agr and Other Two-Component Regulatory Systems

The paradigm of two-component regulatory systems virulence gene regulation is agr, which stands for accessory gene regulator and is schematized in Figure 196-3.⁴⁹ agr functions as a quorum sensing control that reacts to bacterial density, allowing the preferential expression of surface adhesins during the exponential phase of growth (low cell density) and switching to the expression of exoproteins during the postexponential and stationary growth phases (high cell density).^47,49 The switch is composed of two divergent operons (see Fig. 196-3). On the left hand, promoter P2 drives the transcription of a series of components that comprises (1) a transmembrane protein (AgrB); (2) an autoinducing peptide precursor (AgrD), which is processed and exported by membrane-spanning AgrB; (3) a transmembrane sensor (AgrC), which is the cognate receptor of the AgrD-derived autoinducing peptide; and (4) a transcription regulator (AgrA) that can be activated by AgrC. At low cell density (exponential growth phase), the P2 promoter is off and the operon is transcribed at a low level. As cell growth proceeds, the concentrations of both bacteria and extracellular autoinducing peptide increase in the milieu, thereby augmenting the chance of the autoinducing peptide to make contact with its cognate AgrC receptor. On contact, AgrC activates the response regulator AgrA, a process that may involve AgrA dephosphorylation.⁴⁷

Activated AgrA is a DNA-binding protein that turns on the transcription from both promoter P2, generating a positive feedback on the system, and promoter P3, which drives the transcription of δ-hemolysin and of a peculiar effector called RNAIII. RNAIII has a reciprocal effect and activates the expression of several secreted proteins while downregulating the expression of surface-bound factors (see Table 196-1). RNAIII has a complex three-dimensional structure and a long half-life (up to 15 minutes). It regulates gene expression in several ways, including at the translational level by blocking the messenger RNA (mRNA) ribosome-binding site of the target genes.

The S. aureus chromosome encodes for more than 10 two-component regulatory systems involved in both metabolic envi­ronmental control and virulence gene regulation. Important two-component regulatory systems regarding virulence genes include saeR/S (for S. aureus exoproteins),⁵⁰ srrAB (for staphylococcal respiratory response),⁵¹ and arlS (for autolysis-related locus sensor).⁵² saeR/S was identified with transposon mutation in a pleiotropic mutant defective in exoprotein synthesis other than that regulated by agr (e.g., coagulase and nuclease; see Table 196-1).⁵³ saeR/S acts independently of agr and responds to environmental stimuli such as high salt, low pH, glucose, and subinhibitory antibiotic concentrations. srrAB and arlS interfere with growth in microaerobic conditions and autolysis, respectively. srrAB represses the expression of toxic shock syndrome toxin-1 (TSST-1) and protein A in microaerobic conditions,⁵¹ an observation that may be relevant for the pathogenesis of tampon-related toxic shock syndrome (see subsequent discussion).⁴ Both srrAB and arlS interact reciprocally with agr.^48,54

DNA-Binding Proteins

sar stands for staphylococcal accessory regulator.⁵⁴ It is an important locus that encodes a DNA-binding protein, SarA, which positively controls agr (Fig. 196-4)⁸ and maybe also sae and arlS.^47,54 In addition, sar directly regulates adhesin genes (see Table 196-1). The sarA transcripts peak at the end of the logarithmic phase of growth, thus promoting agr expression. Moreover, sarA itself is transcribed downstream of three alternate promoters, which can themselves be regulated by as yet incompletely solved factors.

SarA is the prototype of a growing family of DNA-binding proteins that may drive a number of transcriptional activities, including the expression of housekeeping genes and phage-related genes. sarA homologues include sarR, sarS, sarT, sarV, sarU, sarY, and rot.^48,54 rot stands for repressor of toxins and counters toxin expression by repressing agr.⁵⁵ Inactivation of rot partially restored the agr phenotype in agr-negative mutants, probably by alleviating a repressing effect on the downstream P3 cascade of the agr. This downstream cascade might be the target of several additional regulators that also affect the agr phenotype (see Fig. 196-4).

Sigma factors (σ) are another major mechanism of response to environmental stimuli. In bacteria, σ factors combine with and activate RNA polymerase to transcribe specific sets of genes. S. aureus contains one σ^A and two alternative σ^B and σ^C. Alternative σ^B is important for the microbial response to a variety of stresses, including temperature, energy depletion, and chemical stimuli.⁵⁶ It acts mostly via the global regulatory network but also has some direct effect by activating the expression of coagulase and fibronectin-binding proteins at the early growth phase and downregulating certain secreted proteins in the stationary phase. Mutants overexpressing σ^B were more virulent in experimental endocarditis, probably by increasing the expression of surface adhesins.⁵⁷ But the pathogenic role of σ^B was not shown in other models.

