OUP user menu

Cell envelope stress response in Gram-positive bacteria

Sina Jordan, Matthew I. Hutchings, Thorsten Mascher
DOI: http://dx.doi.org/10.1111/j.1574-6976.2007.00091.x 107-146 First published online: 1 January 2008


The bacterial cell envelope is the first and major line of defence against threats from the environment. It is an essential and yet vulnerable structure that gives the cell its shape and counteracts the high internal osmotic pressure. It also provides an important sensory interface and molecular sieve, mediating both information flow and the controlled transport of solutes. The cell envelope is also the target for numerous antibiotics. Therefore, the monitoring and maintenance of cell envelope integrity in the presence of envelope perturbating agents and conditions is crucial for survival. The underlying signal transduction is mediated by two regulatory principles, two-component systems and extracytoplasmic function σ factors, in both the Firmicutes (low-GC) and Actinobacteria (high-GC) branches of Gram-positive bacteria. This study presents a comprehensive overview of cell envelope stress-sensing regulatory systems. This knowledge will then be applied for in-depth comparative genomics analyses to emphasize the distribution and conservation of cell envelope stress-sensing systems. Finally, the cell envelope stress response will be placed in the context of the overall cellular physiology, demonstrating that its regulatory systems are linked not only to other stress responses but also to the overall homeostasis and lifestyle of Gram-positive bacteria.

  • cell wall
  • signal transduction
  • stress response
  • antibiotic resistance
  • two-component system
  • extracytoplasmic function (ECF) σ factor


Life in the microbial world is characterized by fierce competition, nutritional hardship, and often life-threatening changes of external (i.e. physicochemical) parameters. Adaptive responses of a bacterium to its environment are therefore one defining cornerstone of microbial life in its natural context, irrespective of the individual life style or habitat. Such adaptations require the sensitive monitoring of numerous environmental parameters (input) to orchestrate the activity of intricate and complex regulatory systems that initiate or readjust adequate cellular responses (output) in a continuous balancing act between costs and gain. No surprise then that efficient stress response systems – aimed to maintain the functionality and integrity of the cell under all circumstances – are found embedded in the genomic blueprint of almost any bacterium studied to date (Storz & Hengge-Aronis, 2000).

The Gram-positive cell envelope

One of the crucial cellular structures is the cell envelope and its integrity has to be ensured, at all times and at any costs. A detailed description of the biosynthesis and chemical composition of the Gram-positive cell envelope is beyond the scope of this review, and readers are referred to a number of excellent reviews on this topic (Archibald, 1993; Delcour, 1999; Foster & Popham, 2002). Suffice to say, the Gram-positive cell envelope consists only of two functional layers (compared with three in Gram-negative bacteria) that enclose the cellular contents: a cytoplasmic membrane, surrounded by a thick cell wall. It lacks an outer membrane (and therefore a periplasmic space sensu stricto, see below). The Gram-positive peptidoglycan sacculus – in contrast to its single-layered Gram-negative counterpart – is a three-dimensional multi-layered net-like structure of about 50-nm thickness that can withstand high turgor pressures (up to 20 atm, i.e. more than a racing bike tire!). Owing to the combination of rigid sugar chains perpendicularly crosslinked with flexible peptide bridges, the mesh of this net is a strong, but also elastic stress-bearing structure (Höltje, 1998; Delcour, 1999). It is a highly dynamic super-molecule that undergoes permanent biosynthesis, assembly, maturation, disassembly, and recycling, to allow maintenance of cell shape, cellular growth and division at the same time (Archibald, 1993).

The Gram-positive peptidoglycan sacculus is interspersed with almost equal amounts of teichoic acids (TAs) that can be either tethered to the membrane (lipoteichoic acids) or covalently linked to the sugar backbone of the peptidoglycan sacculus (wall TAs) (Delcour, 1999; Foster & Popham, 2002). TAs are important components of the Gram-positive cell wall, and play a crucial role in defining the physicochemical properties of the envelope: They are poly-anionic, phosphate-rich linear polymers mainly responsible for the overall negative net charge of the Gram-positive cell surface (Bhavsar, 2004). TAs – as peptidoglycan – were generally viewed as essential biopolymers of Gram-positive bacteria. But recent studies in both Staphylococcus aureus and Bacillus subtilis– while clearly verifying the fundamental role of TAs for the overall cell integrity and fitness (Weidenmaier, 2004, 2005) – challenge this assumption by demonstrating the dispensability of wall TAs, at least under laboratory conditions (D'Elia, 2006a, b). In contrast, recent work on lipoteichoic acid biosynthesis in Staphylococcus aureus clearly indicates the essentiality of this biopolymer (Gründling & Schneewind, 2007).

In addition to these basic features of a Gram-positive cell wall, a number of additional cell envelope structures are present in many bacteria, and often play an important role in virulence, antibiotic resistance, colonization, and multicellular differentiation. These structures include extracellular polysaccharide capsules and biofilm matrices (Miyake & Iijima, 2004; O'Riordan & Lee, 2004; Branda, 2005), as well as proteinaceous S-layers (Schäffer & Messner, 2005; Sleytr, 2007). Moreover, the cell walls of mycobacteria and corynebacteria show a unique architecture that includes additional layers, consisting of arabinogalactans and mycolic acid, that surround the cell wall and renders these pathogens impenetrable to many antibiotics (Dover, 2004).

Many central questions on the mechanism and control of cell envelope homeostasis – including aspects of its biosynthesis (such as the Lipid II cycle; see Fig. 1), turnover and overall architecture – still remain largely unanswered, despite its fundamental cellular role and decades of research. The three-dimensional architecture of the sacculus is a matter of an ongoing (and inspiring) debate between two competing – and mutually exclusive – models: peptidoglycan sheets vs. scaffolds (Vollmer & Höltje, 2004; Dmitriev, 2005). Another area of controversy is the presence of a periplasmic space, which is generally viewed as a hallmark feature restricted to Gram-negative bacteria. Contradictory to that, recent advances in electron microscopic techniques indicated the presence of a periplasmic-like space of about 20 nm width in Gram-positive bacteria, located between the membrane and the cell wall (Matias & Beveridge, 2005, 2006). These findings emphasize the limitations in our understanding of Gram-positive cell envelope architecture.

Figure 1

Cell wall biosynthesis and its inhibition by antibiotics. Important steps in cell wall biosynthesis are schematically depicted, and their cellular location is indicated below. GlcNAc, N-acetyl-glucosamine; MurNAc, N-acetyl-muramic acid. Amino acids are symbolized by small grey circles. The steps of cell envelope biosynthesis linked to the cytoplasmic membrane through undecaprenol (symbolized by the waved lines) are referred to as the ‘Lipid II cycle’. This name is based on its central intermediate, Lipid II, which consists of the GlcNAc/MurNAc-pentapeptide building block, covalently linked to the lipid carrier molecule undecaprenol via a pyrophosphate ester bridge. Some cell wall antibiotics relevant for this review are given and placed next to the steps they inhibit. Antibiotics in green sequester the substrate of the given step; those in red inhibit the corresponding enzymatic function. See text for details on their action. This figure was originally based in parts on (Silver, 2003), with modifications.

The cell wall – in addition to its roles as a shape-giving structure, an exoskeleton (to protect the cell from its environment and counteract the turgor pressure), and molecular sieve – is also discussed as a potent endotoxin in bacterial sepsis (Myhre, 2006). Moreover, the envelope also acts as a diffusion barrier (allowing and necessitating selective transport), and a communication interface (mediating information exchange) between the cell and its surrounding. The latter functions are primarily provided by the cytoplasmic membrane and its embedded proteins.

Antibiotics targeting the cell envelope

Because of its many crucial functions, the cell envelope is a prime target for numerous antibiotics, including many with high clinical relevance, that interfere with virtually any step in its biosynthesis (Lazar & Walker, 2002; Koch, 2003; Silver, 2003, 2006; Walsh, 2003). Some of the more common cell wall antibiotics important in the context of this review, as well as their respective major bacterial resistance mechanisms, will be introduced very briefly in the following paragraphs. Their site of interference with cell wall biosynthesis is schematically summarized in Fig. 1. All cell wall antibiotics either directly inhibit the enzymatic activity mediating a cell wall reaction (shown in red in Fig. 1), or sequester the substrate of a given step (indicated in green in Fig. 1) (Silver, 2006).

Fosfomycin inhibits the first committed step, the formation of UDP-N-acetyl-muramic acid from UDP-N-acetyl-glucosamine. It functions as an inactivating structural analog of phosphoenol pyruvate, the cosubstrate of the MurA-catalyzed reaction (Kahan, 1974). Resistance either arises by spontaneous mutations in the transport pathways (Horii, 1999) or is conferred by glutathione/metallothiol transferases that enzymatically inactivate fosfomycin (Suarez & Mendoza, 1991; Bernat, 1997; Cao, 2001).

d-Cycloserine prevents the completion of the pentapetide side chain, the canonical crosslinking agent of the peptidoglycan network. It inhibits both the d-alanine racemase, which converts l-alanine to d-alanine, and the d-alanine/d-alanine ligase, which catalyzes the formation of the corresponding dipeptide (Neuhaus & Lynch, 1964; Lambert & Neuhaus, 1972). Resistance can be either achieved by overexpression of the target proteins or by an efflux pump (Feng & Barletta, 2003; Matsuo, 2003).

A number of antibiotics, including lantibiotics, ramoplanin, vancomycin, or bacitracin, interfere with the Lipid II cycle, the bottleneck of cell wall biosynthesis (Breukink & de Kruijff, 2006). Lipid II is the basic peptidoglycan building block, N-acetyl-glucosamine/N-acetyl-muramic acid-pentapeptide covalently linked to the lipid carrier undecaprenol through a pyrophosphate ester bridge (Delcour, 1999).

Lantibiotics (such as nisin) are polycyclic peptide-derived antimicrobial agents that are ribosomally synthesized and posttranslationally modified to their active forms. They contain the unusual amino acid lanthionine as their name-giving feature and ‘hijack’ Lipid II as a docking molecule to form pores, ultimately resulting in cell lysis (Chatterjee, 2005; Breukink & de Kruijff, 2006). The primary resistance mechanism against these positively charged peptide antibiotics is lowering of the overall negative net charge of the Gram-positive cell wall by d-alanine incorporation into the poly-anionic TAs. Some cases of enzymatic degradation or modification of the lantibiotic have also been reported (Breukink & de Kruijff, 2006).

The overall positive charge and pore formation as the mode of antimicrobial action are shared by the wider family of so-called cationic antimicrobial peptides (CAMPs), which represent an important defense mechanism of the human immune system and have gained a lot of attention in recent years due to their potential as future therapeutics (Brogden, 2005; Hancock & Sahl, 2006; Giuliani, 2007).

Ramoplanin, a nonribosomally synthesized lipoglycodepsipeptide antibiotic (Walker, 2005), inhibits the transglycosylation step of peptidoglycan biosynthesis by binding Lipid II at the extracellular surface of the cytoplasmic membrane (Hu, 2003). No resistance has been reported, so far (Breukink & de Kruijff, 2006).

Vancomycin and other glycopeptide antibiotics, such as teicoplanin, block glycan polymerization and cross-linking by tightly binding to the d-alanyl-d-alanine dipeptide terminus of Lipid II and nascent peptidoglycan (Kahne, 2005). Resistance is gained by reprogramming cell wall biosynthesis, incorporating alternative peptide termini, such as d-alanyl-d-lactate, instead of d-Ala–d-Ala (Walsh, 1996; Healy, 2000).

Bacitracin is a cyclic nonribosomally synthesized dodecylpeptide antibiotic that requires the coordination of a divalent metal ion for its biological activity (Ming & Epperson, 2002). It binds very tightly to undecaprenyl pyrophosphate, thereby preventing the recycling of the lipid carrier by dephosphorylation (Stone & Strominger, 1971; Storm & Strominger, 1973). Four different mechanisms of bacitracin resistance have been described so far: (1) expression of bacitracin-specific ATP binding cassette (ABC) transporters (Podlesek, 1995; Ohki, 2003a; Mascher, 2003a), (2) de novo synthesis of undecaprenyl phosphate (Cain, 1993; Chalker, 2000), (3) expression of alternative undecaprenyl pyrophosphate phosphatases (Cao & Helmann, 2002; Ohki, 2003b; Bernard, 2005), or (4) exopolysaccharide production (Pollock, 1994; Tsuda, 2002).

Penicillin and other β-lactams covalently modify the active site of transpeptidases (which are therefore called penicillin-binding proteins, or PBPs), by mimicking the d-alanyl-d-alanine terminus of the pentapeptide side chain (Strominger & Tipper, 1965). Resistance can be achieved by one of three known mechanisms (Poole, 2004; Wilke, 2005): (1) biosynthesis of β-lactamases that inactivate the antibiotic (Ghuysen, 1991), (2) expression of mutated pbp alleles, so-called mosaic genes, encoding low-affinity PBPs that maintain their physiological function, but show a decreased β-lactam binding (Dowson, 1994; Hakenbeck, 1999), or (3) removal of the antibiotic from its site of action by β-lactam-specific efflux pumps (Poole, 2005).

The majority of classical antibiotics are produced by microorganisms of the soil biosphere (Berdy, 2005), such as bacilli (Stein, 2005; Butcher & Helmann, 2006), fungi (Anke, 1997), and most notably the actinomycetes (McNeil & Brown, 1994; Champness, 2000), presumably to inhibit the growth of competitors. Stress responses and the development of counter strategies, including efficient resistance mechanisms, are therefore widespread survival strategies amongst soil bacteria to succeed in this habitat (D'Costa, 2006). Likewise, pathogenic bacteria encounter antimicrobials as part of the host's defense system, which also leads to the evolution of adequate stress responses and resistance mechanisms for survival. Only very recently, evolutionarily speaking, were pathogenic bacteria suddenly also challenged with antibiotics from the soil biosphere, in the form of novel ‘magic bullets’ for clinical antimicrobial therapy that were initially thought to eradicate the problem of life-threatening bacterial infections once and for all. As it is known now, antibiotic-resistant bacteria emerged faster than novel antimicrobials can be developed and approved for clinical use (Vicente, 2006). The evolution of novel traits of antibiotic resistance mechanisms in human pathogens and their commensal microbial brethren, through a combination of spontaneous beneficial mutations and horizontal gene transfer, happened at breathtaking speed, and sometimes included the transfer of whole functional modules consisting of sensitive antibiotic detection systems together with their respective target genes mediating resistance. This is most notably illustrated by the spread of vancomycin resistance among clinical isolates of enterococci and staphylococci (Palumbi, 2001; Walsh & Howe, 2002). Such traits, in the context of cell wall antibiotics, are one central aspect of bacterial cell envelope stress response.

Cell envelope stress response (CESR)

The physiological role of the cell envelope in combination with the presence of agents and/or conditions that can alter or even destroy this essential cellular structure necessitate that its integrity is closely monitored. The corresponding signal-transducing regulatory systems respond to alterations and dysfunctions of the envelope and induce appropriate counter-measures to repair damage and secure functionality.

Before giving a comprehensive overview on this stress response and the regulatory systems mediating it, one first needs to define the term. In contrast to ‘intuitive’ stresses such as heat or osmotic shock, it is not easy to put a finger on cell envelope stress. Obviously, many stress conditions, including those mentioned above, affect the integrity of the cell envelope one way or another, without being referred to as cell envelope stress. Still, in the broadest sense, regulatory systems mediating CESR respond to envelope perturbating conditions, i.e. detect insults to the cell envelope, which – in Gram-positive bacteria – consists of the cytoplasmic membrane and the murein sacculus, together with the lipids, TAs, and proteins embedded in it.

