Broad-spectrum antimicrobial peptide resistance by MprF-mediated aminoacylation and flipping of phospholipids

Authors


E-mail andreas.peschel@uni-tuebingen.de; Tel. (+49) 7071 2981515; Fax (+49) 7071 293435.

Summary

Bacteria are frequently exposed to cationic antimicrobial peptides (CAMPs) from eukaryotic hosts (host defence peptides) or from prokaryotic competitors (bacteriocins). However, many bacteria, among them most of the major human pathogens, achieve CAMP resistance by MprF, a unique enzyme that modifies anionic phospholipids with l-lysine or l-alanine thereby introducing positive charges into the membrane surface and reducing the affinity for CAMPs. The lysyl or alanyl groups are derived from aminoacyl tRNAs and are usually transferred to phosphatidylglycerol (PG). Recent studies with MprF from Staphylococcus aureus demonstrated that production of Lys-PG only leads to CAMP resistance when an additional flippase domain of MprF is present that translocates Lys-PG and exposes it at the outer leaflet of the membrane. Thus, MprF exerts two specific functions that have hardly been found in other bacterial proteins. MprF proteins are crucial virulence factors of many human pathogens, which recommends them as targets for new anti-virulence drugs. Intriguingly, specific point mutations in mprF cause resistance to the CAMP-like antibiotic daptomycin in a yet unclear way that may involve altered Lys-PG synthesis and/or Lys-PG flipping capacities. Thus, a thorough characterization of MprF domains and functions will help to unravel how bacteria maintain and protect their cytoplasmic membranes.

Introduction

Bacterial membranes fulfil several vital tasks and membrane composition has to be adjusted to prevent cell damage when growth conditions become unfavourable. The main constituents of bacterial membranes are various types of phospholipids whose relative abundance varies profoundly between species and may be different under various environmental conditions and growth phases (Zhang and Rock, 2008). Disruption of pathways leading to biosynthesis of certain phospholipids often has only little effects on viability under optimal laboratory conditions but these mutants have severe growth defects under less favourable conditions such as low osmolarity, high temperature or infection of host organisms (Peschel et al., 2001; Salzberg and Helmann, 2008) The actual roles of the various phospholipids, their biosynthesis, turnover and regulation, have remained incompletely understood. Of note, the same holds true for the nature and function of postulated flippase proteins that are required for translocation of lipids from the inner to the outer leaflet of cytoplasmic membranes.

The most common bacterial phospholipids are phosphatidylglycerol (PG) and cardiolipin (CL, also known as diphosphatidylglycerol) whose head groups are negatively charged (Fig. 1). The zwitterionic phosphatidylethanolamine is found for instance in enterobacteriaceae and bacilli but is absent in several Gram-positive genera such as the staphylococci, streptococci and enterococci (Ratledge and Wilkinson, 1988). Our knowledge concerning bacterial membrane composition is even more limited when it comes to unusual aminoacyl phospholipids such as lysyl-phosphatidylglycerol (Lys-PG) or alanyl-phosphatidylglycerol (Ala-PG). Lys-PG was discovered in Staphylococcus aureus in the early 1960s in the pioneering studies of Marjorie G. Macfarlane and William J. Lennarz (Macfarlane, 1962; Lennarz et al., 1966). Lysine or alanine are linked to PG via ester bonds between the α-carboxyl group of the amino acid and a hydroxyl group of the terminal glycerol moiety of PG (Fig. 1). The process of phospholipid aminoacylation was found to depend on aminoacyl-tRNAs (Lennarz et al., 1966; Gould and Lennarz, 1967). The enzyme MprF catalysing Lys-PG biosynthesis and the functions of this peculiar lipid were discovered almost 40 years later (Peschel et al., 2001). MprF recently turned out to be a bifunctional protein with two unprecedented functions in lipid lysinylation and in translocation of the resulting Lys-PG. This review summarizes the recent achievements and major open questions on the mode of action of MprF and on the impact of the produced lipid on bacterial physiology and host interaction.

Figure 1.

Anionic bacterial phospholipids (PG, CL) and aminoacylated PG variants (Lys-PG, Ala-PG).

