OUP user menu

MprF-mediated biosynthesis of lysylphosphatidylglycerol, an important determinant in staphylococcal defensin resistance

Petra. Staubitz, Heinz. Neumann, Tanja. Schneider, Imke. Wiedemann, Andreas. Peschel
DOI: http://dx.doi.org/10.1016/S0378-1097(03)00921-2 67-71 First published online: 1 February 2004


Frequently bacteria are exposed to membrane-damaging cationic antimicrobial molecules (CAMs) produced by the host's immune system (defensins, cathelicidins) or by competing microorganisms (bacteriocins). Staphylococcus aureus achieves CAM resistance by modifying anionic phosphatidylglycerol with positively charged l-lysine, resulting in repulsion of the peptides. Inactivation of the novel S. aureus gene, mprF, which is found in many bacterial pathogens, has resulted in the loss of lysylphosphatidylglycerol (L-PG), increased inactivation by CAM-containing neutrophils, and attenuated virulence. We demonstrate here that expression of mprF is sufficient to confer L-PG production in Escherichia coli, which indicates that MprF represents the L-PG synthase. L-PG biosynthesis was studied in vitro and found to be dependent on phosphatidylglycerol and lysyl-tRNA, two putative substrate molecules. Further addition of cadaverin, a competitive inhibitor of the lysyl-tRNA synthetases, or of RNase A abolished L-PG biosynthesis, thereby confirming the involvement of lysyl-tRNA. This study forms the basis for further detailed analyses of L-PG biosynthesis and its role in bacterial infections.

  • Staphylococcus aureus
  • Defensin
  • Innate immunity
  • Phospholipid

1 Introduction

The main constituents of bacterial membranes are phospholipids whose composition and relative abundance vary profoundly both between species and under various environmental conditions and growth phases [1,2]. The most common bacterial phospholipids are phosphatidylglycerol (PG) and diphosphatidylglycerol (cardiolipin) whose head groups are negatively charged [3]. The zwitterionic phosphatidylethanolamin (PE) is found, for instance, in Enterobacteriaceae [4] and bacilli [5] but is absent in several bacterial genera such as Staphylococcus[6] and Listeria[7]. Bacterial membranes usually have a profound negative net charge, which distinguishes them from eucaryotic cytoplasmic membranes. Most of the antimicrobial molecules from the human, vertebrate, invertebrate, and even plant host defense systems have cationic properties (cationic antimicrobial molecules, CAMs) that confer a high affinity for the anionic bacterial cell envelope [8]. Many of these CAMs kill bacteria by damaging their membranes either by forming pores, as in the case of defensins and similar peptides [9], or by lipid degradation, as demonstrated for the antimicrobial class IIA phospholipase A2 [10]. Several bacterial strains employ the same concept as eucaryotic cells by producing defensin-like CAMs (bacteriocins) to limit the growth of competing microorganisms in nutrient-poor habitats [11].

Some bacteria modify anionic phospholipids with l-lysine to produce lysylphosphatidylglycerol (L-PG), which imparts a positive net charge onto the cytoplasmic membrane [1,12]. L-PG has been described in several bacterial pathogens, such as Staphylococcus aureus[6] and Listeria monocytogenes[7], or in some soil organisms such as Bacillus subtilis[13]. The role of L-PG production and its molecular basis however, have remained illusive. Recently, an S. aureus mutant lacking L-PG has been identified in a screening for defensin-susceptible transposon mutants [12,14]. The inactivated gene, mprF, encodes a large membrane protein without similarity to proteins of known function. mprF mutants have a higher affinity for cationic peptides, which indicates that lysinylation of phospholipids leads to repulsion of defensin-like peptides. Accordingly, the L-PG-deficient mutants exhibited increased susceptibilities to many CAMs and were inactivated much faster by defensin-producing neutrophils [15,16]. The impact of L-PG-mediated defensin resistance was confirmed by attenuated virulence of the mutant in mice [12]. L-PG deficiency had further consequences, such as increased susceptibility to the glycopeptide antibiotic vancomycin [17] and altered activity of the membrane-associated DnaA protein, which determines the rate of DNA replication initiation [18].

Proteins with significant similarity, in terms of structure and sequence, to MprF have been identified in a great number of Gram-positive and Gram-negative microbial genomes from human, animal, and plant pathogens [8]. Only in a small number of them, however, has the production of lysinylated phospholipids been documented and it is unclear whether these mprF-related genes have similar roles to those seen in S. aureus. Moreover, it has remained uncertain whether MprF represents the L-PG synthase or whether it is only one of several proteins involved in the biosynthetic pathway.

In an attempt to characterize MprF function further we expressed mprF in Escherichia coli. This resulted in L-PG production, indicating that in a heterologous host MprF is sufficient for phospholipid lysinylation and represents the L-PG synthase. Moreover, we determined the substrate requirements for L-PG biosynthesis.

