that have protective roles in Mn homeostasis.Finally, our results emphasize that there are likely multiple ways in which excess Mn can intoxicate cells.
Even the phenotypically similar PS and YS strains appear to have distinct proximal causes for growth inhibition.

Materials and Methods
Bacterial strains and growth conditions.The strains used in this study are listed in Supplementary file Table S1A.Mutations from B. subtilis 168 were moved into B. subtilis CU1065 as the parental WT strain for this study.Cultures were streaked from frozen glycerol stocks onto LB agar plates and grown at 37 °C for 18 hours.Cells were grown in 5 ml of LB broth supplemented when required with antibiotics: MLS (1 µg/ml erythromycin plus 25 µg/ml lincomycin), kanamycin (15 µg/ml), and/or chloramphenicol (10 µg/ml), at 37 °C under shaking conditions of 300 rpm on a gyratory shaker.Once cultures reached an OD600nm ~0.4, 2 µl volume was used as an inoculum into 98 µl of LB dispensed in 96 well plates such that the initial OD600nm was 0.07.These plates were supplemented with appropriate antibiotics, and IPTG/xylose was added as needed.Growth in LB or modified minimal media recipe (11) containing either glucose or malate supplemented with MnCl2 (Sigma) was monitored for up to 24 hr at 37 °C under shaking conditions using Synergy H1 (BioTek Instruments, Inc. VT) plate reader.The minimum inhibitory concentration (MIC) for Mn was defined as the Mn concentration that led to an OD600nm of <0.4 after 8 hr of growth.

Genetic manipulations and strain construction. A) CRISPR editing.
The complementation strain was constructed using a repair template consisting of upstream (of yqgC) and downstream (of sodA) homology regions fused to yqgC-S936-sodA central fragment amplified and stitched using primers listed in the Supplementary file Table S1B.DNA alleles were generated by adaptor-based LFH PCR and were subjected to stitching to generate a repair template.The repair template had compatible SfiI flanking the 5' and 3' ends.Upon digestion with SfiI at 50 °C for 2 hr, a similar restriction digest was performed for pAJS23 plasmid containing guide RNA against erm cassette as described previously (20).
The ligation of the digested repair template into pre-digested pAJS23 was performed and was then transformed into E. coli DH5a selected on Luria-Bertani (LB) medium (Affymetrix) supplemented with 30 µg/ml of kanamycin.Plasmid was isolated and moved into E. coli TG1 strain.The multimeric plasmid was extracted and was then transformed into yqgC-S936-sodA::erm recipient strain grown to OD600nm = ~0.8 in the modified competence (MC) media.
Plasmid DNA was used to transform recipient cells, which were then allowed to recover for 2 hr under aerobic incubation at 30 °C.Transformed Bacillus cells were selected on LBkanamycin plus 0.2% mannose in the agar at 30 °C (until clones appeared, i.e., ~48 hr).The CRISPR plasmid was cured at 45 °C for 2 days on LB agar with repeated passages for all transformants.Clones sensitive to kanamycin and erythromycin were selected, and chromosomal DNA was prepared for PCR analysis.The constructed strains were verified using Sanger sequencing of amplified PCR reactions.B) Overexpression in pPL82 or pAX01.For ectopic expression of Bacillus genes (mneP, menF, dhbC, pabB, aroE, aroH, and menA) and heterologous expression of Staphylococcus (mntE) and E. coli (yiiP/fieF) genes, the coding regions were amplified with primers listed in Supplementary file Table S1B from chromosomal DNA using Phusion high-fidelity DNA polymerase (NEB) and subjected to restriction enzyme digestion, purification, and ligation into pre-digested pPL82/pAX01 vector for propagation in E. coli DH5a on LB supplemented with 100 µg/ml of ampicillin.Prepared recombinant plasmid constructs were transformed into recipient Bacillus strains for insertion at the amyE or lacA locus by double-cross over recombination selected with 10 µg/ml chloramphenicol or 1 µg/ml of erythromycin plus 25 µg/ml of lincomycin).C) Multiple (adjacent) gene deletions.For yqgC-S936-sodA and S936-sodA deletions, upstream and downstream fragments were amplified with long flanking homology PCR and stitched together using joining PCR such that the donor fragment had yqgB-erm-yqgE (DYS) and yqgC-cat-yqgE (DS936-sodA).These fragments were then transformed into CU1065, and integrants were selected with antibiotics.
Disk diffusion and Mn sensitivity assays.Mn sensitivity was evaluated on a minimal medium with malate or glucose as described (11) previously.Briefly, the bottom agar contained 15 ml MM-malate broth agar (with 1.5 % final agar concentrations), and the top soft-agar consisted of 4 ml MM-malate broth agar (with 0.75 % agar final agar concentrations) and 100 µl of mid-exponentially grown cultures (OD600nm of 0.4) as inoculum.
Whatman filter paper number 4 disks (6 mm in diameter) with Mn (10 µl of 10 mM stock) were placed on the plates, and the diameter for the zone of growth inhibition (ZOI) was measured after 18 hr at 37 °C.
Real-Time RT PCR.Total RNA was extracted using a QIAGEN kit from 1.5 ml of midlog (OD600nm =0.4) WT and YS cells, which were grown in LB broth either in the presence or absence of 100 µM Mn.Total RNA was treated with DNase (Ambion) enzyme to further purify and remove traces of DNA.For each reaction, 2 µg of RNA was used for cDNA preparation using High-Capacity reverse transcriptase (Applied Biosystems) amplified with random hexamer primers.Further, for amplicon measurements 10 ng of cDNA was used as a template along with 500 nM of mneP, mneS, and gyrA (control) gene-specific qPCR F/R primers in a 1X SYBR green master mix (Bio-Rad).Threshold and baseline parameters were kept consistent for experiments performed on different days.All Ct mean values for the gene expression were normalized to gyrA (n=3).

