Today, we find ourselves in an urgent need for novel antibacterial drugs, as many important human pathogens have acquired multiple antibiotic resistance factors. Among those, Staphylococcus aureus and S. epidermidis play a major role as the leading sources of nosocomial infections. Recently, it has been suggested to develop therapeutics that attack bacterial virulence rather than kill bacteria. Such drugs are called “antipathogenic” and are believed to reduce the development of antibiotic resistance. Specifically, cell-density-dependent gene regulation (quorum-sensing) in bacteria has been proposed as a potential target. While promising reports exist about quorum-sensing blockers in gram-negative bacteria, the use of the staphylococcal quorum-sensing system as a drug target is now seen in an increasingly critical way. Inhibition of quorum-sensing in Staphylococcus has been shown to enhance biofilm formation. Furthermore, down-regulation or mutation of the Staphylococcus quorum-sensing system increases bacterial persistence in device-related infection, suggesting that interference with quorum-sensing would enhance rather than suppress this important type of staphylococcal disease. The chemical nature and biological function of another proposed staphylococcal quorum-sensing inhibitor, named “RIP”, are insufficiently characterized. Targeting quorum-sensing systems might in principle constitute a reasonable way to find novel antibacterial drugs. However, as outlined here, this approach requires careful investigation in every specific pathogen and type of infection.
Conventional antibiotics work by interfering with fundamental cell processes, such as protein synthesis or DNA replication. The first antibiotic, penicillin, which was discovered in 1928, attacks cell wall synthesis, as do many antibiotics in use today. In the meantime, a series of novel classes of antibiotics has been discovered. However, the frequent use of antibiotics in many countries and the preferred use of antibiotics with broad-range efficacy have promoted the fast spread of bacterial resistance even to modern antibacterials. Antibiotic resistance is particularly problematic in biofilm-associated infections and in staphylococci. Pharmaceutical companies and academia have tried to respond to this dangerous situation using rational drug design and target-oriented drug development. Although this approach has not yet been very successful, it is seen as our last chance to win the battle against multi-resistant bacteria .
More recently, it has been proposed to develop substances that specifically inhibit bacterial virulence . Such “antipathogenic” drugs, in contrast to antibacterial drugs, do not kill bacteria or stop their growth and are assumed not to lead to the development of resistant strains. A very elegant approach consists in the inhibition of regulatory systems that govern the expression of a series of bacterial virulence factors. However, this review will show that the interference with complex regulatory systems requires careful investigation of the physiological consequences in every targeted species. To this end, the inhibition of bacterial cell-density-dependent gene regulation, or quorum-sensing, in the important human pathogens Staphylococcus aureus and S. epidermidis will be discussed.
Pathogenesis of S. aureus and S. epidermidis
S. aureus, S. epidermidis, and other members of the genus Staphylococcus normally colonize human epithelia and mucous membranes . While S. aureus only occurs in a certain percentage of carriers, S. epidermidis is ubiquitous among the human population. Carriage of S. aureus, particularly nasal carriage, predisposes to S. aureus infection . S. epidermidis is the most frequent cause of nosocomial infections, followed by S. aureus, which together account for a majority of nosocomial infections . Neither species normally causes disease without penetration of the epithelial barrier of the skin or mucosal surfaces. Once inside the human body, S. aureus tends to cause more severe infection than S. epidermidis, as it has a large arsenal of virulence factors, which includes exotoxins, superantigens, exfoliative toxins, hemolysins, and Panton–Valentine leukocidin . The types of diseases caused by S. aureus range from abscesses and furuncles to scalded skin syndrome, sepsis, and toxic shock syndrome. The more severe diseases are only caused by a subset of S. aureus strains that synthesize the corresponding virulence factor, such as toxic shock syndrome toxin. Among the mentioned aggressive virulence factors, S. epidermidis only synthesizes delta-hemolysin . Its main virulence determinant is considered to be biofilm formation. However, we have only begun to investigate virulence factors of S. epidermidis and further virulence determinants will certainly be discovered in this species. For example, a complex of pro-inflammatory peptides called phenol-soluble modulin, which includes the delta-hemolysin (PSM gamma), has recently emerged as a novel determinant of S. epidermidis virulence . Nevertheless, human infection by S. epidermidis is still considered an “accident” in the lifestyle of this bacterium and accordingly, S. epidermidis infection is chronic and characterized by the intention of the bacteria to remain in relative silence and evade the human immune response.
