Introduction

Non-proteinaceous low-molecular-weight (LMW) thiols play numerous functions in cellular physiology, but they are primarily recognized for maintaining a reduced intracellular environment and repairing oxidized proteins. Glutathione (GSH) is one of the most abundant and extensively studied LMW thiols (Fig. 1). Intracellular GSH is present at low millimolar concentrations and is maintained predominantly in its reduced form by GSH reductase (Fahey et al. 1978). GSH has roles in preventing osmotic and acidic stress, as well as protecting cells from intoxication by methylglyoxal, chlorine compounds and metal ions (Masip et al. 2006; Helbig et al. 2008a, b; Potter et al. 2012; Gutiérrez-Escobedo et al. 2013).

Fig. 1
figure 1

Structures of the major low-molecular-weight thiols found in Bacteria. The structures of glutathione, mycothiol, bacillitiol, coenzyme A and cysteine are shown

Full size image

A number of Gram-positive bacteria do not produce GSH (Fahey et al. 1978; Fahey 2013). In Actinomycetes, mycothiol (MSH) serves as the most abundant LMW thiol (Fig. 1). Strains unable to produce MSH displayed increased sensitivities to oxidative and nitrosative stressors, alkylating agents and several antibiotics (reviewed in Newton et al. 2008).

In some Firmicutes, such as Bacillus and Staphylococcus species, bacillithiol (BSH) is a major LMW thiol together with coenzyme A and cysteine (Fig. 1) (Gaballa et al. 2010). The structure of BSH was elucidated in 2009 (Newton et al. 2009) and shortly thereafter the genes involved in its biosynthesis were identified (Gaballa et al. 2010). The first step in BSH biosynthesis is catalyzed by the glycosyltransferase BshA, which produces N-acetylglucosaminylmalate (GlcNAc-Mal) from UDP-N-acetylglucosamine and l-malate. BshB then removes the acetyl group to generate N-glucosaminemalate (GlcN-Mal). The last step is catalyzed by BshC, which adds cysteine to produce BSH. In characterized Bacillus spp., the accumulation of BSH varied from 2 to 5 mM (Sharma et al. 2013), depending on the growth conditions, while lower levels (~1 mM) were detected in Staphylococcus aureus (Posada et al. 2014).

The use of strains defective in BSH biosynthesis and biochemical analyses using chemically or chemoenzymatically synthesized BSH has yielded significant insight into the functions of this LMW thiol. Similar to GSH, BSH was found to protect against peroxide intoxication (Posada et al. 2014; Handtke et al. 2014), thiol/disulfide stress (Gaballa et al. 2013; Posada et al. 2014) and hypochlorite stress (Pöther et al. 2013; Chi et al. 2013). BSH has also been shown to be involved in the detoxification of the thiol-reactive antibiotics fosfomycin (Roberts et al. 2013) and rifamycin (Newton et al. 2012), as well as the electrophilic metabolite methylglyoxal (Chandrangsu et al. 2014).

Bacillithiol has a role in Fe–S cluster biogenesis

Recent studies have found that BSH has a role(s) in Fe–S cluster biogenesis. S. aureus strains lacking BSH displayed phenotypic abnormalities when cultured in chemically defined medium, which were corrected by (i) genetic complementation, (ii) supplementing the growth media with Fe, or (iii) supplementing the growth media with the amino acids leucine (Leu) and isoleucine (Ile) (Rosario-Cruz et al. 2015). A Bacillus subtilis bshA mutant strain displayed similar growth phenotypes when cultured in chemically defined media (Fang and Dos Santos 2015), which were corrected by supplementing the growth medium with (i) GlcNAc-Mal, (ii) BSSB, the oxidized form of BSH, (iii) casamino acids, or (iv) Fe (Fang and Dos Santos 2015). In all cases, the authors speculated that the growth phenotypes were the result of defects in biochemical pathways that require Fe- or Fe–S cluster-dependent enzymes for functionality.