A central aspect of the Gram-negative definition – sensing and responding to damaged proteins in the extracytoplasmic compartments, collectively known as the cell envelope – is neither applicable nor helpful, due to the fundamental differences in cell envelope architecture between Gram-negative and -positive bacteria. This definition, which certainly does not cover all aspects of CESR in Gram-negative bacteria, is based primarily on the way that CESR was discovered and subsequently studied in Escherichia coli and other Gram-negative bacteria. Following suit, a seemingly more precise definition for CESR in Gram-positive bacteria could again be based on the approach by which CESRs and most of the regulatory systems involved have been identified and studied.

With very few exceptions, such as the CseABC-σE system of Streptomyces coelicolor, the model signaling systems orchestrating the Gram-positive envelope stress response were initially identified by one of three approaches. (1) They turned out to be responsible for an antibiotic resistance phenotype in (spontaneous) mutants, as exemplified by the CiaRH system of Streptococcus pneumoniae (Guenzi, 1994). (2) They were identified in the course of global gene expression (DNA microarray) studies to characterize the response of an organism when challenge with a cell wall antibiotic. This approach was used to decipher the complex regulatory network orchestrating CESR in the Gram-positive model bacterium B. subtilis (Cao, 2002b; Mascher, 2003a; Pietiäinen, 2005). (3) Their potential role was identified during phenotyping of systematic mutational libraries to elucidate the role of signal transducing systems in a given organism (Hancock & Perego, 2004).

An approach-driven definition could therefore state that ‘The CESR of a Gram-positive bacterium consists of those signal-transducing regulatory systems (and their regulons) that are involved in sensing and responding to the presence of cell wall antibiotics and other envelope perturbating conditions.’ Obviously, such a narrow definition is prone to generating blind spots, and the authors are aware of the potential pitfalls. Moreover, many – if not most – cell envelope stress-responding regulatory systems play an integral role in the overall stress physiology and often normal growth as well, as will be addressed towards the end of this review. This study therefore offers both definitions as suggestions to be evaluated, challenged and subsequently improved by alternative experimental approaches. The analysis of conditional lethal or overexpressing mutants in cell wall biosynthesis genes by transcriptomics, as exemplified by recent work on the cell wall stress stimulon of Staphylococcus aureus (McAleese, 2006; Sobral, 2007), is one very promising example of such an alternative approach that harbors great potential towards that goal.

Regulatory principles orchestrating CESR in Gram-positive and -negative bacteria

While the architecture of the cell envelope – and therefore the definition of the corresponding stress – differs greatly between Gram-positive and -negative bacteria, the regulatory principles orchestrating the corresponding stress responses are remarkably similar. CESR in the Gram-negative model organism Escherichia coli and related bacteria is well investigated and has been reviewed recently (Raivio, 2005; Ruiz & Silhavy, 2005; Rowley, 2006). It is orchestrated by one alternative σ factor, three two-component systems (TCS) and the phage-shock protein response. For in-depth information on these systems, readers are referred to the cited review articles.

The essential extracytoplasmic function (ECF) σ factor σE has been intensively studied in Escherichia coli. It is induced in the presence of misfolded (outer membrane) proteins in the periplasm which can accumulate during growth at elevated temperatures (Ades, 2004; Alba & Gross, 2004; Rowley, 2006). It controls a large regulon including proteins that act directly on misfolded periplasmic proteins or are involved in the synthesis of lipopolysacchrides (Rhodius, 2006).

The CpxAR TCS is activated by elevated external pH, misfolded periplasmic proteins or changes in the lipid composition of the inner membrane (Ruiz & Silhavy, 2005). It is subject to a negative feedback regulation exerted by the periplasmic protein CpxP (Buelow & Raivio, 2005; Fleischer, 2007). Its regulon consists of more than 100 proteins, and partly overlaps with that of σE (De Wulf, 2002).

The RcsBC TCS, together with the unstable auxiliary regulator RcsA, represents a complex phosphorelay system that is involved in regulating capsule biosynthesis, biofilm formation, and the expression of additional periplasmic and membrane proteins (Majdalani & Gottesman, 2005).

Little is known about BaeRS, the third envelope stress-sensing TCS of Escherichia coli. It protects the cell from perturbations of the envelope caused by the presence of indole or misfolded proteins, acting in conjunction with the Cpx system (Raffa & Raivio, 2002). Furthermore, it regulates the expression of a multidrug-efflux pump, thereby conferring resistance to antimicrobial compounds, including β-lactam antibiotics (Nagakubo, 2002; Hirakawa, 2003).

The physiological role of the PspA-mediated phage-shock protein response is less well understood. It is induced by various stress conditions such as filamentous phage infection (hence the name), heat shock, osmotic shock, exposure to organic solvents and proton ionophores as well as long incubation under alkaline conditions. Strains defective in the Psp system show only minor physiological aberrations, for instance poor stationary phase survival, increased motility, slower protein secretion, and some defects in maintaining the membrane potential (i.e. proton motif force) when stressed (Model, 1997; Darwin, 2005).

The same regulatory principles orchestrate the Gram-positive CESR, as outlined schematically in Fig. 2. Because envelope stress occurs outside the cytoplasm, all systems comprise transmembrane sensory components that detect their stimulus in the extracellular space. As for the Gram-negative bacteria, TCS and ECF σ factors are at the core of the Gram-positive envelope stress response. Both systems are functionally analogous in that they consist of two proteins, a membrane-anchored sensory component (sensor histidine kinase (HK) or anti-σ factor, respectively) and a cytoplasmic transcriptional regulator (response regulator (RR) or ECF σ factor, respectively). In both cases, the regulator is usually kept inactive in the absence of inducing conditions. Upon perceiving envelope stress by the sensory component, the regulators become activated and direct (normally up-regulate) the expression of their target genes. The two regulatory principles differ in the mechanism by which the sensor and regulator proteins communicate with one another.

Figure 2

Regulatory principles orchestrating CESR in Gram-positive bacteria. From left to right: ECF σ factors, two-component systems (HK, histidine kinase; RR, response regulator), BlaR1/MecR1 system, BcrR. Sensor proteins are shown in green, inhibitor in blue, transcriptional regulators in red. Arrows indicate activation, T-shaped lines repression. CM, cytoplasmic membrane. See text for details.

In the case of TCS, activation of the RR by its cognate sensor HK is based on the transfer of a phosphoryl group from a donor phospho-histidine in the C-terminal transmitter domain of the HK to an acceptor aspartate in the N-terminal receiver domain of the RR (Parkinson, 1993).

In contrast, communication between anti-σ factor and its corresponding ECF σ factor is based on direct protein–protein interactions. In the absence of stress conditions, the anti-σ factor tightly binds the ECF σ factor, thereby keeping it inactive. Under inducing conditions, the ECF σ factor is released, either by a conformational change or by regulated proteolysis of the anti-σ factor. The σ factor becomes available for recruitment by RNA polymerase core enzyme to redirect transcription initiation to its specific target promoters, ultimately resulting in the upregulation of its regulon (Helmann, 2002).

Two additional, altogether different envelope stress-responsive signal transduction mechanisms have been described in Firmicutes bacteria. The BlaRI/MecRI systems are specific for β-lactam antibiotics. Here, the transmembrane sensory protein harbors an extracytoplasmic penicillin-binding domain and a cytoplasmic zinc metalloprotease activity. Upon antibiotic binding, the protease becomes active and cleaves the corresponding regulator, a repressor protein. This proteolytic inactivation relieves the transcriptional inhibition, resulting in the expression of the target gene, a β-lactamase (Fuda, 2005). The bacitracin-specific BcrR protein is a unique membrane-anchored sensor/transcriptional regulator of Enterococcus faecalis that activates bacitracin resistance determinants in response to extracellular presence of this compound (Manson, 2004).

A distinct phage-shock protein system is absent in Gram-positive bacteria. Instead, its core component, the PspA protein, seems to be embedded in TCS-mediated envelope stress response, at least in the genera Bacillus and Listeria (Jordan, 2006).

The regulatory network orchestrating CESR in B. subtilis: a case study

Over several years, a detailed picture of the CESR emerged for B. subtilis, the model orgaism for the Firmicutes bacteria. Many of the underlying studies addressed the response of this Gram-positive model bacterium to cell wall antibiotics by applying transcriptomics approaches to identify the corresponding stimulons (=all genes that are differentially expressed, usually up-regulated, in the presence of a specific stimulus). Subsequent work identified the regulatory systems that orchestrate this response, thereby dissecting the stimulons into discrete regulons (Fig. 3). Initial work from the Helmann laboratory aimed to identify inducers of the ECF σ factor σW among cell envelope-perturbating agents (Cao, 2002b). Vancomycin was identified as the strongest stimulus and used for subsequent in-depth transcriptional profiling. In addition to σW, three other ECF σ factors were found to be induced by vancomycin, namely σM, σV, and σY, the last two being induced only weakly (Cao, 2002b).

Figure 3

The regulatory network of CESR in Bacillus subtilis. The same symbols and color-code was applied as in Fig. 2. Dotted lines indicate cross-regulation. The antibiotic specificity for each system is shown above. Bac, bacitracin; CAP, CAMPs (note that individual peptides will only induce a certain subset of the regulators indicated); Cep, cephalosporin C; Fos, fosfomycin; Nis, nisin; Van, vancomycin. It should be noted that because of the known regulatory overlap between ECF σ factors, an clear assignment of the inducer spectrum, is not as unambiguously possible as indicated in this figure. In contrast, the ‘crosstalk’ observed for BceRS-like TCS is most likely an experimental artifact without relevance in vivo. See text for details.

Another study from the same group analyzed the bacitracin stress response, which is orchestrated by an even larger number of signal transducing systems (Mascher, 2003a). In addition to σM and the σB-mediated general stress response (Price, 2002), three TCSs respond to the extracellular presence of bacitracin. The LiaRS TCS is strongly induced by both vancomycin and bacitracin. The paralogous TCSs BceRS and YvcPQ both specifically regulate the expression of an ABC transporter, which – in case of the Bce system – confers high level bacitracin resistance (Mascher, 2003a; Ohki, 2003a). A second bacitracin resistance determinant, the undecaprenyl pyrophosphate phosphatase BcrC (Bernard, 2005) under the dual control of two ECF σ factors, σM, and σX (Cao & Helmann, 2002; Ohki, 2003b), was also induced by bacitracin.

More recently, the CESR of B. subtilis was exploited further with regard to β-lactams, d-cycloserine, fosfomycin, and CAMPs (Hutter, 2004; Pietiäinen, 2005). The transcriptional profiles of the first three compounds were part of a broader panel aimed to apply stimulon patterns for predicting the mechanism of action of unknown compounds, and not analyzed in detail (Hutter, 2004). In contrast, the work on CAMPs provided further insights and helped to complete the picture of the regulatory network orchestrating CESR in B. subtilis. Challenges with two naturally occurring cationic peptides, human LL-37 and porcine PG-1, and a synthetic analog, poly-l-lysine, provoked distinct response patterns, orchestrated by three ECF σ factors, σM, σW, and σX, the LiaRS TCS and another BceRS homolog, the YxdJK TCS (Pietiäinen, 2005). The latter specifically responded to LL-37 only, without conferring resistance against this compound. This study also indicated a significant and surprising amount of cross-dependency between TCS- and ECF-mediated responses, even though no direct regulatory overlap has ever been observed. Deletion of ECF σ factors strongly reduced the TCS-mediated response to CAMPs. This observation is reminiscent of the results obtained earlier for the alkaline shock response, where a similar link between σW and the LiaRS-dependent gene expression was already observed (Wiegert, 2001). The reason for this interference in signal transduction is unclear at the moment.

The transcriptomics approaches described above identified four TCS that respond to some aspect of cell envelope stress in B. subtilis, with three being induced by bacitracin alone. During the in-depth analysis of the regulatory network orchestrating the bacitracin response, it was noticed that all TCS involved share some overall similarities with regard to the domain architecture of their HKs. These membrane-anchored sensor kinases are characterized by a very short N-terminal input domain, consisting solely of two deduced transmembrane helices with hardly any periplasmic linker (<10 amino acids for most) in between (Mascher, 2003a). Comparative genomics analysis revealed that these so-called intramembrane-sensing HKs are widespread and conserved in Firmicutes bacteria, but can also be found in the Actinobacteria (Hutchings, 2004; Mascher, 2006). Two conserved subgroups can be distinguished in Firmicutes bacteria, and both groups are involved in mediating CESR in B. subtilis and other Firmicutes bacteria (Mascher, 2006): (1) LiaRS-like three-component systems, and (2) BceRS-like TCS that are functionally linked to ABC transporters.

Signal-transducing systems orchestrating CESR in Gram-positive bacteria

The following paragraphs provide an overview of current knowledge on signal-transducing systems orchestrating CESR in Gram-positive bacteria from both the low G+C (Firmicutes) and the high G+C (Actinobacteria) branch. These systems will be grouped and described according to their mechanism of signal transduction, as outlined in Fig. 2. Based on their overall role in orchestrating overall envelope integrity, three major groups can be loosely distinguished.

Cell wall antibiotic-specific detoxification modules

This group includes (in the order of appearance) the BceRS-like TCS (with the exception of ApsXRS and VirRS), VanRS, BlaR1/MecR1, and BcrR. All of these systems respond to only one specific compound or a group of closely related compounds, as is the case for the glycopeptide-specificity of some VanRS systems. They regulate only one target locus that is often driven by a single promoter. This target locus is located next to the regulatory unit and usually confers high-level resistance to its inducers. Many of these systems are characterized by a high degree of genetic mobility and they are often encoded on plasmids or transposable elements. These detoxification modules are self-sufficient regulatory units that are normally not interlinked with any other aspect of cellular physiology. They are switched off under normal growth conditions and are only induced in the presence of their inducers.

Cell wall antibiotic-responsive systems involved in maintaining envelope integrity and counteracting envelope damages

This group includes the ECF σ factors, LiaRS-like TCS, ApsXRS/VirRS, CroRS/CesRK-, LisRK-, and CiaRH-like TCS. These systems are normally induced by a variety of envelope-damaging agents. Very often – but not necessarily, e.g. B. subtilis LiaRS – they mediate antibiotic resistance, usually against more than one drug. While these systems are highly conserved, their target genes often are not, and their regulons show a remarkable diversity. This is best exemplified by the LiaSR-VraSR-CesSR TCS, which – while being highly homologous to one another, based on sequence conservation and inducer spectrum – bear little or no similarity within their regulons (Kuroda, 2003; Jordan, 2006; Martinez, 2007). The function of most of these systems seems to be the maintenance of envelope integrity under stress conditions that affect its functionality, rather than mediating a specific antibiotic resistance. Therefore, they play a homeostatic role, in contrast to the ON–OFF switches described above. This modulating aspect is even more pronounced within the last group.

Regulators of cell envelope integrity that do not respond to cell wall antibiotics

The three systems belonging to this group – YycFG-like, LytSR-like and MtrAB-like TCS – are only indirectly linked to cell envelope stress. They were included in this review because of their fundamental role for orchestrating envelope integrity. These systems do not respond to damages caused by the presence of cell wall antibiotics. Instead, two of them are essential for survival under normal growth conditions. YycFG has even been described as a regulatory antithesis to a cell envelope stress-responsive system. These systems seem to be switched on under normal growth conditions, reflecting a healthy state of cell envelope metabolism and integrity. Perturbations appear to lead to a shut-off of cell wall metabolism or cell division (YycFG), or more drastically to a bacterial form of programmed cell death in case of LytSR.

Cell envelope stress-sensing ECF σ factors

The ECF (Group 4) σ factors belong to the σ70 family of bacterial σ factors (Lonetto, 1994; Helmann, 2002). They share a number of common features: (1) They are small proteins, harboring only two of the four conserved regions of the primary σ factors, namely region 2 and region 4; (2) They recognize promoters with a highly conserved ‘AAC’ motif in the −35 region; (3) They are usually cotranscribed with their cognate anti-σ factor, a transmembrane protein; (4) They are often involved in regulating functions associated with some aspect of the cell envelope or transport processes (Helmann, 2002; Butcher, 2008).