Quorum-sensing control and possible applications in drug development
Quorum-sensing describes the regulation of gene expression in response to increasing cell density. It enables the bacteria to adapt to the altered environmental conditions at a high density of population, such as a low supply of nutrients. To signal and sense the status of cell density (“quorum”), bacteria use specific secreted signal molecules, also called pheromones or autoinducers. The pheromone is always produced at a certain level. When its extracellular concentration reaches a threshold, a signal transduction system is activated, which ultimately leads to target gene activation. Often, pheromone production is stimulated in an autoregulatory fashion to ascertain a fast response to the increase in population density. The best-studied quorum-sensing system in gram-negative bacteria is the LuxR–LuxI system, which was first discovered in Vibrio fischeri. This system is widespread in gram-negatives and uses an N-acyl-homoserine lactone (AHL) as quorum-sensing signal .
Among the many virulence factors and phenotypes that are controlled by quorum-sensing, a focus has been on biofilm formation . Over recent years, biofilm formation has been recognized as an important virulence mechanism of many bacterial pathogens. It has been estimated that 60% of all microbial infections involve biofilms . Advantages of life in a biofilm comprise protection from host defense and antibiotics, increased availability of nutrients by sequestration to nutrient-rich areas, and cooperative benefits of life in a community in general .
A recent example of how quorum-sensing inhibitors might be used to control infection is the development of furanone-based compounds to inhibit biofilm formation by P. aeruginosa[11,12]. P. aeruginosa is the most frequent opportunistic pathogen encountered in cystic fibrosis patients. Furanones, halogenated and acylated furan structures, are naturally found in the marine alga Delisea pulchra and have antimicrobial and quorum-sensing-inhibitory effects on several gram-negative bacteria . They function by competing with the AHL pheromone for its receptor . The natural D. pulchra furanone has only little inhibitory activity on the P. aeruginosa quorum-sensing system. However, synthetic analogs of the D. pulchra furanones have been produced to efficiently block P. aeruginosa biofilm formation . It can be expected that furanone-based drugs may be designed to block homologous quorum-sensing systems in many gram-negative bacteria.
Another recently discovered quorum-sensing system, called the luxS system, seems to be even more widespread, as it also is present in some gram-positive bacteria . It uses an autoinducer named AI-2. However, it is not clear yet if it governs the expression of virulence determinants.
With the possible exception of the luxS system, quorum-sensing systems in gram-positive bacteria reveal far more diversity than in gram-negative bacteria. However, similar to gram-negatives, gram-positives use quorum-sensing control to regulate a variety of processes, including competence development (in several Streptococcus species), sporulation (in Bacillus subtilis), antibiotic biosynthesis (in Lactococcus lactis), and virulence (in S. aureus and S. epidermidis) . The gram-positive systems differ from the gram-negative AHL systems in 2 main factors. First, the signal substances are usually peptides, which frequently have posttranslational modifications. Second, the signal substances often do not diffuse into the cell to bind to a cytoplasmic protein, but signal through binding to a membrane-located part of a two-component system. The cytoplasmic response regulator part of the two-component system then activates target genes by binding to DNA. Not much is known about quorum-sensing control of virulence, and particularly biofilm formation, in gram-positive bacteria, with the exception of S. aureus, S. epidermidis, and some streptococci. For example, the com system in Streptococcus mutans regulates biofilm architecture and might therefore represent a target for quorum-sensing blockers . Obviously, the heterogeneity of quorum-sensing systems in gram-positive bacteria requires in-depth investigation of quorum-sensing control in every strain.
Quorum-sensing control in Staphylococcus: the agr system
Function of the agr system
The staphylococcal quorum-sensing system, named agr for accessory gene regulator, consists of a two-component system for signal transduction (AgrA and AgrC), a prepheromone protein (AgrD), and of AgrB, which presumably is responsible for maturation and export of the post-translationally modified pheromone peptide  (Fig. 1). The pheromone peptide is 7–9 amino acids in length and has a thiolactone linkage between a central conserved cysteine residue and the C-terminal carboxy group (Fig. 2). The thiolactone ring structure is absolutely necessary for biological activity [19,20]. The pheromone binds to the histidine kinase AgrC in a not yet completely understood fashion. Several agr subgroups exist that are characterized by a specific pheromone amino acid sequence, in which only the thiolactone structure (or lactone structure, in one case) is conserved. Interestingly, specific S. aureus and S. epidermidis agr subgroups are predominantly associated with certain diseases [21,22]. Importantly, the staphylococcal agr pheromones, in contrast to the gram-negative AHLs, show the phenomenon of cross-inhibition: pheromones of the self induce, whereas pheromones of non-self inhibit the quorum-sensing response [23,24].