LeuCD and IlvD are dehydratase enzymes that require Fe–S clusters for the biosynthesis of the branched-chain amino acids (BCAA) Leu and Ile, respectively. S. aureus and B. subtilis cells lacking BSH displayed decreased LeuCD and IlvD activities when compared to their respective wild-type (WT) strains (Fang and Dos Santos 2015; Rosario-Cruz et al. 2015). These findings may explain why limiting for either Leu or Ile in the growth media exacerbated the growth abnormalities of the BSH-deplete cells.

The decreased activities of Fe–S cluster-dependent enzymes in B. subtilis and S. aureus cells lacking BSH were not limited to enzymes involved in BCAA biosynthesis. The activities of the Fe–S cluster-dependent enzymes aconitase (AcnA) and glutamate synthase (GOGAT or GltBD) were also decreased in cell-free lysates of bshA mutants. Supplementing the culture medium with Fe increased the enzymatic activities of GOGAT, AcnA and LeuCD in a B. subtilis bshA mutant (Fang and Dos Santos 2015), but it had no observable effect on the activities of these enzymes in a S. aureus bshA mutant (Rosario-Cruz et al. 2015).

Reactive oxygen species (ROS) can damage Fe–S clusters resulting in decreased activities of enzymes that require these cofactors for catalysis (Jang and Imlay 2007). Two findings suggest that the decreased Fe–S enzyme activities were not the result of ROS damage in BSH-deplete cells. First, the addition of Fe(II) to cell-free lysates generated from either B. subtilis or S. aureus bshA mutant strains had no effect on AcnA enzyme activity, suggesting that the absence of BSH did not cause an increased abundance of repairable Fe–S clusters in AcnA (Fang and Dos Santos 2015; Rosario-Cruz et al. 2015). Second, the Leu- and Ile-dependent growth abnormalities of S. aureus BSH-deplete cells persisted when cultured in the absence of dioxygen (Rosario-Cruz et al. 2015).

In S. aureus, Fe–S clusters are synthesized on the SufBCD scaffold before being transferred to Fe–S cluster carrier molecules, which then traffic the cofactors to target proteins. Efforts in creating strains with mutations in the sufCDSUB operon have been unsuccessful (Bae et al. 2004; Chaudhuri et al. 2009; Fey et al. 2013; Valentino et al. 2014, Mashruwala et al. 2015), suggesting that these gene products are essential and that Suf is the sole Fe–S cluster biosynthetic machinery system in S. aureus. Genetic and biochemical analysis by our group supports the hypotheses that the S. aureus Nfu (Mashruwala et al. 2015) and SufA (Rosario-Cruz et al. 2015) proteins are Fe–S cluster carriers. Strains lacking both Nfu and BSH had exacerbated growth defects, including glutamate/glutamine auxotrophies and displayed AcnA and GOGAT enzymatic activities that were significantly lower than strains lacking only Nfu or BSH (Rosario-Cruz et al. 2015). These results further support the hypothesis that BSH has a role in Fe–S cluster biogenesis and suggest that BSH functions independently of Nfu. Importantly, overexpression of nfu or sufA allowed for partial or full recovery of the growth phenotypes and enzymatic defects of the bshA mutant strain (Rosario-Cruz et al. 2015), providing genetic evidence that the role of BSH in Fe–S cluster biogenesis has functional overlap with Fe–S cluster carriers (Fig. 2).

