Scs system links copper and redox homeostasis in bacterial pathogens

Abstract

The bacterial envelope is an essential compartment involved in metabolism and metabolites transport, virulence, and stress defense. Its roles become more evident when homeostasis is challenged during host–pathogen interactions. In particular, the presence of free radical groups and excess copper in the periplasm causes noxious reactions, such as sulfhydryl group oxidation leading to enzymatic inactivation and protein denaturation. In response to this, canonical and accessory oxidoreductase systems are induced, performing quality control of thiol groups, and therefore contributing to restoring homeostasis and preserving survival under these conditions. Here, we examine recent advances in the characterization of the Dsb-like, Salmonella-specific Scs system. This system includes the ScsC/ScsB pair of Cu⁺-binding proteins with thiol-oxidoreductase activity, an alternative ScsB-partner, the membrane-linked ScsD, and a likely associated protein, ScsA, with a role in peroxide resistance. We discuss the acquisition of the scsABCD locus and its integration into a global regulatory pathway directing envelope response to Cu stress during the evolution of pathogens that also harbor the canonical Dsb systems. The evidence suggests that the canonical Dsb systems cannot satisfy the extra demands that the host-pathogen interface imposes to preserve functional thiol groups. This resulted in the acquisition of the Scs system by Salmonella. We propose that the ScsABCD complex evolved to connect Cu and redox stress responses in this pathogen as well as in other bacterial pathogens.

The cell envelope of Gram-negative bacteria is where the uptake of nutrients and efflux of unnecessary or toxic species occur. It is also the first barrier against multiple external stressors. Several physiological processes such as energy generation and cell division depend on its integrity (1). The bacterial envelope is composed of two proteolipid bilayers: the inner membrane (IM) enclosing the cytoplasm and the outer membrane (OM) with lipopolysaccharide exposed to the external milieu. Between these two membranes is the periplasm, a protein-rich compartment that harbors a peptidoglycan layer (2). Many bacterial virulence factors reside in this compartment or depend on envelope-located systems for their secretion (3, 4). Thus, preserving its homeostasis is critical for virulence.

Pathogens are exposed to host-generated reactive oxygen and/or nitrogen species (ROS and RNS, respectively), key components of the innate immune system (5, 6, 7) (Fig. 1). ROS and RNS are also produced as common inflammation by-products in mammalian tissues (8). Two enzymes, the membrane-associated NADPH-dependent oxidase (Phox) and the inducible nitric oxide synthase (iNOS), catalyze the conversion of O₂ into superoxide (O₂^-) and the oxidation of L-arginine to L-citrulline and nitric oxide (NO), respectively, in the host (9). The reactive O₂^- is rapidly converted into hydrogen peroxide (H₂O₂) and hydroxyl radicals (HO·) within the bacterial cell. At the same time, NO reacts with O₂^- and H₂O₂ to produce a variety of RNS such as peroxynitrite (ONOO⁻), perhaps the most potent non-specific biocidal agent known (10, 11). These species react with bacterial lipids and proteins, affecting the synthesis and translocation of major envelope structural components as peptidoglycans, lipopolysaccharides, and lipoproteins (12, 13, 14). This causes a rapid metabolic breakdown that eventually results in cell death.

Figure 1.

Handling of copper and redox stress in Salmonella envelope. Cu ions, ROS, and RNS are part of the host’s defense against this pathogen. Cu-transporters and chaperones are depicted in different tones of blue, cuproproteins in orange, and redox enzymes in green. The regulatory systems (in red) and their regulatory networks are shown. The flux, generation, and chemical transformation of ROS/RNS are indicated with arrows. Arrows with question marks represent the influx and efflux of Cu ions carried out by uncharacterized transporters. Inactivation of proteins with Cys residues involves not only Cu(I) but also ROS and RNS.