The effector molecule of the agr system is a regulatory RNA, called RNAIII, whose synthesis is dependent on agr activation and driven by the P3 promoter of the agr system . In contrast, synthesis of RNAII, which encodes the agr protein components, is driven by the P2 promoter. It is not known how RNAIII regulates the transcription of target genes.
agr controls virulence factors
Almost all known aggressive virulence factors of S. aureus are up-regulated by agr, which has made agr well-known as a regulator of virulence . Infrequent exceptions are the enterotoxins A and K. Accordingly, animal infection models of acute virulence, e.g. subcutaneous abscess formation and endocarditis, demonstrated that agr is required for S. aureus virulence [26,27]. In contrast, colonization factors such as MSCRAMMs (microbial surface components recognizing adhesive matrix molecules) are regulated by agr in S. aureus in an opposing fashion. This type of regulation makes sense, as low agr activity during the beginning of an infection allows for the expression of colonization factors. High agr activity at progressed infection states then causes the expression of aggressive virulence factors and degradative exoenzymes, allowing the bacteria to acquire nutrients from host tissue, and eventually to escape from the abscess and spread to other infection sites.
With the possible exception of some proteases and lipases, whose role in virulence is still unspecified, not much is known about S. epidermidis virulence factors and their regulation by quorum-sensing [2,28]. This has changed somewhat with the recent discovery of the pro-inflammatory PSM complex. The biologically active components of this complex are several amphipathic peptides, which have a variety of pro-inflammatory properties. For example, they cause chemotaxis of neutrophils, delayed apoptosis, and the expression of cytokines in several cells of macrophage lineage [7,29,30]. The production of PSM peptides is strictly regulated by the agr system, resulting in the complete absence of PSM peptides in an isogenic agr mutant of S. epidermidis. The agr mutant did not promote cytokine production, and only induced chemotaxis to a very reduced extent. These findings demonstrated that agr has a dramatic impact on the pro-inflammatory capacity of S. epidermidis, most likely due to the regulation of PSM expression. Presumably, the significance of quorum-sensing regulation of the pro-inflammatory PSM peptides consists in creating a well-balanced level of inflammation during infection. Whereas during initial states of S. epidermidis infection, low agr activity suppresses the production of PSMs and thus allows for immune evasion, PSM expression at high cell density attracts immune cells that contribute to tissue degradation and thus produce nutrients for the starved bacterial population.
agr and biofilm formation
While the sizeable number and significance of virulence factors that are regulated by agr suggest that agr represents an excellent target for quorum-sensing blockers used to control S. aureus and S. epidermidis infection, more recent discoveries raise some doubt about this approach . Colonization factors such as MSCRAMMs and the complex phenotype of biofilm formation are down-regulated by agr in S. aureus. Furthermore, the dysfunction of agr is correlated with persistent bacteremia in S. aureus. Moreover, in several infection models, agr expression did not seem necessary for the development of disease [35,,37].
As S. epidermidis infections tend to be chronic rather than acute, blocking of the quorum-sensing response seems to be even less appropriate in this pathogen. In fact, we recently demonstrated that quorum-sensing blockers cause increased biofilm formation of S. epidermidis in vitro  (Fig. 3(a)). Moreover, a clinical biofilm-forming S. epidermidis strain showed increased biofilm formation in vitro, increased colonization of epithelial cells and most importantly, increased persistence in a rabbit model of device-related infection  (Fig. 3(b)). The likely effector molecules that influence biofilm formation and are regulated by agr in S. epidermidis are: (i) the abundant surface-attached autolysin AtlE, which presumably contributes to cell surface hydrophobicity and initial attachment to plastic surfaces, (ii) the PSM peptides, which appear to inhibit S. epidermidis biofilm formation in secondary stages, and (iii) S. epidermidis MSCRAMMs, although the quorum-sensing control of the latter is merely speculative . Interestingly, permanent elimination of the quorum-sensing response by mutation of agr occurs naturally in S. epidermidis. Furthermore, agr mutants are found significantly more frequently among S. epidermidis strains isolated from the infection of indwelling medical devices, suggesting that disabling agr enhances pathogen success during this most important type of S. epidermidis infection .
RAP and TRAP: a second quorum-sensing system in S. aureus?
The nature and function of two substances, “RAP” and “RIP”, have recently been controversially discussed . Problematic aspects are therefore summarized in the following.