Mariner-transposon (mTn) mutagenesis. The YS cells containing pMarA plasmid
were grown in LB broth with erythromycin (1 μg/ml) at 30 °C till OD600 reached ~1.0, and cells were plated onto Mn gradient plate made up of MM containing glucose at 42 °C.
Colonies that grew after overnight incubation near the high levels of Mn were subcultured in LB supplemented with kanamycin (15 μg/ml) at 37 °C to confirm the presence of mTn.Chromosomal DNA was purified from these DYS-Tn cells and was subjected to TaqαΙ restriction enzyme digestion at 37 °C for 2 h, followed by overnight ligation of cohesive ends to generate a circular transposon-gDNA chimeric library.The PCR reaction from ends of the mTn performed on the chimeric library was further subjected to Sanger' sequencing analysis at Cornell BRC facility to identify the mTn insertion site within the genome (Supplemental Table S1D).
Whole genome re-sequencing.Suppressors for YS strains were isolated using Mndisk diffusion assay performed using either LB-with 150 µM Mn present in the broth, 100 µM Mn supplemented in a plate condition, or a disk diffusion performed in a MMmalate agar.For MM-malate agar, suppressors were picked from the clearance zone.
All the suppressors were tested multiple times for the increased resistance to Mn and genomic DNA was extracted, purified, and subjected to Illumina sequencing (San Diego, CA) by MiGS or SeqCenter (Pittsburgh).The Illumina sequence reads were trimmed, mapped, aligned, and analyzed for sequence variants against the reference genome sequence NC_000964 using CLC genomics workbench software (Supplementary File Table S2).The nucleotide changes (relative to reference) found in both the parental WT (CU1065) and the YS (unstressed) strains were ignored, and only newly arising nucleotide variants found in the suppressor strains are indicated.Determination of intracellular metal content.ICP-ES analysis was performed using modifications of a previously described procedure (11).Briefly, strains were grown in 5 LB broth to an OD600nm of 0.4 at 37 ˚C under aerobic growth conditions.Cultures were treated with or without 150 µM Mn and further grown for 30 min. 2 ml of samples were collected by centrifugation and washed twice with 5 ml Chelex-treated PBS, resuspended in 0.5 ml PBS, and weighed.Samples were digested in double distilled HNO3 in a carbon heat block using a Vulcan 84 automated sample digestion unit b-galactosidase plates and enzyme activity.All bacterial cultures were grown to a mid-log phase, streaked onto LB agar containing 100 µg/ml of X-gal and Mn (0-100 µM), and plates were incubated at 37 ˚C overnight.Blue color development for different strains was noted by capturing images.
Metabolite extraction and measurements.All strains were grown in LB at 37 ˚C until the culture reached an OD600 of 0.4 and then treated with or without 0.15 mM (final concentration) of MnCl2 (Sigma-Aldrich) for 60 min at 37 ˚C.Cells were pelleted and quenched by resuspending in chilled 0.6 ml of mixtures of acetonitrile:methanol:water (40:40:20).These were further lysed using 0.1 mm Zirconia beads in a Precellys homogenizer (Bertin Instruments).Freshly lysed suspensions were centrifuged at ~12,000 rpm for 8 min at 37 ˚C.The supernatants were passed through a 0.22 µm SpinX tube filter (Sigma-Aldrich).Extracted metabolites were separated on a Cogent Diamond Hydride Type C column of 1200 liquid chromatography coupled to an Mass 6220 TOF spectrophotometer (Agilent).Ion abundances of metabolites were estimated using Profinder 8.0 and log2 fold changes were analyzed and plotted using Gene Cluster 3.0 and Java Treeview.