ECF σ factors involved in orchestrating CESR of B. subtilis

The genome of B. subtilis encodes seven ECF σ factors (Helmann & Moran, 2002), at least three of which, σM, σW, and σX, play a role in orchestrating CESR.

σ W is the best-understood ECF σ factor in B. subtilis (Helmann, 2006). It is induced by a number of cell wall antibiotics, such as vancomycin, cephalosporin C, and the CAMPs LL-37 and PG-1, but also by alkaline shock (Wiegert, 2001; Cao, 2002a; Pietiäinen, 2005). Promoter consensus searches, in combination with in vivo and in vitro approaches identified ∼30 target promoters, controlling about 60 genes, that are at least partially expressed in a σW-dependent manner (Huang, 1999; Cao, 2002a). It was postulated that σW controls an ‘antibiosis’ regulon, based on the inducer spectrum and the putative function of many of its target genes (Helmann, 2002). This hypothesis was subsequently confirmed by studies demonstrating that σW-controlled genes confirm resistance against fosfomycin (Cao, 2001), as well as a number of antimicrobial compounds synthesized by closely related Bacillus species (Butcher & Helmann, 2006). Moreover, σW also provides intrinsic immunity to the antimicrobial protein SdpC (Butcher & Helmann, 2006), which is produced by sporulating B. subtilis cells to lyse nonsporulating sibling cells in a process termed cannibalism (Gonzalez-Pastor, 2003; Ellermeier, 2006).

The sigW gene is cotranscribed with rsiW, encoding the cognate membrane-anchored anti-σ factor. A direct protein–protein interaction between σW and RsiW was verified by yeast two-hybrid analysis (Yoshimura, 2004). Similar to Escherichia coliσE, σW activation is based on the regulated proteolytic degradation of its cognate anti-σ factor RsiW. Three consecutive proteolytic steps cleave first the extracytoplasmic, then the membrane-spanning, and finally the cytoplasmic domains of RsiW (Fig. 4). This proteolytic cascade is initiated by PrsW-dependent site-1 cleavage (Heinrich & Wiegert, 2006; Ellermeier & Losick, 2006). PrsW, a membrane-anchored novel protease with five transmembrane helices, is therefore the prime candidate for the real sensory module in σW-dependent signaling. Subsequently, site-2 cleavage by RasP-mediated regulated intramembrane proteolysis, homologous to Escherichia coli RseP, generates a soluble N-terminal fragment of RsiW (Schöbel, 2004). The latter is degraded by the cytoplasmic ClpXP proteolytic complex (Zellmeier, 2006), thereby ultimately releasing the active σ factor, again similar to the mode of σE-activation in Escherichia coli (Ades, 2004).

Figure 4

Signal transduction mediated by ECF σ factors. The proteolytic cascade leading to anti-σ factor degradation and ECF σ factor activation is illustrated. For color-code, see legend to Fig. 1. The stimulus is represented by the red arrow. Additional proteins are shown in yellow. See text for further details. This figure is based on a presentation kindly provided by Thomas Wiegert.

σ X was the first ECF σ factor to be studied in detail in B. subtilis. Its gene is cotranscribed with rsiX, encoding the corresponding anti-σ factor, and primarily expressed in the logarithmic and early stationary phase (Huang, 1997). Interaction between the σ factor: anti-σ pair could be demonstrated by the yeast two-hybrid system (Yoshimura, 2004). Its regulon consists of ∼30 genes, organized in 15 transcriptional units (Cao & Helmann, 2004). To date, the major physiological role of σX is the modulation of the overall envelope net charge, due to the regulation of the dltABCDE and the pssA-ybfM-psd operons. The products of the dlt operon introduce positively charged amino groups into TAs, thereby lowering the overall negative net charge of the cell wall (Perego, 1995). A comparable role is exhibited by PssA/Psd, which together catalyze the synthesis of the neutral lipid phosphatidylethanolamine. Because the cytoplasmic membrane has a negative net charge due to the abundance of anionic phospholipids, increased incorporation of neutral lipids therefore also lowers the overall negative net charge (Cao & Helmann, 2004). Altering the overall net charge of the cell envelope has been shown to affect both autolysis and resistance to CAMPs. Consequently, a sigX mutant has an increased rate of autolysis and is more sensitive to CAMPs (Cao & Helmann, 2004).

σ M is also cotranscribed with its negative regulators, encoded by yhdL and yhdK that together form the anti-σ complex (Horsburgh & Moir, 1999): a direct protein–protein interaction between σM and the N-terminal fragment of YhdL could be demonstrated, as well as specific interactions between the two membrane proteins YhdL and YhdK (Yoshimura, 2004). The sigM-yhdLK operon is maximally expressed in early to mid-logarithmic growth phase (Horsburgh & Moir, 1999). It is induced by cell wall antibiotics, such as bacitracin, vancomycin, or fosfomycin, and also by acidic pH, heat, ethanol, and superoxide stress (Thackray & Moir, 2003), and confers resistance to bacitracin and high salinity (Horsburgh & Moir, 1999; Cao & Helmann, 2002; Mascher, 2003a). In the B. subtilis strain W23, σM is induced by phosphate depletion and involved in TA biosynthesis (Minnig, 2005). A study of 12 σM target promoters in the reference strain W168 has been published recently (Jervis, 2007), but preliminary in vitro data identified many more target genes (John Helmann, pers. commun.).

A significant amount of regulatory overlap between the target genes of all three ECF σ factors has been demonstrated both in vivo and in vitro, with many promoters being recognized by two, or even all three ECF σ factors (Huang, 1998; Qiu & Helmann, 2001; Cao & Helmann, 2002). It can be envisioned that target gene discrimination (from the available pool of genes preceded by an ECF-type promoter) is therefore a combined result of the timing of expression of individual ECF σ factors during the life cycle and the presence of specific inducing conditions, rather than promoter selectivity based on sequence preference alone. This overlap not only is restricted to the expression of individual target genes, but also manifests itself in the regulation of complex phenotypes, such as overall cell envelope integrity, pellicle formation, and colony morphology (Mascher, 2007). Moreover, in B. subtilis W23 the concerted action of σX and σM is required for septum formation and cell wall biosynthesis (Minnig, 2003). Both ECF σ factors together regulate the synthesis of wall TAs, which consist of poly(ribitol phosphate) in this strain, in contrast to the poly(glycerol phosphate)-containing TAs of the sequenced reference strain B. subtilis W168 (Lazarevic, 2002).

ECF σ factors involved in orchestrating CESR in other Gram-positive bacteria

Disappointingly little is known about ECF σ factors beyond B. subtilis. A quick glance on the distribution and presence of ECF σ factors in Gram-positive genomes shows that these regulators are abundant in most Actinobacteria, whereas their presence is very heterogeneous in the Firmicutes bacteria (Staron, 2007) (Tables 1 and 2). All Bacillus species are rich in ECF σ factors, while these regulators are almost absent in the lactic acid bacteria and other cocci (Table 1). In addition to Streptomyces coelicolorσE, which represents a special case that will be described in detail later in this review, only one study addressed the role of ECF σ factors in Gram-positive CESR, so far. The genome of Bacillus licheniformis, a close relative of B. subtilis, contains nine ECF σ factors, of which six are direct orthologs to the B. subtilis proteins σM, σV, σW, σX, σY, and σylaC (Wecke, 2006). A homolog of the seventh ECF σ factor from B. subtilis, σZ, is absent in the genome of B. licheniformis. The three novel ECF σ factors were designated σecfG, σecfH, and σecfI. In-depth transcriptional profiling demonstrated that seven of the nine ECF σ factors respond to cell envelope stress in this organism: six were significantly induced by vancomycin (σM, σV, σW, σX, σY, and σecfH), and three by bacitracin (σM, σV, and σY) (Wecke, 2006). The contribution of these regulators to mediating resistance against these antibiotics has not yet been explored.

View this table:
Table 1

Distribution and conservation of (potential) cell envelope stress-sensing regulatory systems in the genomes of Firmicutes bacteria

OrganismSize (Mb)ECFBlaR1Two-component systems with
ΣIntramembrane-Periplasmic-sensing HK
Bacillus anthracis5.23–5.51647–501 (I)12211 (I)1/1/(1)
B. cereus5.43–5.8413–1845–501 (I)53311 (I)1/1/1
B. clausii4.341401 (I)411 (I)1/1/1
B. halodurans4.29451 (I)511 (I)1/–/1
B. licheniformis4.229234–351 (I)41 (I)1/1/1
B. subtilis4.2171331 (I)31 (I)1/1/1
B. thuringiensis5.311650–511 (I)33311 (I)1/1/1
Clostridium acetobutylicum4.13340411–/–/(1)
C. difficile4.3215471–/–/–
C. perfringens2.96–3.26418–242(1)–/–/(1)
C. tetani2.87102821(1)1/–/1
Desulfitobacterium hafiense5.7320–237051–/–/1
Enterococcus faecalis3.362181 (II)11111 (I)1/1/–
Geobacillus kaustophilus3.593271 (I)311 (I)–/–/–
Lactobacillus acidophilus1.9981111 (I)–/–/–
L. brevis2.3410111 (I)–/–/–
L. casei2.92175111 (I)–/–/–
L. delbrueckii bulgaricus1.8616–7111 (I)–/–/–
L. gasseri1.89511 (I)–/–/–
L. johnsonii1.999111 (I)–/–/–
L. plantarum3.3515111 (I)–/–/(1)
L. sakei 23K1.88101 (II)1111 (I)–/–/–
L. salivarius1.836111 (I)–/–/–
Lactococcus lactis2.37–2.617–91 (II)111 (II*)–/–/(1)
Leuconostoc mesenteroides2.08101 (I)–/–/–
Listeria innocua3.091171 (I)1111 (I)–/–/–
L. monocytogenes2.91–2.94115–161 (I)1111 (I)–/–/–
Moorella thermoacetica2.6331611 (I)1/–/–
Oceanobacillus iheyensis3.639202 (I/II)111 (I)1/–/(1)
Pediococcus pentosaceus1.8381111 (I)1/1/–
Staphylococcus aureus2.74–2.9214–171 (II)211 (I)1/1/1
S. epidermidis2.56–2.641161 (II)211 (I)1/1/1
S. haemolyticus2.691161 (II)311 (I)1/1/1
S. saprophyticus2.58111 (II)1(I)1/1/1
Streptococcus agalactiae2.13–2.21120–221 (II)111 (II)1/1/(1)
S. mutans2.03141 (II)111 (II)1/1/–
S. pneumoniae2.04–2.16141 (II)11 (II)–/–/–
S. pyogenes1.84–1.9412–141 (II)11 (II)–/–/–
S. thermophilus1.8–1.868–101 (II)1111 (II)–/–/(1)
Thermoanaerobacter tengcongensis2.696201 (I)1/–/–
  • * Information on genome size and total numbers of TCS is derived from the MiST (Ulrich & Zhulin, 2007) and Genome Atlas (Pedersen, 2000; Hallin & Ussery, 2004; Kiil, 2005b) databases at http://genomics.ornl.gov/mist/ and http://www.cbs.dtu.dk/services/GenomeAtlas/index.php, respectively.

  • Numbers of ECF σ factors are derived from a comprehensive ECF database and classification system, and the σ factor census published in Genome Atlas (Kiil, 2005a; Staron, 2007).

  • Identification of specific TCS in bacterial genomes is based on genomic blast searches (blastp) of either the complete proteins at http://www.ncbi.nlm.nih.gov/sutils/genom_table.cgi or – where applicable – signature domains thereof, such as the extracytoplasmic senor domains of HK (i.e. for CiaRH), to increase the specificity of the search. Moreover, genomic context conservation was also used as a criterion, where applicable (LiaRS, BceRS, VanRS, YycFG, LytSR), using the MicrobesOnline (Alm, 2005) or the SEED databases at http://www.microbesonline.org/ and http://theseed.uchicago.edu/FIG/index.cgi, respectively.

  • § Identification of TCS with intramembrane-sensing histidine kinases is derived from a comparative genomics analysis (Mascher, 2006). Two groups of LiaRS-like TCS can be distinguished, based on their genomic context. Group I, liaIH(G)FSR; Group II, liaFSR.

  • The two groups of YycFG-like TCS differ in HK domain architecture and overall genomic context (Ng & Winkler, 2004): Group I, extended genomic context conservation of three to four genes, including homologs of yycI and yycH. The N-terminus of YycG-like HK of this group consists of two deduced transmembrane helices that flank a large periplasmic domain. Group II lacks homologs of yycI and yycH, and the N-terminal domain of the HK consist of a single transmembrane helix. *The L. lactis YycFG-like TCS has a group II genomic context, but a unique HK domain architecture, consisting of two deduced transmembrane helices, but lacking an extracellular domain in between.

  • Annotation for the presence of the LytSR-type holin-/antiholin systems: LytSR-like TCS/LrgAB-homologs (antiholin)/CidR-CidAB homologs (holin). Presence of CidAB-homologs without a discernable CidR homolog encoded adjacent on the chromosome is indicated by (1). Genomic context analysis was performed using the databases described above ().

View this table:
Table 2

Distribution and conservation of (potential) cell envelope stress-sensing regulatory systems in actinobacterial genomes

OrganismSize (Mb)ECFTwo-component systems (with)
Acidothermus cellulolyticus2.4411241
Arthrobacter FB245.078291
Arthrobacter aurescens5.23221
Bifidobacterium adolescentis2.091141
Bifidobacterium longum2.261911
Corynebacterium diphtheriae2.497111
Corynebacterium efficiens3.155131
Corynebacterium glutamicum3.285131
Corynebacterium jeikeium2.48791
Frankia CcI35.43103211
Leifsonia xyli xyli2.583211
Mycobacterium avium5.4815181
Mycobacterium avium paratuberculosis4.8315181
Mycobacterium bovis4.35–4.3710131
Mycobacterium leprae3.27251
Mycobacterium smegmatis6.992351
Mycobacterium tuberculosis4.40–4.4210131
Mycobacterium ulcerans5.63161
Mycobacterium vanbaalenii6.4918391
Nocardia farcinica6.29173311
Nocardioides JS6145.2918431
Propionibacterium acnes2.563131
Rhodococcus RHA19.7017401
Rubrobacter xylanophilus3.236181
Streptomyces avermitilis9.123672211
Streptomyces coelicolor8.6742877111
Thermobifida fusca3.649281
  • * Information on genome size and total numbers of TCS is derived from the MiST (Ulrich & Zhulin, 2007) and Genome Atlas (Pedersen, 2000; Hallin & Ussery, 2004; Kiil, 2005b) databases at http://genomics.ornl.gov/mist/ and http://www.cbs.dtu.dk/services/GenomeAtlas/index.php, respectively.

  • Numbers of ECF σ factors are derived from a comprehensive ECF database and classification system, and the σ factor census published in Genome Atlas (Kiil, 2005a; Staron, 2007).

  • Identification of specific TCS in bacterial genomes is based on genomic blast searches (blastp) of either the complete proteins at http://www.ncbi.nlm.nih.gov/sutils/genom_table.cgi or – where applicable – signature domains thereof, such as the extracytoplasmic senor domains of HK (i.e for CiaRH), to increase the specificity of the search. Moreover, genomic context conservation was also used as a criterion, where applicable (LiaRS, BceRS, VanRS, YycFG, LytSR), using the MicrobesOnline (Alm, 2005) or The SEED databases at http://www.microbesonline.org/ and http://theseed.uchicago.edu/FIG/index.cgi, respectively.

  • § Identification of TCS with intramembrane-sensing histidine kinases (IM-HK) is derived from a comparative genomics analysis (Mascher, 2006).

