Thiol-based redox switches in the major pathogen Staphylococcus aureus

Enzymatic detoxification systems

Functions of bacillithiol in detoxification and antibiotics resistance

LMW thiols are non-proteinogenous thiol compounds that are present in millimolar concentrations in the bacterial cytoplasm and function in maintenance of redox homeostasis (Chandrangsu et al. 2018; Loi et al. 2015; Van Laer et al. 2013). While the tripeptide glutathione (GSH) is utilized as LMW thiol by most Gram-negative bacteria, Staphylococcus species and other firmicutes produce bacillithiol (BSH) as alternative LMW thiol, which functions as GSH surrogate to maintain redox homeostasis (Chandrangsu et al. 2018). BSH derivatives are present in Thermus thermophiles and in phototrophic Chlorobiaceae (Hiras et al. 2018; Newton and Rawat 2019; Norambuena et al. 2018) and widely distributed in archaea (Rawat and Maupin-Furlow 2020). In addition, S. aureus uses coenzyme A (CoASH) as alternative LMW thiol, which is maintained in its reduced state by the CoASH disulfide reductase Cdr (delCardayre and Davies 1998; delCardayré et al. 1998).

Bacillithiol (BSH, Cys-GlcN-malate) is an α-anomeric glycoside of L-cysteinyl-d-glucosamine with l-malic acid of 398 Da (Figure 3), which is synthesized by the glycosyltransferase BshA, the deacetylase BshB and the Cys ligase BshC in S. aureus (Chandrangsu et al. 2018; Gaballa et al. 2010; Newton et al. 2009). Using biophysical methods, the BSH standard redox potential was determined as E^0’(BSSB/BSH) of −221mV (Sharma et al. 2013). However, Brx-roGFP2 measurements inside S. aureus revealed a more negative BSH redox potential (E_BSH) ranging from −282 to −295 mV during the growth (Loi et al. 2017). Under oxidative stress provoked by H₂O₂ and HOCl stress, BSH is oxidized to BSH disulfide (BSSB) as shown in vitro and in vivo in S. aureus (Dickerhof et al. 2020). At higher levels, HOCl leads to BSH sulfonamide formation and overoxidation to BSH sulfonic acids (Dickerhof et al. 2020). The reaction of HOCl with BSH is very fast and occurs with second order rate constants of 6 x 10⁷ M⁻¹s⁻¹ (Dickerhof et al. 2020).

Figure 3:

Bacillithiol functions in detoxification, antibiotics resistance and virulence in S. aureus. BSH is involved in detoxification of ROS, RES, HOCl, RSS and confers resistance to the redox active antibiotics fosfomycin, allicin, lapachol and AGXX® in S. aureus. (1) ROS, HOCl and AGXX® oxidize BSH to bacillithiol disulfide (BSSB), which is recycled by the BSSB reductase YpdA. BSSB could further react with HS⁻ generating BSH persulfides (BSSH), which are regenerated by the dioxygenase CstB. (2) Allicin reacts with BSH to S-allylmercaptobacillithiol (BSSA) as another substrate of YpdA. (3) The S-transferase BstA conjugates RES to BSH, forming BS-electrophiles (BS-R), which are cleaved by the amidase BshB to GlcN-Mal and mercapturic acids (Cys-SR), followed by their export. (4) The S-transferase FosB conjugates BSH to fosfomycin for its inactivation. (5) HOCl leads to S-bacillithiolations of proteins, which are reversed by the Brx/BSH/YpdA/NADPH pathway. (6) The Brx/BSH/YpdA/NADPH pathway is important under infections in S. aureus. (7) BSH functions in FeS cluster assembly in S. aureus.

The reduced state of BSH is maintained by the NADPH-dependent flavin disulfide reductase YpdA in S. aureus, which functions as a BSSB reductase (Figure 3) (Linzner et al. 2019; Mikheyeva et al. 2019). YpdA acts in the Brx/BSH/YpdA/NADPH redox pathway to regenerate S-bacillithiolated proteins under oxidative stress in S. aureus as outlined in the next sections (Linzner et al. 2019).

Importantly, BSH was identified as virulence factor in clinical MRSA isolates, such as COL and the highly virulent USA300 strain. BSH promotes survival inside murine macrophages and neutrophils in human whole-blood infection assays (Posada et al. 2014; Pöther et al. 2013; Rajkarnikar et al. 2013). Moreover, S. aureus isolates of the NCTC8325-4 lineage evolved mutations in the bshC gene, leading to the lack of BSH synthesis (Newton et al. 2012; Pöther et al. 2013). Thus, the S. aureus strain SH1000 of the NCTC8325-4 lineage was impaired in survival in infection assays compared to the SH1000 bshC repaired strain (Pöther et al. 2013). Interestingly, the S. aureus bshA, ypdA and brxAB mutants showed similar defects in survival inside J774.1 macrophages and neutrophils, indicating that the Brx/BSH/YpdA/NADPH system is essential under infections to regenerate BSH and protein thiols (Linzner et al. 2019; Mikheyeva et al. 2019; Pöther et al. 2013; Posada et al. 2014). Future studies should investigate the physiological role of BSH and the Brx/BSH/YpdA/NADPH redox pathway in S. aureus in murine infection models with defects in the NADPH oxidase and myeloperoxidase to elucidate if BSH homeostasis provides protection against the oxidative burst of neutrophils and macrophages.

Phenotype analyses of BSH-deficient mutants revealed important functions of BSH in the defense against ROS, HOCl, electrophiles, metals, xenobiotics, fosfomycin, rifampicin, AGXX®, allicin and lapachol in S. aureus (Figure 3)(Chandrangsu et al. 2018; Linzner et al. 2020; Loi et al., 2018a, 2019).

BSH is an important cofactor for the thiol-S-transferase FosB, which confers fosfomycin resistance. FosB catalyzes the ring cleavage of fosfomycin leading to formation of the BS-fosfomycin-conjugate (Figure 3) (Chandrangsu et al. 2018; Roberts et al. 2013). Both fosB and bshA mutants were similarly susceptible under fosfomycin stress in S. aureus, supporting the function of FosB as a BSH-dependent thiol-S-transferase (Posada et al. 2014; Rajkarnikar et al. 2013). In addition, the BSH-dependent S-transferase BstA was shown to conjugate BSH to electrophiles, such as 1-chloro-2,4-dinitrobenzene, monobromobimane and the antibiotic cerulenin in S. aureus (Newton et al., 2011,2012; Perera et al. 2014). BstA belongs to the widely conserved DinB/YfiT superfamily of S-transferases in BSH-, MSH- and GSH-producing bacteria (Newton et al. 2011; Perera et al. 2014). The resulting BS-electrophile conjugates are cleaved by the BSH S-conjugate amidase BshB to GlcN-Mal and mercapturic acids (CysSR), which are subsequently exported (Figure 3) (Newton et al. 2011). Thus, the thiol-S-transferases FosB and BstA are important determinants of antibiotics resistance in S. aureus.

Additionally, BSH and the Brx/BSH/YpdA/NADPH pathway were shown to confer protection against the naphthoquinone lapachol in S. aureus (Linzner et al. 2020). Lapachol acts via the oxidative mode, resulting in ROS formation, which leads to increased thiol oxidations and an oxidative shift in the E_BSH in S. aureus. Thus, the bshA, brxAB and ypdA mutants were more sensitive in growth and survival assays under lapachol treatment, indicating that the Brx/BSH/YpdA/NADPH pathway is essential to restore thiol homeostasis under lapachol stress (Linzner et al. 2020).

Similarly, BSH was shown to promote resistance to the antimicrobial surface coating AGXX® (Largentech GmbH, Berlin), which is composed of Ag⁺ and Ru⁺ ions and can be electroplated on medical devices and implants (Clauss-Lendzian et al. 2018; Gupta et al. 1999; Guridi et al. 2015; Heiss et al. 2017; Loi et al. 2018a; Vaishampayan et al. 2018). The Ag⁺ and Ru⁺ ions form a micro-galvanic cell, resulting in H₂O₂ and OH· generation due to the electric field. In support of ROS production, AGXX® caused strong oxidative, quinone and metal stress responses in the S. aureus transcriptome (Loi et al. 2018a). AGXX® induced an oxidative shift in the E_BSH and protein S-bacillithiolation of GapDH. The bshA mutant was strongly impaired in growth and survival under AGXX® stress, indicating an important role of BSH in protection against ROS generated by AGXX^® (Loi et al. 2018a). Thus, BSH is a beneficial antioxidant in S. aureus, to ensure survival during host-pathogen interactions and under ROS-producing antimicrobials.