Periplasmic enzymes such as Cu-Zn superoxide dismutases (SodC), thiol oxidoreductases of the Dsb family that repair cysteine (Cys) residues, and proteases that degrade damaged proteins, are the first mechanisms used by Gram-negative bacteria to counteract oxidative stress (12, 15, 16, 17, 18, 19). Furthermore, some pathogens like Salmonella enterica and Proteus mirabilis are adapted to survive in oxidative environments, and this was accomplished by acquiring additional gene products that strengthen their response capabilities (see below). The CpxR/CpxA two-component system, together with other transcriptional regulators, coordinates cell envelope redox stress response (4, 20, 21). Instead of sensing ROS/RNS directly, these regulatory systems detect the damage caused by the reactive species. For instance, activation of the CpxR/A pathway depends on a Cys-rich sensor whose folding and/or insertion into the membrane is altered by redox stress (22).

Copper ions are required as a prosthetic group in various essential enzymes, but can also be potentially cytotoxic. For simplicity, we will refer to copper mainly as Cu⁺, as it is the oxidation state being bound, transported, and distributed across the cell. Mechanisms of Cu⁺ toxicity include: (i) participation in Fenton-like reactions with ROS, (ii) displacement of redox-active Fe ions from Fe-S centers, (iii) binding to non-cognate sites on macromolecules, (iv) oxidation of sulfhydryl groups, particularly in the periplasm of Gram-negative bacteria (v) and protein aggregation (23, 24, 25, 26, 27, 28) (Fig. 1). Periplasmic Cu⁺ overload also affects key processes such as folding and transit of lipoproteins to the OM or amino acid uptake (29, 30). In fact, Cu⁺ is a component of the innate immune system against bacterial pathogens (31, 32, 33). Phagocytic cells actively accumulate Cu⁺ in endocytic vesicles used to intoxicate intracellular pathogens (34). Cu⁺ is also mobilized into the bladder to restrict colonization by uropathogens such as Escherichia coli or P. mirabilis (35, 36). Furthermore, exposure of uropathogenic E. coli to sub-lethal Cu⁺ concentrations led to increased expression of the periplasmic SodC and cytoplasmic SodA enzymes, likely counteracting the metal-generated oxidative stress (37).

Although the connection of Cu⁺ dyshomeostasis with redox stress is well known (25, 33, 38, 39), it is still unclear how the cell envelope handles and distributes this redox active metal while maintaining the required redox balance (40, 41). Defense against Cu⁺ excess in Gram-negative species relies on the Cu⁺ sensing cytoplasmic CueR regulator and the periplasmic-sensing two-component systems (CusR/S, CopR/S, CpxR/A) that control the expression of membrane Cu⁺ transporters, metal chaperones, oxidases, and species-specific Dsb-like proteins such as the ScsABCD system (32, 39, 40). In particular, Salmonella ScsABCD is required for resistance to both Cu⁺ and H₂O₂ stress, two common stressors this pathogen handles during the infection cycle (38, 42). Expression of this non-canonical Dsb system is induced by periplasmic Cu overload via activation of CpxR/A (38). Recent advances in the biochemical characterization of the ScsC/ScsB pair of thiol-oxidoreductases and identification of likely periplasmic substrates, together with a better definition of the role of associated proteins, ScsA and ScsD, support a revision of the evidence informing future studies of these proteins. In addition, we also discuss here the acquisition of this genetic locus by other Gram-negative pathogens and its putative integration into the CpxR/A regulon. Finally, we consider the relevance of the Scs system during infection.