Before the mode of action of agr and the structure and function of agr pheromones became clear, a not further characterized protein named RAP was reported to activate RNAIII synthesis . Since then, the agr system was investigated in detail and it was shown that the agr pheromones are the only substances that activate the agr system in S. aureus. More recent investigation showed that RAP is the ribosomal protein L2 , which appears in the supernatant of S. aureus for unknown reasons. RAP lacks any known secretion signal and its appearance in the supernatant is difficult to explain other than by cell lysis. RAP has also been reported to trigger the phosphorylation of a cytoplasmic protein named TRAP . However, this would require a novel, yet unknown signal transfer mechanism across the cytoplasmic membrane. The RAP–TRAP system has been proposed as a second quorum sensing system in S. aureus that communicates with the agr system . However, many details of this system, for example signal production (secretion of RAP) and signal transduction (interaction of two proteins inside and outside of the cell) are unexplained. Furthermore, the signal level of the RAP–TRAP system (extracellular RAP concentration) appears to be dependent on cell lysis rather than on cell density, as would be required for quorum-sensing control. With the possible exception of the yet insufficiently characterized luxS system, the agr system is thus here considered the only quorum-sensing system in staphylococci.
In 1995, an RNAIII-inhibiting peptide (“RIP”) was found in what was then thought to be S. aureus culture filtrate . The strain later turned out to be a coagulase-negative Staphylococcus, likely Staphylococcus warneri. This explained the inhibiting activity, as most coagulase-negative agr pheromones inhibit S. aureus agr. Further, the partially determined amino acid sequence of the peptide is very similar to that of the S. warneri agr pheromone. However, the structure of “RIP” has not been completely determined. Instead, a peptide was synthesized (YSPWTNF), in which a tryptophane was placed at the position normally occupied by a cysteine in agr pheromones . A ring structure, which independent groups have shown to be absolutely required for agr activation [19,20]c was not introduced in this peptide. The linear peptide YSPWTNF was reported to have agr-inhibiting activity . However, this could not be reproduced in a series of laboratories .
More recent studies on the biological properties of the peptide also used the amide form (YSPWTNF-NH2) [46,,,49]. All peptide forms, including the insufficiently characterized and impure initial preparation (“native RIP”), which are chemically different, have been called “RIP”. “RIP” (in most cases, YSPWTNF-NH2 was used) has since been shown to inhibit a series of staphylococcal infections, including biofilm-associated infection, and in vitro biofilm formation [47,,49]. As a consequence, it has been suggested to coat indwelling medical devices by “RIP” to inhibit colonization by S. aureus and S. epidermidis. As inhibition of the agr system leads to increased rather than decreased biofilm formation [33,38,39], the biofilm-inhibiting activity of “RIP” contrasts its inhibitory activity on RNAIII production. Importantly, it has not been shown that the same, defined form of “RIP” combines quorum-sensing inhibitory and biofilm-inhibitory activity. Furthermore, the used “RIP” concentration (∼10–50 μg/ml) is in a range, in which peptides with a certain degree of amphipathy can inhibit adhesion merely by their detergent-like properties [33,39]. This might well explain the observed activity of RIP during biofilm formation. Coating of indwelling medical devices by detergent-like molecules to circumvent bacterial adhesion has been studied . As long as there is no clear evidence that inhibition of biofilm formation by a defined form of “RIP” is caused by a regulatory process that involves quorum-sensing, rather than by the physico-chemical properties of “RIP”, the use of “RIP” to coat implants does not represent an advance over previously suggested coating substances. Taken together, the structure of RIP and the biological properties of the different forms of RIP will have to be investigated in more detail before a conclusive answer can be given to the question if RIP has potential as an antistaphylococcal drug.
Interference with quorum-sensing control for the treatment of Staphylococcus infection is problematical. The agr-type pheromones and derivatives might have some potential use for certain types of acute S. aureus infection. However, there always is the caveat that they will increase colonization and biofilm formation due to the intrinsic properties of the agr system. In other words, altering the structure of such derivatives can overcome the problem of species specificity but never that of encountering colonization-enhancing side effects. Furthermore, systemic application of the agr pheromones is difficult, as they are peptide-based and most likely are rapidly degraded inside the human body. If pursued, rational drug design based on the agr pheromones must include the effort to stabilize the structure. It will also be necessary to gain more detailed insight into the role of agr in different types of infection.
As the example of furanone quorum-sensing blockers in gram-negative bacteria tells us, in-depth investigation of quorum-sensing control may lead to the development of novel antipathogenic drugs. Possibly, the mechanism of quorum-sensing inhibition of biofilm formation in Staphylococcus might be a basis for drug development aimed at controlling biofilm-associated infection. Furthermore, the widespread luxS system might represent a putative target for the interference with quorum-sensing in gram-negative and gram-positive bacteria. However, it remains to be analyzed in every single case if the diversity and complexity of the regulatory systems that control virulence factor expression provide the necessary basis for antibacterial drug development.