Small RNAs and Endoribonuclease III

Small RNAs (sRNA) are increasingly recognized as major players in regulation of gene expression. They act mainly at the translational level via antisense hybridization with mRNA, where they can either alter mRNA stability, hide ribosome-binding sites (RBS) from ribosome recognition, or conversely reveal RBS that are hidden in secondary mRNA structures by unfolding these very structures. Alternatively, sRNA can also bind regulatory DNA-binding proteins, thus sequestering them from their original gene regulatory function. At least 12 regulatory sRNA have been identified in S. aureus,⁵⁸ but many more are likely to come. The most prominent of them are RNAIII, which orchestrates the agr response, and RNAI, which regulates the replication of multiresistance plasmid pSK41.

In symmetry, post-transcriptional expression is also modulated by direct RNA alteration via endoribonuclease III (RNase III). This RNA double-stranded endonuclease plays a critical role in RNA processing and decay. It was recently shown to modulate post-transcriptional expression by various mechanisms, including turnover of transcribed and nontranscribed RNAs, as well as maturating the 5′ untranslated region (5′UTR) of the mRNAs of the cold-shock protein cpsA and maybe the protein A spa genes, to increase their stability and translation.⁵⁹

This intricate network underlines the subtlety of the bacterial response to environmental stresses. The regulatory network must be considered as a metabolic hub that integrates both external and internal information and responds in the most appropriate way. Some of these circuitries are likely to complement each other. Interruption of one of them may cause compensation by others, thus introducing biases in the observed phenotype. In this complex system, agr appears to be a central switch toward which many other regulators converge (see Fig. 196-4).

Role in Pathogenesis

The intuitive agr-based model suggests that scattered growing bacteria produce primarily adhesins, promoting tissue colonization, whereas installed organisms that form dense populations switch to the production of hydrolytic enzymes and toxins for the purpose of feeding and escaping host defenses.^8,47 Accordingly, inactivation of the function of agr alone decreased pathogenicity in experimental models of tissue destruction (e.g., subcutaneous abscesses), where exoprotein production is likely to be important.⁶⁰ On the other hand, agr inactivation did not influence much the course of experimental endocarditis, where bacterial surface adhesins are critical for valve colonization.⁶¹ Indeed, although agr-negative mutants are hampered in exoprotein production, they are still fully equipped with surface-bound colonizing determinants (see Table 196-1). In contrast, inactivation of sar decreased infectivity in experimental endocarditis⁶¹ because in addition to its effect on agr expression (see Table 196-1; see Fig. 196-3), sar also acts directly on expression of surface-bound fibronectin-binding protein A, which promotes experimental endocarditis.^62,63

In addition, in vivo gene expression revealed a further level of complexity.⁴⁸ For instance, although sar transcripts were detected in infected vegetations during experimental endocarditis, they were expressed from both P1 and P2 promoters, rather than only from the P1 promoter as observed in vitro.⁶⁴ Likewise, in vivo expression of several genes appeared dissociated from their control regulator as described in vitro. Although agr positively regulates TSST-1 in vitro (see Table 196-1), the toxin was still expressed by an agr-negative mutant in a rabbit model of toxic shock syndrome in vivo.⁶⁵ This may result from alternative regulation by other regulators that act either downstream of the agr locus or directly on the tss gene promoter. Eventually, agr-negative mutants can be recovered from clinical samples. Such agr-negative clinical isolates, and agr-negative laboratory mutants, have an increased ability to form biofilms, which are found in chronic infections such as osteomyelitis and device infections.⁶⁶

Hence, the pathogenic implication of regulatory circuitries cannot be drawn merely from in vitro observations. In vivo experimentation reveals the plurality of S. aureus infection forms, which may be variously altered by novel antivirulence therapies. For instance, inhibition of the agr loop by acting on the autoinducing peptide impedes acute tissue destruction⁶⁰ but might promote biofilm formation and chronic infection.

Ecologic and Epidemiologic Implication of agr

Genetic and functional experiments revealed the existence of at least four agr groups in S. aureus, which were characterized by specific variations in all three AgrB, AgrD, and AgrC proteins (Fig. 196-5).⁶⁷ Whereas the autoinducing peptide of a given agr group stimulated signaling in other strains sharing the same agr group, it either cross-inhibited (e.g., group I and group IV) or cross-activated (e.g., group I and group II) members of other groups. This suggests that certain antagonistic agr groups could be mutually exclusive with attempts to simultaneously colonize the same niche. However, studies regarding this hypothesis gave conflicting results. In particular, patients with cystic fibrosis colonized with S. aureus can successfully harbor organisms from two antagonistic agr groups.

Although agr and other global regulators control the timely expression of pathogenic genes, they are not bona fide pathogenic factors themselves. The agr locus has homologues in numerous nonpathogenic staphylococci. A phylogenic study of nonpathogenic CoNS indicated that variations in agr genes followed parallel variations in species-specific rRNA genes.⁶⁸ In fact, agr groups diverged very early on during the evolution of staphylococci (see “Comparative Genomics and Evolution”) and represent a lineage marker of strains that evolve in distinct environments rather than strategy to exclude potential competitors. Thus, global regulators were originally meant to control the expression of useful metabolic genes. How exogenous virulent genes, which were acquired later, succeeded in taking advantage of such systems remains a fascinating question of evolutionary genetics.