Periplasmic redox and Cu-stress handling by Salmonella

Salmonella encounters ROS/RNS stress during infection. Its intestinal colonization triggers inflammation with a concomitant generation of these reactive species (5). The pathogen also faces the simultaneous presence of ROS, RNS, and Cu⁺ excess (Fig. 1) in the lumen of the endocytic vesicles within phagocytic cells (5, 43). Salmonella employs two periplasmic Cu-Zn-Sod enzymes (Fig. 1) to remove toxic superoxide ions from the periplasm: SodCII, an E. coli SodC ortholog, and the horizontally acquired SodCI, encoded in the Gifsy-2 prophage along with other virulence-factors (16, 44). These homologous enzymes have different structural/functional properties and their expression depends on different transcription factors. The dimeric SodCI is tethered to the peptidoglycan and shows a catalytic rate higher than that of the soluble monomeric SodCII (45). Deletion of sodCI increases Salmonella sensitivity to both superoxide and NO (46) suggesting that SodCI prevents the formation of peroxynitrite through degradation of superoxide radicals. Supporting its role in pathogenesis, sodCI transcription is upregulated by the PhoP/PhoQ two-component system, a major coordinator of Salmonella virulence (47). Thus, its deletion significantly attenuates macrophage survival and mice infection (48, 49).

The bacterial response to Cu⁺ is also critical for Salmonella virulence (40, 50). It depends on the cytoplasmic Cu⁺-sensing transcriptional regulator CueR and a set of envelope proteins expressed under its control (Fig. 1): the Cu⁺ transporting P-type ATPase CopA, the multicopper oxidase CueO, and the Cu⁺-binding protein CueP (50, 51, 52, 53, 54). CueP expression also depends on the CpxR/A system (55) and this periplasmic protein is required for Salmonella survival and replication inside macrophages (50, 56). Some Salmonella serovars harbor a second Cu⁺ transporting P-type Cu⁺-ATPase, GolT, and a cytoplasmic Cu⁺-chaperone GolB, both controlled by a CueR-like regulator, GolS (Fig. 1) (52, 57). In vitro, CueP can acquire Cu⁺ from either CopA or GolT, and participates in the metallation of SodCI and SodCII (54, 58). However, a ΔcueP ΔcopA ΔgolT triple mutant retains fully functional SodCI and SodCII (59). On the other hand, CueP can reduce Cu²⁺ to metallic Cu⁰ (56).

Salmonella harbors the canonical DsbA/DsbB and DsbC-G-E/DsbD oxidoreductase systems that maintain thiol groups in the oxidative periplasmic environment (Fig. 1). The structure and biochemical function of these major complexes are discussed below. Salmonella also has many horizontally acquired genes coding for Dsb-like proteins (17). Among them, DsbL/DsbI is a DsbA/DsbB-like system required for activation of AssT, a periplasmic aryl-sulfate sulfotransferase encoded in the same chromosomal locus than dsbL and dsbI (60). SrgA is an alternative DsbA-like oxidase identified in the Salmonella virulence plasmid (61). This enzyme interacts with DsbB and assists in the assembly of the plasmid-encoded Pef-fimbriae and the secretion of other virulence factors (61, 62). Accordingly, srgA transcription is coregulated with the fimbria-coding genes (61). The Dsb-like disulfide isomerase and thiol oxidase BcfH also contribute to fimbrial biogenesis (63).

Salmonella scsABCD is a chromosomal operon coding for three non-canonical Dsb-like proteins, ScsB, ScsC and ScsD, and ScsA, a putative membrane-protein ScsA of unknown function (38, 42, 64) (Fig. 1). scsABCD transcription is induced by Cu stress and depends on CpxR/A (38). The scs locus was initially identified by its ability to suppress the Cu-sensitivity of an E. coli mutant lacking nlpE, the Cpx OM envelope stress sensor coding gene (64). The inclusion of both scsABCD and cueP as part of the Salmonella Cpx regulon (38, 55) might account for the increased sensitivity of the cpxRA mutant to both Cu⁺ and Au⁺ (65) (Checa and Soncini, unpublished results). Under Cu-stress, the Scs proteins could provide feedback control to the CpxR/A system, as deletion of the scs locus results in an increased Cu-dependent activation of the CpxR/A pathway (38). This resembles a phenotype observed in a dsbA deletion strain in E. coli, also a member of the Cpx regulon in this species (66).