Quantification of Siderophores. B. subtilis cultures were aerobically grown overnight
in MM-malate (1 µM of Fe) at 37 °C with and without 100 µM added Mn. Optical density at 600nm was noted, and cells were harvested by centrifugation at 5000 rpm for 10 min to retrieve supernatant, and to 1 mL supernatant 200 µl of 10 mM FeCl3 (Sigma) dissolved in 100 mM HCl was added.This supernatant fraction was then neutralized by the addition of 100 µl of 1 M Tris-HCl (pH 8.0), and the absorbance at 510 nm was recorded.A standard curve was prepared with DHBA (Sigma) under similar conditions to extrapolate levels of siderophores in unknown culture samples.All 510 nm values were normalized to culture OD at 600nm (n=3).Color development for different reaction tubes was captured using an iPhone XS (Apple, CA).Table 1.Intracellular Mn levels (µg g -1 cells dry weight).based siderophore yield (absorbance 510 nm/ OD600nm) was determined for the indicated strains with and without over-expression of MST enzymes (menF, dhbC, and pabB).Cells were grown overnight in MM-malate with or without 0.2 mM IPTG with 1 µM ferric ammonium citrate, harvested and then resuspended in the same medium with no added iron and growth continued for 5 hrs.Supernatant was used in determination of DHBA levels after colorimetric derivatization with excess FeCl3.The level of siderophores/DHBA were extrapolated from DHBA standard curve treated in a similar manner (n=2: the difference in siderophore production among all strains were compared to the fur mutant grown in liquid minimal media malate (MMM) without Mn using a oneway ANOVA test with Tukey's post-hoc analysis where ****p<0.0001,***p= or <0.001, and **p<0.01,with n.s.no significant difference).In the YS deletion strain, transcription of mneP and mneS is affected leading to Mn sensitivity.Mn intoxication inhibits Mg-dependent enzymes including MST enzymes.However, over-production of MST enzymes rescues YS cells, but not the PS strain, reinforcing the idea that the proximal cause of growth inhibition differs between these two phenotypically similar strains.

(
http://www.qtechcorp.com/),followed by the addition of 60/40 nitric/perchloric acid and incubation at 150 °C.Samples were cooled down and resuspended in deionized 18 MOhm water with 5% HNO3.ICP analysis was run on an Agilent 7700 series inductively coupled argon plasma-optical emission spectrometer (ICP-ES) housed at the US Department of Agriculture-ARS Robert W. Holley Center for Agriculture and Health in Ithaca, NY (mean ± standard deviation [SD]; n=3 independent biological replicates, each containing three technical replicates).

Table 2 : Intracellular Mg, Zn, and K levels. Mg levels (µg g -1 ) Zn levels (µg g -1 ) K levels (mg g -1 )
(11)nderstand the processes failing under such conditions we tested the effect of (i) PS suppressors of Mn intoxication isolated previously(11); Since YS phenocopies PS, with reduced mneP and mneS gene expression.(ii) newly isolated spontaneous suppressors from the Mn zone of inhibition for YS background, with Tukey's post hoc analysis; n.s. is the difference of non-significant value).Note the disk diameter is 6.5 mm.d) schematic view of site of mTn insertion in yazB is shown for YS Mn resistant mutant leading up to possible upregulation of mneP/S genes.e) disk diffusion assay was used to assess Mn sensitivity in various strains by monitoring diameter of clearance around Mn saturated disk (10 µl of 100 mM Mn) on MM-malate agar (MMA) plates supplemented with 0.8 % malate (n=4) **** is adjusted p < 0.0001 using a one-way ANOVA test with Tukey's post hoc analysis; n.s. is the difference of non-significant value.f) transcript levels of mneP and mneS were monitored in YS and YS yazB strains.Cells were grown in LB till OD600nm = 0.4 and then cells were incubated with or without 150 µM Mn for 30 min, cells were harvested, and RNA was purified.The expression was normalized to the gyrA transcript.Values shown are mean ± SD from 3 independent experiments.