Investigating the genetic regulation of the ECF sigma factor σ^S in Staphylococcus aureus

Staphylococcus aureus is an exceedingly virulent and successful pathogen, capable of causing a wide range of infections, from relatively benign skin lesions to life threatening septicemia. With an overwhelming ability to adapt to its environment, S. aureus has become the most common cause of both hospital and community acquired infections, and is believed to be the leading cause of death by a single infectious agent in the United States [1],[2]. The threat posed by this organism to human health is further heightened by the rapid and continued emergence of multi-drug resistant isolates [1]-[4]. The success of S. aureus as a pathogen can, in part, be attributed to its many immune-evasion strategies, mediated by secreted and surface-associated virulence determinants. These virulence determinants have multi-faceted roles, allowing for adhesion, immune subversion and dissemination.

The regulation of gene expression is a fundamental property to all living systems that allows them to adapt and respond to different environmental conditions. Regulation can occur at the level of transcription, translation and post-translation, with the primary point being at the level of transcription. In bacteria, transcription is catalyzed by RNA polymerase (RNAP) whose promoter recognition and selectivity is influenced by a variety of transcription factors, with the most important group being sigma factors. In order for prokaryotes to initiate transcription, a sigma factor is required to associate with core RNAP. Once associated, sigma factors play a role in promoter recognition, and promoter melting to form the transcriptional open complex. All prokaryotes have a primary sigma factor, σ^A, which mediates transcription of the majority of genes, including those with housekeeping functions (e.g. replication machinery, cell division). In addition to this, alternative sigma factors exist, which control subsets of genes that are involved in specialized cellular functions or stress responses (e.g. oxidative stress, heat shock, etc.) [5]. In 1994, Lonetto et al. described a subfamily of alternative sigma factors known as extracytoplasmic function (ECF) sigma factors; and over the past two decades there has been a large number of these elements identified [5]-[7]. Indeed, ECF sigma factors now represent the most numerate sub-family of these enzymes, with many bacteria possessing multiple such factors; so that this group outnumbers all other types of alternative sigma factors combined [5]. In addition to their growing number, their role in virulence is become increasingly apparent. In Pseudomonas aeruginosa, ECF sigma factors play an important role in iron uptake pathways, alginate secretion, and the expression of virulence factors, all of which contribute to pathogenesis. With such a complex life style, it would not be surprising if several ECF sigma factors were present in the genome of S. aureus. However, to date, only one has been identified, σ^S, which forms the basis of this work [7].

Our group has previously shown that purified σ^S is able to bind core-RNAP and initiate transcription from its own promoter. A sigS mutant was found to be more sensitive to lysis by Triton X-100, and is outcompeted in long-term growth experiments by the parent under both standard conditions and in the presence of chemical stressors. Using a murine model of septic arthritis, it was also found that sigS is required for full virulence of S. aureus[7]. More recently, our group demonstrated that sigS expression is not observed under standard conditions except in the highly mutated strain, RN4220 [8]. However, sigS expression is induced in the presence of a number of chemical stressors that include DNA damaging and cell wall targeting agents. In addition, this element is upregulated upon challenge by components of the immune system, and during phagocytosis by murine macrophage-like cells [8].

The control of alternative sigma factor expression has been increasingly documented, with a number of novel genetic regulatory mechanisms identified [9],[10]. A recent study regarding sigH in S. aureus, demonstrates that the expression of this element can only be induced by a chromosomal gene duplication rearrangement that occurs spontaneously, generating a new chimeric sigH gene driven by the promoters of either nusG or rplK[9]. A second, post-transcriptional mechanism of regulation also occurs, through an upstream inverted repeat that suppresses the translation of σ^H protein [9]. With regards to the regulation of ECF sigma factors, a number of studies have described unique mechanisms of regulation, including for an ECF sigma factor that is closely related to σ^S, σ^M of B. subtilis. In a recent study it was shown that inactivation of a conserved membrane protein, yfhO, leads to an increase in activity of a putative bactoprenol glycosyltransferase. This subsequently leads to activation of σ^M[10], as the shortage of available bactoprenol impacts cell envelope integrity, which σ^M is believed to play a role in protecting [11].