Although the entire scsABCD operon is induced by Cu overload, only ScsB, ScsC, and ScsD are required for Cu⁺ resistance, while they play minor roles in H₂O₂ tolerance (38). On the other hand, ScsA is required for H₂O₂ detoxification but does not provide any protection against Cu (38, 42). The presence of Scs proteins in the Salmonella periplasm made the DsbC/DsbD system dispensable for Cu resistance (38), differing from E. coli, which lacks the Scs system. Since Cu and ROS are simultaneously present in the host intracellular environment (23, 67), the Scs proteins are thought to be important under these conditions, contributing to Salmonella virulence. The Scs system might also compensate for the low dsbA transcription detected at the intracellular niche (68, 69). Although the precise contribution of the Scs system to virulence has not been established and could be complex, wild-type replication levels inside macrophages were reported for Δscs strains (42), except when the cellular-culture medium was supplemented with 100 μM CuSO₄ (30). In the presence of Cu²⁺, the Δscs mutant also showed a significant reduction in the expression and secretion of virulence factors required for cell invasion (42). Finally, impaired proliferation of a ΔscsA mutant strain was reported during host’s cortisol outburst, a phenomenon that promotes intracellular bacterial replication in the intestinal tract (70).

The canonical Dsb system

Cysteine is a redox-active amino acid that frequently appears in pairs forming S-S bonds in periplasmic proteins (71). This stabilizes protein structure and prevents oxidation/nitrosylation of their side chains (72). Still, Cys in its reduced form is required in periplasmic proteins either for catalytic activity or to coordinate metal ions such as Cu⁺ (24, 72). In this oxidative environment, the Cys redox status is usually maintained by two major thiol oxidoreductase complexes, DsbA/DsbB and DsbC/DsbD (17) (Figs. 2 and 3). DsbA and DsbC are the substrate-interacting periplasmic enzymes. DsbB and DsbD are their IM-partners that drive electrons through the membrane. These Dsb proteins harbor domains with a thioredoxin-like fold and distinctive CX_2/3C motifs that are required for their redox activities.

Figure 2.

De novo disulfide bond formation catalyzed by the DsbA/DsbB pair. Three-dimensional structures of DsbA (PDB 1A2I) and DsbB (PDB 2ZUQ) are drawn in ribbon form (in blue). All catalytic Cys residues from CX_2/3C motifs are represented in yellow spheres. Thiol groups or S-S bonds are depicted in magenta. UQ/UQH₂ is the ubiquinone/ubiquinol pool. Magenta dashed arrows indicate the electron flow. Solid black arrows show the overall redox reactions.

Figure 3.

Reactions catalyzed by DsbC. A, isomerization of non-native disulfide bonds in envelope proteins. B, reduction of the Cys groups on protein substrates, a reaction that requires DsbD to recycle DsbC. The structures of DsbC (PDB 1EEJ), DsbDα (PDB 1JPE), DsbDγ (PDB 2FWF), and TrxA (PDB 6H1Y) are drawn in ribbon form (in blue). The DsbDβ structure predicted by AlphaFold 2 is shown in gray. Catalytic Cys, thiol groups, or S-S bonds are shown in Figure 2. Dashed arrows indicate either nucleophilic attack or electron flow, as shown in the figure. Solid black arrows illustrate the redox reactions.