2 Materials and methods

2.1 Bacterial strains, plasmids, and growth conditions

S. aureus SA113 (ATCC 35556) [19] and E. coli DH5α[20] are commonly used laboratory strains. They were grown in LB broth (1% tryptone, 0.5% yeast extract, 0.5% NaCl) or BM broth (LB supplemented with 0.1% K2HPO4 and 0.1% glucose) unless otherwise noted. A DNA fragment of 2523 bp encoding the mprF gene from S. aureus SA113 was amplified by polymerase chain reaction (PCR) and inserted into pUC19 downstream of the tac promoter. Primers for amplification of mprF (GATTTATAACAGAAAGGATCCGAGGAGGTGTGAAAAAATGAATCAGGA, CGTTTTTTATAAAGAATTCTTTATAGAAACCAAAAAATC) contained suitable restriction sites for cloning in the plasmid polylinkers. The upstream primer was also modified to optimize the Shine Dalgarno sequence. The PCR-derived DNA fragment was sequenced to confirm a correct sequence. Standard methods for molecular cloning, transformation, and sequencing were used [20].

2.2 Isolation and detection of lipids by thin-layer chromatography (TLC)

Staphylococci were grown overnight in BM broth containing 0.25% glucose for 14 h, washed once in sodium acetate buffer (20 mM, pH 4.5), and disrupted in the same buffer using glass beads and a Disintegrator S (Biomatic GmbH, Rodgau, Germany) as described recently [12]. E. coli cells were grown overnight in LB for 14 h, washed once in sodium acetate buffer, and disrupted in the same buffer with a Sonifier B12 (Branson, Danbury, CT, USA). Polar lipids were extracted by the Bligh-Dyer procedure [21], vacuum-dried, and dissolved in chloroform/methanol (2:1, by vol.). TLC was carried out as described previously [6]. Briefly, appropriate amounts of lipid extracts were spotted onto silica 60 F254 HPTLC plates (Merck, Darmstadt, Germany) and developed with chloroform/methanol/water (65:25:4, by vol.). All glycerolipids were visualized with molybdatophosporic acid spray (Merck, Darmstadt, Germany) followed by charring at 120°C and treatment with ammonia vapor to improve the contrast. Phospholipids or amino group-containing lipids were selectively stained with Molybdenum Blue (Sigma, St. Louis, MO, USA) or ninhydrin spray (0.3 g ninhydrin dissolved in 100 ml 1-butanol and 3 ml 100% acetic acid), respectively.

2.3 In vitro L-PG biosynthesis

In order to isolate tRNAs, S. aureus wild-type was grown in 3 l of LB broth until OD578 of 1.5–2 was reached and harvested. Bacteria were resuspended in 20 ml water, added to 50 ml boiling water, and boiled for 15 min. Insoluble material was removed by centrifugation. The bacterial lysate was vacuum-dried and dissolved in 10 ml water. Aliquots (1 ml) were mixed with 2.5 ml QRL1 buffer from an RNA/DNA kit (Quiagen, Hilden, Germany) and the tRNAs were further purified on columns from the kit and precipitated, as recommended by the supplier. Crude cell-free extracts from the S. aureus wild-type and the mprF mutant grown to OD578 of 1.5 in BM broth were prepared, as described previously [22] using the FP120 cell disintegrator (Thermo Savant, Holbrook, NY, USA). Bacteria were suspended and disrupted in Tris–maleate buffer (100 mM Tris, pH 7) on ice. Residual intact cells were removed by centrifugation (2000×g, 15 min, 4°C). The protein content in cell extracts was determined using the DC Protein-Assay kit (Bio-Rad Laboratories, Hercules, CA, USA). The basic compounds included in the reaction mixtures were 100 mM Tris–maleate buffer, 0.625 mM MgCl2, 5 mM ATP, 31.5 μM 14C-labeled l-lysine (318 μCi μmol−1, Amersham, Little Chalfont, UK), 15 μg PG (Sigma), 1.7–1.8 mg protein from the cell lysate, and 24 μg S. aureus tRNAs in a final volume of 200 μl. In some experiments, the concentration of certain ingredients was varied and the potential inhibitor cadaverin or RNase A (Qiagen) was added, as noted in the individual figure legends. Samples were incubated without 14C l-lysine for 15 min at 37°C, followed by 1 h at 30°C in the presence of 14C l-lysine.

Reactions were stopped by addition of 400 μl 2 M NaCl in 0.01 M HCl, 600 μl methanol, and 600 μl chloroform. Lipids were extracted as described above and the organic phase was washed once with 0.9% NaCl. The combined organic phases were analyzed for radioactivity in a scintillation counter or vacuum-dried and separated on TLC plates, as described above. Radioactive lipids were visualized by autoradiography using X-ray films.