LiaFSR-like three-component systems: conserved cell envelope stress-sensors in Firmicutes bacteria

LiaFSR-homologous three-component systems are widespread amongst the Firmicutes bacteria, with the noteworthy exception of the genera Lactobacillus and Clostridium (Table 1). Two groups can be distinguished, based on the genomic context of the corresponding loci (Jordan, 2006; Mascher, 2006) (Fig. 5a). LiaRS-homologs in bacilli (Group I) are embedded in the liaIH-(G)FSR locus, with a liaG-like gene only present in B. subtilis, B. licheniformis, and Bacillus halodurans. Group II only shows a conservation of the liaFSR locus. In Listeria species, liaIH- and liaFSR-like genes are organized as two separate operons, but both still seem to be under the transcriptional control of the LiaRS systems (Jordan, 2006). Therefore, they are also listed as group I (Table 1). This difference in genomic context also seems to indicate a different cellular role for these systems. Based on the available data, TCS embedded in liaIH-(G)FSR-like loci primarily – maybe even exclusively – regulate their own expression, whereas Group II LiaRS-like TCS seem to orchestrate much larger regulons and seem to represent the primary cell envelope stress-responding systems in the corresponding organisms. This hypothesis is supported by data from B. subtilis LiaRS, Staphylococcus aureus VraSR, and Lactococcus lactis CesSR.

Figure 5

Cell envelope stress-sensing two-component systems in Firmicutes bacteria. (a) LiaFSR-like three-component systems. (b) BceRS-BceAB-like TCS–ABC transporter connection. Symbols and color-code as before. See text for details. This figure was taken from (Mascher, 2006), with modifications.

LiaRS of Bacillus subtilis

LiaRS was originally identified as part of the complex regulatory network orchestrating the bacitracin stress response in B. subtilis (Mascher, 2003a). It also strongly responds to the external presence of other cell wall antibiotics that interfere with the lipid II cycle, such as ramoplanin, vancomycin, or CAMPs (Mascher, 2004; Pietiäinen, 2005). LiaRS-dependent gene expression is also induced by alkaline shock, detergents, ethanol, phenol, organic solvents, and secretion stress, albeit to a lesser extent (Petersohn, 2001; Wiegert, 2001; Mascher, 2004; Hyyryläinen, 2005; Pietiäinen, 2005; Tam le, 2006). Moreover, it was recently shown that LiaRS is intrinsically induced by a peptide produced by B. subtilis itself and encoded within the yydFGHIJ operon (Butcher, 2007). Its activity is influenced by the density of the negative net charge of the cell envelope. The LiaRS system responds more strongly to the CAMP LL-37 and secretion stress in cells defective in the Dlt system, which has an overall higher negative net charge due to its inability to incorporate d-alanine into its TAs (Hyyryläinen, 2007).

The LiaRS TCS is functionally and genetically linked to a third protein, LiaF, which acts as a strong inhibitor of LiaR-dependent gene expression (Jordan, 2006) (Fig. 5a). The LiaRS-LiaF three-component system is conserved by sequence, genomic context, and function in Gram-positive bacteria with a low G+C content (Jordan, 2006; Mascher, 2006). The lia locus consists of six genes, liaIH-liaGFSR. A basal expression level of the last four genes, liaGFSR, encoding the LiaFSR three-component system and a putative membrane-anchored hypothetical protein, LiaG, is ensured by a weak constitutive promoter upstream of liaG. In contrast, expression of the liaIH operon from PliaI is completely LiaR-dependent (Jordan, 2006).

In B. subtilis, only two promoters are known to be regulated by the LiaRS TCS: the liaI promoter (PliaI) and the yhcY promoter (Jordan, 2006), with PliaI being the primary target. PliaI is tightly regulated: in the absence of a stimulus, it is virtually switched off, while addition of bacitracin results in an about 200-fold induction (Mascher, 2003a, 2004). In contrast to PliaI, an LiaR-dependent PyhcY-activity was only observed in a liaF mutant, i.e. in the absence of the LiaRS-inhibitor protein (Mascher, 2004; Jordan, 2006).

PliaI is also induced in the absence of exogenous stimuli at the onset of the stationary phase (Jordan, 2007). This time point in the B. subtilis life cycle is characterized by the initiation of a complex regulatory cascade that allows Bacillus to adapt to worsening living conditions, which can ultimately lead to the formation of dormant endospores (Msadek, 1999; Phillips & Strauch, 2002; Errington, 2003). It was demonstrated that PliaI is directly repressed by binding of the transition state regulator AbrB within the promoter sequence, thereby acting as a roadblock to prevent premature PliaI activity during logarithmic growth. AbrB repression is released during the transition state by Spo0A, the master regulator of sporulation and PliaI is induced by an unidentified endogenous stimulus, resulting in the expression of the liaIH operon. While AbrB-binding is sufficient to inhibit the endogenous growth-dependent induction of PliaI, it can be bypassed completely by exogenous induction with cell wall antibiotics. Taken together, LiaRS-dependent gene expression is embedded in transition phase regulation in B. subtilis, and the activity of its primary target promoter, PliaI, is controlled by at least five regulators (Jordan, 2007).

In contrast to the detailed knowledge on the mechanism of LiaFSR-dependent signal transduction, the physiological role of its primary target genes, liaIH, is largely unknown. LiaI is a small hydrophobic protein of unknown function with two putative transmembrane helices. LiaH is a member of the phage-shock protein family. While the strong induction of liaIH, and to a lesser degree also liaGFSR, by cell envelope stress is well documented, mutational analyses of the lia locus so far failed to identify strong phenotypes associated with them, and deletion of lia genes did not alter the sensitivity of the corresponding mutants to the known inducers of the Lia system.

VraSR of Staphylococcus aureus

VraSR is the best-understood LiaRS homolog, so far. It was originally identified as an upregulated locus in a vancomycin-resistant Staphylococcus aureus (VRSA) strain, compared with a sensitive (VSSA) strain (Kuroda, 2000). VraSR was also upregulated in vancomycin intermediate resistant Staphylococcus aureus (VISA) (McAleese, 2006). The vraSR genes are cotranscribed with two upstream genes, termed orf1 and the liaF-homolog yvqF, and a VraR-dependent auto-induction of this four gene operon was demonstrated (Yin, 2006). The VraSR system is strongly induced by a number of cell wall antibiotics, including vancomycin, teicoplanin, β-lactams, bacitracin, and d-cycloserine, but not by general stresses, such as heat, osmotic shock or pH shifts (Kuroda, 2003). It also responds to sublethal perturbations of cell wall biosynthesis caused by the depletion of pbpB, encoding an essential PBP crucial for peptidoglycan crosslinkage, and murF, its gene product catalyzing the last cytoplasmic step in peptidoglycan precursor biosynthesis, addition of the d-alanyl-d-alanine dipeptide to the UDP-linked MurNAc-tripeptide (Gardete, 2006; Sobral, 2007). Surprisingly, preliminary data indicate that Staphylococcus aureus VraSR – in contrast to the LiaRS system of B. subtilis– does not respond to CAMPs such as LL-37 (Pietiäinen, 2007).

Knock-out of vraSR resulted in increased susceptibility towards all of its inducers, with the exception of d-cycloserine, and fosfomycin (Kuroda, 2003; Gardete, 2006). Transcriptional profiling identified 46 genes that were induced by vancomycin in a VraSR-dependent manner, including its own locus and a number of genes encoding functions involved in cell envelope biogenesis, such as murZ (peptidoglycan monomer precursor biosynthesis), pbp2, sgtB (peptidoglycan polymerization), or tagA (TA biosynthesis) (Kuroda, 2003).

CesSR of Lactococcus lactis

CesSR was originally described as LlkinD/LlrD in a systematic analysis of six TCS from Lactococcus lactis strain MG1363. These authors were unable to generate an insertion mutant of llkinD, encoding the HK, but the corresponding RR mutant MGRrD showed an increased salt-/osmosensitivity (O'Connell-Motherway, 2000). Recently, it was shown to respond to the extracellular presence of the lactococcal bacteriocin Lcn972 and renamed CesSR (Martinez, 2007) – an unfortunate choice, since it should not be confused with another cell envelope stress-sensing TCS, CesRK of Listeria monocytogenes (Kallipolitis, 2003), to which it bears no similarity (see further below for details). Transcriptome analysis revealed that the expression of 26 genes was significantly upregulated in the presence of Lcn972, of which 23 responded in a CesSR-dependent manner. Many of these genes encode putative membrane or stress-related proteins. As with all LiaRS-homologs, the corresponding locus of Lactococcus lactis is subject to positive autoregulation, and includes a third liaF-homologous gene, yjbB (llmg1650), which is located directly upstream of cesSR. The CesSR TCS is also induced by other cell wall antibiotics, such as bacitracin, vancomycin, and plantaricin C. CesR disruption results in a slight increase in susceptibility to bacitracin, nisin, and plantaricin C, all of which interfere with the Lipid II cycle (Martinez, 2007).

One of the most strongly induced genes of the CesSR regulon is spxB, formerly yneH, one of seven paralogs of Lactococcus lactis that are homologous to B. subtilis Spx (Nakano, 2003). SpxB expression and subsequent interaction with RNA polymerase leads to upregulation of oatA, ultimately resulting in increased O-acetylation of peptidoglycan, rendering the cells more resistant to autolysis and lysozyme treatment (Veiga, 2007). The authors of this study hypothesize that this novel resistance mechanism is induced upon CesS-dependent sensing of cell envelope stress, such as peptidoglycan hydrolysis caused by the presence of lysozyme.

LiaFSR-homologs in other Firmicutes bacteria

Vancomycin treatment of both a sensitive and a tolerant strain of Streptococcus pneumoniae resulted in an upregulation of the SP0385-0387 locus, encoding the liaFSR-homologous TCS03 (Haas, 2005).

Enhanced nisin resistance in Listeria monocytogenes is associated with increased expression of hpk1021 and pbp2229, encoding a LiaS-homologous HK and a PBP, respectively (Gravesen, 2001). It could be demonstrated that disruption of both genes abolished the nisin resistance phenotype, and that pbp2229 expression depends on HPK1021 (Gravesen, 2004). Moreover, a mutant harboring an in-frame deletion of the corresponding RR, RR1022, showed a slightly increased ability to invade Cos-1 fibroblast cells compared to the isogenic wild type strain (Williams, 2005).

In-depth transcriptional profiling of the CESR in B. licheniformis revealed that the LiaFSR-homologous YvqFEC system is strongly induced by bacitracin and nisin. Weaker induction was observed for vancomycin and d-cycloserine (Wecke, 2006).

Systematic inactivation and subsequent phenotypic characterization of TCS in Enterococcus faecalis revealed that a mutant of the LiaR-homologous protein RR03 shows increased bacitracin sensitivity (Hancock & Perego, 2004), indicative for a role of the corresponding TCS in counteracting cell envelope stress.

Taken together, LiaFSR and its homologs are conserved cell envelope stress-sensing three-component systems in Firmicutes bacteria that play a crucial role in responding to, and counteracting damages caused by the extracellular presence of cell wall antibiotics and other perturbations of cell envelope integrity. While this signaling module is clearly conserved, its output is not. It shows a remarkable variability with regard to regulon size and cellular role, indicating that its regulatory features have been ‘used’ by evolution and adapted to the physiological requirements and life style of the individual organism.

BceRS-like detoxification modules: a TCS–ABC transporter connection conserved in Firmicutes bacteria

As mentioned above, a regulatory relationship between TCS and ABC transporters has been described already some years ago in the Bacillus/Clostridium group, and the authors demonstrated a TCS-dependent expression of the ABC transporter genes for B. subtilis (Joseph, 2002). A recent analysis of Gram-positive genomes demonstrates the predominance of such TCS–ABC transporter modules in these two genera (Table 1). Moreover, this group completely overlaps with a conserved subgroup of TCS harboring intramembrane-sensing HKs (Mascher, 2003a; Mascher, 2006). Work from B. subtilis established that in these detoxification modules the TCS respond to the extracellular presence of antimicrobial compounds. Upon activation, they upregulate the expression of ABC transporters encoded by neighboring, usually downstream, genes (Joseph, 2002; Ohki, 2003a; Joseph, 2004) (Fig. 5b). The ABC transporter then facilitates removal of the harmful drug, thereby removing the initial stimulus of the system, which finally shuts down again (Mascher, 2006). In contrast to the Lia-like systems, the TCS of the detoxification modules are not autoregulated.

BceRS and its paralogs in Bacillus subtilis

The genome of B. subtilis harbors three TCS–ABC transporter modules, encoded by bceRS-bceAB (formerly ytsABCD), yvcPQ-yvcRS, and yxdJK-yxdLM (Joseph, 2002; Mascher, 2006).

The BceRS–BceAB system is the best understood of these detoxification modules (Fig. 5b). The BceRS TCS specifically responds to the extracellular presence of bacitracin. Its activation leads to binding of phosphorylated BceR to an inverted repeat upstream of the bceA promoter, resulting in a strong up-regulation of bceAB expression. The encoded ABC transporter is an efficient bacitracin resistance determinant and is thought to facilitate its removal (Mascher, 2003a; Ohki, 2003a). Recently, it was demonstrated that the HK BceS alone is unable to sense bacitracin. Instead, the corresponding ABC transporter BceAB is crucial for bacitracin perception (Bernard, 2007). Moreover, ATP binding/hydrolysis by the nucleotide-binding subunit BceA is a prerequisite for stimulus perception, indicating that BceRS responds to bacitracin transport by BceAB, rather than the extracellular presence of this antibiotic (Rietkötter and Mascher, unpublished observation).

Only very little is known about the other two systems. The YxdJK–YxdLM system has been analyzed genetically, without gaining insights into its biological role (Joseph, 2004). Again, the expression of yxdLM, encoding the ABC transporter, is strictly dependent on the corresponding RR YxdJ. More recently, the human CAMP LL-37 was identified as a strong inducer of yxdLM expression (Pietiäinen, 2005). Therefore, it is tempting to speculate that YxdJK–YxdLM functions as a CAMP-specific detoxification module. But the exact nature of its substrate remains to be identified.

Induction of yvcRS expression was shown to be induced by bacitracin and nisin (Mascher, 2003a; Hansen, 2007). While the bacitracin induction is weak and presumably indirect (Rietkötter and Mascher, unpublished observation), the nisin-dependent upregulation could point towards a role of this module in mediating resistance against some lantibiotic. But again, this assumption needs to be verified.

BceRS-like TCS in other Gram-positive bacteria

Relatively little is known about these systems beyond B. subtilis, despite their overall abundance. Two bacitracin resistance modules have been described. The BacRS-BcrABC module confers bacitracin self-resistance in the producing strain B. licheniformis ATCC10716 (Podlesek, 1995; Neumüller, 2001). In Streptococcus mutans, a similar module is encoded by the mbrABCD locus (Tsuda, 2002). Here, the genes encoding the TCS, mbrCD, are located downstream of the genes for the ABC transporter MbrAB (Fig. 5b). The genome of B. licheniformis DSM13 encodes four TCS–ABC transporter modules, two of which –ytsABCD and yxdJ-Bli04268-70 – are induced by bacitracin. The second system also responds to nisin (Wecke, 2006). Recently, the GraRS TCS of Staphylococcus aureus was described to regulate expression of the VraFG ABC transporter, encoded by genes located directly downstream of graSR. This module is involved in resistance to vancomycin and – more pronounced – CAMPs, such as polymyxin B, gallidermin, or the human lysozyme-derived peptide LP9 (Herbert, 2007; Meehl, 2007). A graRS mutant was also shown to be more resistant to the muraminidase activity of lysozyme and biofilm-negative (Herbert, 2007). The genes of another related system, GtcRS of B. brevis, are located next to the grs operon encoding the multienzymes involved in the biosynthesis of the peptide antibiotic gramicidin (Turgay & Marahiel, 1995). While no functional characterization has been carried out, a role of this TCS in gramicidin autoimmunity seems likely.