Occurrence of aminoacyl phospholipids and MprF proteins

Lys-PG occurs in many Gram-positive bacteria including firmicutes and actinobacteria, as well as certain proteobacteria [for reviews see (Ratledge and Wilkinson, 1988; Roy, 2009; Roy et al., 2009)]. In contrast, Ala-PG appears to be less abundant. Certain bacteria produce only Lys-PG [e.g. staphylococci (Peschel et al., 2001)] or Ala-PG [e.g. Pseudomonas aeruginosa (Klein et al., 2009)] while others produce both, Lys-PG and Ala-PG [e.g. Clostridium perfringens (Johnston et al., 2004)]. Some studies also reported about the occurrence of lysine- or alanine-modified cardiolipin (e.g. Listeria monocytogenes, Vagococcus fluvialis), which usually occur along with the corresponding PG variants and may be side products of MprF proteins with relaxed acceptor substrate specificity (Fischer and Arneth-Seifert, 1998; Thedieck et al., 2006). Ornithyl-PG has also been described in some instances (Khuller and Subrahmanyam, 1970).

MprF was identified when its inactivation rendered a S. aureus transposon mutant susceptible to a wide range of cationic antimicrobial peptides (CAMPs) leading to the name ‘multiple peptide resistance factor’ (MprF) (Peschel et al., 2001). Defensins, cathelicidins, kinocidins and related CAMPs are essential components of the antimicrobial warfare arsenal in humans, vertebrate and invertebrate animals, and even plants (Hancock and Sahl, 2006). Moreover, many bacteriocins including lantibiotics, lipopeptides and unmodified antimicrobial peptides represent CAMPs (Cotter et al., 2005). Although peptide structures vary, overall structural features (cationic, amphipathic properties) and modes of action (damage of microbial membrane-associated processes) are shared by most of these peptides (Yeaman and Yount, 2007). The cationic properties of CAMPs impart strong affinities to the negatively charged bacterial ‘standard’ lipids PG and CL. However, lysinylation of PG leads to a lipid with positive net charge, which reduces the bacterial affinity for CAMPs and renders bacteria more tolerant to CAMPs (Fig. 2; Table 1) [for reviews (Nizet, 2006; Peschel and Sahl, 2006)].

Figure 2.

Mode of action of S. aureus MprF. Lys-PG is synthesized from Lys-tRNA and PG by the synthase domain of MprF. Lys-PG can only neutralize the outer surface of the membrane upon translocation to the outer cytoplasmic membrane leaflet, which is facilitated by the flippase domain of MprF. (Ernst et al., 2009).

Table 1.  Phenotypic changes associated with mprF inactivation in different bacterial species.
SpeciesAminoacyl phospholipidPhenotypic changes associated with mprF inactivation leading toReferences
Reduced susceptibility to cationic antimicrobialsAttenuated host interactionOther changes
  1. N.D., not determined.

Bacillus anthracisLys-PGProtamine, LL-37, HNP1N.D. Samant et al. (2009)
Bacillus subtilisLys-PGNisin, daptomycinN.D. Salzberg and Helmann (2008); Hachmann et al. (2009)
Listeria moncytogenesLys-PG, Lys-CLGallidermin, HNP1, 2Invasion of epithelial cells; infection in mice Thedieck et al. (2006)
Mycobacterium tuberculosisLys-PGLysozyme, vancomycin, polymyxin BSurvival in macrophages; infection in mice and guinea pigsAltered membrane potentialMaloney et al. (2009)
Pseudomonas aeruginosaAla-PGProtamineN.D.Susceptibility to CrCl3, sodium lactate, cefsulodinKlein et al. (2009)
Staphylococcus aureusLys-PGHNP1-3, LL-37, tachiplesin, protegrins, magainin, indolicidin, gallidermin, nisin, group IIA phospholipase A2, platelet microbicidal proteins, vancomycin, gentamicin, daptomycinSurvival in neutrophils, infection of mice and rabbitsDistinct changes of membrane proteomePeschel et al. (2001); Koprivnjak et al. (2002); Kristian et al. (2003); Nishi et al. (2004); Staubitz et al. (2004); Weidenmaier et al. (2005); Ernst et al. (2009); Sievers et al. (2010)
Rhizobium tropiciLys-PGPolymyxin BNodulation in bean rootsAcid tolerance deficiencyVinuesa et al. (2003); Sohlenkamp et al. (2007)