Furthermore, BSH functions in detoxification of RSS and RNS in S. aureus, such as H₂S and nitroxyl (HNO), the latter being an intermediate of NO· and H₂S (Peng et al. 2017a,b). H₂S caused widespread persulfidations in the proteome and thiol metabolome of S. aureus as revealed by elevated amounts of BSH persulfide (BSSH), CoASH persulfide (CoASSH) and cysteine persulfide (CysSSH) after exposure to Na₂S and HNO (Figure 3) (Peng et al. 2017a,b). Among the targets for persulfidations are the glycolytic GapDH and the major virulence regulator MgrA, which both were redox-controlled and inactivated by persulfidations of the active site Cys residues. The reversal of persulfidations was controlled by the novel thioredoxins TrxP and TrxQ (Peng et al. 2017bPeng et al. 2017).

In addition, BSH is crucial for resistance towards allicin and polysulfanes from garlic in S. aureus and Bacillus subtilis (Borlinghaus et al. 2014; Chi et al. 2019; Loi et al. 2019). The diallylthiosulfinate allicin decomposes upon heating to diallyl polysulfanes with up to seven sulfur chains (Münchberg et al. 2007; Tocmo et al. 2017). Allicin and diallyl tetrasulfane (DAS4) caused a strong oxidative and sulfur stress response in the transcriptome and widespread S-thioallylations of redox sensitive enzymes and transcriptional regulators in the proteome (Figure 2) (Chi et al. 2019; Loi et al. 2019). BSH-deficient mutants of S. aureus and B. subtilis are strongly impaired in growth and survival under allicin and DAS4 treatment (Arbach et al. 2019; Chi et al. 2019; Loi et al. 2019). In S. aureus, allicin caused a strong oxidative shift in the E_BSH, which is caused by BSH depletion and formation of S-allylmercaptobacillithiol (BSSA) (Figure 3). BSSA was identified as substrate of the BSSB reductase YpdA to regenerate BSH by production of allyl thiol (Loi et al. 2019). In addition, the disulfide reductase MerA was shown to function in reduction of allicin to allyl thiols and the Brx/BSH/YpdA/NADPH pathway was important for regeneration of S-thioallylated proteins in S. aureus. Consistent with their biochemical functions in allicin detoxification, the bshA, brxAB and ypdA mutants were susceptible to allicin stress in S. aureus (Loi et al. 2019). In B. subtilis, the toxicity of polysulfanes was increased with longer sulfur chains (Arbach et al. 2019). Exposure to DAS3 and DAS4 caused strong depletion of BSH and cysteine, which was accompanied by increased formation of allyl thiols (Arbach et al. 2019), suggesting detoxification of polysulfanes in B. subtilis via related pathways as revealed in S. aureus (Loi et al. 2019).

During host-pathogen interactions, S. aureus has to cope with nutritional immunity caused by restriction of iron and other metal ions by host proteins (Maresso and Schneewind 2006; Marchetti et al. 2020). Thus, S. aureus must develop strategies to maintain Fe²⁺ homeostasis to ensure the biosynthesis of FeS clusters and Fe²⁺-containing enzymes, such as catalase and superoxide dismutase (Maresso and Schneewind 2006; Marchetti et al. 2020). BSH plays an important role in FeS-cluster biogenesis in S. aureus (Figure 3). Specifically, BSH-deficient mutants were impaired in growth under iron starvation (Rosario-Cruz and Boyd 2016; Rosario-Cruz et al. 2015). BSH mutants showed growth defects in the absence of branched chain amino acids due to decreased synthesis of FeS cluster-containing enzymes, involved in leucine or isoleucine biosynthesis (Rosario-Cruz and Boyd 2016; Rosario-Cruz et al. 2015). It was suggested that BSH functions in FeS cluster assembly and transport to target proteins independently of the SufA and Nfu carrier proteins (Rosario-Cruz and Boyd 2016; Rosario-Cruz et al. 2015). However, the role of BSH in FeS cluster assembly is still unknown. It might be possible that BSH coordinates FeS cluster biogenesis together with the bacilliredoxins BrxAB as shown for GSH and glutaredoxins in eukaryotes (Lill 2020).

Redox regulation of proteins by protein S-bacillithiolation

During infections, S. aureus encounters the highly bactericidal HOCl as part of the oxidative burst by activated neutrophils (Winterbourn and Kettle 2013; Winterbourn et al. 2016). HOCl stress leads to formation of BSH mixed protein disulfides, termed as protein S-bacillithiolations, which are widespread across Bacillus and Staphylococcus species (Chi et al., 2011, 2013; Lee et al. 2007). These S-bacillithiolated proteins include metabolic and antioxidant enzymes, translation factors and redox sensing regulators, which respond to oxidative stress (Figure 4) (Chi et al., 2011,2013; Imber et al., 2018a, 2019). S-bacillithiolation was proposed to function in redox regulation of HOCl-specific transcription factors and to prevent irreversible overoxidation of vulnerable Cys thiols to sulfonic acids (Imber et al. 2019). In B. subtilis, the OhrR repressor was shown to sense HOCl and organic hydroperoxides (OHP) by S-bacillithiolation at its lone Cys15 residue, resulting in expression of the OhrA peroxidredoxin, which confers OHP and HOCl resistance (Chi et al. 2011; Lee et al. 2007).

Figure 4:

Physiological roles of S-bacillithiolations of GapDH, AldA and SarZ under HOCl stress in S. aureus. The OhrR-type repressors MgrA and SarZ were identified as HOCl-sensitive using OxICAT and S-thioallylated at Cys13 by allicin in S. aureus, indicating their possible redox regulation by S-bacillithiolation under HOCl stress (Imber et al. 2018a; Loi et al. 2019). SarZ controls the ohr peroxiredoxin, efflux pumps for antibiotics (norB, tet38), virulence factors and metabolic genes. SarZ confers resistance to OHP and antibiotics and contributes to virulence (Chen et al. 2011). HOCl stress causes S-bacillithiolation of the glycolytic GapDH and the aldehyde dehydrogenase AldA in S. aureus, resulting in the switch from glycolysis to the pentose phosphate pathway for NADPH regeneration (Deng et al. 2013; Imber et al. 2018a). AldA might be involved in methylglyoxal detoxification under HOCl stress. Abbreviations are: Pgi, glucose 6 phosphate isomerase; Pfk, phosphofructokinase; FbaA/B, fructose-1,6-bisphosphate aldolase; GapDH, glyceraldehyde 3-phosphate dehydrogenase; Pgk, phosphoglycerate kinase; Pgm, phosphoglycerate mutase; Eno, enolase; Pyk, pyruvate kinase; Zwf, glucose 6 phosphate dehydrogenase; TpiA, triose phosphate isomerase; MgsA, methylglyoxal synthase.

In S. aureus, two OhrR homologs SarZ and MgrA are important virulence and antibiotics resistance regulators and share the conserved N-terminal Cys residue of the OhrR family, which senses oxidative stress via thiol oxidation (Figure 4) (Chen et al. 2011; Hillion and Antelmann 2015). The crystal structure of SarZ was resolved with Cys13 in the reduced, Cys-SOH and S-thiolated form with benzene thiol (Poor et al. 2009). Only the S-thiolated SarZ resulted in conformational changes in the DNA-binding helix-turn-helix motif, resulting in dissociation of SarZ from the target gene promoter (Poor et al. 2009). Increased thiol oxidation of MgrA and SarZ was found under HOCl stress in the redox proteome of S. aureus, supporting the hypothesis that both might be redox-controlled by S-bacillithiolation in vivo (Imber et al. 2018a). Furthermore, MgrA was regulated by persulfidation under sulfide stress, which impacts virulence factor secretion in the secretome of S. aureus (Peng et al. 2017bPeng et al. 2017).

Additionally, conserved S-bacillithiolated proteins were identified as metabolic enzymes with redox active catalytic centers, such as the methionine synthase MetE, the inosine monophosphate (IMP) dehydrogenase GuaB, the glyceraldehyde 3-phosphate dehydrogenase GapDH and the aldehyde dehydrogenase AldA (Figure 4) (Chi et al. 2011; Imber et al., 2018a, 2019). MetE represents the most abundant S-bacillithiolated protein in B. subtilis cells, which forms the BSH mixed disulfide at Cys730 in the active site Zn²⁺ center, resulting in inactivation of the enzyme and methionine auxotropy under HOCl stress. Moreover, the enzymes SerA, PpaC, MetI and YxjG were S-bacillithiolated under HOCl stress, which act in the same pathway as MetE (Chi et al. 2011; Imber et al., 2018a, 2019).