Two unusual examples of BceRS-like TCS were described in Staphylococcus epidermidis and Listeria monocytogenes. These TCS, while sharing the overall sequence similarities and genomic context conservation (Mascher, 2006), differ from most TCS of this group by regulating more than one operon encoding an ABC transporter. ApsXRS is a three-component antimicrobial peptide-sensing system in Staphylococcus epidermidis that is conserved in other staphylococci (Li, 2007). The TCS ApsRS and the third protein, ApsX, are crucial for CAMP sensing although the exact function of ApsX is still unknown. This three-component system responds to a wide range of structurally unrelated CAMPs, including the α-helical LL-37, the bridged β-defensin 3, the His-rich histatin and the bacterial lantibiotic nisin. Preliminary results indicate that the HK ApsS might sense these CAMPs by direct binding to the nine amino acid short extracellular loop, which has a high density of negatively charged amino acids (Li, 2007). But the role of the neighboring ABC transporter, encoded by vraFG, in stimulus perception has not been addressed so far.

A homologous system, VirRS, was described as a TCS critical for the virulence of Listeria monocytogenes (Mandin, 2005). It lacks an ApsX homolog and its genomic context differs from the aps system. But the regulons of both systems are almost identical in size and functions. Both regulate the expression of the dlt operon, the mprF gene, and an operon encoding an ABC transporter. The first two loci are involved in lowering the net surface charge of the cell envelope, by incorporating d-alanine into TAs, and lysine into membrane lipids, respectively. Consequently, both systems play an important role for CAMP resistance (Thedieck, 2006; Li, 2007).

Taken together, the available data so far identified two subgroups of BceRS-homologous TCS in Firmicutes bacteria. The first group – exemplified by BceRS, MbrAB, YtsAB, and GtcRS – is involved in mediating drug-specific resistance against peptide antibiotics, such as bacitracin, by regulating the expression of a neighboring ABC transporter that facilitates removal. The second group includes ApsXRS and VirRS, and eventually also GraRS, which represent important CAMP-specific detoxification systems. These systems – in addition to inducing the expression of an ABC transporter – lower the overall negative net charge of the cell envelope in response to the extracellular presence of a variety of structurally unrelated CAMPs.

Additional TCS involved in orchestrating CESR in Firmicutes bacteria

Three additional TCS have been implicated in orchestrating CESR in Firmicutes bacteria. In contrast to the TCS mentioned so far, these systems are characterized by periplasmic-sensing HKs, based on the presence of a significantly large extracellular input domain, flanked by two deduced transmembrane helices (Mascher, 2006b). Interestingly, all of these systems are associated with β-lactam resistance, while none of the cell envelope stress-responsive TCS with intramembrane-sensing HKs – with the noteworthy exception of S. aureus– shows any link to this class of antibiotics. But such a hypothesized connection between sensor domain architecture and antibiotic action – while fascinating – remains purely speculative. To date, the role of the input domains for stimulus perception has not been investigated for any of the TCS described below.

CroRS/CesRK-like TCS

These systems have been investigated to some extent in Enterococcus faecalis (CroRS) and Listeria monocytogenes (CesRK), but close homologs are also present in the genome sequences of some bacilli and clostridia (Table 1). The TCS CroRS was found to play a crucial role in the intrinsic β-lactam resistance of Enterococcus faecalis. Its deletion resulted in a 4000-fold increase in susceptibility to ceftriaxone, a third-generation cephalosporin (Comenge, 2003). A croR mutant was also significantly more sensitive to bacitracin, other second- and third-generation cephalosporins, such as cefuroxime and cefotaxime, and vancomycin (Hancock & Perego, 2004), indicating a more general role of this TCS in CESR. Additionally, a croR mutation led to significant growth defects and alterations in cell morphology (Le Breton, 2003). Expression of the croRS operon was induced by an even wider variety of cell wall antibiotics, including fosfomycin, d-cycloserine, moenomycin and ramoplanin, in addition to those mentioned above (Comenge, 2003). Recently, it was also shown to be induced by bovine bile and sodium dodecyl sulfate (Solheim, 2007). The physiological role of CroRS – including the mechanism by which it confers antibiotic resistance – is still unclear, despite the recent identification of two target loci, encoding a secreted protein, SalB, and the putative glutamine/glutamate ABC transporter GlnQHMP (Muller, 2006; Le Breton, 2007).

The CesRK system is one of two-cell envelope stress-responsive TCS described in Listeria monocytogenes (the second one, LisRK, is addressed below). It was originally shown to play a role in virulence and ethanol tolerance (Kallipolitis & Ingmer, 2001). Subsequently, the same authors identified a CesR-target gene, orf2420, which is located directly downstream of the cesRK operon. They could show that its expression is induced in a CesR-dependent manner by ethanol, lysozyme, β-lactams, vancomycin, and – to a weaker degree – also bacitracin and d-cycloserine (Kallipolitis, 2003). This broad inducer spectrum is therefore quite similar to the one described above for the CroRS system. Moreover, deletion mutants in either cesR or cesK showed an increased susceptibility to all β-lactam antibiotics tested, including penicillin and first-, second-, or third-generation cephalosporins (Kallipolitis, 2003).

LisRK-like TCS

Homologous systems are present in many, but not all Firmicutes genomes (Table 1). A link between these systems and CESR has so far only been reported for LisRK of Listeria monocytogenes. It was originally described in the context of its role in virulence and stress tolerance: A lisK mutant was 10-fold less virulent than its parental wild type strain, and displayed a growth-phase dependent response to acidic conditions (Cotter, 1999). Moreover, a lisR mutant was more sensitive to the presence of ethanol and oxidative stress (Kallipolitis & Ingmer, 2001), and LisRK has also been shown to be important for osmotolerance (Sleator & Hill, 2005).

A lisK mutant was also much more susceptible to numerous cephalosporins, but no other antibiotics, and, surprisingly, more resistant against the lantibiotic nisin (Cotter, 2002). The authors of this study demonstrated that three genes previously identified as being upregulated in spontaneous nisin resistant mutants of Listeria monocytogenes (Gravesen, 2001) were expressed in a LisRK-dependent manner: absence of LisK resulted in the complete loss of expression of all three genes (Cotter, 2002). Intriguingly, one of these genes, lmo1022, encodes the LiaS-homologous protein HPK1022 of Listeria monocytogenes, indicative for a regulatory cascade orchestrating CESR in this organism. Moreover, subsequent work by Gravesen (2004) demonstrated that the second LisK-dependent gene, pbp2229, is expressed in an RR1021-dependent manner. Therefore, the observed LisRK-dependent expression of pbp2229– and most likely also the observed effects of a lisK mutant on nisin and cephalosporin resistance – could be an indirect effect mediated by the LisRK-dependent expression of the LiaRS homologs of Listeria monocytogenes, encoded by lmo1021–1022.

In contrast, the contribution of LisRK to both osmotolerance and virulence can be attributed to its recently described control of htrA expression (Stack, 2005). HtrA is a highly conserved extracellular serine protease important for the degradation of misfolded proteins that accumulate under stress conditions and can interfere with normal cell functioning (Pallen & Wren, 1997; Kim & Kim, 2005). A crucial role of HtrA for virulence, heat resistance, and salt-/osmotolerance has already been described for Listeria monocytogenes (Wonderling, 2004; Stack, 2005). Disruption of other LisRK-homologous TCS, i.e. Staphylococcus aureus ArlRS, Streptococcus pyogenes CovRS/CsrRS, Lactobacillus acidophilus LBA1524/1525, resulted in remarkably similar patterns of affected phenotypes, including general stress sensitivity, increased autolysis, decreased proteolytic activity and reduced virulence in the pathogens, but no further link to cell envelope stress in particular (Fournier & Hooper, 2000; Fournier, 2001; Dalton & Scott, 2004; Azcarate-Peril, 2005; Gryllos, 2007). Rather, LisRK-homologous TCS seem to represent general stress response systems. Future research will hopefully reveal, whether the described regulatory link between LisRK and the Lia-like TCS encoded by lmo1021/1022 is a singular coincidence of Listeria monocytogenes and the only link between LisRK-like TCS and CESR.

CiaRH-like TCS

These TCS are closely related to LisRK (Cotter, 2002) and are restricted to the genus Streptococcus (Table 1). The CiaRH TCS of Streptococcus pneumoniae was first identified in spontaneous mutants resistant to the third-generation cephalosporin cefotaxime. A point mutation in the gene encoding the corresponding HK, ciaH, was found to increase cefotaxime resistance and abolish competence for genetic transformation (Guenzi, 1994). This mutation was thought to switch the HK in a constitutive kinase ‘ON’ mode, thereby activating the corresponding RR LiaR, resulting in increased expression of its target genes (Giammarinaro, 1999; Mascher, 2003b). Interestingly, a recent in-depth analysis of CiaR-dependent gene expression indicates that the CiaRH system is already highly active under normal growth conditions in the wild type, and that the ‘ON’ T230P mutation, close to the invariant histidine residue, only results in a moderate further increase in CiaR activity (Halfmann, 2007), a behavior reminiscent of the YycFG system described below.

Expression of the ciaRH operon is weakly induced by prolonged incubation with sublethal amounts of both vancomycin and penicillin (Haas, 2005; Rogers, 2007). CiaRH plays an important role in mediating resistance to lytic processes, such as cell wall antibiotic-induced lysis or competence-induced stationary phase autolysis (Dagkessamanskaia, 2004; Mascher, 2006a). Moreover, CiaRH-dependent regulation is also linked to the virulence of Streptococcus pneumoniae (Throup, 2000; Sebert, 2002; Ibrahim, 2004).

Target genes controlled by CiaR have been identified in three independent studies (Sebert, 2002; Mascher, 2003b; Dagkessamanskaia, 2004). All three studies demonstrated a down-regulation of the competence regulon under conditions, when the Cia system is highly active. The link between CiaRH and regulation of genetic competence is well-documented (Guenzi, 1994; Giammarinaro, 1999; Echenique, 2000; Martin, 2000; Mascher, 2003b; Dagkessamanskaia, 2004; Sebert, 2005), but still poorly understood. One CiaR-target gene, htrA, encoding an extracellular serine protease, seems to play a central role in this connection (Sebert, 2005). HtrA is also crucial for the CiaRH-dependent virulence (Sebert, 2002), reminiscent of the listerial LisRK system described above, where regulation of HtrA expression also significantly contributes to the phenotypes associated with this TCS. Moreover, the Cia system seems to be involved in TA metabolism and could be important for controlling the overall negative net charge of the cell envelope by modulating the expression of the dlt operon and parts of the lic locus (Mascher, 2003b). These two loci are important for incorporating d-alanine and choline, respectively, into pneumococcal TAs (Zhang, 1999; Fischer, 2000; Kharat & Tomasz, 2006; Kovács, 2006). A recent in-depth reinvestigation of CiaR-dependent gene expression identified five small noncoding RNAs as novel members of the CiaRH regulon that seem to be at least partially responsible for the CiaRH-mediated resistance to cellular lysis (Halfmann, 2007), but their exact physiological role remains to be identified.

The CiaRH system of Streptococcus mutans has also been investigated to some extent. Inactivation of ciaH resulted in the diminished production of the lantibiotic mutacin I. This mutation also negatively affects competence development, acid stress tolerance and biofilm formation (Qi, 2004). Interestingly, these phenotypes were not observed in a ciaR mutant, indicative of a yet-to-be discovered CiaH-dependent, but CiaR-independent, regulatory pathway in this organism (Ahn, 2006). This study indicated significant differences in the CiaRH-dependent regulation between Streptococcus pneumoniae and Streptococcus mutans, despite the overall similarities with regard to the phenotypes involved.

VanRS-mediated vancomycin resistance: a bridge between high-G+C producers and low G+C pathogens

The VanRS TCS is the only cell envelope stress system found in both the low- and high-GC Gram-positive bacteria (Tables 1 and 2). VanRS regulates expression of inducible vancomycin resistance in response to vancomycin and other closely related glycopeptide antibiotics (Hong, 2008). As mentioned in the beginning, these antibiotics kill Gram-positive bacteria by binding to the d-Ala–d-Ala termini on the stem peptides of cell wall precursors and preventing cross-linking of the mature cell wall (Barna & Williams, 1984). When the mechanism of action of vancomycin was first elucidated, it was suggested that pathogenic bacteria would never become resistant because the precursors themselves would have to be altered. In fact, the highly inducible vancomycin resistance genes encode enzymes that do exactly that. They reprogram the cell wall precursors such that the stem peptides terminate with d-Ala–d-Lac, a substrate for which vancomycin has 1000-fold lower affinity (Bugg, 1991).

Vancomycin is produced by the actinomycete Amycolotopsis, formerly Streptomyces, orientalis (McCormick, 1955). Other actinomycetes produce similar glycopeptide antibiotics, including Streptomyces toyocaensis and Actinoplanes teichomyceticus, which synthesize the sugarless glycopeptide A47934 and the lipidated glycopeptide teicoplanin, respectively (Pootoolal, 2002; Sosio, 2004). Vancomycin was approved for clinical use in 1956 and was effectively used for over 30 years (Palumbi, 2001) before the first vancomycin resistant enterococci were isolated in the early 1980s (Bugg, 1991). Two major types of inducible vancomycin resistance have been identified in the enterococci, named VanA and VanB (Arthur & Quintiliani, 2001). The VanA resistance genes are carried on transposon Tn1546 and confer resistance to vancomycin and to the lipidated glycopeptide teicoplanin. They are easily transferable between strains as evidenced by the recent acquisition of VanA resistance by MRSA from a coisolate of Enterococcus faecalis (Weigel, 2003). VanB strains are resistant to vancomycin but sensitive to teicoplanin (Courvalin, 2005). These are the genes found in the glycopeptide-producing actinomycetes, which must carry resistance to avoid autotoxicity. These are now widely accepted to be the source of all inducible glycopeptide resistance in pathogenic bacteria (Davies, 1994).

Inducible vancomycin resistance has been well studied in the model actinomycete Streptomyces coelicolor (Hong, 2004, 2005; Hutchings, 2006b). Although Streptomyces coelicolor does not produce any glycopeptide antibiotics, it is a close relative of the producing organisms and, presumably, the acquisition of resistance genes has given Streptomyces coelicolor a selective advantage in its soil environment. Inducible vancomycin resistance in Streptomyces coelicolor is encoded by seven genes organized in four transcriptional units, all of which are expressed VanRS-dependently (Fig. 6a). The transmembrane histidine sensor kinase VanS senses vancomycin, probably by binding it directly, and transmits the signal via phosphorylation to its RR VanR (Hutchings, 2006b). The activated VanR then binds to and activates all four vancomycin-dependent promoters. This includes the vanHAX genes that encode enzymes which reprogram the cell wall; VanH is a lactate dehydrogenase, VanA is a d-Ala–d-Lac ligase and VanX is a d-Ala–d-Ala dipeptidase. The vanHAX genes, together with vanRS, are found in all bacteria with inducible vancomycin resistance (Fig. 6b).

Figure 6

VanRS-like vancomycin resistance modules in Actinobacteria and Firmicutes bacteria. (a) Mechanism of signal transduction mediated by the VanRS TCS. VanR is readily phosphorylated by the cellular pool of acetyl phosphate. In the absence of the appropriate stimulus, i.e. extracellular vancomycin, VanS functions as a VanR-specific phosphatase, thereby preventing premature and unspecific induction of VanR-dependenet gene expression (Hutchings, 2006b). (b) Genomic context conservation of the van locus in both high-G+C and low-G+C Gram-positive bacteria. Symbols and color-code as before. See text for details.