Under physiological conditions, the DsbA/DsbB system assists in de novo disulfide-bond formation when newly synthesized proteins reach the periplasm (Fig. 2). The periplasmic DsbA functions as a monomeric oxidase that oxidizes the thiol groups of its substrates to form disulfide bonds (73, 74). As a result, the catalytic CX₂C in DsbA becomes reduced. The IM component DsbB re-oxidizes the DsbA active-site residues and transfers the electrons to quinone carriers that feed the respiratory chain (75, 76) (Fig. 2). DsbB has four transmembrane segments and two periplasmic regions, each containing a pair of Cys residues (77, 78). In the N-terminal periplasmic domain of the E. coli protein, Cys41 and Cys44 are part of the thioredoxin-like domain that interacts with the reduced quinone carrier. Cys104 and Cys130 in the C-terminal periplasmic region perform DsbA oxidation (Fig. 2). A transient Cys41/Cys130 disulfide bond interaction enables the electron flow between both terminal regions of DsbB. E. coli DsbA/DsbB substrates include proteins requiring the formation of S-S bonds between non-consecutive Cys; for instance, LptD and BamA required for lipopolysaccharide or β-barrel protein insertion into the OM, or FtsN, an IM divisome component (79, 80, 81, 82). Other DsbA/DsbB substrates participate in the assembly and/or delivery of virulence factors such as flagella, adhesins, and secretion systems, or in stress signaling and activation of the major regulatory envelope homeostasis pathways (83, 84). A mechanistic example of how this Dsb system is linked to redox homeostasis is NlpE, the OM sensor of the CpxR/A system. NlpE is a large lipoprotein with five Cys and two periplasmic domains, each harboring an S-S bond in its final structure (85). Perturbation of S-S bond formation blocks NlpE trafficking to the OM. The IM-retained misfolded NlpE interacts with CpxA, inducing its autokinase activity, phosphorylation of CpxR, and expression of the Cpx regulon (29, 66). The thiol oxidase DsbA probably provides feedback control to the CpxR/A system, because its coding gene is part of the Cpx regulon in E. coli, but also in Salmonella and other bacteria (86, 87, 88). Similarly, activation by redox stress of the Rcs system, a phospho-relay also contributing to envelope homeostasis, involves two proteins exposing Cys to the periplasm, the OM lipoprotein RcsF and the IM system repressor IgaA (89, 90, 91, 92, 93). RcsF usually remains in the OM trapped in a complex with BAM, the protein incorporation β-barrel assembly machinery, and porins. This prevents RcsF interaction with IgaA. Perturbation in S-S bond formation releases RcsF from the above-mentioned complex and allow it to bind IgaA. This finally results in the activation of the Rcs-regulatory pathway.

DsbC can act both as an isomerase and as a reductase (82, 94, 95). The first activity is essential to repair dysfunctional S-S bonds on periplasmic proteins having an odd number of Cys residues (Fig. 3A). Also, DsbC maintains free-SH groups preventing their oxidation/nitrosylation and catalyzes the reduction of dysfunctional S-S bonds (Fig. 3B). In fact, an E. coli dsbC null mutant is highly sensitive to Cu⁺ (96). DsbC is a V-shaped homodimer, where each protomer contains a thioredoxin domain with a CX₂C active site (97, 98). Catalysis involves the nucleophilic attack of the N-terminal Cys98 of DsbC and the formation of a mixed-disulfide complex with the substrate (97). Resolution of this intermediate and product release involves the attack of either an alternative Cys in the misfolded substrate (Fig. 3A), or the Cys101 of DsbC (Fig. 3B) (97, 99). A DsbC CX₂S mutant cannot resolve the transient enzyme/substrate disulfide complex. Consequently, this mutant has been employed to trap and identify some of the DsbC target proteins (81, 100). Recycling of oxidized DsbC to its reduced form is mediated by DsbD (Fig. 3B), via its periplasmic CX₃C-containing α-domain (101). This large IM reductase accepts electrons from cytoplasmic thioredoxins, apparently through a cascade of redox exchange reactions involving multiple Cys in both the transmembrane helixes and the periplasmic γ-domain (101, 102, 103, 104, 105).

E. coli and Salmonella also harbor DsbG, a DsbC homolog, which forms an alternative redox pair with DsbD (Fig. 4). DsbG is required to preserve the catalytic Cys residues of L,D-transpeptidases involved in linking peptidoglycan to the Braun lipoprotein Lpp during peptidoglycan maturation (95). Like DsbC, the active form of DsbG is a dimer (106). Differences in a linker helix length and the pattern of hydrogen bonding at the dimer interface between these homologous proteins might contribute to substrate discrimination (98, 107). The protein environment surrounding the CX₂C motifs in DsbG and DsbC also differs. This likely change the pK_a and redox potentials of these enzymes, affecting their corresponding activities. Despite this, DsbC and DsbG can partially substitute each other in vitro and also in vivo (95, 108), except under Cu⁺ stress (96). Another DsbD substrate is the inner membrane-anchored protein DsbE (CcmG), which is involved in cytochrome c biogenesis (104, 109) (Fig. 4). Like DsbC and DsbG, DsbE has a periplasmic thioredoxin-like domain with a CX₂C motif (110, 111).