The glycolytic GapDH displayed 29% increased thiol oxidation under HOCl stress and was the most abundant S-bacillithiolated protein in S. aureus cells, representing 4% of the total Cys proteome (Imber et al., 2018a). GapDH S-bacillithiolation occurs under oxidative stress at the active site Cys151, resulting in enzyme inactivation and metabolic switching from glycolysis to the pentose phosphate pathway (Deng et al. 2013; Imber et al. 2018a). Interestingly, GapDH inactivation by S-bacillithiolation proceeds faster compared to its inhibition by overoxidation in the absence of BSH in vitro (Imber et al. 2018a). Thus, S-bacillithiolation functions in redox regulation of GapDH activity and can prevent overoxidation of the active site (Imber et al. 2018a). Using molecular docking, BSH was modeled into the Cys151 active site of the apo- and holoenzymes. Interestingly, BSH occupies two different positions in the GapDH active site depending on the NAD⁺ cofactor. While the BSH moiety is located in the active site with Cys151 in the attacking state position in the holoenzyme, S-bacillithiolation of the apoenzyme occurs with Cys151 in the resting state in the absence of the NAD⁺ cofactor (Imber et al., 2018a,2019). However, S-bacillithiolation of GapDH did not cause major structural changes.

Similarly, the NAD⁺-dependent aldehyde dehydrogenase AldA showed 29% increased thiol oxidation at its active site Cys279 and was inactivated by S-bacillithiolation under HOCl stress in S. aureus (Imber et al. 2018a,b). Computational chemistry revealed similar attacking and resting state positions of Cys279 upon S-bacillithiolation of the AldA holo- and apoenzymes, respectively, as shown for GapDH. In addition, S-bacillithiolation of AldA elicits no major structural and conformational changes (Imber et al., 2018b,2019). AldA was shown to catalyze oxidation of formaldehyde, methylglyoxal, acetaldehyde and glycolaldehyde in vitro, but its physiological substrate in vivo is unknown. Transcription of aldA is elevated under oxidative and electrophile stress, such as formaldehyde, methylglyoxal, HOCl, allicin and AGXX®, and depends of an unknown redox regulator. Phenotype analyses revealed that AldA is required for survival under HOCl stress, but dispensable under aldehyde stress. Thus, the physiological role of AldA under HOCl stress and its regulation remain to be elucidated in future studies (Imber et al. 2018b).

Apart from GapDH and AldA, GuaB is another NAD⁺-containing enzyme with a highly conserved Cys308 active site that forms an adduct with the IMP substrate. GuaB is the most conserved target for S-glutathionylation, S-bacillithiolation and S-mycothiolation across prokaryotes and eukaryotes (Imber et al. 2019; Loi et al. 2015). Based on the different conformations of Cys308 in the substrate bound enzyme or in the apoenzyme structures, the BSH moiety may adopt a similar position in the GuaB active site with Cys308 in the attacking or resting state, respectively, as shown for GapDH and AldA (Imber et al., 2018a,b,2019). Altogether, S-bacillithiolations were shown to occur at accessible active site Cys residues of redox sensitive metabolic enzymes or transcriptional regulators. While S-bacillithiolation of GapDH and AldA does not lead to conformational changes, major structural rearrangements in the DNA binding HTH motif are required to accommodate redox regulation of the 1-Cys type OhrR-family repressors.

Reversal of protein S-bacillithiolation by the Brx/BSH/YpdA/NADPH pathway

The bacilliredoxins BrxA, BrxB and BrxC were previously identified as S-bacillithiolated in B. subtilis and Staphylococcus carnosus under HOCl stress, which suggested their role in the reversal of protein S-bacillithiolations (Chi et al. 2013). BrxA and BrxB are paralogs of the DUF1094 family with a Trx fold and a CGC active site motif, which co-occur together with the BSH synthesis enzymes in BSH-producing bacteria (Gaballa et al. 2014). However, the redox potential of BrxA is relatively positive with −130 mV, and rather in the range of disulfide isomerases (Derewenda et al. 2009). BrxC has a TCPIS active site motif suggesting its possible function as monothiol Brx.

The functions of BrxA and BrxB as bacilliredoxins in reduction of S-bacillithiolated proteins have been first demonstrated for the S-bacillithiolated OhrR and MetE proteins in B. subtilis (Gaballa et al. 2014). DNA binding assays revealed the reactivation of OhrR after de-bacillithiolation of OhrR-SSB with the BrxB resolving Cys mutant protein, but not with the BrxB active site mutant. The removal of BSH from S-bacillithiolated MetE in B. subtilis cell extracts was shown with both BrxA and BrxB using BSH specific Western blots and mass spectrometry (Gaballa et al. 2014). In S. aureus, Brx activity assays were conducted with purified GapDH as most abundant S-bacillithiolated protein in vivo (Imber et al. 2018a). The S. aureus BrxA homolog catalyzed the de-bacillithiolation of GapDH, as shown by reactivation of the glycolytic activity of GapDH to oxidize the glyceraldehyde 3-phosphate substrate to 1,3-bis-phoshoglycerate with generation of NADH. While the BrxA resolving Cys mutant could restore GapDH activity, the BrxA active site mutant did not. The de-bacillithiolation reaction could be monitored using non-reducing BSH specific Western blots (Imber et al. 2018a). These results established the function of BrxA and BrxB in reversal of S-bacillithiolations in vitro.

Furthermore, the flavin disulfide reductase YpdA was phylogenetically associated with the BSH biosynthesis enzymes and bacilliredoxins (Gaballa et al. 2010). Transcriptome analyses revealed increased transcription of ypdA, brxA, brxB, bshA, bshB and bshC under oxidative and electrophile stress, such as H₂O₂, HOCl, diamide, AGXX®, lapachol, allicin and azurophilic granule proteins in S. aureus (Linzner et al. 2020; Loi et al., 2018a,b,2019; Palazzolo-Ballance et al. 2008). These data provide evidence for a functional Brx/BSH/YpdA/NADPH redox cycle in S. aureus (Figure 3), which is co-regulated under redox stress conditions. YpdA was characterized as a NADPH-dependent BSSB reductase in S. aureus to regenerate BSH homeostasis during recovery from oxidative stress and under infections (Linzner et al. 2019; Mikheyeva et al. 2019). Moreover, the S. aureus brxAB, bshA and ypdA mutants were more sensitive in growth and survival under HOCl stress and in macrophage infection assays in vivo (Linzner et al. 2019; Mikheyeva et al. 2019; Pöther et al. 2013; Posada et al. 2014). In addition, the ypdA mutant showed a 2-fold increased basal level of BSSB compared to the wild type (WT). Overproduction of YpdA, in turn, resulted in higher BSH levels (Mikheyeva et al. 2019). While the basal BSH levels of the S. aureus COL WT and the ypdA mutant were determined in a similar range of 1.5–1.9 µmol/g raw dry weight (rdw), the basal BSSB levels were measured as ∼0.05 µmol/g and ∼0.09 µg rdw in WT and ypdA mutant cells, respectively (Linzner et al. 2019). Thus, the BSH/BSSB ratio was determined as 35:1 for the WT and decreased to 17:1 in the ypdA mutant under non-stress conditions. Based on the high BSH levels in both WT and ypdA mutant strains, the basal E_BSH was not affected under non-stress conditions along the growth curve as measured using the Brx-roGFP2 biosensor, which is described in more detail in the following section (Linzner et al. 2019). However, due to the increased BSSB levels, the ypdA mutant was impaired to regenerate the reduced E_BSH upon recovery from oxidative stress, supporting the function of YpdA as BSSB reductase especially under oxidative stress in S. aureus (Linzner et al. 2019; Mikheyeva et al. 2019).

These results confirm previous findings, which revealed unchanged cytosolic GSH levels, but increased GSSG amounts in a yeast mutant deficient for the glutathione disulfide reductase Glr (Morgan et al., 2011,2013). Grx-roGFP2 biosensor measurements showed that the basal GSH redox potential (E_GSH) was only slightly increased by ∼20 mV in the yeast glr mutant, whereas the H₂O₂ response was much stronger in the glr mutant with an impaired recovery of the reduced state (Morgan et al. 2011). It was further shown that Grx2 and Trx2 can compensate for GSSG reduction in the absence of Glr (Morgan et al. 2013). However, in S. aureus BrxA did not show BSSB reductase activity in vitro and the BSSB levels were not affected in the brxAB mutant, indicating that BrxAB cannot replace YpdA in regeneration of BSH (Linzner et al. 2019; Mikheyeva et al. 2019).