MprF-mediated resistance towards antimicrobials is not restricted to Staphylococci. In fact, MprF homologues can be found in most bacterial kingdoms and appear to be most abundant in firmicutes, actinobacteria and proteobacteria except for enterobacteria. Interestingly, some archaea also harbour genes for MprF-related proteins that are most related to the proteobacterial proteins and probably result from lateral gene transfer events (Weidenmaier et al., 2005; Roy, 2009; Roy and Ibba, 2009). The prevalence of MprF over millions of years strongly indicates that it has remained an effective defence mechanism against antimicrobial peptides produced by competing microorganisms and animals (Peschel and Sahl, 2006).

MprF proteins are integral membrane proteins with a characteristic structure. A large hydrophobic domain is found at the N-terminus. It consists of 14 transmembrane domains (TMDs) in S. aureus and many other bacteria (Fig. 3) but may contain as few as six TMDs in actinobacteria or may even lack TMDs entirely in some exceptional cases (Roy and Ibba, 2009). The similarity of these N-terminal domains is limited between unrelated bacteria. A well-conserved hydrophilic cytoplasmic domain forms the C-terminus of most MprF proteins (Fig. 3). The two MprF domains do not reveal similarity to proteins of known function.

Figure 3.

Proposed domain structure (top) and topology (bottom) of the S. aureus MprF according to the SOSUI algorithm (http://bp.nuap.nagoya-u.ac.jp/sosui/). The Lys-PG synthase domain is shown in blue. The exact extension of the Lys-PG flippase has not been elucidated yet.

Aminoacyl PG biosynthesis by MprF proteins

Lys-PG and Ala-PG synthase activities were discovered in bacterial crude extracts in the 1960s. The lysyl acceptor and donor substrates turned out to be PG and aminoacyl-tRNAs (Lennarz et al., 1966; Nesbitt and Lennarz, 1968), respectively, and these early findings were recently confirmed with recombinant MprF proteins (Staubitz et al., 2004; Roy and Ibba, 2008, 2009; Klein et al., 2009). Thus, Lys-PG biosynthesis belongs to the few examples of non-ribosomal pathways that employ activated amino acids from aminoacyl tRNAs (RajBhandary and Soll, 2008). MprF represents a new class of transesterase with no detectable relatedness to other transesterases. Various recombinant MprF proteins differ profoundly in specificity for aminoacyl tRNAs. For example, the S. aureus and P. aeruginosa MprFs only synthesize Lys-PG or Ala-PG respectively (Staubitz et al., 2004; Klein et al., 2009). In contrast, Enterococcus faecium MprF2 exhibits rather relaxed specificity for the donor substrate and produces both, Ala-PG and Lys-PG along with small amounts of arginyl PG (Roy and Ibba, 2009). On the other hand, the L. monocytogenes MprF is less strict in its specificity for the acceptor substrate and generates both, Lys-PG and Lys-CL (Thedieck et al., 2006). The molecular basis for strict or broad substrate specificity of MprF proteins are unknown. It should be noted that some studies on MprF substrate specificity are based on heterologous expression in Escherichia coli and experimental evidence for biosynthesis of more than one lipid product in the natural producer has to be obtained in the future. Nevertheless, the impact of different parts of aminoacyl tRNAs for utilization by MprF proteins has been studied to some extent. Truncated lysyl- or alanyl-tRNAs consisting of only the acceptor and T stems were sufficient for synthesis of the corresponding aminoacyl phospholipid indicating that the aminoacyl moiety is the primary determinant of aminoacyl-tRNA recognition by MprF (Roy and Ibba, 2008). However, the tRNA also plays a role in recognition by MprF because alanine coupled to a tRNACys was reported not to be a substrate for Ala-PG synthesis (Gould et al., 1968). How the acceptor substrate PG is recognized and how specific the recognition is remains even less clear. MprF proteins from S. aureus, C. perfringens, Enterococcus faecalis and others produce Lys-PG or Ala-PG in E. coli despite the fact that the fatty acids of PG are quite different between these organisms, indicating that recognition of the acceptor substrate by MprF proteins relies largely on the head group rather than the fatty acids of PG (Staubitz et al., 2004; Sohlenkamp et al., 2007; Ernst et al., 2009; Roy and Ibba, 2009).