In this study we expand on our previous work exploring the environmental induction of sigS expression in S. aureus, by focusing at the genetic level, so as to understand the regulatory circuits involved in controlling the expression of this gene.

The average bacterial genome is estimated to encode at least six ECF sigma factors [5],[33]. S. aureus is a highly successful pathogen with complex regulatory circuits that, interestingly, encodes only one ECF sigma factor, σ^S. This sigma factor has previously been shown to be important in the stress and virulence response of S. aureus[7]. In addition, we have also shown that sigS is differentially regulated across a number of S. aureus wild-type strains, which would indicate a network of regulation may exist, controlling the expression of this element. In earlier work by our group we have investigated the regulation of sigS by environmental signals in response to external stress; therefore in this study, we investigated the genetic mechanisms that regulate this transcription factor. We observed a number of findings that appear to correlate with our previous works regarding the function of σ^S, detailed as follows (and in Figure 7).

To explore direct regulators of sigS expression, we used a biotin pull down assay in conjunction with the sigS promoter, and identified CymR as specifically binding to this region of DNA. Using qPCR we were able to demonstrate that sigS expression actually increased in a cymR mutant, indicating that this regulator is acting as a repressor. CymR belongs to the poorly characterized Rrf2 family of proteins, and has been shown to be the master regulator of cysteine metabolism in S. aureus[28],[34]. Functionally, CymR binds directly to the promoter sequences of eight transcription units corresponding to 18 genes, and represses the transcription of elements involved in pathways leading to cysteine formation. Interestingly, in the presence of reactive oxygen species (ROS), it has been shown that there is derepression of several direct CymR targets [8],[28],[34]. We ourselves have previously shown that sigS expression is inducible during oxidative stress, and that sigS mutants are more sensitive to ROS. This would therefore suggest a link between cysteine metabolism, CymR, oxidative stress and sigS expression in S. aureus.

We identified further repressors of sigS expression in S. aureus using transposon mutagenesis as well as an MNNG based mutagenesis screen, with the most profound effects observed resulting from the inactivation of LacR. In S. aureus, the lacR gene encodes the transcriptional regulator for the lactose utilization system [27]. Because this gene encodes a transcriptional regulator, we were interested to see if the protein binds to the promoter region of sigS. Using EMSAs we determined that the regulatory affect we observe for sigS by LacR is indirect. It is unclear why the S. aureus LacR protein may function to regulate genes other than those involved in lactose utilization; however, it does share homology with deoR of E. coli, which is involved in repressing genes that are a part of the catabolism of deoxyribonucleosides. Interestingly, preliminary work by our group suggests that sigS may be involved in the de novo assembly of nucleotides (Miller and Shaw, unpublished observation). As such, it is possible that LacR may have retained an evolutionary role in nucleoside catabolism, which explains why it might connect to sigS expression in S. aureus. In addition, we also observed that when the response regulator arlR is disrupted there is an increase in sigS expression. This response regulator is involved in autolysis, adhesion and virulence of S. aureus. Interestingly, arlR mutants display an increase lysis when exposed to Triton X-100, which is a phenotype that we have previously described for sigS mutants [7],[35]. Interestingly, an array performed on an arlR mutant indicates that this response regulator directly and/or indirectly controls expression of at least 114 genes. These include several genes identified in this study, such the lacABCDFEG operon, which is negatively regulated by lacR; ebhAB, which encodes an adhesin identified in our transposon screen, and the two component system kdpDE. In the context of this latter element we identified mutations in the sensor histidine kinase, kdpD as negatively influencing sigS expression. This gene is part of the two component system, KdpDE, which is involved in sensing and responding to potassium levels in the cell [36]. There has been an increase in evidence that demonstrates KdpDE is an adaptive regulator involved in the virulence and intracellular survival of pathogenic bacteria. Specifically, KdpDE has been shown to regulate genes involved in responding to oxidative stress, which we have previously shown sigS to be involved in [37].