YpdA is a flavin disulfide reductase, which contains a conserved Cys14 residue in the glycine-rich Rossmann-fold NADPH binding domain (GGGPC₁₄G) (Bragg et al. 1997; Mikheyeva et al. 2019). Using NADPH coupled electron transfer assays, we demonstrated that YpdA consumes NADPH in the presence of BSSB, but not with other LMW thiol disulfides, such as GSSG, cystine or CoAS₂, confirming its function as BSSB reductase in S. aureus in vitro (Linzner et al. 2019). The BSSB reductase activity of YpdA was dependent on the conserved Cys14, which is unique in YpdA homologs and represents a novel active site. In support of its active site function, Cys14 of YpdA was identified as HOCl-sensitive with >10% increased oxidation using the redox proteomics approach OxICAT in S. aureus (Imber et al. 2018a). Thus, Cys14 might be S-bacillithiolated by BSSB, and regenerated by electron transfer from NADPH via the FAD co-factor. Moreover, YpdA works in concert with BrxA and BSH to establish a functional Brx/BSH/YpdA/NADPH redox pathway for reduction of S-bacillithiolated proteins in S. aureus. In biochemical assays with the complete Brx/BSH/YpdA/NADPH pathway, BrxA was sufficient for the complete de-bacillithiolation of GapDH. In contrast, YpdA cannot regenerate S-bacillithiolated proteins, but functions instead in BSSB reduction to complete the Brx/BSH/YpdA/NADPH cycle (Linzner et al. 2019). In this redox cycle, the BrxA active site Cys attacks the S-bacilithiolated protein, resulting in Brx-SSB formation (Gaballa et al. 2014). Brx-SSB is resolved with BSH, leading to BSSB formation, which is recycled by YpdA to BSH (Linzner et al. 2019). However, structural analyses are required to resolve the detailed catalytic mechanism of YpdA in BSSB reduction. Future studies should be directed to identify the substrates of the Brx/BSH/YpdA/NADPH pathway under infection conditions inside host cells, which promote virulence, survival and persistence of S. aureus inside the host.

Monitoring the changes in the E_BSH using the Brx-roGFP2 biosensor

To monitor E_BSH changes inside S. aureus, BrxA were fused to the redox sensitive green fluorescent protein (roGFP2), generating a genetically encoded Brx-roGFP2 biosensor (Loi et al. 2017; Loi and Antelmann 2020). Upon oxidation, the two Cys residues in roGFP2 form a disulfide, influencing the spectral properties of roGFP2 (Meyer and Dick 2010; Schwarzländer et al. 2016). Specifically, the 488 nm excitation maximum is decreased and the 405 nm maximum increased, resulting in ratiometric changes in the 405/488 excitation ratio, which is quantified as oxidation degree (OxD). The fusion of Brx to roGFP2 facilitates the equilibration of roGFP2 with the BSH/BSSB redox pair enabling specific measurements of the E_BSH in S. aureus (Loi and Antelmann 2020; Loi et al. 2017; Meyer and Dick 2010; Schwarzländer et al. 2016). ROS exposure causes oxidation of cellular BSH to BSSB, which reacts with the Brx active site Cys of the Brx-roGFP2 probe, leading to Brx-SSB formation. The BSH moiety of Brx-SSB is transfered to the coupled roGFP2, which rearranges to the roGFP2 disulfide resulting in ratiometric changes of the excitation spectrum (Figure 5). Assuming that roGFP2 and the BSH/BSSB pair are in equilibrium, E_BSH is equal to the calculated E_roGFP2. Consequently, E_BSH can be calculated based on the OxD values of the Brx-roGFP2 biosensor and the previously determined EroGFP2o′ of −280 mV (Dooley et al. 2004), according to the Nernst equation as described (Loi et al. 2017; Morgan et al. 2011).

Figure 5:

Principle, ratiometric changes and results of the Brx-roGFP2 biosensor in S. aureus. (A) The Brx active site thiolate of the Brx-roGFP2 biosensor reacts with BSSB, resulting in Brx-SSB formation, transfer of BSH to roGFP2, and re-arrangement to the roGFP2 disulfide (Loi et al. 2017). (B) Brx-roGFP2 oxidation results in ratiometric changes at the 405 and 488 nm maxima of the excitation spectrum, shown reduced (blue) and oxidized (red). The 405/488 excitation ratio is used for calculation of the biosensor oxidation degree (OxD). (C) S. aureus cells respond differentially to 100 mM H₂O₂, 100 µM HOCl, 50 µM allicin and 100 µM lapachol (Linzner et al. 2020; Loi et al. 2017; Loi et al. 2019). (D) The bshA mutant showed a fully oxidized basal redox state and (E) the ypdA mutant was impaired in recovery of the reduced E_BSH under oxidative stress (Linzner et al. 2019; Loi et al. 2017).

We have used the Brx-roGFP2 biosensor to monitor the changes in the E_BSH during the growth, under oxidative stress and antbiotics treatments in S. aureus isolates and in different mutant backgrounds (Loi and Antelmann 2020; Loi et al. 2017). First, purified Brx-roGFP2 was shown to respond fast and specific to low levels of 10–100 µM BSSB, but not to other LMW disulfides in vitro. The changes in E_BSH were determined in S. aureus COL bshA, brxAB and ypdA mutants along the growth and under oxidative stress to investigate the impact of the Brx/BSH/YpdA/NADPH pathway on the basal E_BSH and in response to oxidative stress (Linzner et al. 2019; Loi et al. 2017). Brx-roGFP2 measurements along the growth revealed a highly reducing basal level E_BSH in the range from −282 to −295 mV in the S. aureus COL WT, which was similar in the ypdA and brxAB mutants (Linzner et al. 2019; Loi et al. 2017). However, in the bshA mutant the biosensor was fully oxidized (Figure 5) (Loi et al. 2017). This indicates an impaired redox balance in the BSH-deficient strain, which might be caused by ROS increase. Interestingly, the ypdA mutant showed a strong Brx-roGFP2 oxidation after exposure to HOCl and H₂O₂, but was unable to regenerate the reduced state of E_BSH during recovery from oxidative stress due to its higher BSSB levels (Figure 5) (Linzner et al. 2019). In addition, the ypdA mutant was impaired in H₂O₂ detoxification as measured using the Tpx-roGFP2 biosensor (Linzner et al. 2019). These results confirmed the function of YpdA to regenerate reduced BSH after oxidative stress.

In addition, Brx-roGFP2 oxidation was analyzed after treatment of S. aureus COL WT with oxidants, different antibiotics and ROS-producing antimicrobials to investigate E_BSH changes (Loi et al., 2017,2018a, 2019). Brx-roGFP2 responds very fast to low levels of 50–100 µM HOCl, resulting in complete oxidation and slow recovery of the reduced E_BSH in S. aureus COL. Due to the high level of H₂O₂ resistance, Brx-roGFP2 was less reactive to 100 mM H₂O₂ and the S. aureus cells were able to recover faster their reduced E_BSH (Loi et al. 2017). Determination of the E_BSH changes of S. aureus COL after infection inside THP-1 macrophages revealed about 87% oxidation after 1 h, supporting that S. aureus experiences the oxidative burst (Loi et al. 2017). However, Brx-roGFP2 did not respond to sub-lethal concentrations of different antibiotics classes, such as rifampicin, fosfomycin, ampicillin, oxacillin, vancomycin, aminoglycosides and fluoroquinolones in S. aureus log phase cells (Loi et al. 2017).

In contrast, the Brx-roGFP2 biosensor was highly oxidized by the redox active antimicrobials AGXX®, allicin and lapachol resulting in an oxidative shift of E_BSH in S. aureus (Loi et al., 2018a, 2019, Linzner et al. 2020). Exposure of S. aureus COL to sub-lethal concentrations of AGXX® leads to strong biosensor oxidation, which is caused by OH· and other ROS generated by AGXX® (Clauss-Lendzian et al. 2018; Loi et al. 2018a). Allicin causes S-thioallylations of BSH, resulting in BSSA formation, which oxidizes Brx-roGFP2 similar as BSSB (Loi et al. 2019). Thus, 50–100 µM allicin lead to fast and complete biosensor oxidation within 10 min in S. aureus, which was reversible with DTT, supporting the strong thiol reactivity of allicin (Figure 5) (Loi et al. 2019). The response of the Brx-roGFP2 biosensor was different for the naphthoquinone lapachol, which acts via the redox cycling mode in S. aureus (Linzner et al. 2020). Upon exposure to 100 µM lapachol, Brx-roGFP2 oxidation was slowly increased within 2 h with no recovery of reduced E_BSH (Linzner et al. 2020). This indicates constant generation of H₂O₂ by lapachol as measured with the Tpx-roGFP2 biosensor. In conclusion, Brx-roGFP2 is a valuable tool to screen for ROS-production and intracellular redox potential changes by redox active antibiotics in S. aureus.