The synthesis of aminoacyl phospholipids does not require full-length MprF proteins. The hydrophilic C-terminus represents the actual aminoacyl PG synthase and it seems to constitute an active enzyme in in vitro assays with purified Bacillus subtilis and C. perfringens MprF (Roy and Ibba, 2009). However, in vivo Lys-PG or Ala-PG production can only be observed when a minimal portion of the hydrophobic N-terminal domain of MprF is present (Ernst et al., 2009; Roy and Ibba, 2009). At least six of the 14 TMDs of the S. aureus MprF are necessary to maintain in vivo enzyme activity (Fig. 3) (Ernst et al., 2009). It was suggested that this hydrophobic part might be required for positioning of the enzyme close to the acceptor molecules (Roy and Ibba, 2009). However, it remains unclear why the soluble part of MprF is sufficient for Lys-PG or Ala-PG production in vitro but not in vivo. It is possible that the disturbed membrane properties of membrane vesicles used for in vitro assays render the enzyme more independent of appropriate membrane insertion.

An additional lysyl-tRNA synthetase domain was found to be fused to the Lys-PG synthase domain in certain actinobacterial MprFs and was shown to be essential for Lys-PG synthesis in these enzymes (Maloney et al., 2009). It was speculated that a special lysyl-tRNA synthetase dedicated to Lys-PG synthesis might lead to better supply of donor substrate at the membrane and a tighter regulation of MprF.

Translocation of Lys-PG by the flippase domain of MprF

Many MprF proteins encompass an additional N-terminal hydrophobic domain that is dispensable for Lys-PG biosynthesis. In S. aureus and many other bacteria the N-terminus consists of 14 TMDs, only the last six of which are required for Lys-PG production (Fig. 3) (Ernst et al., 2009). The available software tools for topology prediction of membrane proteins concur largely with the number of TMDs in a given MprF protein but they predict quite different locations of TMDs and connecting loops. Thus, topological investigations are required to determine the most likely structure of MprF.

Only recently it was discovered that the N-terminal hydrophobic part of the S. aureus MprF facilitates Lys-PG transport from the inner leaflet of the membrane where Lys-PG is produced to the outer leaflet where it can exert its role in repulsing CAMPs (Fig. 2) (Ernst et al., 2009). This finding was based on a number of observations that led to the conclusion that the MprF N-terminus represents the first known bacterial phospholipid flippase: (i) While the first eight TMDs of S. aureus MprF were dispensable for full-level Lys-PG production, CAMP resistance was only achieved when the N-terminal domain was present. (ii) The separated N-terminal domain coexpressed in trans with the Lys-PG synthase still led to CAMP resistance indicating that it has a distinct function. (iii) The N-terminal domain conferred CAMP resistance only in the presence of Lys-PG but not alone. (iv) Most of the Lys-PG produced in the absence of the N-terminal domain was found in the inner leaflet of the cytoplasmic membrane and was evenly distributed between the two leaflets only when the N-terminal domain was present. This latter finding was based on covalent modification of Lys-PG by the membrane-impermeable fluorescent dye fluorescamine and subsequent quantification of labelled (outer leaflet) and unlabeled (inner leaflet) Lys-PG. Altogether these data indicated that the S. aureus MprF and probably most of the other MprFs with an extended N-terminal domain represent bifunctional proteins with separable Lys-PG synthase and Lys-PG flippase domains (Ernst et al., 2009).

This interesting observation raises several questions that require more detailed analyses in the future. It is currently not clear if the two domains are somewhat overlapping or not and if they need to interact physically even if expressed in trans. Moreover, it is unclear if the Lys-PG flippase requires energy to mediate asymmetric distribution of Lys-PG between the two leaflets. The N-terminal domain does not contain any amino acid motives known to bind and hydrolyse ATP but energy might come from the proton-motive force. Of note, recent studies differed with respect to a symmetric or asymmetric distribution of Lys-PG (Mishra et al., 2009; Yang et al., 2009a; 2010). More detailed studies will be necessary in the future to answer the question if MprF can only facilitate an even distribution of Lys-PG in the inner and outer leaflet of the membrane or if it can actively generate an asymmetric distribution.