Fig. 2
figure 2

Working model for the roles of bacillithiol (BSH) in Fe–S cluster biogenesis and intracellular metal ion buffering. BSH is proposed to have a role in the carriage of Fe–S clusters. BSH was shown to act as a Zn buffer in vivo and additional data suggest that BSH also has roles in the processing of alternate metals such as Fe, Cu and Mn, but further experimentation is necessary to determine whether BSH also acts as a buffer for these metals

Full size image

In B. subtilis, the SufBCD system is also predicted to be the sole Fe–S cluster biosynthetic system (Albrecht et al. 2010). In a B. subtilis bshA mutant strain, SufC accumulated to lower levels than in the WT, but SufB accumulated to similar levels in both strains (Fang and Dos Santos 2015). SufC is required for Suf function and decreased Fe–S cluster synthesis could account for the decreased Fe–S cluster-dependent enzyme activity observed in the B. subtilis BSH-deplete cells. The addition of the Fe chelator 2,2-dipyridyl (DIP) increased SufC expression in BSH-deplete cells, while AcnA activity was slightly lowered and GOGAT activity was unaffected. While the addition of Fe(II) corrected both the growth phenotypes and decreased Fe–S cluster-dependent enzyme activities of BSH-deplete cells, the effect of Fe addition on SufC accumulation was not reported.

BSH has roles in metal buffering

Metal ions such as zinc (Zn), iron (Fe), manganese (Mn) and copper (Cu) are required for the functionality of a number of metalloproteins and, therefore, it is thought that all cells require one or more of these metals in trace amounts. The use of these metals has both favorable and unfavorable consequences; the metals aid in the proper function of metabolism, yet elevated intracellular metal accumulation can result in intoxication. As a result, bacteria may use different mechanisms to prevent metal intoxication including the use of (i) transcriptional regulators that sense and respond to changes in intracellular metal concentrations, (ii) proteins that export metal ions from the cytosol, (iii) molecules that chelate or sequester metals, and (iv) molecules that chaperone or traffic metals to their target proteins within the cell.

Studies have found that BSH plays a role in metal ion buffering and, thereby, protects cells from metal ion intoxication (Fig. 2). Ma et al. discovered that BSH functions in buffering the intracellular Zn pool in B. subtilis (Ma et al. 2014). The authors found that BSH bound Zn(II) with high affinity in vitro, predominantly in a 2:1 ratio, and most likely using both the thiolate and carboxylate moieties as ligands. The authors also noted that the thiol group of BSH could coordinate Co(II). Challenging B. subtilis with Zn(II) led to decreased accumulation of Zn in the bshC mutant strain when compared to the WT, which was most likely due to increased expression of the CadA and CzcD metal efflux systems. Moreover, the absence of BSH resulted in increased sensitivity to Zn(II) and Cd(II) in strains defective in the export of these metals. Nearly, ~2/3 of the intracellular Zn was found to associate with LMW molecules in the WT strain. The authors proposed that the presence of BSH allows for increased accumulation of Zn and the absence of BSH reduces the Zn buffering capacity of the cell.

The authors also noted that the expression of cadA was increased after diamide treatment, which causes disulfide stress. Data presented suggest that upon depletion of the reduced form of BSH, Zn is released from the LMW pool, which binds to CzrA leading to cadA derepression. The authors also showed that the addition of BSH facilitated the removal of Zn from the CzrA in vitro, and propose this as a possible mechanism by which BSH facilitates Zn dissociation from CzrA in vivo upon the transition from Zn excess to normal conditions. These findings highlight the importance of BSH in the integration of the cellular networks involved in Zn homeostasis and thiol metabolism.

Bacillus subtilis and S. aureus strains lacking BSH display sensitivity to various metals including Cu(I), Cu(II), Cr(IV), Cd(II) and As(III) (Rajkarnikar et al. 2013; Ma et al. 2014; Fang and Dos Santos 2015). Ma et al. found that the expression of the cadA and czcD genes, which aid in Cd efflux, was not induced in the BSH-deplete strain upon Cd(II) treatment, suggesting that BSH is required for the induction of these systems under Cd stress. The authors hypothesize that Cd mobilizes Zn from BSH, which can then metalate CzrA resulting in derepression of the cadA and czcD genes. A bshC mutant has a lower Zn buffering capacity and, therefore, the amount of Zn released under Cd(II) treatment was presumably lower than that of the WT strain, resulting in decreased cadA and czcD derepression. Similarly, As(III) sensitivity was observed in the bshC mutant strain and a lower level of induction of the arsenic resistance system (arsR) occurred in cells lacking BSH when compared to the WT strain. The authors hypothesized that As(III) mobilization within the cell might be limited in the absence of BSH, thereby affecting As(III) loading into ArsR resulting in decreased expression of arsR. Alternatively, it was postulated that Zn released from BSH upon treatment with As(III) might somehow facilitate AsrR dissociation from the arsR promoter (Ma et al. 2014).