Redox regulation of proteins by protein S-CoAlation

S. aureus produces millimolar CoASH and encodes a CoASH disulfide reductase Cdr, which reduces CoASH disulfide (CoAS₂) back to CoASH in vitro (delCardayre and Davies 1998,1998). Thus, before the discovery of BSH, CoASH was proposed to function as LMW thiol in S. aureus (delCardayre and Davies 1998; Tsuchiya et al. 2018). CoASH is an important thiol cofactor, which activates sugar metabolites of the central carbon catabolism by formation of energy-rich CoA thioesters, such as acetyl-CoA and succinyl-CoA to ensure ATP generation (Jackowski and Rock 1986). Recently, evidence was provided that CoASH can function in redox modification of proteins via S-CoAlations, which may substitute for S-bacillithiolation in the absence of BSH in S. aureus (Gout 2019; Tsuchiya et al. 2018). S-CoAlated proteins were enriched using pulldown with monoclonal anti-CoASH antibodies and identified by mass spectrometry (Gout 2019; Tsuchiya et al. 2018). In S. aureus, about 356 S-CoAlated proteins were identified under diamide stress, including conserved redox sensitive metabolic and antioxidant enzymes (GapDH, AldA, GuaB, Trx, AhpC, Tpx) and transcriptional regulators (SarR, CtsR, AgrA, PerR, SarS), which harbor active site Cys residues and overlap with S-bacillithiolated proteins under HOCl stress (Gout 2019; Tsuchiya et al. 2018).

Furthermore, protein S-CoAlation of the active site Cys151 of GapDH resulted in enzyme inactivation, which was reversible by DTT reduction (Tsuchiya et al. 2018). Computational chemistry proposed that the ADP moiety of CoASH occupies the NAD⁺ binding pocket in the apoenzyme, facilitating S-CoAlation of Cys151 of GapDH (Tsuchiya et al. 2018). However, the physiological role of S-CoAlation in relation to sugar catabolism remains to be elucidated in S. aureus. Based on its metabolic role, enzymes involved in the activation of pyruvate and succinate by CoASH, such as pyruvate dehydrogenase (Pdh) and succinyl-CoA-synthetase (SucCD) could be redox-controlled by S-CoAlation, which may impact the metabolic flux through the TCA cycle. Similarly, the pathways for reversal of S-CoAlations and the physiological role of Cdr under oxidative stress and infections are important subjects of future studies in S. aureus.

PerR as Fur-family metal-based peroxide sensor

The peroxide specific PerR repressor belongs to the Fur family of metalloregulators and was characterized in B. subtilis and S. aureus (Faulkner and Helmann 2011; Pinochet-Barros and Helmann 2018). PerR negatively controls the adaptive response to H₂O₂ and confers H₂O₂ resistance in both bacteria (Faulkner and Helmann 2011; Horsburgh et al. 2001a; Ji et al. 2015; Mongkolsuk and Helmann 2002). The members of the PerR regulon function in ROS detoxification and iron storage, including genes for catalase, peroxiredoxins and thioredoxin reductase (katA, ahpCF, bcp, trxB), the iron storage ferritin and miniferritin (ftnA, dps), the ferric uptake (fur) repressor and the FeS cluster machinery (sufCDSUB) in S. aureus (Figure 6) (Horsburgh et al. 2001a). PerR binds to a conserved inverted repeat sequence ATTATAATTATTATAAT in the promoter region of the PerR regulon genes (Horsburgh et al. 2001a). In S. aureus, PerR is required for virulence in murine skin abscess and Caenorhabditis elegans infection models. In contrast, the catalase KatA is dispensable for pathogenicity (Ji et al. 2015; Horsburgh et al. 2001a).

Figure 6:

Redox regulation of the PerR repressor by iron-catalyzed His oxidation in response to H₂O₂ and by a putative thiol switch under HOCl stress in S. aureus. PerR senses H₂O₂ by Fe²⁺ catalyzed oxidation in the regulatory Asp/His site, leading to HO• formation and His37/His91 oxidation to 2-oxo-histidines, resulting in derepression of the PerR regulon (Ji et al. 2015; Lee and Helmann 2006). HOCl treatment leads most likely to PerR inactivation by a thiol switch in the structural Zn²⁺ site as identified in Bacillus subtilis cells using mass spectrometry in vivo (Chi et al. 2011). PerR controls genes encoding catalase, peroxiredoxins and thioredoxin reductase (katA, ahpCF, bcp, trxB), Fe-storage ferritin and miniferritin (ftnA, dps), the FeS cluster machinery (sufCDSUB) and the ferric uptake (fur) repressor in S. aureus. This figure is adapted from (Hillion and Antelmann 2015).

The structures of B. subtilis and S. aureus PerR proteins contain two metal binding sites each, a structural Zn²⁺ site coordinated by four Cys residues and the regulatory Fe²⁺ or Mn²⁺ site, which includes His and Asp residues (Faulkner and Helmann 2011; Pinochet-Barros and Helmann 2018). The PerR repressors sense H₂O₂ by metal-catalyzed histidine oxidation in the regulatory Fe²⁺ binding sites, leading to their inactivation and derepression of transcription of the PerR regulons in B. subtilis and S. aureus (Ji et al. 2015; Lee and Helmann 2006; Pinochet-Barros and Helmann 2018). While the His and Asp residues in the regulatory site can bind both Fe²⁺ or Mn²⁺ as corepressors, only the Fe²⁺-bound PerR can sense H₂O₂ by iron-catalyzed His oxidation. Specifically, Fe²⁺ in the regulatory site is oxidized by H₂O₂ in a Fenton reaction to generate OH·, leading to oxidation of His37 or His91 to form 2-oxo-histidine in the B. subtilis and S. aureus PerR proteins (Duarte and Latour 2010; Ji et al. 2015; Lee and Helmann 2006). However, the PerR repressor of S. aureus was hypersensitive to low H₂O₂ levels, which might explain its predominant Mn²⁺-dependent repressor activity under aerobic conditions (Ji et al. 2015). This hypersensitivity of PerR towards H₂O₂ could be responsible for the higher basal expression of the PerR regulon genes, resulting in a high level of H₂O₂ resistance in S. aureus.

However, the PerR regulon was strongly induced under oxidative and electrophile stress, such as HOCl, diamide, MHQ, lapachol, allicin and AGXX® in S. aureus (Fritsch et al. 2019; Loi et al., 2018a,b, 2019; Linzner et al. 2020). Thus, PerR might sense various redox active compounds by a thiol switch mechanism in its structural Zn²⁺ site. In support of this hypothesis, a PerR intramolecular disulfide was identified by mass spectrometry in HOCl-treated B. subtilis cells (Chi et al. 2011). Thus, PerR might employ different redox sensing mechanisms to respond to H₂O₂ and other redox active compounds, that induce oxidative and electrophile stress responses.

The MgrA/SarZ/SarA-family of virulence and antibiotic regulators

S. aureus encodes several global MarR-type transcription factors, which control virulence, antibiotics and oxidative stress resistance, including the Multiple gene regulator MgrA, SarZ and the Staphylococcal accessory regulator SarA (Ballal et al. 2009; Chen et al. 2009; Kaito et al. 2006; Poor et al. 2009; Truong-Bolduc et al. 2005, 2008). MgrA and SarZ belong to the MarR/OhrR-family, which sense ROS by thiol switch mechanisms via a conserved single Cys residue (Chen et al. 2011). MgrA is a global virulence regulator, which controls expression of >300 genes, involved in virulence, autolysis, antibiotics resistance, capsule and biofilm formation (Luong et al. 2006). Among the MgrA regulon genes are α−toxin (hla), coagulase (coa), protein A (spa), large surface proteins (ebh, sraP and sasG), fibrinogen-binding protein (fnb), extracellular protease, nuclease (nuc), autolysins (lytM and lytN), multidrug efflux pumps (norA, norB and tetAB), virulence regulators (agr, lytRS, arlRS, sarS and sarV) and capsule biosynthesis (cap5) (Ingavale et al. 2005; Kaatz et al. 2005; Luong et al. 2006; Truong-Bolduc et al. 2005, 2008). In DNase-I footprinting experiments, purified MgrA protein was shown to protect the conserved nucleotide sequence (A/T)GTTGT, which was repeated thrice upstream of the sarV promoter (Crosby et al. 2016; Manna et al. 2004). MgrA is important for virulence of S. aureus as shown in several infection models, including murine abscess, septic arthritis, sepsis, murine bacteremia and endocarditis models (Chen et al. 2006; Jonsson et al. 2008; Li et al. 2019).

The second MarR/OhrR-type regulator SarZ regulates a large regulon involved in virulence, antibiotic resistance, cellular metabolism and the ohr peroxiredoxin as main target gene (Kaito et al. 2006). Purified SarZ was shown to bind rather non-specific to the promoter regions of hla, asp23 and agr, since a specific SarZ recognition motif could not be identified (Kaito et al. 2006). Both MgrA and SarZ sense oxidative stress by thiol switches, which involve the conserved redox sensing Cys13 residues. Thiol oxidation of MgrA and SarZ by organic hydroperoxides and H₂O₂ leads to formation of the Cys13-SOH (Chen et al., 2006,2009; Poor et al. 2009). Further oxidation of SarZ with the synthetic benzene thiol resulted in S-thiolated SarZ, which causes conformational changes in the HTH motif, leading to dissociation of SarZ from the promoter DNA as resolved in the crystal structure (Figure 4) (Poor et al. 2009).