It remains unclear why bacteria require a dedicated flippase for Lys-PG. It is possible that the bacterial (yet unknown) house keeping flippases are specific for ‘standard’ anionic phospholipids while a cationic lipid such as Lys-PG may require a special translocator. Many MprF homologues are lacking the N-terminal extension and appear to represent only aminoacyl-PG synthases. Accordingly, these bacteria probably use other flippase proteins that may be able to translocate Lys-PG or Ala-PG. Of note, M. tuberculosis and related actinomycetes have N-terminally truncated, probably monofunctional MprFs and they encode additional proteins with similarity to the MprF flippase domain lacking the synthase domain. Thus, aminoacyl PG-mediated CAMP resistance appears to depend on either one bifunctional or two monofunctional proteins in different bacteria.

Regulation of MprF expression and aminoacyl PG biosynthesis

The reported percentages of aminoacyl PG in relation to the total amounts of phospholipids are quite different in individual bacterial species and may vary strongly with changing environmental conditions. The Lys-PG content of M. tuberculosis can be as low as 0.3% (Maloney et al., 2009) while that of S. aureus can reach 38% (Short and White, 1971). Nevertheless, the amounts of Ala-PG in P. aeruginosa can change from 0.5% to 6% when growth conditions change from neutral to acidic conditions indicating that the importance of the lipid may vary with the pH and that its biosynthesis is regulated (Klein et al., 2009). Along this line, Lys-PG amounts have been shown to increase during acidification in numerous bacterial species (Vinuesa et al., 2003; Roy, 2009). Moreover, Lys-PG amounts changed with the growth phase in S. aureus with the highest levels in the exponential phase (Ernst et al., 2009) maybe because dividing bacteria are more susceptible to CAMPs than resting bacteria (Kristian et al., 2007). How the production of aminoacyl PG may be controlled, either on the protein level by allosteric regulation of MprF or on the transcriptional level has hardly been studied.

In staphylococci, mprF and two additional CAMP resistance mechanisms encoded by the dltABCD and vraFG genes are upregulated by the Aps/GraRSX system in response to the presence of antimicrobial peptides (Li et al., 2007a,b; Otto, 2009). This system constitutes a two-component system composed of a membrane-inserted sensor histidine kinase (Aps/GraS), a transcriptional response regulator that becomes activated upon phosphorylation by the sensor protein (Aps/GraR), and a third protein of unknown function (Aps/GraX) found exclusively in Staphylococci. The Aps/GraS sensor contains two TMDs and a very short extracellular loop of nine amino acids, which varies in composition between strains (Li et al., 2007a,b). The loop bears several anionic residues and interacts with CAMPs such as hBD3, tachyplesin, nisin, LL-37, magainin, histatin and bervinin, which leads to activation of Aps/GraS (Li et al., 2007b). Of note, the loops of the S. aureus and S. epidermidis Aps/GraS proteins differ in numbers of negatively charged amino acids, which results in different spectra of CAMPs that can activate the sensors in the two species. Interestingly, Aps/GraRSX also upregulates genes involved in lysine biosynthesis (Li et al., 2007a). Thus, regulation by Aps/GraRSX might be a strategy to limit the energetic burden of lysine biosynthesis when antimicrobial peptides are not encountered and lower Lys-PG levels are sufficient for survival.

The Aps/GraRSX system is highly conserved in Gram-positive bacteria including L. monocytogenes, Clostridium difficile, Bacillus anthracis, Staphylococcus haemolyticus and Streptococcus pneumonia (Otto, 2009). In fact, the L. monocytogenes mprF and many other virulence-associated genes belong to a regulon controlled by the Aps/GraRS-related VirRS system (Mandin et al., 2005).

Roles of aminoacyl phospholipids in bacterial physiology and immune evasion

MprF mutants are hardly impaired in growth and exhibit little phenotypic changes under standard laboratory conditions (Peschel et al., 2001; Thedieck et al., 2006; Salzberg and Helmann, 2008; Klein et al., 2009). The absence of one of the major membrane lipids was expected to have a profound impact on the stability and turnover of membrane proteins. However, a recent membrane proteome analysis in S. aureus revealed that the absence of Lys-PG did not lead to a massive alteration but distinct upregulation or downregulation of only 1.5% or 3.5% of quantifiable proteins respectively (Sievers et al., 2010). Lys-PG deficiency had no major impact on the abundance of lipid-biosynthetic enzymes but affected significantly the amounts of cell envelope stress-sensing regulatory proteins such as SaeS and MsrR suggesting that these proteins may interact with Lys-PG (Sievers et al., 2010). A number of additional phenotypes have been described for certain mprF mutants as described in Table 1.