Figure 4.

The alternative DsbD partners, DsbG and DsbE (CcmG) and the reactions they catalyze. The structures of DsbG (PDB 1V58) and DsbE (CcmG) (PDB 2B1K) DsbDα (PDB 1JPE), DsbDγ (PDB 2FWF) and TrxA (PDB 6H1Y) are drawn in ribbon form (in blue). Structures of DsbDβ and the N-terminal region of DsbE predicted by AlphaFold 2 are shown in gray. Catalytic Cys, thiol or sulfoxide groups, S-S bonds, and electron flow are shown as in Figure 2.

The Salmonella Scs system and its paralogues in other pathogens

The scs locus is present in all available S. enterica and Salmonella bongori genomes, as well as in other enterobacterial species like Citrobacter, Serratia, Yersinia, Klebsiella, Cedecea, Raoultella, Enterobacter, Pantoea, Providencia, Proteus, and a swine diarrheagenic E. coli TA206 strain, as well as in the gamma proteobacteria Aeromonas (Fig. 5). In these species, a similar operon array is conserved. The first gene, scsA, is followed by scsB, scsC, and scsD. However, scsC and scsD are inverted in Vibrio sp. (Fig. 5). We have reported putative CpxR-binding sites in the promoter of most scs-like operons (38), suggesting both the presence of Cpx-dependent regulation and a Scs-mediated feed-back control. Curiously, in Cronobacter sakazakii a scsB homolog is next to a gene coding for a protein with ∼60% identity to Salmonella DsbG (Fig. 5). Incomplete scs operons, lacking either scsA or scsC, are present in Shewanella sp. and Photobacterium profundum, respectively. Caulobacter crescentus, a member of the alphaproteobacterial class and an opportunistic pathogen, encodes scsB and scsC homologs (112), while Haemophilus influenzae and Neisseria meningitidis only harbor a scsD homolog in their genomes (Fig. 5).

Figure 5.

The scs locus from Salmonella and paralogue species. The figure shows the chromosomal region of organisms harboring scsABCD homologs (analysis performed on the “Seed Viewer”, http://pubseed.theseed.org). Only a representative strain from either S. enterica or S. bongori is included in the figure, but all Salmonella genomes have the scsABCD locus. The S. enterica scsABCD promoter is indicated with a black arrow. The identified (black boxes) or predicted (gray boxes) direct repeats corresponding to the CpxR-binding sites upstream of scsABCD promoter are shown. The full names of analyzed strains are Salmonella enterica subsp. enterica serovar Typhimurium str. LT2; Salmonella bongori 12,149; Citrobacter sp. 30_2; Cedecea davisae DSM 4568; Escherichia coli TA206; Serratia sp. AS13; Yersinia sp; Raoultella ornithinolytica B6; Enterobacter sp B509; Klebsiella sp 1_1_55; Hafnia alvei ATCC 51873; Erwinia billingiae Eb661; Aeromonas sp 159; Pantoea sp GM01; Proteus mirabilis HI4320; Providencia stuartii MRSN 2154; Vibrio sp MED222; Photobacterium profundum SS9; Cronobacter zakazakii ATCC BAA-894; Shewanella sp W3-18-1; Haemophilus influenzae R2846; Neisseria meningitidis FAM18.