The third redox sensing MarR-type regulator SarA controls genes involved in the oxidative stress defense (sodB, trxB) and virulence, including α- and ß-hemolysins (hla, hlb), toxic shock syndrome toxin 1 (tst), staphylococcal enterotoxin B (entB), protein A (spa) and fibronectin binding protein (fnb) (Chan and Foster 1998). SarA binds to a conserved operator sequence ATTTGTATTTAATATTTATATAATTG located upstream of the −35 promoter region of several SarA-regulated virulence genes (Chien et al. 1999). SarA contains a redox sensing Cys9 in its dimer interface and may use a thiol switch for regulation (Ballal and Manna 2009, 2010).

Furthermore, MgrA, SarZ and SarA were shown to be reversibly redox-controlled by Cys phosphorylation by the serine/threonine kinase (Stk1) and phosphatase (Stp1) (Sun et al. 2012). However, we identified the Cys13 peptides of MgrA and SarZ as HOCl-sensitive with >10% increased thiol oxidation under HOCl stress in S. aureus (Imber et al. 2018a). Furthermore, SarZ and MgrA were S-thioallylated under allicin stress in S. aureus (Loi et al. 2019). Thus, SarZ and MgrA might be redox-controlled by S-bacillithiolation under HOCl stress in S. aureus.

HypR as Rrf2-family redox sensor of HOCl stress

The Rrf2-family HypR repressor was identified as a novel redox sensor of HOCl stress in S. aureus, which uses a thiol switch mechanism (Loi et al. 2018b). Rrf2-family transcription factors are widespread in prokaryotes and regulate diverse functions, such as FeS cluster biogenesis (IscR), cysteine biosynthesis (CymR) and NO· detoxification (NsrR, RsrR) (Karlinsey et al. 2012; Mettert and Kiley 2015; Nakano et al. 2014; Partridge et al. 2009; Remes et al. 2015; Soutourina et al. 2009, 2010). HypR senses strong oxidants, such as HOCl, diamide, AGXX® and allicin stress (Loi et al. 2018a,b, 2019). HypR negatively controls expression of the hypR-merA operon, which was most strongly upregulated under HOCl stress in the transcriptome (Loi et al. 2018b) (Figure 7). In addition, the hypR-merA operon was highly induced in S. aureus during phagocytosis assays with neutrophils (Voyich et al. 2005), suggesting important functions as HOCl defense mechanism. The HypR repressor was shown to bind to a 12-3-12 bp inverted repeat sequence TAATTGTAACTA-N₃-CAGTTACAATTA in the hypR-merA promoter region, which is conserved across staphylococci.

Figure 7:

Redox regulation of the Rrf2-family repressor HypR under HOCl stress in S. aureus. HypR of S. aureus resembles a 2-Cys type redox regulator, which senses HOCl stress by intersubunit disulfide formation between Cys33 and Cys99′, leading to its inactivation and derepression of transcription of the hypR-merA operon (Loi et al. 2018b). The flavin disulfide reductase MerA is involved in HOCl and allicin detoxification and confers resistance to HOCl, allicin and infection conditions.

HypR controls the NADPH-dependent flavin disulfide reductase MerA, which confers resistance towards HOCl, allicin and AGXX® stress as well as under infection conditions inside murine macrophages J774A.1 (Loi et al. 2018a,b, 2019). We hypothesize that MerA could be involved in HOCl reduction to promote survival under infections inside macrophages and neutrophils. In addition, the coupled transcription in the hypR-merA operon suggests that MerA could be the redox partner of HypR to recycle oxidized HypR during recovery from oxidative stress. The detailed functional analyses of MerA in terms of HOCl resistance are important goals of our current research.

The HypR structure contains three Cys residues, including Cys33, Cys99 and Cys142, but only Cys99 is conserved in Rrf2 homologs (Loi et al. 2018b). While Cys33 is essential for redox sensing of HOCl stress, Cys99 is required for DNA binding activity of HypR in S. aureus. Under HOCl stress, HypR is oxidized to intersubunit disulfides between Cys33 and Cys99′ of opposing subunits in the HypR dimer, resulting in inhibition of its repressor activity and derepression of transcription of the hypR-merA operon (Loi et al. 2018b). Thus, the thiol switch model of HypR resembles that of a typical 2-Cys type redox sensing regulator. Future studies are directed to elucidate the structural changes of HypR upon oxidation, leading to its inactivation.

The MarR-family regulators QsrR and MhqR sense quinones

S. aureus and B. subtilis both encode two quinone-sensing MarR-type repressors, including the redox sensing MarR/DUF24-family regulator QsrR (or YodB) and the MarR-type repressor MhqR (Figure 8) (Chi et al. 2010; Fritsch et al. 2019; Ji et al. 2013; Leelakriangsak et al. 2008; Töwe et al. 2007). QsrR controls genes involved in the detoxification of quinones, such as ring-cleavage dioxygenases (catE, catE2), quinone and nitro reductases (azoR1, frp, and yodC) (Fritsch et al. 2019; Ji et al. 2013). A conserved inverted repeat with the consensus GTATAN₅TATAC was identified as the QsrR operator in the promoter region of the QsrR regulon genes (Ji et al. 2013). The MhqR repressor negatively controls transcription of the mhqRED operon, encoding the dioxygenase MhqE and the phospholipase/carboxylesterase MhqD. MhqR was shown to bind to a 9-9 bp inverted repeat sequence TATCTCGAA-aTCGAaATA in the promoter region upstream of mhqR (Fritsch et al. 2019). The quinone and nitroreductases (AzoR1, Frp, YodC) function in quinone reduction to redox stable hydroquinones (Figure 8) (Antelmann et al. 2008). Since benzoquinone alkylates protein and LMW thiols (Liebeke et al. 2008), the dioxygenases CatE, CatE2 and MhqE might catalyze ring cleavage of hydroquinones and quinone-S-adducts (Tam le et al. 2006).

Figure 8:

Redox sensing of quinones and antibiotics by the MarR-type repressors QsrR and MhqR in S. aureus. QsrR senses quinones by thiol-S-alkylation of Cys5, leading to derepression of dioxygenases (catE, catE2) and quinone reductases (frp, azoR1, yodC) (Ji et al. 2013). In addition, QsrR might sense oxidative stress by a thiol switch via intersubunit disulfide formation in S. aureus as revealed for the homologous YodB repressor in B. subtilis (Chi et al. 2010; Lee et al. 2016). Quinone sensing by MhqR involves most likely quinone binding to the conserved ligand pocket, resulting in upregulation of the mhqRED operon, which codes for the phospholipase/carboxylesterase MhqD and the dioxygenase MhqE (Fritsch et al. 2019). Since benzoquinone alkylates protein and LMW thiols in B. subtilis (Liebeke et al. 2008), we hypothesize that dioxygenases catalyze ring cleavage of hydroquinones and quinone-S-adducts. Both QsrR and MhqR confer independently resistance to quinones and the antimicrobials pyocyanin, ciprofloxacin and rifampicin.

In S. aureus and B. subtilis, the MhqR and QsrR regulons are most strongly upregulated by methylhydroquinone (MHQ) in the transcriptomes and contribute independently to quinone resistance (Fritsch et al. 2019; Ji et al. 2013; Leelakriangsak et al. 2008; Töwe et al. 2007). Interestingly, the MhqR and QsrR regulons of S. aureus confer resistance to antimicrobials with quinone elements, such as pyocyanin, rifampicin and ciprofloxacin (Fritsch et al. 2019; Ji et al. 2013). Pyocyanin is a toxic pigment produced by P. aeruginosa. Thus, MhqR and QsrR provide resistance of S. aureus towards pyocyanin to survive respiratory co-infections with P. aeruginosa in cystic fibrosis patients (Noto et al. 2017).

Furthermore, the MhqR regulon conferred tolerance to lethal H₂O₂ stress and was important for long-term survival of S. aureus inside macrophages (Fritsch et al. 2019). Similarly, the QsrR regulon mediates resistance to killing by macrophages in infection assays, indicating a crucial role of quinone detoxification pathways for virulence and survival of S. aureus inside host cells (Ji et al. 2013). During infections, the QsrR and MhqR regulons could be involved in detoxification of host-derived catecholamines, which are produced in macrophages and neutrophils to enhance the inflammatory response (Flierl et al. 2007). Infection-relevant electrophilic quinones could further arise in neutrophils from MPO-catalyzed serotonin oxidation to tryptamine-4,5-dione, which might be detoxified by the QsrR and MhqR regulons to enhance survival of S. aureus (Ximenes et al. 2009). Phenotype analyses further revealed an increased respiratory chain activity and higher ATP levels in the S. aureus mhqR mutant, pointing to the physiological role of the MhqR regulon to maintain the respiratory menaquinones in their reduced state (Fritsch et al. 2019). In support of this notion, the MhqR regulon of B. subtilis was important for the formation of cell-wall free L-forms, which requires reduced quinones for continued respiration and contributes to survival and resistance (Kawai et al. 2015).