The most conspicuous phenotype found in Lys-PG-deficient bacterial mutants is the strongly increased susceptibility to CAMPs such as host defence peptides (e.g. α and β-defensins, LL-37, platelet-derived antimicrobial peptides) and bacteriocins (e.g. gallidermin, nisin) (Table 1). Depending on the individual antimicrobial peptide and bacterial strain, deletion of mprF led to up to 30-fold decreased minimal inhibitory concentrations (Peschel et al., 2001). In addition, Lys-PG production led to resistance to cationic antimicrobial mammalian enzymes such as lysozme or antimicrobial group IIA phospholipases A2 in M. tuberculosis or S. aureus respectively (Koprivnjak et al., 2002; Maloney et al., 2009). Of note, high or low amounts of Lys-PG do not necessarily correlate with the levels of CAMP resistance beyond a certain threshold level. Accordingly, a recent study demonstrated that even a threefold to fourfold increase of Lys-PG amounts did not lead to increased resistance to CAMPs indicating that a basic level of Lys-PG is sufficient for full-level CAMP resistance (Ernst et al., 2009).

It remains unclear if the presence of Lys-PG or Ala-PG has different consequences for CAMP resistance. While Lys-PG has a positive net charge, Ala-PG is a zwitterionic lipid (Fig. 1) with a less pronounced impact on the net charge of the membrane surface and on CAMP repulsion than Lys-PG. Nevertheless, the fact that bacteria produce Ala-PG either alone or in combination with Lys-PG indicates that it should have a distinct impact on membrane properties and may be a means to fine-tune the charge of the membrane in response to environmental challenges as recently suggested (Roy and Ibba, 2008).

In accord with increased susceptibility to major antimicrobial phagocyte molecules S. aureus mprF mutants are more rapidly inactivated by human neutrophils than the parental strains (Peschel et al., 2001; Kristian et al., 2003). Moreover, the virulence of S. aureus, L. monocytogenes and M. tuberculosis mprF mutants were shown to be attenuated in in vivo infection model (Peschel et al., 2001; Thedieck et al., 2006; Maloney et al., 2009). Reduced virulence was also observed with a S. aureus apsS mutant with impaired expression of mprF (Li et al., 2007a). All these findings indicate that mprF is crucial for survival of major human pathogens in mammalian hosts and may represent a promising target for new antimicrobial strategies. Lys-PG seems to have a similar impact on rhizobia-plant interaction as a Rhizobium tropici mprF mutant exhibited reduced capacity to induce root nodulation in beans (Vinuesa et al., 2003).

MprF point mutations leading to resistance to the antibiotic daptomycin

MprF and Lys-PG are known to affect the susceptibility of S. aureus to cationic antibiotics such as the glycopeptide vancomycin and the aminoglycoside gentamicin (Nishi et al., 2004). Similar observations were recently made for the lipopeptide antibiotic daptomycin (Ho et al., 2008; Jung et al., 2008), which shares cationic properties and a membrane-associated target (currently unknown) with CAMPs upon binding of calcium ions (Straus and Hancock, 2006; Ho et al., 2008; Jung et al., 2008). Daptomycin has recently been approved as an antibiotic of last resort against methicillin- and vancomycin-resistant S. aureus. However, a number of reports on spontaneously daptomycin-resistant S. aureus strains that occurred during therapy have raised the spectre of spreading S. aureus clones with resistance to almost all available antibiotics (Murthy et al., 2008; Baltz, 2009).