We also identified a number of other genes as negatively influencing sigS expression that would appear to corroborate previous studies performed by our group. For example, we have shown that sigS expression is increased in the presence of DNA damaging agents [8]. Herein we identified insertions in genes that are involved in DNA replication, recombination and repair pathways as influencing the expression of sigS. Specifically, we identified a mutation in SACOL1413, which is a putative helicase demonstrating homology to Snf2 family of proteins that are thought to function in DNA damage repair [38]. Again, connected with this, we have demonstrated that cells deficient in sigS are more sensitive to a variety of DNA damaging agents, and that expression of sigS is similarly increased in the presence of such stressors. A mutation in the DNA mismatch repair protein, hexA, was also identified in our screen. HexA is a major component of the methyl mismatch repair system [39]. In S. aureus, disruption of hexA leads to a higher frequency of strains acquiring random SNPs [39]. Thus, it is plausible that, in the absence of hexA, the increased level of natural mutation would instigate a DNA-damage like-response, which seems to involve σ^S in the S. aureus cell. In addition, disruption of nupC, which encodes a nucleoside permease, increases expression of sigS. In the closely related bacteria B. subtilis, NupC is responsible for transporting all nucleosides other than hypoxanthine and guanine, which are then catabolized to purine and pyrimidine bases, and returned to the nucleotide pool for DNA biosynthesis [40]. This is of significant interest, as preliminary work by our group indicates that sigS mutants have severely diminished expression of nucleotide biosynthesis genes (Miller and Shaw, unpublished observation). This would suggest that in the absence of nupC, sigS expression is upregulated to aid nucleotide biosynthesis to replace those not being transported into the cell.

We also identified insertions in two genes that are involved in cell wall biosynthesis as having negative influence on sigS expression. This is of interest as we have previously reported that sigS expression is induced in the presence of antibiotics that targeting the cell wall, and that sigS mutants are more sensitive to this kind of stress [8]. The first of these two elements is ebh, a cell wall associated fibronectin-binding protein. Interestingly, cells deficient in ebh have been found to be more susceptible to lysis by Triton X-100, a phenotype that we also observed with a sigS mutant [8],[41]. Ebh, which is localized over the entire cell surface, is the largest protein in S. aureus, and deletion of this factor from the cell has serious consequences to cell wall stability [41],[42]. Thus, it is conceivable that in the absence of ebh, the expression of sigS would be increased in an effort to combat cell wall instability. Further to this, inactivation of femB also resulted in an increase in sigS expression. FemB is involved in cell wall biosynthesis, and femB mutants of S. aureus have been shown to have reduced glycine content of their peptidoglycan, leading to decreased cell wall stability and an increase in sensitivity to cell wall targeting antibiotics, such as β-lactams [43]. This data very closely mirrors our previous reports with σ^S, explaining why cells deficient in femB might upregulate sigS.

There were also six insertions identified in the repressor screen for genes involved in amino acid biosynthesis. These include disruptions in the amino acid metabolism genes asd (aspartate-semialdehyde dehydrogenase), ald1 (alanine dehydrogrenase), alr2 (alanine racemase) and argJ (glutamate N-acetyltransferase). Connected to this, we have previously shown that sigS expression is increased in the presence of amino acid limiting media [8]. Moreover, in an earlier study, we have also demonstrated the importance for a functional σ^S during extended starvation [7]. As such, it appears possible that disruption in the normal flow of amino acids in the S. aureus cells necessitates that activity of σ^S. We have also identified insertions in genes involved in alanine biosynthesis, including, SACOL1434. Further to this, in B. subtilis, alr2 is responsible for the interconversion of L-isomer and D-isomer of alanine [44],[45]. Interestingly, the ability for bacteria to synthesize D-alanine is essential for the biosynthesis of the cell wall in B. subtilis[45]. Therefore, the disruption of alr2 could lead to an increase in sigS expression via two mechanisms: amino acid limitation and cell wall instability.