The MhqR repressors of B. subtilis and S. aureus do not use thiol-based mechanisms for redox sensing of quinones (Fritsch et al. 2019; Töwe et al. 2007). Since many MarR-type regulators harbor conserved ligand binding pockets, MhqR might sense quinones by direct binding to a ligand pocket, leading to derepression of the mhqRED operon in S. aureus (Grove 2017; Perera and Grove 2010; Wilkinson and Grove 2006).

In contrast to MhqR, QsrR contains a conserved redox sensing Cys5 and two non-conserved Cys30 and Cys33 residues (Ji et al. 2013). The redox sensing mechanism of QsrR was shown to involve thiol-S-alkylation of the Cys5 residue, resulting in QsrR inactivation (Figure 8) (Ji et al. 2013). The detailed mechanism and structural changes of QsrR upon S-alkylation have been resolved for reduced and menadione-bound QsrR. QsrR alkylation leads to movement and rotation of the α4 and α4′ DNA binding helices from 106 to 117°, which is incompatible with DNA binding and causes dissociation of QsrR from the promoter (Ji et al. 2013). However, this Cys-alkylation model has been resolved with the QsrR Cys30,33 double mutant protein. The B. subtilis YodB repressor was previously shown to sense diamide and quinones via intersubunit disulfide bond formation between the conserved Cys6 and the C-terminal Cys101′ or Cys108′ of opposing subunits in the YodB dimer in vivo (Chi et al. 2010).

Recent crystal structure analyses have resolved two mechanisms for YodB inactivation, the S-alkylation of Cys6 by methyl-p-benzoquinone and the thiol switch between Cys6 and Cys101′ under diamide stress (Lee et al. 2016). Quinones cause S-alkylation at Cys6 with minor structural changes by 3 Å rotations of the α4 and α4′ domains, similar as shown for QsrR (Lee et al. 2016). In contrast, diamide leads to Cys6-Cys101′ intersubunit disulfide formation with large structural rearrangement, leading to complete dissociation from the DNA (Lee et al. 2016). Transcriptome analyses revealed that the QsrR regulon is strongly upregulated by quinones and other redox active compounds, such as MHQ, HOCl, AGXX®, allicin and lapachol in S. aureus (Fritsch et al. 2019; Linzner et al. 2020; Loi et al. 2018a,b, 2019). Thus, the two Cys QsrR repressor most likely senses quinones and strong oxidants by S-alkylation and disulfide formation, respectively, which remains to be elucidated.

It is possible that QsrR/YodB and MhqR sense different quinone-related redox signals, controlling the detoxification/reduction of oxidized quinones and ROS by QsrR/YodB as overarching thiol-based regulator, while MhqR could be additionally involved in trapping and regulation of ring cleavage of reduced and oxidized quinones. QsrR/YodB uses a thiol switch mechanism to respond to ROS, which oxidize respiratory menaquinones and require quinone detoxification/reduction pathways to restore respiration. MhqR could rather sense accumulating quinones to induce additional quinone detoxification pathways when the QsrR/YodB regulon is overwhelmed.The crosstalk of both systems was confirmed by lower expression of QsrR/YodB regulons in the mhqR mutants and vice versa in B. subtilis and S. aureus (Fritsch et al. 2019; Leelakriangsak et al. 2008; Töwe et al. 2007). The investigation of the kinetics, specificities, crosstalks and reversibilities of both systems are interesting goals of future mechanistic studies.

The thiol-based redox sensor Spx and its YjbH adapter

The thiol-based redox sensor Spx belongs to the arsenate reductase (ArsC) family, representing an unusual transcriptional activator lacking the HTH motif (Nakano et al. 2003, 2005; Zuber 2009). In S. aureus, transcription of spx is positively regulated by the ArlRS two-component system (Crosby et al. 2020). Spx is activated in response to oxidative stress by thiol oxidation of its CXXC redox switch motif to an intramolecular disulfide (Nakano et al. 2005). The Spx disulfide interacts with the α-C-terminal domain of the RNA polymerase and activates transcription of the Spx regulon, which is involved in the thiol redox homeostasis and oxidative stress resistance, including thioredoxin and thioredoxin reductase (trxA, trxB) (Figure 9) (Nakano et al. 2003, 2005; Zuber 2004, 2009). Thus, the spx mutant shows increased susceptibility to oxidative, heat and salt stress in S. aureus (Pamp et al. 2006; Wang et al. 2010). Furthermore, Spx is involved in pia-dependent biofilm formation and controls expression of the icaABCD operon. Spx represses cspA transcription, which inhibits the activation of SigB leading indirectly to downregulation of aureolysin (aur) and staphyloxanthin biosynthesis (crtOPQMN operon) (Austin et al. 2019; Donegan et al. 2019).

Figure 9:

Redox control of the oxidative stress regulator Spx in S. aureus. Under control conditions, Spx is complexed by the YjbH adapter and targeted to the ClpXP machinery for proteolytic degradation. Oxidative stress causes possibly thiol oxidation and self-aggregation of YjbH, resulting in Spx stabilization (Engman and von Wachenfeldt 2015). Spx is activated by thiol oxidation to an intramolecular disulfide. Oxidized Spx binds to the αCTD of RNAP leading to activation of transcription of trxA, trxB, trfA, cspA and icaABCD in S. aureus, which impacts redox homeostasis, virulence factor expression, antibiotics resistance and biofilm formation. This figure is adapted from (Hillion and Antelmann 2015).

The post-translational control involves the proteolytic degradation of Spx by the ClpXP system with the help of the YjbH adapter. Under control conditions, YjbH binds Spx, which is targeted to the ClpXP machinery resulting in Spx degradation (Engman et al. 2012; Garg et al. 2009; Pamp et al. 2006). Oxidative stress causes YjbH self-aggregation possibly due to thiol oxidation of its Zn²⁺ redox switch motif leading to Spx stabilization and increased Spx protein levels (Figure 9) (Engman and von Wachenfeldt 2015). Apart from proteolysis, YjbH controls antibiotic resistance in S. aureus. The yjbH mutant was resistant to ß-lactam antibiotics due to increased PBP4 expression and peptidoglycan cross-linking (Gohring et al. 2011). Spx controls the mecA homolog trfA in S. aureus. While the trfA mutant was susceptible to oxacillin and glycopeptide antibiotics, trfA upregulation in the yjbH mutant conferred antibiotics resistance (Jousselin et al. 2013). In addition, Spx activates the MazEF toxin-antitoxin system, which promotes dormancy and antibiotic tolerance (Panasenko et al. 2020). The lack of the extracellular proteases, such as aureolysin in the yjbH mutant led to hypervirulence in a systemic mouse infection model and enhanced colonization of the kidney and the spleen in a murine sepsis model (Austin et al. 2019; Kolar et al. 2013).

The TetR-family regulator and electrophile sensor GbaA

The TetR-family regulator GbaA (Glucose-induced biofilm accessory protein A) was characterized as a novel thiol switch and monothiol electrophile sensor (Ray et al. 2020). GbaA controls glucose-induced biofilm formation by the polysaccharide intracellular adhesion (PIA), which is composed of poly-N-acetyl glucosamine (PNAG) as biofilm matrix in S. aureus (Ray et al. 2020; You et al. 2014; Yu et al. 2017). GbaA negatively controls two operons of unknown functions, encoding a putative glyoxalase, the NAD⁺ dependent epimerase/dehydratase NmrA, the short chain dehydrogenase/oxidoreductase GbaB, an amidohydrolase and an α, ß-fold hydrolase (Figure 10) (Ray et al. 2020; You et al. 2014; Yu et al. 2017). The GbaA repressor was shown to bind to a palindromic sequence AAACGGAGAGTTATCCGTTT in the upstream promoter region of gbaA. The gbaA mutant showed strongly enhanced biofilm formation in a super-biofilm elaborating clinical isolate S. aureus TF2758 (Yu et al. 2017), and biofilm formation was dependent on GbaB (Ray et al. 2020). GbaB was suggested to catalyze the oxidation of an alcohol to a sugar aldehyde in the PNAG biosynthesis pathway, but the physiological role and the inducer of the GbaA regulon are unknown (You et al. 2014; Yu et al. 2017).