Most of the spontaneously daptomycin-resistant S. aureus clones turned out to have point mutations in a particular domain of MprF, which brought this protein into the focus of extensive research activities. Mutations in other genes such as rpoB, rpoC, yycG or altered expression of dltABCD have also been reported (Baltz, 2009) but the point mutations in mprF have been the most often and most consistently found ones. Of note, MprF cannot be the target for daptomycin as mprF deletion mutants are hypersusceptible to daptomycin (Ernst et al., 2009). On the contrary, all the reported mutations appear to lead to a gain of function of the protein that interferes with the ability of daptomycin to inactivate the bacteria. Expression of mprF-specific antisense RNA was recently shown to re-establish susceptibility to daptomycin in daptomycin resistant strains thereby supporting the important role of MprF (Rubio et al., 2010).

Certain point mutations were repeatedly found in independent studies with clinical isolates (Julian et al., 2007; Murthy et al., 2008; Yang et al., 2009b) or with strains cultivated in broth with increasing daptomycin concentrations (Friedman et al., 2006; Kosowska-Shick et al., 2009). Most of them accumulated in a specific region of the N-terminal, hydrophobic part at the junction of the Lys-PG synthase and flippase domains (Fig. 3). It remains to be elucidated if these point mutations lead to increased Lys-PG production, to increased Lys-PG flipping to the outer leaflet of the cytoplasmic membrane or to another change in membrane properties. Some recent studies presented conflicting data that favour either of the possibilities (Jones et al., 2008; Yang et al., 2009a; 2010) indicating that MprF-mediated daptomycin resistance is a complex and maybe multifactorial process. Thus, it remains to be investigated in more detail how MprF leads to daptomycin resistance.

Perspectives

MprF proteins and their lipid products have attracted increasing interest for a number of reasons. First, MprF turned out to be a crucial and well-conserved virulence factor in many major human pathogens, including Gram-positive and Gram-negative bacteria, which raises the possibility to consider MprF as a target for new anti-virulence drugs that would not directly kill bacteria but would render them susceptible to a wide range of antimicrobial host factors (Weidenmaier et al., 2005). There is hope that such therapies would impart less selection pressure and would be less prone to induce resistance in bacterial pathogens (Smith and Romesberg, 2007). For this purpose a thorough understanding of the mode of action and structure of the functional domains of MprF are required.

Second, the discovery that MprF of S. aureus contains a second domain that is indispensable for mediating CAMP resistance and is required for flipping of Lys-PG to the outer leaflet of the cytoplasmic membrane sheds new light on the mechanisms used by bacteria to translocate lipids between membrane leaflets. Bacterial proteins have been implicated in the flipping of cell wall precursor lipid II, of the lipoteichoic acid precursor glucosyldiacylglycerol or the lipopolysacharid intermediate lipid A (van Dam et al., 2007; Grundling and Schneewind, 2007; Ruiz et al., 2009) but a dedicated flippase for phospholipids has so far not been described in bacteria. It remains unclear if MprF translocates Lys-PG in a unidirectional or bidirectional fashion and if it requires energy. In eukaryotes, proteins have been defined that translocate lipids in an ATP-dependent, unidirectional fashion to the exoplasmic leaflet or in a bidirectional, no energy-consuming manner (Daleke, 2007). Most of these studies are based on modification of lipids in the outer leaflets of membranes by enzymes or fluorescent dyes (Pomorski and Menon, 2006) in a similar way as recently used for MprF (Ernst et al., 2009). Only in rare cases flippase function has been studied with purified proteins reconstituted into defined proteoliposomes. Given the difficulties in purifying large membrane proteins such as MprF such strategies remain a particularly challenging task for more detailed investigations on the mode of action of the Lys-PG flippase of MprF.

Third, the involvement of MprF in resistance to daptomycin, an antibiotic of last resort for the treatment of multiresistant staphylococcal and enterococcal infections puts urgency to investigations on how mutations in MprF confer daptomycin resistance. It will be crucial to elucidate if these notorious point mutations lead to altered activity of one of the functional domains of MprF or to other membrane alterations. Answering these questions will also lead to insights into the action of the Lys-PG flippase and synthase domain.

The bacterial cytoplasmic membrane hosts many vital cellular processes and it will be crucial to understand how the variability and plasticity of phospholipids, the major membrane constituents, contribute to bacterial integrity and fitness during changing environmental conditions.

Acknowledgements

We thank our colleagues and collaborators for helpful discussion and continuing support. Our research is funded by grants from the German Research Foundation (SFB766, TR34) and the German Ministry of Education and Research (SkinStaph) to AP.

Ancillary