In the positive transposon screen, we again identified a number of mutations in genes that are involved in amino acid biosynthesis. These include ilvB (the large subunit of acetolactate synthase), trpF (phophoribosylanthranilate isomerase), brnQ3 (a branch-chain amino acid transporter), and SACOL1476 (a putative amino acid permease) that lead to a decrease in sigS expression. Thus, akin to that suggested above, it is plausible that without the ability to import and/or synthesize amino acids, S. aureus is unable upregulate sigS expression and repair the DNA damage caused by MMS. We also identified a mutation in pepA3, which encodes a putative glutamyl aminopeptidase. Glutamyl aminopeptidases are involved in cleaving glutamic and aspartic amino acids from the N-terminus of proteins. The pepA gene of E. coli encodes aminopeptidase A, a homohexameric multifunctional protein with aminopeptidase and DNA-binding activities, the latter being associated with mechanisms of transcriptional control and DNA recombination. Research has demonstrated that PepA is involved in the transcriptional regulation of the carAB operon, which encodes carbamoylphosphate synthetase; this enzyme catalyzes the ATP-dependent synthesis of carbamoyl phosphate from glutamine [46]. This process is the first committed step in pyrimidine and arginine biosynthesis. As we observed a decrease in sigS expression in a pepA3 mutant, it may suggest a link between amino acid and pyrimidine biosynthesis, and sigS. Interestingly, we have also observed that sigS may be involved in both amino acid biosynthesis and pyrimidine biosynthesis ([8], Miller and Shaw unpublished observations).

In a global effort to identify regulators of sigS expression, two Tn551 transposon screens were performed. A consideration with such approaches is that we have previously shown that Tn551 has particular preference for a 108 kb hotspot region of the genome that extends from SACOL1411 to SACOL1490 [17]. In order to ensure that the findings generated in our study were specifically connected to transposon disruption, and not an artifact of transposition, we first set out to validate the role of genes identified in modulation of sigS expression. In our repressor transposon screen, we identified 51 unique insertion sites, 57% (29 mutants) of which fall within the known Tn551 hotspot. In order to ensure that the sigS repression observed in these mutants directly connected to transposon insertion, we validated all Tn551 insertions possible by parallel creation of mutants using the Nebraska Transposon Library. This is a collection of bursa aurealis transposon mutants in almost all non-essential S. aureus gene [32]. Of the 40 mutants available in the NTML for the repressor screen, only 20 retained phenotype when recreated using the bursa aurealis mutants, with only 9 of 17 from the hotspot region. In our transposon screen for activators of sigS expression, we identified 35 unique insertion sites, 51% (18 mutants) of which fell within the Tn551 hotspot. The NTML contained 21 mutants to test for recapitulation, with 11 proving to retain phenotype upon recreating mutations. In this regard 11 of these mutants were originally in the hotspot, with 8 bursa aurealis transposon mutants validating alteration in sigS expression. Interestingly, we also identified 5 transposon mutants that were common to both screens. However, upon validation, none of the mutants retained phenotypes as both repressors and activators; with one of them proving to have only negative effects on sigS expression, whilst the remaining four were positive effectors. Overall, these findings suggest caution should be applied to performing transposon mutagenesis screens without significant efforts exerted to validate findings. It also suggests that, despite the presence of a Tn551 hotspot in the S. aureus genome, mutations identified within are no more, or less, likely to be artifacts for the phenotypes assessed.

Herein, we present evidence that sigS expression is controlled by a number of factors within the S. aureus cell that are involved in amino acid biosynthesis, DNA-damage repair and cell wall structural integrity. In addition, we have identified a number of transcriptional regulators that influence sigS expression, at least two of which (CymR and KdpE) do so by direct binding to the promoter of this gene. The findings of the present study correlate with our previous works on σ^S, which show a role for this regulator in the DNA-damage response, and in the resistance to cell wall targeting antibiotics and amino acid starvation [7],[8]. Taken together, these results confirm our contention that sigS expression is tightly regulated within the S. aureus cell in a complex and multifactorial manner. A continued exploration of the role of this enigmatic protein remains a primary focus of our laboratory’s research endeavors.