ScsB forms a functional redox pair with the periplasmic component ScsC in Salmonella, C. crescentus, and P. mirabilis (112, 113, 114) (Fig. 5). ScsB and its ortholog define a new group within the DsbD protein family (112, 113). These IM proteins have eight transmembrane segments and two periplasmic domains (α and γ) with the archetypical CX_2/3C motif (Fig. 6). Considering the similarity between ScsB and DsbD at the IM region (Figs. 3 and 6), it was proposed that TrxA, the cytoplasmic enzyme that feeds electrons to DsbD, is also the electron donor of ScsB (Fig. 6). Differences between ScsB and DsbD are more evident at the periplasmic α-region. In this region, DsbD has a single immunoglobulin-like fold (Fig. 3) (115), while the ScsBα (Fig. 6) contains two of these domains in tandem (112, 116). Only the N-terminal subdomain, named α-A, includes the redox-active site (CX₃C motif) that primes the ScsC activity and is conserved among the ScsB proteins (113). The other immunoglobulin-like subdomain, α-B, is more variable and has been suggested to modulate the interaction with other redox partners. For instance, C. crescentus ScsB interacts with ScsC and other proteins involved in the cell envelope reductive pathway such as PprX, a periplasmic peroxiredoxin, and PrxL, a peroxiredoxin-like protein (112). The possibility of crosstalk between the Salmonella ScsB/ScsC and DsbD/DsbC systems was analyzed in vitro (114). While the DsbDα domain transfers electrons to oxidized ScsC, ScsBα cannot catalyze the reduction of DsbC.

Figure 6.

The Salmonella Typhimurium Scs system. The structure of ScsC (PDB 4GXZ) and TrxA (PDB 6H1Y) are shown in ribbon form (in blue). The two subdomains (A and B) from the ScsBα region predicted by AlphaFold 2 are colored pink and purple, respectively. Structures of the β and γ domains of ScsB, ScsA and ScsD predicted by AlphaFold 2 are shown in gray. Yellow spheres represent the catalytic C composing the CX_2/3C motifs. Inserts highlight the active sites of ScsC and ScsD with Cys residues colored in yellow. Other conserved residues important for the redox activity of ScsC are highlighted, as well as the conserved P135 residue of ScsD. Dashed arrows indicate reported (black) or hypothetical (magenta) electron transfer between Scs proteins.

All ScsC homologs have the characteristic thioredoxin fold, the CX₂C catalytic motif, and various conserved residues (117). Salmonella ScsC Gly₁₈₆, Pro₁₇₅, and Thr₁₇₄ are also present in all homologs (Fig. 6), including E. coli DsbG. These residues are proposed to stabilize the reduced form of the protein (117). Lys₇₀ and Lys₉₆ residues that stabilize the thiolate form of the nucleophilic Cys₆₆ in the catalytic motif of Salmonella ScsC (117), are conserved in most of ScsC homologs, except in C. crescentus, Shewanella, and Pantoea sp. However, the ScsC-like proteins show significant heterogeneity in their oligomerization state. Proteus and Caulobacter ScsC form homotrimers connected by the long N-terminal α-helix, while the Salmonella homolog lacking this region is a monomer (116, 118) (Fig. 6). These distinct structures seem to be associated with different roles. Caulobacter and Proteus ScsC have disulfide isomerase activity (112, 116, 118). By contrast, Salmonella ScsC, initially proposed as a disulfide reductase in vitro (117), was shown to facilitate disulfide bond formation of the periplasmic arginine sensor ArtI under stress conditions (30). Because ArtI only has two Cys, it was speculated that ScsC may also repair these oxidized residues under redox stress. SodCI and SodCII are putative Salmonella ScsC substrates, as are other proteins involved in amino acid sensing and import, including ArgT, GltI, and HisJ (30). Residues around the Cys in the CX₂C motif likely influence the redox potential of periplasmic proteins. Both ScsC and the disulfide isomerase DsbG have identical CPYC motif at their active sites and similar low redox midpoint, −132 and −129 mV, respectively (117). Nevertheless, Salmonella ScsC shows a low pKa (3.4), comparable to that of the thiol oxidase DsbA (3.5). Thus, ScsC might have both oxidase and reductase activities in different conditions.