Two genes, vanJ and vanK, are found in the resistance gene clusters of Streptomyces but are not present in the VanA or VanB gene clusters in vancomycin-resistant enterococci (VRE). vanJ encodes a membrane protein of unknown function and is not required for vancomycin resistance (Hong, 2004). vanK encodes a Fem (factor effecting methicillin resistance) protein that replaces the activity of FemX following reprogramming of the cell wall in response to vancomycin (Hong, 2005). In the absence of vancomycin, FemX is an essential enzyme that adds the glycine cross-bridge to the stem peptides of the cell wall precursors. However, after vancomycin-induced cell wall reprogramming FemX cannot recognize d-Ala–d-Lac-containing precursors, and VanK is required for cross-linking of the reprogrammed cell wall (Hong, 2005).

The Streptomyces VanS proteins are unusual in that they have very small extracellular sensor domains consisting of <30 amino acids. This compares to sensor domains that are 103 and 37 amino acids in the enterococcal VanSA and VanSB, respectively. The sensor domains of Streptomyces VanS proteins are so small that they could be included in the family of intramembrane-sensing HKs, which are notable for their lack of extracellular sensing domains. These proteins are proposed to sense intramembrane stress via their transmembrane domains (Mascher, 2006b). However, genetic evidence, including the switching of sensor domains between Streptomyces coelicolor and the glycopeptide producer Streptomyces toyocaensis, suggests that this domain is required for activation and that it determines the substrate specificity of the VanS proteins. In addition, the Streptomyces VanS proteins are only induced by glycopeptide antibiotics with closely related structures, suggesting that they bind the drugs directly rather than sensing a perturbation in the cell envelope (Hutchings, 2006b).

The VanS protein is a bifunctional enzyme that switches between phosphatase and kinase activities in the absence or presence of vancomycin. In a ΔvanS strain of Streptomyces coelicolor, the four van promoters are always switched on (Hutchings, 2006b). Similarly, null mutations in Enterococcus faecium VanS lead to constitutive activation of the vancomycin resistance genes (Depardieu, 2003). This is because, in the absence of VanS, the VanR protein can be phosphorylated in vivo by the small molecule phosphodonor acetyl phosphate (Hutchings, 2006b). The enterococcal protein has also been shown to be activated by acetyl phosphate when the system was reconstituted in Escherichia coli (Haldimann, 1997). The phosphatase activity of VanS in the absence of the signal suppresses the level of phospho-VanR and keeps vancomycin resistance genes switched off. This may be because cells containing d-Ala–d-Lac in their cell walls grow more slowly and appear weaker than wild-type cells, at least in Streptomyces.

While most of the efforts to understand the VanRS system have focused on the VanA and VanB isolates of enterococci and actinomycetes, four other systems have also been identified in the enterococci. These are the constitutively expressed VanD-type resistance genes, which confer resistance by replacing the terminal d-Ala with d-Lac, as in types VanA and VanB (Casadewall & Courvalin, 1999), and the comparatively rare VanC, VanE, and VanG isolates, which contain d-alanyl-d-serine (d-Ala–d-Ser) ligases instead of a d-Ala–d-Lac ligase (Courvalin, 2005). The substitution of d-Ser for d-Ala results in low-level resistance to vancomycin but not teicoplanin. The origins of VanC-, VanE-, and VanG-type resistance are not clear since no glycopeptide producers make d-Ala-d-Ser containing cell wall precursors. It seems likely that the resistance genes in these clusters were acquired from several different sources (Courvalin, 2005).

Two-component systems involved in regulating cell envelope integrity

A number of TCS have been proposed to play important roles in maintaining cell wall integrity and homeostasis, but are not directly implicated in responding and/or counteracting cell envelope stress. These include essential TCS such as YycFG-like (Firmicutes) and MtrAB-like (Actinobacteria). Additionally, the LytSR-dependent holin-/antiholin system of Staphylococcus aureus, which is also conserved in many Firmicutes bacteria, will be described.

YycFG-like TCS

YycFG of B. subtilis was the first Gram-positive TCS to be described as essential for survival under normal laboratory conditions (Fabret & Hoch, 1998). It is restricted to and conserved in the Firmicutes bacteria, possibly excluding some Clostridia, where often no clear homologs can be found (H. Szurmant, pers. commun.) (Table 1). Its essentiality has been verified in a number of these bacteria (Lange, 1999; Clausen, 2003; Hancock & Perego, 2004; Williams, 2005), with the noteworthy exception of Lactococcus lactis (O'Connell-Motherway, 2000).

Based on its genomic context and the domain architecture of the corresponding HK, YycG, two major groups of YycFG-like TCS can be distinguished. Group I is found in most Firmicutes bacteria, with the exception of the streptococci. It is characterized by an extended genomic context conservation, with three to four genes, including yycH and yycI and its homologs, located directly downstream of the yycFG operon. The N-terminus of the YycG HK consists of two deduced transmembrane helices that flank a large periplasmic domain. In this group, both HK and RR are essential. Group II is restricted to streptococci. Here, homologs of yycI and yycH are missing, and the N-terminal domain of the sensor kinases consists of a single transmembrane region and lacks a significant extracellular domain. In group II, only the RR was shown to be essential (Ng & Winkler, 2004). Again, Lactococcus lactis is the exception to the rule. It shows a group II genomic context and a unique input domain architecture for the sensor kinase LlkinC consisting of two transmembrane regions, but lacking a periplasmic domain.

Considering its essentiality and high degree of conservation, there is a surprising variation in the corresponding regulons controlled by the YycFG orthologs from different Firmicutes bacteria. But as a common denominator, many gene products are involved in cell wall homeostasis, cell membrane integrity, and cell division processes (Fukuchi, 2000; Howell, 2003; Dubrac & Msadek, 2004; Mohedano, 2005; Ng, 2005; Bisicchia, 2007; Dubrac, 2007). The reason for its essentiality either could be traced back to the expression of a single target gene, i.e. in Streptococcus pneumoniae (Ng, 2003), or is polygenic by nature, as is the case for B. subtilis. Here, it is based on disrupted cell wall metabolism caused by alterations in the expression of a number of YycF target genes (Bisicchia, 2007).

The eponymous YycFG system of B. subtilis is the best-understood TCS of this group with regard to its regulation (Szurmant, 2007a). The system seems to be at least partially activated during normal growth, and no inducing conditions have so far been identified. Its activity is regulated by YycI and YycH through direct protein–protein interactions with the sensor kinase YycG (Szurmant, 2005, 2007b). Both proteins are peripherally bound to the cytoplasmic membrane and harbor large periplasmic domains that are similar in fold, but not in primary amino acid sequence (Szurmant, 2006; Santelli, 2007). Surprisingly, the transmembrane helix of each protein is sufficient to perform the regulatory role described for YycH and YycI (Szurmant, 2007c).

To control essential processes via means of a signal transduction system appears counterintuitive, at first glance. However, the observation that both the depletion of the YycFG system as well as an overactive YycFG system, as evidenced in the yycH and yycI deletion strains or when yycF is overexpressed, has negative effects on growth and cell wall integrity underscores the importance of tightly controlling the expression of genes in the regulon (Fabret & Hoch, 1998; Szurmant, 2005, 2007b). It is for this reason that the YycFG system can be considered a regulator of cell wall homeostasis rather than just cell wall metabolism (H. Szurmant, pers. commun.).

A link between YycFG-like systems and envelope stress response only emerged recently. In Staphylococcus aureus, upregulation of yycFG expression or nucleotide insertion in yycG resulted in a decreased susceptibility to vancomycin and daptomycin, respectively (Friedman, 2006; Jansen, 2007). Moreover, YycFG-depleted cells are more resistant to detergent- and lysostaphin-induced lysis (Dubrac, 2007). The authors of the latter study suggest that this phenotype is due to the role of YycFG (renamed WalRK) in positively affecting cell wall degradation by up-regulating autolysin synthesis in Staphylococcus aureus. Accordingly, YycFG-depletion results in increased peptidoglycan crosslinking and glycan chain length (Dubrac, 2007).

Based on the findings of their recently published, and impressively thorough, analysis on the essential nature of the YycFG system of B. subtilis, Bisicchia (2007) suggested that the signal sensed by this system might be the antithesis of cell envelope stress, i.e. is derived from normal cell wall metabolism. This hypothesis is supported by the similar expression pattern of YycF-target genes in vancomycin-treated cells, i.e. cell envelope stress conditions, and under conditions of YycFG-depletion. Moreover, the YycFG system is active under normal growth conditions and shuts down towards stationary phase or during cell envelope stress (Cao, 2002b; Howell, 2003, 2006; Bisicchia, 2007). Interestingly, this cell envelope stress-dependent inhibition of YycFG activity seems to be antibiotic-specific: bacitracin-treatment does not affect YycFG-dependent gene expression (Mascher, 2003a; Bisicchia, 2007). Thus, the type of cell envelope stress negatively affecting YycFG activity might ultimately give a clue to the stimulus sensed by this essential TCS. Alternatively, one could of course imagine that the YycFG system does not need a positive stimulus for activity at all. Rather, its ‘default setting’ might be the ON mode (which would also explain its essentiality), with some stimuli, such as growth arrest or cell envelope stress, resulting in the inhibition of the YycFG system. Future work will hopefully expand the understanding of this fascinating TCS.

MtrAB-like TCS

In Corynebacterium glutamicum a single TCS, MtrAB, has been implicated in the CESR (Möker, 2004). This system is conserved throughout the Actinobacteria (Table 2) (Hoskisson & Hutchings, 2006) but appears to serve different functions in different genera. In Mycobacterium tuberculosis the MtrA RR is essential while the sensor kinase, MtrB, is nonessential (Zahrt & Deretic, 2000). This suggests that MtrA can be activated via cross-talk in the absence of its cognate sensor kinase. The only known target of MtrA in M. tuberculosis is the dnaA gene, which encodes the essential master regulator of DNA replication, DnaA (Fol, 2006). This could explain why MtrA is essential, although it is not yet clear whether dnaA expression is absolutely dependent on MtrA. Also, other targets may exist in M. tuberculosis.

In C. glutamicum, both mtrA and mtrB are dispensable for viability although an ΔmtrAB mutant strain has defects in osmoprotection and its cell envelope (Möker, 2004). Direct target genes include the putative cell wall peptidase-encoding genes mepA and nlpC, which are repressed by MtrA, and the solute carrier genes betP and proP, which are activated by MtrA (Möker, 2004; Brocker & Bott, 2006). MtrB does not appear to be activated by monovalent cations in an osmostress-dependent manner (Möker, 2007a). However, compounds such as amino acids, sugars, and polyethylene glycol were recently shown to activate the protein through interaction with its cytoplasmic domain. These solutes are proposed to change the hydration state of the protein and shift it into the active state (Möker, 2007b). If correct, this represents yet another novel mechanism of signal sensing via a ‘simple’ TCS where envelope stress might be detected via its periplasmic sensor domain and osmotic stress via its cytoplasmic kinase domain.

Interestingly, in all available actinobacterial genome sequences the mtrAB genes are clustered with a third gene, lpqB (Hoskisson & Hutchings, 2006) (Fig. 7b). This gene encodes a putative lipoprotein, which has been described as a signature protein of the Actinobacteria (Gao, 2006). The lpqB gene might be essential in C. glutamicum (Brocker & Bott, 2006) and it seems likely that LpqB acts as an accessory protein to the MtrB sensor kinase (Hoskisson & Hutchings, 2006).

Figure 7

The σE-CseABC cell envelope stress module of Streptomyces coelicolor. (a) Mechanism of signal transduction mediated by the σE-CseABC system. (b) Genomic context conservation of lipoproteins and TCS in Actinobacteria. With the exception of the mtrAB-lpqB system, all loci are found in the genome of Streptomyces coelicolor. Symbols and color-code as before. See text for details.

LytSR of Staphylococcus aureus

TCS with LytS-like HKs are common in Firmicutes bacteria (Table 1) (Mascher, 2006b). These kinases are characterized by a conserved N-terminal input domain consisting of six deduced membrane-spanning helices (Anantharaman & Aravind, 2003). The only studied example is LytSR of Staphylococcus aureus, which was originally identified to be involved in regulating autolysis and murein hydrolase activity in this organism (Brunskill & Bayles, 1996b). It is induced by gramicidin, carboxyl cyanide 3-chlorophenylhydrazone (CCCP), and valinomycin – but not by nigericin, indicating that the membrane potential, but not ΔpH, generated by the proton motif force is sensed by LytS (Patton, 2006). Furthermore, its expression was affected in VISA isolates (McAleese, 2006). Its autophosphorylation and subsequent activation of the cognate RR LytR leads to an induction of lrgAB transcription, an operon located directly downstream of lytSR (Brunskill & Bayles, 1996a). Expression of LrgAB negatively affects extracellular murein hydrolase activity and inhibits penicillin-induced killing, while lrgAB mutations show the opposite effect, i.e. increased murein hydrolase activity and decreased penicillin tolerance (Groicher, 2000). LrgA is thought to function as an antiholin that – when expressed during stationary phase – prevents the formation of murein hydrolase export channels and thereby mediates penicillin tolerance (Brunskill & Bayles, 1996a; Bayles, 2000).

But the LytSR-dependent pathway represents only one half of a complex regulatory switchboard that controls autolysis, antibiotic tolerance, stationary phase survival, and biofilm formation of Staphylococcus aureus (Bayles, 2007). The second half is encoded by the cidABC operon, a homolog and functional counterpart of LrgAB (Rice, 2003, 2005). Expression of the holin-like protein CidA enhances murein hydrolase activity and negatively affects penicillin tolerance, i.e. it shows the exact opposite effect to LrgAB, as described above. Furthermore, it was recently shown that cidA-controlled cell lysis plays an important role during biofilm formation (Rice, 2007).

Taken together, work from the Bayles laboratory over the last decade has revealed that LytSR, together with the cytoplasmic regulator, CidR (Yang, 2006), orchestrates a holin (CidAB)/antiholin (LrgAB) system that controls extracellular murein hydrolase activity (i.e. autolysis) and therefore ultimately a bacterial type of programmed cell death (Bayles, 2000, 2003, 2007; Rice & Bayles, 2003). Both the lytSR-lrgAB and the cidR-cidAB modules can be found in the genomes of many Firmicutes bacteria, and at least the cidAB operon is also present in numerous Gram-negative bacteria (Bayles, 2007), indicative of a conservation of this holin/antiholin mechanism in regulating autolysis (Table 1).

This process is somehow reminiscent of the murein hydrolase-dependency of penicillin-induced lysis observed for Streptococcus pneumoniae (Tomasz, 1970, 1988; Lopez, 1990; Moreillon, 1990). Here, it has also been proposed that such a deregulated autolysis observed in bacteria challenged with cell wall antibiotics like β-lactams or vancomycin is a manifestation of bacterial programmed cell death (Lewis, 2000; Henriques Normark & Normark, 2002).

The σE-CseABC module: a unique regulatory link between ECF σ factors and TCS orchestrates CESR in streptomycetes

The σE-CseABC pathway is unique in incorporating an ECF σ factor, a TCS, and a novel accessory lipoprotein into a single signal transduction pathway (Paget, 1999a, b; Hutchings, 2006a). It is also unusual among envelope stress ECF σ factors in that it is not under the control of an anti-σ factor. This system is conserved in all of the streptomycete genomes sequenced to date (Table 2) and appears to be the major pathway for sensing cell envelope stress in the genus Streptomyces. It is important to note that StreptomycesσE is not homologous to Escherichia coliσE. Expression of the sigE-cseABC operon is regulated by a novel three-component system consisting of a sensor kinase, CseC, a RR, CseB, and an accessory lipoprotein, CseA (Fig. 7). However, the sigE promoter is not transcribed by RNA polymerase containing σE. CseC is proposed to bind a cell wall precursor or breakdown product since expression of the sigE operon can be induced by a wide range of cell envelope-specific compounds, including antibiotics such as bacitracin or vancomycin, and muramidases such as lysozyme (Hong, 2002).