Figure 10:

Redox sensing of electrophiles by the TetR-family biofilm regulator GbaA. The GbaA repressor is oxidized by different oxidants to the Cys55-Cys104 intramolecular disulfide, which does not affect the structure and DNA binding activity of GbaA. Alkylation of GbaA by electrophiles causes structural changes and loss of DNA binding, resulting in derepression of transcription of the GbaA regulon (Ray et al. 2020). GbaA controls two operons involved in biofilm formation, including a glyoxalase, the NAD⁺-dependent epimerase/dehydratase NmrA, DUF2316, the short chain dehydrogenase/oxidoreductase GbaB, the GbaA repressor, an amidohydrolase and an α, ß-fold hydrolase.

Interestingly, the GbaA regulon was induced by different redox active compounds and antibiotics, including RNS, AGXX®, allicin, erythromycin, colistin, vancomycin and most strongly by the electrophiles methylglyoxal and N-ethylmaleimide (NEM) (Loi et al. 2018a, 2019; Mäder et al. 2016; Ray et al. 2020; Schlag et al. 2007Loi et al. 2018, 2019; Mäder et al. 2016; Ray et al. 2020; Schlag et al. 2007). GbaA was shown to function as a thiol switch and an electrophile sensor in vitro, which involves the conserved Cys55 and Cys104 residues in the DNA binding and regulatory domains (Ray et al. 2020). GbaA is oxidized to an intramolecular disulfide in response to diamide, BSSB, GSSG and GSNO in vitro, which does not affect its DNA binding activity (Figure 10). However, treatment of the single Cys55A or Cys104A mutant proteins with NEM and BSSB resulted in NEM-alkylation and S-bacillithiolation of the remaining Cys leading to inactivation of the GbaA repressor (Figure 10) (Ray et al. 2020). In addition, Cys104 was found more reactive and accessible towards electrophiles, indicating that Cys104 is the redox sensing Cys and most likely modified by electrophiles or BSSB in vivo (Ray et al. 2020). Thus, the redox sensing mechanism of GbaA differs from other electrophile sensors, such as YodB, which senses diamide and quinones via thiol switch and S-alkylation mechanisms, both leading to structural changes and inhibition of its DNA binding activity (Chi et al. 2010; Lee et al. 2016; Leelakriangsak et al. 2008). While GbaA resembles a 2-Cys type regulator, only S-thiolation or S-alkylation causes structural changes (Ray et al. 2020). However, the physiological electrophile and functions of the GbaA regulon members under oxidative and electrophile stress and biofilm formation are unknown.

Since the GbaA regulon is required for glucose-induced biofilm formation, host-derived oxidation products of glucose, such as the α,β-dicarbonyls glucosone, glyoxal, methylglyoxal, glycolaldehydes, or other reactive glycolytic intermediates could be physiological electrophiles sensed by GbaA. These glucose oxidation products accumulate in human tissues, blood and activated neutrophils causing glycation of lysine, arginine and cysteine in proteins (Vetter 2015). In neutrophils, the MPO reaction product HOCl causes formation of chloramines, which are degraded to acrolein and p-hydroxyphenylacetaldehyde as further infection-related electrophiles (Hazen et al. 1998). Reactive aldehydes might increase in S. aureus during the switch to biofilm formation and anaerobic fermentation, which relies on high glycolytic activity for ATP production. In the next section, we discuss the SrrAB system, which senses oxygen levels and controls the switch to anaerobiose and biofilm formation via the SrrB redox switch (Tiwari et al. 2020; Yarwood and Schlievert 2000). Thus, GbaA and SrrAB might be responsible to sense different signals during biofilm formation and the switch from aerobic to anaerobic conditions to allow biofilm matrix synthesis, removal of toxic aldehydes and metabolic adaptation.

Another infection-relevant electrophile could be the major immunoregulatory metabolite itaconate, which inhibits the succinate dehydrogenase resulting in metabolic rewiring during inflammatory macrophage activation (Lampropoulou et al. 2016). Itaconate reacts as electrophile with GSH and induces electrophile stress in host cells (Bambouskova et al. 2018). In P. aeruginosa lung infections, host-derived itaconate reprograms bacterial metabolism and induces biofilm formation via enhanced exopolysaccharide matrix biosynthesis (Riquelme et al. 2020). Thus, itaconate could be a candidate electrophile to inactivate the GbaA repressor to promote PNAG synthesis and biofilm formation in S. aureus, which remains to be investigated.

Furthermore, GbaA belongs to the TetR-family regulators, which are often inactivated by direct binding of antibiotics as ligands to the C-terminal regulatory domain (Cuthbertson and Nodwell 2013). Since GbaA responds to various antibiotics and allicin (Loi et al. 2019; Mäder et al. 2016), a direct binding of antibiotics and electrophiles coupled with ROS-induced S-thiolation or S-alkylation of Cys104 could possibly inactivate GbaA in vivo. In summary, S. aureus can be exposed to a plethora of oxidants and electrophiles present during host-pathogen interactions that might be sensed by GbaA and other redox sensors, which enable host adaptation. Thus, future studies should shed light on the signals and redox chemistry of host-pathogen interactions.

The SrrAB two-component system as redox sensor of oxygen levels

The Staphylococcal respiratory regulator (SrrAB) two-component system consists of the sensor histidine kinase SrrB and the DNA-binding response regulator SrrA (Yarwood and Schlievert 2000). The SrrAB regulon is implicated in long-term biofilm stability, anaerobic growth, oxidative and nitrosative stress resistance and contributes to the virulence and protection against neutrophil killing in S. aureus (Kinkel et al. 2013; Mashruwala and Boyd 2017; Pragman et al. 2004; Ulrich et al. 2007). SrrAB regulon members include genes for the biogenesis of cytochromes and terminal oxidases (ctaB, cydAB, qoxABCD), heme biosynthesis (hemACDX), anaerobic fermentation (pflAB, adhE, nrdDG), and RNS resistance (hmp, scdA) (Figure 11) (Kinkel et al. 2013; Mashruwala and Boyd 2017; Oogai et al. 2016; Yarwood and Schlievert 2000). In addition, the SrrA response regulator binds to the agr, spa, srrAB, icaA, RNAIII and tst promoters and is involved in virulence regulation, directly and indirectly via Agr (Pragman et al. 2004; Ulrich et al. 2007). However, no consensus sequence of a potential SrrA recognition motif could be identified in the promoter regions of the target genes. Furthermore, SrrAB positively influences expression of dps, scdA, ahpCF and katA, which function in iron storage, FeS-cluster repair and H₂O₂ detoxification leading to an oxidative stress resistance (Mashruwala and Boyd 2017). The crosstalk between Nos and SrrAB mediates the switch from aerobic to anaerobic energy metabolism, since SrrAB co-regulates narG and nirB, which are involved in nitrate respiration and nitrite transport (James et al. 2019).

Figure 11:

Sensing of oxygen by the SrrAB two-component system. SrrAB senses oxygen availability via the redox state of respiratory menaquinones and by a thiol switch in its sensor kinase SrrB (Tiwari et al. 2020). Aerobic conditions lead to menaquinone oxidation and intramolecular disulfide formation between Cys464 and Cys501 in SrrB, resulting in SrrB inactivation. During anaerobiosis, the two Cys residues in SrrB are reduced directly by reduced menaquinones or indirectly by a ligand binding to the PAS domain of SrrB. Reduced SrrB is active as kinase, leading to autophosphorylation and phosphorylation of the response regulator SrrA. SrrA-P positively controls biofilm formation via the icaABCD operon and represses tst gene transcription. Anaerobic respiration and fermentation are activated by SrrAB, while virulence factors are repressed as indicated.

S. aureus utilizes SrrAB for adaptation to lower oxygen availability during anaerobiosis and changes in the respiratory flux (Yarwood and Schlievert 2000). SrrB senses the reduced state of the respiratory menaquinone pool and acts as kinase and phosphatase via the PAS domain to regulate the phosphorylation state of the response regulator SrrA (Kinkel et al. 2013; Mashruwala et al. 2017; Tiwari et al. 2020). Two cysteines (Cys464 and Cys501) are present in the ATP-binding catalytic domain of SrrB, which are oxidized to an intramolecular disulfide under aerobic conditions inhibiting the autokinase activity of SrrB (Figure 11) (Tiwari et al. 2020). Under anaerobic conditions, the reduced menaquinone pool leads to reduction of the disulfide in SrrB, directly or indirectly via a PAS domain/ligand complex, leading to a 40% increase in autophosphorylation of SrrB, which activates SrrA, affecting biofilm formation and tst toxin expression (Tiwari et al. 2020). SrrAB senses indirectly increasing ROS levels during aerobic growth and induces the oxidative stress response. Thus, SrrAB contributes to the protection of FeS cluster enzymes, such as aconitase and facilitates the growth under aerobic conditions in S. aureus (Mashruwala and Boyd 2017). In support of this notion, the SrrB disulfide under aerobic conditions was important for pathogenesis of S. aureus in an endocarditis model of infection (Tiwari et al. 2020).