Salmonella ScsC and ScsB, and C. crescentus ScsC were shown to bind Cu⁺ and Cu²⁺ in vitro (114, 118). Both Salmonella ScsC and ScsB from Salmonella can also transfer Cu⁺ to CueP with similar efficiency (114). However, neither the specificity of metal binding nor the unidirectionality of Cu⁺ transfer has been tested. Also, the relevance of Cu transfer and its relationship with the redox activity of these enzymes await elucidation.

Salmonella scsD encodes a predicted IM-anchored protein with a periplasmic thioredoxin-fold domain. Its AlphaFold 2 structure (Fig. 6) resembles that of the membrane-linked thiol:disulfide interchange proteins such as DsbE/CcmG and TlpA, involved in cytochrome c and aa3 biogenesis, respectively (110, 119). Apparently, ScsD (Fig. 6) and its homologs have a conserved proline residue separated by 61 residues from their putative redox-active CX₂C motif. Interestingly, this Pro residue stabilizes the oxidoreductase motif of DsbE/CcmG-like proteins (111). Considering that TlpA is a substrate of Caulobacter ScsB (112), DsbE is another DsbD partner in E. coli (109) (Fig. 4), and the function of both Salmonella ScsC and ScsD under Cu⁺ stress depends on ScsB (38), we propose that ScsD is also a ScsB redox partner (Fig. 4).

Finally, Salmonella ScsA is the only Scs protein having a CX₂C motif but lacking the thioredoxin-like fold and probably the redox activity associated with this domain. Interestingly, it has no homology to any structurally characterized protein. Contrary to the other Scs proteins, its location within the bacterial envelope could not be determined (42). Some in silico predictions indicate that ScsA might be an integral IM protein with two transmembrane helixes (Fig. 6), while others showed a lipobox-like sequence in the N-terminal region that might direct ScsA to the OM. If ScsA is an OM protein, the CX₂C motif located at the N-terminus of the putative cleavage site is expected to be removed during processing. In the hypothetical case that ScsA remains in the IM, this motif would be hidden in the membrane (Fig. 6). Interestingly, all identified ScsA homologs carry this predicted lipobox sequence, and those more closely related to the Salmonella ScsA also conserve the CLAC array within the CX₂C motif. In addition, a peroxidase signature sequence, TAQRMAGLHAL (64), is detected in most ScsA homologs in a region predicted to be exposed to the periplasm. Taking this into account and given the role of ScsA in H₂O₂ tolerance (38, 42), it is tempting to speculate that this protein functions as a periplasmic peroxidase. The observation that expression of the Scs proteins avoided H₂O₂-mediated carbonylation of envelope proteins (42) further supports this idea. Based on the requirement of Cu stress for scsA transcription (38, 42), ScsA might prevent the generation of ROS species derived from Cu overload.

Concluding remarks

Counteracting redox and Cu stress in the envelope of bacterial pathogens is required to establish a successful infection. The collected evidence indicates that the canonical Dsb systems might not satisfy the extra demands that the host-pathogen interface imposes to preserve functional thiol groups. Then, it is logical that novel Dsb-like systems were acquired during evolution by intracellular pathogens like Salmonella or uropathogens like P. mirabilis. In fact, the Salmonella Scs system emerges as the linker between Cu and redox stress. The role of three of its components in disulfide bond generation and/or isomerization of envelope proteins might be assumed based on their structural features. However, their specific chemical function and biological roles require experimental confirmation. In addition, the biochemical function of ScsA and its membrane topology remain to be determined. Importantly, the specific requirements of the ScsABCD system at the host/pathogen interface, as well as its relevance during the infection cycle of Salmonella and other pathogens, also need further inquiries.

Data availability

All data are contained within this paper.

Conflict of interest

The authors declare that they have no conflicts of interest with the contents of this article.

Reviewed by members of the JBC Editorial Board. Edited by Joan B. Broderick