BSH also appears to have a role in Fe homeostasis. When cultured in a chemically defined medium, S. aureus BSH-deplete cells had increased transcription of genes necessary for Fe uptake, which are under the transcriptional control of the ferric uptake regulator (Fur) (Rosario-Cruz et al. 2015). This result suggests that BSH depletion leads to non-metalated Fur and derepression of genes involved in Fe acquisition. Further analysis found that the overall Fe load in the WT and BSH-deplete cells was statistically indistinguishable. Moreover, the growth phenotypes displayed by the bshA mutant were corrected by supplementing the growth medium with Fe and exacerbated by the addition of DIP (Rosario-Cruz et al. 2015 and Z. Rosario-Cruz and J.M. Boyd, unpublished results). As previously mentioned, the growth phenotypes were also suppressed by overexpression of sufA, which encodes for an A-type protein (Vinella et al. 2009; Rosario-Cruz et al. 2015). In addition to their roles as Fe–S cluster carriers, A-type proteins bind Fe(II) and Fe(III) in vitro (Ding and Clark 2004; Mapolelo et al. 2012) and evidence supports a role for these proteins in Fe donation for Fe–S cluster synthesis in vivo (Ding et al. 2004; Landry et al. 2013). Perhaps increasing the concentration of SufA improves the growth of BSH-deplete cells by increasing bioavailable Fe for processes such as Fe–S cluster synthesis.

Bacillus subtilis cells lacking BSH had a slight, but statistically significant decrease in total cellular Fe (Fang and Dos Santos 2015). These strains also displayed decreased activities of the mononuclear Fe-dependent enzymes threonine dehydrogenase and quercetin 2,3-dioxygenase. Supplementing the growth media with Fe(II) corrected the growth abnormalities of the B. subtilis BSH-deplete cells and increased Fe–S cluster-dependent enzyme activity. The addition of Fe(II) to cell-free lysates did not alter the activities of mononuclear Fe-dependent enzymes, but the effect of Fe supplementation to the growth medium on their activities was not reported. The addition of DIP increased the expression of SufC in BSH-deplete cells and, thereby, possibly increased the output of Suf. In the WT strain, DIP treatment did not increase SufC expression providing further evidence that BSH has a role in Fe homeostasis.

Bacillus subtilis BSH-deplete cells had a slight increase in intracellular Mn when compared to the WT strain (Fang and Dos Santos 2015). Upon challenge with paraquat, which results in superoxide stress, both the WT and bshA mutant accumulated Mn, but the accumulation was greater in the BSH-deplete cells. Mn accumulation occurred concomitant with increased Mn-dependent superoxide dismutase (Sod) activity. In Escherichia coli, Mn aids in combating oxidative stress by replacing Fe in mononuclear Fe enzymes, thereby maintaining protein functionality by shifting towards a Mn-centered metabolism (Sobota and Imlay 2011). In addition, free intracellular Mn can eliminate superoxide stress in B. subtilis by a yet uncharacterized mechanism (Inaoka et al. 1999). This may explain why the B. subtilis WT and BSH-deficient cells have increased levels of both labile Mn and protein-associated Mn under paraquat stress (Fang and Dos Santos 2015). Why BSH-deplete cells have increased basal levels of Mn and increased Mn uptake upon paraquat treatment remains unknown.