Transcription of the sigE operon is absolutely dependent on CseB with ΔsigE and ΔcseB strains sharing an identical phenotype, namely an altered cell wall and hypersensitivity to lysozyme (Paget, 1999a). These phenotypes can be restored to wild-type by including high concentrations of magnesium in the growth medium, again suggestive of a cell wall defect (Paget, 1999a). RNA polymerase holoenzyme containing σE is proposed to activate genes involved in cell envelope repair and homeostasis although, to date, only two target promoters have been identified; the cwg operon, which encodes biosynthetic enzymes for a putative cell wall glycan (Hong, 2002) and the P2 promoter of the hrdD gene, which encodes a homolog of the housekeeping σ factor HrdB (Paget, 1999a). Null mutations in cwg or hrdD do not result in the characteristic phenotype of the ΔsigE strain, suggesting that additional σE targets exist.

Intriguingly, the sigE promoter is constitutively transcribed at a basal level, but this transcription is absolutely dependent on CseB (Paget, 1999b). This suggests either that the CseBC TCS is leaky or, perhaps more likely, that constant turnover of the cell wall results in constitutive low-level activation of CseC and, ultimately, sigE. Disruption of the cseA gene led to a fivefold increase in this basal activity, suggesting that CseA negatively regulates the sigE promoter. However, since it is an extracytoplasmic lipoprotein, it must do this from outside the cell and it seems likely that CseA negatively modulates the sensor domain of CseC (Hutchings, 2006a). CseA has no homologs outside of the streptomycetes and, so far, there are no clues as to how it might modulate signal sensing by CseC.

Transcription of sigE is still inducible by cell envelope-specific antibiotics in a ΔcseA strain and is induced to a higher level than in the wild-type (Hutchings, 2006a). This suggests that CseA somehow reduces the activity of CseC, perhaps by reducing signal binding or by trapping it in the ‘OFF’ state. Interestingly, a null mutation in cseC could not be isolated, indicating that this gene might be essential. This is despite the fact that ΔsigE-, ΔcseA- and ΔcseB-null mutants were readily obtained. One possible explanation is that, as in the case of Streptomyces coelicolor VanR, CseB can be activated via cross-talk in the absence of CseC, and this leads to a lethal overexpression of σE and its target genes. This hypothesis is supported by the fact that cseC can be disrupted in the acetyl phosphate-deficient Δpta ackA strain (M.I. Hutchings, unpublished data). Clearly, there is still much to learn about this signal transduction pathway, not least the mechanism of action of CseA and the regulon of genes controlled by σE.

Beyond ECF and TCS: novel mechanisms of transmembrane signal transduction in CESR

In this section, two unusual mechanisms of sensing and responding to extracellular stress caused by cell wall antibiotics will be discussed: (1) Inducible β-lactam resistance in Staphylococcus aureus, B. licheniformis and some other Firmicutes bacteria, orchestrated by the BlaR1/MecR1 systems, and (2) acquired bacitracin resistance in Enterococcus faecalis, mediated by the membrane-anchored transcriptional regulator BcrR.

The BlaR1/MecR1-dependent β-lactam resistance modules

High-level β-lactam resistance in MRSA strains of Staphylococcus aureus results from the expression of a β-lactamase, encoded by blaZ, and the mecA gene, encoding the PBP 2A (PBP2a) (Rosato, 2003). Transcription of both genes is controlled by the BlaR–BlaI, and MecR–MecI regulatory systems, respectively, which are remarkably similar in structure and function. These systems were covered in two recent and comprehensive review articles (Fuda, 2005; Wilke, 2005), and will therefore be described relatively briefly.

Both systems consist of a transmembrane sensor protein, BlaR1 or MecR1, and a cognate cytoplasmic transcriptional repressor, BlaI/MecI, respectively. The BlaR1/MecR1 proteins harbor three functionally important domains, (1) an extracellular input domain that directly interacts with the β-lactam antibiotic, (2) a cytoplasmic zinc protease domain that is important for the proteolytic inactivation of the cognate repressor protein, and (3) four membrane-spanning helices that are crucial for the process of intramolecular transmembrane signal transduction between the input and the protease domain. The repressors BlaI/MecI consist of two functional domains, namely dimerization and the DNA-binding domain (Fuda, 2005). Knowledge stems from work on both systems in Staphylococcus aureus and a homologous system for β-lactamase expression in B. licheniformis. Related systems are also present in the genomes of some other bacilli and staphylococci (Table 1). Induction of blaZ/mecA expression in response to the extracellular presence of β-lactams is the result of a four-step process. For reasons of simplicity, this process will be described for BlaR1/BlaI, but similar steps are involved in the MecR1/MecI system (Fuda, 2005).

The initial step of stimulus perception comprises an irreversible acylation of an active site serine in the extracellular sensor domain of BlaR1 in the presence of β-lactams (Kerff, 2003; Wilke, 2004; Marrero, 2006). This is followed by an again irreversible decarboxylation of an active site lysine carbamate (Thumanu, 2006).

Intramolecular signal transduction is propagated through the transmembrane α-helices to the cytoplasmic metalloprotease domain, presumably involving noncovalent interactions between the input domain and the extracellular loop 2, which are affected by β-lactam acylation (Hanique, 2004). While the exact mechanism of this novel way of signal transduction is not yet fully understood, two important events that are associated with it have been identified, so far. (1) A changed modality of the integral membrane proteins from antiparallel β-barrel to the α-helix bundle state, involving a realignment of the four transmembrane helices. This mechanism is supported by the observed significant conformational changes in the α-helices of the BlaR1 sensor domain upon acylation (Hardt, 1997; Golemi-Kotra, 2003). (2) Proteolytic cleavage of the key signaling components, BlaR1/BlaI, i.e. autocleavage of the cytoplasmic metalloprotease domain of BlaR1 and the subsequent induction of metalloprotease-dependent repressor cleavage (Zhang, 2001).

Autocleavage activates the intracellular Zn-metalloprotease domain of BlaR1, as a result of sensor domain acylation in response to extracellular β-lactams. This leads to the proteolytic cleavage and subsequent dissociation of the dimeric repressor BlaI, thereby releasing it from its operator sequences (Gregory, 1997; Zhang, 2001; Filee, 2003; Garcia-Castellanos, 2003; Safo, 2005). Removal of the repressor results in the transcription of the divergent operons blaR1-blaI (coding for the signal transducer and repressor, respectively) and blaZ, encoding the β-lactamase. This leads to β-lactam inactivation and hence resistance.

BlaR1 is thought to (auto-)activate signal transduction only once. Therefore, intact BlaR1 needs to be continuously replenished for sensing β-lactams. Once the extracellular concentration of this inducer diminishes, BlaR1 is no longer autoactivated. As a result, proteolytic cleavage of BlaI ceases and the intracellular concentration of intact repressor increases. This leads to repressor dimerization, its DNA binding and ultimately resuppression of gene expression (Zhang, 2001; Fuda, 2005). The same overall mechanism holds true for mecR1-mecI/mecA expression, with one major difference: BlaR1 activation initiates blaZ expression within minutes, whereas MecR1 acylation induces PBP2a expression over hours (Ryffel, 1992; McKinney, 2001). The reason for this dramatic temporal difference, despite the virtual identical mechanisms of activation, remains to be elucidated.

BcrR – a membrane-anchored transcriptional regulator mediating bacitracin resistance

A second unusual signal transducer was recently identified in Enterococcus faecalis (Manson, 2004). Here, the expression of a bacitracin-specific ABC transporter, homologous to B. subtilis BceAB or B. licheniformis BcrAB, is mediated by a novel membrane-anchored transcriptional regulator, designated BcrR. This protein harbors an N-terminal helix–turn–helix DNA-binding motif of the XRE family, and four putative membrane-spanning helices in the C-terminus. The bcrR gene is constitutively expressed and located directly upstream of bcrABD, encoding the ABC transporter and an undecaprenol kinase-homolog BcrD. BcrR is required for the bacitracin-specific induction of bcrABD expression and is only activated by bacitracin (Manson, 2004). Other cell wall-active antimicrobials have no effect on the induction of bcrABD. Subsequent work from the same group verified that BcrR, anchored to the membrane, indeed binds to a direct repeat in the bcrA promoter region and is necessary for transcription initiation (Gauntlett, 2007). Activation of bcrA expression does not depend on the presence of bcrABD, suggesting that bacitracin is directly sensed by BcrR. Moreover, the membrane-spanning helices of BcrR appear to harbor the crucial residues needed for bacitracin sensing (G. Cook, pers. commun.).

Taken together, the data available at the moment demonstrates that BcrR is indeed a novel membrane-anchored sensor/regulator protein that specifically responds to the presence of extracellular bacitracin and mediates the induction of bacitracin resistance determinants.

Beyond cell wall antibiotics: common themes and additional functions of cell envelope stress-sensing systems

This study has so far addressed regulatory systems responding to some aspect of cell envelope stress primarily in the light of our narrow initial definition, i.e. mainly in light of their induction by cell wall antibiotics. One of the most important questions, of course, is the physiological role that these systems play. What is their function, their purpose? Are they solely involved in responding to cell envelope stress, or do they play a physiological role beyond CESR?

The analysis of most cell envelope stress-responsive systems – with the exception of the cell wall antibiotic-specific detoxification modules (i.e. BceRS-/VanRS-like TCS, BlaR1/MecR1, and BcrR) – has revealed links to other aspects of bacterial physiology and differentiation, as already noticed throughout the previous paragraphs on individual cell envelope stress-sensing systems. Owing to the crucial nature of maintaining envelope integrity, these bounds are often tied to the most prominent features/characteristics of the life style of a given species. The overall picture emerging is that cell envelope stress networks are regularly linked to, sometimes deeply embedded in, differentiation and global stress response cascades. Here, some of the more prominent connections are briefly highlighted.

CESR and osmotic stress response

One of the major functions of the cell envelope is to counteract the high internal osmotic pressure, and osmotic stress conditions definitely strain the envelope. Therefore, a link between cell envelope and osmotic stress response seems to be almost natural at first glance. Surprisingly, this is not the case. Most cell envelope stress-responsive systems are blind to classical osmotic shock situations, i.e. hyperosmotic shock exerted by the presence of high concentrations of osmotically active solutes such as salts and glucose (Morbach & Krämer, 2002). The only regulatory systems described in this regard are in fact only very loosely linked to envelope stress response. The MtrAB system of C. glutamicum has clearly been linked to osmotic stress response, since it regulates the expression of osmoprotective functions, such as solute carrier proteins (Möker, 2004). Likewise, the impact of the listerial LisRK TCS on CESR seems to be indirect and mediated through the regulation of a LiaRS-like system, as described above. The observed link between LisRK and osmotic stress response has been attributed to the LisRK-dependent regulation of HtrA (Sleator & Hill, 2005). This protein is implicated in degrading misfolded proteins under various stress conditions (Wonderling, 2004; Sleator & Hill, 2005; Stack, 2005), indicating that LisRK is not involved in mediating osmoprotection. Overall, the behavior of LisRK and its orthologs – none of which has been demonstrated to participate in CESR – seems to be that of a more general stress-response system.

CESR and secretion stress

Control of HtrA expression seems to represent a critical intersection between CESR and a number of other phenotypes, including virulence. As mentioned above, HtrA is a highly conserved periplasmic serine protease, which degrades misfolded proteins as a result of stress conditions such as heat shock (Pallen & Wren, 1997; Kim & Kim, 2005). Its regulation is at the core of CESR in Escherichia coli, where HtrA expression is controlled by both σE and the CpxAR TCS (Danese & Silhavy, 1997). HtrA and its homologs are also controlled by cell envelope stress-sensing systems in a number of Gram-positive bacteria, such as Streptococcus pneumoniae CiaRH (Sebert, 2002, 2005; Mascher, 2003b), Listeria monocytogenes LisRK, and Staphylococcus aureus VraSR (Kuroda, 2000), often accounting for phenotypes associated with their activity. Interestingly, the orthologous system of the latter in B. subtilis, LiaRS – while not regulating htrA expression – is strongly induced by secretion stress exerted by the overexpression of heterologous proteins such as B. amyloliquefaciens AmyQ, Lactococcus lactis USP45, and Escherichia coliβ-lactamase (Hyyryläinen, 2005; Trip, 2007). In this organism, expression of HtrA and its ortholog, HtrB, is tightly controlled by the CssSR TCS (a homolog of the cell envelope stress-sensing TCS CpxAR of Escherichia coli), which is specifically induced by secretion stress (Hyyryläinen, 2001; Darmon, 2002). In addition to activating the LiaRS TCS, secretion stress also induced the ECF σ factor σX (Hyyryläinen, 2005), further emphasizing that envelope and secretion stress cannot be separated easily. Remarkably, this is in good agreement with the data on CESR in Escherichia coli, which is defined as counteracting misfolded proteins in the periplasm (Ruiz & Silhavy, 2005). This now brings the envelope stress responses in both Gram-positive and -negative bacteria close together, despite the overall differences in envelope architecture and the initial approaches to study the respective stress responses, as outlined in the introduction.

CESR and differentiation

A number of cell envelope stress-sensing systems are linked to, or embedded in cellular differentiation programs. This, also, makes perfect sense, considering that the envelope is the prime structural component of the bacterial cell, and hence of crucial importance in (multi)cellular differentiation. ECF σ factors of B. subtilis have recently been shown to play an important role in biofilm formation and multicellular differentiation (Kobayashi, 2007; Mascher, 2007). Moreover, both σW and the LiaRS TCS of B. subtilis are embedded in transition state regulation (Qian, 2002; Jordan, 2007), a complex regulatory differentiation program that ultimately leads to the formation of highly resistant endospores (Msadek, 1999; Phillips & Strauch, 2002; Errington, 2003). A role in biofilm formation has also been described for the holin/antiholin system of Staphylococcus aureus (Rice, 2007). It seems reasonable to assume that a number of additional cell envelope stress sensors are important in regulating differentiation processes.

CESR and virulence

Basically, any regulatory system mediating CESR in a pathogen that has been investigated in this regard has been shown to be important for virulence. Again, this is not too surprising, given the crucial role of the Gram-positive cell envelope for numerous aspects of virulence, including host cell contact, adhesion, penetration, displaying virulence factors at the cell surface, and as an antigen for immune system discovery and evasion (Sriskandan & Cohen, 1999; Ginsburg, 2002; Boneca, 2005; Myhre, 2006). A detailed discussion of the link between virulence, the cell envelope and regulatory systems mediating CESR is beyond the scope of this review. But it is important to stress that examples from almost any conserved group of regulatory systems described in this review (with the exception of the specific detoxification systems, i.e. ‘classical’ BceSR-like TCS, VanRS, or BlaR1/MecR1) have been described or are implicated in playing a role in pathogenicity. This includes Staphylococcus aureus VraSR (LiaRS-like), Listeria monocytogenes VirRS and most likely also Staphylococcus epidermidis ApsXRS (BceRS-like), Listeria monocytogenes CesRK, LisRK-like TCS in pathogenic bacteria, and Streptococcus pneumoniae CiaRH.

One aspect linking CESR to both antibiotic resistance and virulence simultaneously is its role in regulating cell envelope net charge. As mentioned before, the Gram-positive envelope has an overall negative net charge, due to the high density of anionic groups both in membrane lipids and TAs. While this trait is important for the functionality of the envelope, it also makes it an easy target for CAMPs that are produced both by other bacteria, i.e. lantibiotics, and by the immune system (Brogden, 2005; Hancock & Sahl, 2006; Giuliani, 2007). Lowering the overall negative net charge by incorporating neutral or positively charged groups is therefore an efficient mechanism to mediate unspecific resistance against both groups of antimicrobial compounds, as described in ‘Introduction.’ The most common strategy to achieve this resistance, the incorporation of d-alanine into TAs, is described in detail in the next paragraph.