Both S. aureus and B. subtilis cells lacking BSH are sensitive to Cu(II) intoxication (Rajkarnikar et al. 2013). B. subtilis bshA mutants are also sensitive to Cu(I) intoxication (Fang and Dos Santos 2015). Copper toxicity could be, in part, a result of a Cu-catalyzed Fenton-type reaction resulting in the production of hydroxyl radicals and DNA damage (Gunther et al. 1995). In E. coli, copper intoxication does not appear to be a result of increased DNA damage (Macomber et al. 2007), but Cu does inactivate Fe–S cluster-dependent enzymes (Macomber and Imlay 2009). Consistent with these findings, Cu(I) treatment affected AcnA and GltBD activities in the B. subtilis WT strain. However, the activities of these enzymes were unaffected by Cu(I) challenge in the bshA mutant strain (Fang and Dos Santos 2015). We have also noted that S. aureus BSH-deplete cells display increased generation times and lower growth yields when Cu(II) is provided in the growth media, but these phenotypes persist in the absence of dioxygen (Z. Rosario-Cruz and J.M. Boyd, unpublished results). These findings suggest that the S. aureus Cu(II)-dependent growth defects are not the result of ROS accumulation.

Conclusions and future directions

The discovery of BSH as one of the major LMW thiols in some Gram-positive bacteria is relatively recent and many questions about its physiological roles remain unanswered. Genetic studies suggest that the role of BSH in Fe–S cluster biogenesis has functional overlap with Fe–S cluster carrier molecules, but the exact role of BSH in Fe–S cluster biogenesis is currently unknown. GSH has been shown to act as a ligand for Fe–S cluster binding in vitro (Qi et al. 2012). GSH molecules can also act as ligands for Fe–S cluster binding in conjunction with monothiol glutaredoxins (Feng et al. 2006). In both scenarios, it is thought that GSH participates as an Fe–S cluster carrier in vivo. Further genetic and biochemical studies are necessary to determine whether BSH, like GSH, can ligate an Fe–S cluster and/or ligate an Fe–S cluster in conjunction with a protein partner. It will then be necessary to examine whether BSH or the protein-BSH complex can transfer the bound Fe–S clusters to apo-proteins.

BSH can bind Zn(II) and Co(II) in vitro and data suggest that it serves as an intracellular Zn buffer in vivo (Ma et al. 2014). Studies by others in the field imply that BSH also has roles in the processing of alternate metals, such as Fe, Mn, As, Cd and Cu, but it is currently unknown if BSH directly ligates these metals. Additional physiological and biochemical studies will be necessary to determine if BSH can ligate these metals and to uncover additional role(s) for BSH in metal ion homeostasis and potentially detoxification. It will also be interesting to learn more about the role(s) of BSH in altering the affinities of ArsR and CzrA for DNA.

Bacteria face a number of challenges in their environments including metal limitation. Not surprisingly, human pathogenic bacteria face metal limitation within our bodies (reviewed in Hood and Skaar 2012). Additionally, human macrophages bombard phagocytized bacteria with Cu, which disrupts Fe–S cluster biogenesis, to aid in bacterial clearance (Johnson et al. 2015). Pathogenic bacteria, such as S. aureus, are notorious for their ability to adapt to various stress environments, allowing them to successfully cause a wide range of infections (Tuchscherr and Löffler 2015). Having a better understanding of the role of gene products, metabolites and/or regulatory proteins required to cope with stressful conditions during host infection will aid in identifying new antimicrobial targets. In some bacterial pathogens, such as S. aureus, BSH may play various roles during the infection process such as serving as an intracellular reservoir of metals, thereby allowing functionality of metal-dependent proteins during periods of metal starvation. BSH may also serve as a buffering agent to prevent the detrimental effects of metal-induced intoxication. In summary, not only does BSH have roles in combating oxidative stress and detoxification of thiol-reactive antibiotics, but it also participates in Fe–S cluster biogenesis and metal homeostasis, which are processes required for successful bacterial colonization and infection. Moreover, BSH is absent in mammals, making BSH biosynthesis an attractive antimicrobial target. Future studies using models of infection will be necessary to elucidate the roles of BSH in microbial pathogenesis.