Conservation and regulation of core functions embedded in Gram-positive CESR stimulons

As noted above, the importance of CESR is reflected by the degree of conservation of its regulatory systems in Gram-positive bacteria (Tables 1 and 2). But even when these systems are highly conserved, their target genes often are not. On the other hand, even though the diversity within individual cell envelope stress stimulons is very large, some ‘marker’ loci that are almost invariantly linked to the Gram-positive envelope stress response can be identified, which are regulated by a whole variety of signal transducing systems, in different organisms. The two functionally most important examples are the Dlt system and alternative PBPs.

The Dlt system

The core of TAs, the second most abundant polymer of the cell wall in low-G+C Gram-positive bacteria, is typically comprised of linear poly(glycerol phosphate) or poly(ribitol phosphate) chains. TAs can be either covalently linked to the cell wall (wall TAs) or tethered in the cytoplasmic membrane via a terminal glycolipid moiety (lipoteichoic acid) (Fischer, 1988, 2000; Delcour, 1999; Foster & Popham, 2002; Neuhaus & Baddiley, 2003). Because of the anionic nature of the phosphate ester bridges, this molecule is the major source of the negative net charge of the Gram-positive envelope. Hence, TAs have an important impact on a number of biological processes, including autolysis (Fedtke, 2007), binding of cations and surface proteins (Hughes, 1973; Briese & Hakenbeck, 1985; Jonquieres, 1999), adhesion (Abachin, 2002; Weidenmaier, 2004), host colonization and biofilm formation (Gross, 2001), stimulation of the immune response (Morath, 2002; Grangette, 2005), and virulence. For most of these processes, the d-alanylation of Tas – mediated by the activity of the Dlt system – has been demonstrated to be of crucial importance (Neuhaus & Baddiley, 2003). It lowers the overall negative charge of the envelope, thereby affecting the interaction with various compounds and structures on or near the surface, including the aforementioned CAMPs and host cell surfaces. d-Alanylation renders the Gram-positive cell more virulent and more resistant to CAMPs and other aspects of the host's immune defense (Peschel, 1999; Abachin, 2002; Collins, 2002; Kovács, 2006).

The Dlt system is encoded by the dltABCD operon, which is present in all Firmicutes bacteria sequenced so far. Because of its crucial role in cell envelope integrity in general and CAMP resistance in particular, it is a typical component of the Gram-positive CESR (Table 3). Its induction in the presence of cell envelope stress is regulated by a plethora of different systems, again covering most of the conserved regulator groups described in this review. In B. subtilis, σX regulates envelope net charge by inducing expression of the dlt and pss-psdA operons, the latter incorporating neutral phospholipids into the cytoplasmic membrane, thereby enhancing the overall effect (Cao & Helmann, 2004). The BceRS-like TCS VirRS and ApsXRS orchestrate dlt expression in Listeria monocytogenes and Staphylococcus epidermidis, respectively (Mandin, 2005; Li, 2007). In Streptococcus pneumoniae, the dlt operon is part of the CiaR regulon (Mascher, 2003b). A unique TCS regulates the expression of the dlt operon in Streptococcus agalactiae (Poyart, 2001). It is encoded by the dltRS genes located directly upstream of dltABCD. d-Alanine deficiency activates transcription of all six genes in a DltR-dependent manner. The DltRS system has not yet been investigated with regard to cell envelope stress. But its assumed role in maintaining an appropriate level of d-alanine esters in the LTA of Streptococcus agalactiae seems indicative for such a function.

View this table:
Table 3

Overview of transcriptomics studies on the cell envelope stress response in Firmicutes bacteria

OrganismStimulusConditionsSignal transducing systemsTarget genes/remarksReference
Bacillus licheniformisBacitracin300 μg mL−1, 10 minYvqEC, YtsAB, YxdJK2σM, σV, σYpbpE, pbpX, (dlt); SLWecke (2006)
Penicillin G100 μg mL−1, 10 minσM, σVNA; SLWecke (2006)
Vancomycin1 μg mL−1, 10 minYvqECσM, σV, σW, σYpbpE, pbpX, dlt; SLWecke (2006)
Bacillus subtilisBacitracin100 μg mL−1, 5 minBceRS, LiaRS, YvcPQσMSLMascher (2003a)
CAMP: LL-371.5 μM/20 minLiaRS, YxdJKσM,pbpE, dlt; SLPietiäinen (2005)
CAMP: PG-150 nM/20 minLiaRSσW, σXpbpE, pbpX, dlt; SLPietiäinen (2005)
CAMP: PLL1 mM/20 minσM, σWpbpE, dlt; SLPietiäinen (2005)
Cefalexin0,25 μg mL−1, 10 minHutter (2004)
Cefotaxime1 μg mL−1, 10 minHutter (2004)
Cefoxitin0,5 μg mL−1, 10 minHutter (2004)
D-Cycloserine16 μg mL−1, 10 minHutter (2004)
Fosfomycin256 μg mL−1, 10 minHutter (2004)
Oxacillin0.25 μg mL−1, 10 minHutter (2004)
Penicillin G4 μg mL−1, 10 minHutter (2004)
Secretion stress(expression of AmyQ)LiaRS, CssRSσXpbpA, dltHyyryläinen (2005)
Triton X-11464 μg mL−1, 10 minLiaRSσW, σYpbpE, (pbpX), dltHutter (2004)
Vancomycin2 μg mL−1, 3/10 minLiaRS, BceRSσM, σV, σW, σYpbpE; 10x MICCao & Helmann (2002)
Lactococcus lactisLcn97220 AU mL−1, 9minCesSRMartinez (2007)
Niss vs. Nisr0.04 vs. 3 μg mL−1CesSRdltC, pbpAKramer (2006)
Staphylococcus aureusBacitracin6.7 U mL−1, 60 minVraRSNApbpB, pbpDUtaida (2003)
D-Cycloserine150 μg mL−1, 60 minVraRSNApbpBUtaida (2003)
TeiR vs. TeiSVraRS, LytRS, ArlRSNApbpA-D, dltRenzoni (2006)
murF depletionVraRSNApbpASobral (2007)
Oxacillin1.2 μg mL−1, 60 minVraRSNApbpBUtaida (2003)
pbpB depletionVraRSNANAGardete (2006)
Vancomycin10 μg mL−1, 10 minVraRS, SaeRSNApbpBKuroda (2003)
Vancomycin (MSSA)10 μg mL−1, 5 minVraRSNAMcCallum (2006)
Vancomycin (VISA)8 μg mL−1, 30 minVraRSNAMcAleese (2006)
VISA vs. VSSAGraRSNACui (2005)
WT vs. VRSAVraRSNApbpBKuroda (2000)
Streptococcus pneumoniaePenicillin0.03 μg mL−1, 60 minCiaRH, CtsRNARogers (2007)
Vancomycin5 μg mL−1, 10 minCiaRH, TCS03, TCS11NAVans vs. VanrHaas (2005)
  • * CAMP, cationic antimicrobial peptide; Lcn972, lactococcin; Nis, nisin; Tei, teicoplanin; MSSA, methicillin-sensitive S. aureus; VISA/VSSA/VRSA, vancomycin-intermediate resistant/sensitive/resistant S. aureus; WT, wild type

  • From the analyses of ([148]Hutter et al., 2004) only the 10 min time points were considered. The 40 and 80 min time points were not taken into account for generating this table.

  • Induction of the regulatory systems listed was either determined based on up-regulation of their target genes (TCS), and/or auto-induction (ECF σ factors); NA, not applicable.

  • § Only pbp genes or the dlt operon were exemplarily chosen (see text for details).

  • NA, not applicable; MIC, minimal inhibitory concentration; SL, sub-lethal antibiotic concentration.

Expression of the dlt operon is induced by various stress and growth conditions in other Firmicutes bacteria, indicating the presence of regulators yet to be identified. In Lactococcus lactis, the dlt operon is upregulated in spontaneous nisin-resistant strains, compared to the nisin-sensitive wild type (Kramer, 2006). In Staphylococcus aureus, expression of the dlt operon has been shown to be controlled by the agr (for accessory gene regulator) locus (Dunman, 2001), which is important for the virulence of this organism, but also linked to cell envelope stress regulation. Dlt expression is induced by high concentrations of cations (Koprivnjak, 2006). Moreover, dlt-upregulation was observed in teicoplanin-resistant strains, relative to the sensitive wild type (Renzoni, 2006). In Streptococcus mutans, its expression is growth phase dependent and can be modulated by carbohydrates (Spatafora, 1999).

Alternative PBPs

PBPs play a key role as both targets of, and resistance mechanisms against β-lactam antibiotics (Goffin & Ghuysen, 2002; Poole, 2004; Macheboeuf, 2006). Because the β-lactam mode of action involves a covalent inactivation of the catalytic center of PBPs, and therefore their sequestration away from their normal substrates, one mechanism to gain β-lactam resistance is to simply increase the amount of PBPs in order to resume cell wall biosynthesis. Therefore, it is not surprising that upregulation of PBP expression is a common strategy in Gram-positive envelope stress responses. As was the case for the dlt operon, again many different regulatory systems participate in mediating upregulation of pbp gene expression. Genes encoding PBPs have been identified in the regulons of B. subtilisσX and σW (Cao, 2002a; Cao & Helmann, 2004), and were also expressed in a presumably ECF-dependent manner in B. licheniformis (Wecke, 2006). The VraR-dependent expression of pbp2 has been studied in detail (Kuroda, 2003; Yin, 2006). A PBP-like protein was also found amongst the target genes of the orthologous CesSR system of Lactococcus lactis (Martinez, 2007), and pbp2229 of Listeria monocytogenes was expressed in a LisRK/Lmo1021-dependent manner (Cotter, 2002; Gravesen, 2004). A PBP-like function is also embedded in the VanD-like vancomycin resistance gene cluster of Enterococcus faecium (Casadewall & Courvalin, 1999). The MecR1-dependent expression of PBP2a in Staphylococcus aureus was already described above. A number of additional PBPs, for which regulators have not yet been identified, were shown in the course of DNA microarray-based global transcriptome analyses to be upregulated as part of cell envelope stress stimulons (Table 3).

Conclusions and outlook

The advent of the genomics era has significantly advanced one's understanding of the Gram-positive CESR in recent years. Global transcriptome analyses (see Table 3) have not only generated large lists of genes, but have also helped to identify regulators that sense and respond to cell envelope stress. This approach has worked particularly well for both B. subtilis and Staphylococcus aureus. These studies helped to unravel the complex regulatory network orchestrating CESR in B. subtilis (Cao, 2002b; Mascher, 2003a; Pietiäinen, 2005) and defined the cell wall stress stimulon of Staphylococcus aureus (Kuroda, 2003; Utaida, 2003; Gardete, 2006; Sobral, 2007). For Staphylococcus aureus, transcriptomics was also used to compare the expression patterns between various antibiotic-sensitive and -resistant clinical isolates, in order to understand the mechanism of glycopeptide resistance (Kuroda, 2000; Cui, 2005; McAleese, 2006; McCallum, 2006; Renzoni, 2006). These approaches helped gain some insight into the core cell wall stress stimulon of Staphylococcus aureus, identified the key regulator, VraSR, and allowed a first glimpse into the complex interdependence of differential gene expression that connects virulence, antibiotic resistance and overall maintenance of envelope integrity. It also indicated that additional layers of regulation exist that could not be accounted for by transcriptional regulation alone. Needless to say, these studies raised more questions than they offered answers.

Despite significant progress, many of the differentially expressed CESR genes still cannot be assigned to specific regulons, even in the two best-studied organisms. For other organisms, such as B. licheniformis and Streptococcus pneumoniae, the first transcriptome datasets will hopefully serve as starting points for subsequent in-depth studies. Surprisingly, there is still not a single global analysis on CESR in any actinobacterial species, despite the great potential these bacteria offer in studying signal transduction and adaptation. Hopefully, this situation will change in the near future.

While a number of specific antibiotic resistance genes have been identified, most notably within the bacitracin stimulon of B. subtilis (Cao & Helmann, 2002; Mascher, 2003a; Ohki, 2003a, b), this does not seem to be the rule. In most cases, such as VISA, the antibiotic resistance trait is less specific or pronounced. Because of their clinical relevance, a number of comparative global transcriptome analyses have been performed for VISA strains (Kuroda, 2000; Mongodin, 2003; Cui, 2005; McAleese, 2006). However, no overall conserved gene expression pattern emerged, and none of the target genes could be attributed to VISA strains specifically. Instead, the VISA trait appears to stem from an overall adjustment of global gene expression and therefore cell physiology, based on – at least in part – the accumulation of individual point mutations (Hiramatsu, 2001; Hiramatsu, 2002; Gemmell, 2004). One striking feature of many VISA strains is their thickened cell wall (Cui, 2000; Cui, 2003). Its contribution to vancomycin resistance is based on reducing the free diffusion of vancomycin while penetrating through the cell wall to its target site. Despite this strong phenotype and a wealth of transcriptome data, little is known about the underlying regulation. This single example should point out that transcriptomics – while a powerful technique – has limitations and is only one aspect of studying CESR.

Table 3 also indicates a more technical issue associated with studying and comparing CESR based on global transcriptional profiling. There is huge variation in the overall experimental conditions between different analyses, including strains, growth conditions, antibiotic concentrations, and induction times. While there is no such thing as ‘the right way’ to do such analyses, there are some precautions that need to be taken into account. Choosing sublethal antibiotic concentrations is mandatory, of course. But what about comparing a sensitive and a resistant strain: should the same absolute concentrations be used for both, or should the concentration be adjusted relative to the minimal inhibitory concentration for each given strain? Arguments can be found for both ways, but they will give profoundly different results. Likewise, the choice of incubation times before stopping transcription will strongly affect the information that can be extracted from such an experiment. Short incubation times of 3–10 min will allow conclusions to be drawn on the initial stress response, and will also catch very transient responses. In contrast long incubation times of 30–60 min, i.e. two generation for fast-growing bacteria, will more likely reflect the overall long term adjustments in gene expression towards a given condition. The specific conditions should be chosen carefully depending on the questions being asked. Comparisons between data sets are only meaningful if the same experimental conditions are used.

The link between cell envelope and secretion stress illustrates that CESR systems cannot – and should not – be viewed separated from other stress responses. It also demonstrates the limitations and pointlessness of the desire to classify and generate strict term definitions, including the one introduced in this review for cell envelope stress. Evolution works in functions and necessities, not names and categories, and the multitude of functions is reflected by a plethora of mechanisms to respond to the stress exerted upon any organism. Nevertheless, the progress made in the field of CESR is impressive, especially during the last five years, thanks mostly to the advent of -omics technologies.

With this review it was intended to present a first comprehensive overview of the conservation and diversity of the Gram-positive CESR. This is a dynamic and exciting field of research that has made rapid progress in recent years. Given the important role that cell wall antibiotics continue to play in treating bacterial infections, this progress is of immense clinical relevance, to increase the understanding of antibiotic action and the mechanisms of antibiotic resistance. Despite the rapid advances in recent years it is still the very beginning and a great deal more remains to be discovered. However, the tools and techniques to effect these discoveries are at hand. These are fascinating times indeed, to study the bacterial CESR.


The authors are indebted to John Helmann, Anna-Barbara Hachmann, and Tina Wecke for critical reading of the manuscript and their thoughtful comments. Moreover, the authors would like to thank Kenneth Bayles, Reinhold Brückner, Gregory Cook, Susanne Morbach, and Hendrik Szurmant, for their expertise and comments on individual parts of the manuscript and sharing of unpublished information. The authors also gratefully acknowledge the constructive criticism of the two anonymous reviewers, most notably the inspiring and thought-provoking comments of reviewer 2 on defining CESR. Work in the authors' labs was supported by grants from the Deutsche Forschungsgemeinschaft (MA2837) and the Fonds der Chemischen Industrie (to T.M.), and from the Royal Society, the Research Councils UK and the Biotechnology and Biological Sciences Research Council (to M.I.H.).


  • Editor: Fritz Unden