3 Results and discussion

3.1 MprF is sufficient to confer L-PG synthesis in E. coli

In order to characterize the function of MprF in a heterologous host, we expressed the S. aureus mprF gene in E. coli. The available enterobacterial genome sequences, including that of E. coli, are lacking mprF-related genes. Accordingly, E. coli does not produce L-PG. Nevertheless, it contains considerable amounts of PG, the putative L-PG precursor. The S. aureus mprF was cloned in plasmid pUC19 under the control of the tac promoter. E. coli DH5α (pUCmprF) produced a new lipid that comigrated in TLC with L-PG from S. aureus (Fig. 1). Positive staining with ninhydrin and Molybdenum Blue confirmed the presence of amino and phosphate groups. Taken together, these results demonstrate that MprF is necessary and sufficient to produce L-PG. Although we cannot rule out that unknown E. coli proteins could contribute to L-PG synthesis it is very likely that MprF represents the L-PG synthase.

Figure 1

L-PG production in mprF-expressing E. coli. Polar lipids from S. aureus wild-type (1), E. coli without plasmid (2), with pUC19 (3), or with pUC-mprF (4) were extracted and separated by TLC. Amino groups containing lipids were visualized with ninhydrin. The positions of L-PG and of PE (only in E. coli) are indicated. Small amounts of two unidentified lipids were also detected.

L-PG-producing E. coli strains were as susceptible to CAMs as E. coli without L-PG (data not shown), which is in contrast to the large difference in CAM susceptibility of S. aureus with and without MprF. One reason for this lack of phenotype in E. coli may be the relatively low L-PG amounts produced.

3.2 L-PG synthesis requires PG and lysyl-tRNA

Synthesis of L-PG is likely to require energy in the form of ATP or an activated form of lysine. Studies from the 1960s have provided evidence that the lysine group is derived from a lysyl-tRNA and transferred to PG [23]. In order to investigate the substrate requirements in MprF-mediated L-PG synthesis, we monitored incorporation of 14C-labeled l-lysine into the lipid fraction using crude cell extracts from S. aureus wild-type or mprF mutant in the presence of putative substrates and possible inhibitors. Radioactivity was only incorporated by the cell extracts from S. aureus wild-type and not by those from the mprF-deficient mutant. The amount of incorporated lysine increased with the amount of cell extract added (Fig. 2A). TLC analysis revealed only one radioactive lipid spot, which exhibited the same migration behavior as L-PG (Fig. 2B) indicating that the reaction had, in fact, yielded L-PG. Efficient in vitro L-PG synthesis depended on the presence of PG as omission of PG resulted in 97% reduced L-PG production (Fig. 2C).

Figure 2

Dependence of L-PG synthesis on PG and lysyl-tRNA. A: Increasing amounts of cell-free extracts of S. aureus wild-type (black circles) or ΔmprF (open circles) were incubated in the presence of 14C-labeled l-lysine, PG, purified tRNAs, and ATP. Samples were extracted subsequently with methanol and chloroform and the radioactivity incorporated into the organic phase was determined by liquid scintillation. B: Autoradiograph of a TLC plate on which lipids from the organic phase had been separated. C: In vitro L-PG production in the presence or absence of PG or tRNAs. D: In vitro L-PG production after addition of RNase A or of the lysyl-tRNA synthetase inhibitor cadaverin. In samples with RNase A, purified tRNAs were omitted. Means together with S.D. of at least three independent experiments are shown in A, C, and D.

In order to confirm that a lysyl-tRNA was the source of the activated lysine, tRNAs purified from S. aureus were added to the reaction mixture. The required lysyl-tRNA synthetase was assumed to be present in the cell lysate. Addition of tRNA resulted in a slight increase in L-PG production (Fig. 2C). Since a tRNA would be required only in catalytical amounts very small concentrations as present in the lysate should be sufficient for efficient L-PG synthesis (Fig. 3). When RNase A, which degrades tRNAs very quickly and efficiently (data not shown), was added, L-PG synthesis was no longer detectable (Fig. 2D), thereby providing strong evidence for the involvement of a tRNA. Addition of cadaverin, a competitive inhibitor of lysyl-tRNA synthetases [24], led to a dose-dependent inhibition of L-PG synthesis, which is in agreement with the notion that a lysyl-tRNA is an essential substrate in L-PG biosynthesis.

Figure 3

Putative pathway of l-lysine transfer into PG. The staphylococcal cell envelope, composed of a thick cell wall (light gray) and the plasma membrane (dark gray) is shown. While PG is negatively charged and favors interactions with positively charged defensins or other CAMs, L-PG is cationic and repulses CAMs from the membrane surface. MprF confers L-PG biosynthesis by modification of PG with l-lysine (l-Lys) to produce L-PG. The lysine group is most probably derived from a lysyl-tRNA.

The use of aminoacyl tRNAs as activated precursors for the incorporation of amino acids into biomolecules is a rare and unusual feature but it is not unique to L-PG synthesis. The inter-peptide bridges in staphylococcal peptidoglycans are formed by five glycine residues, which are derived from a glycyl-tRNA [25]. In this process, a dedicated tRNA, which is not a substrate in protein synthesis, is employed. In contrast, the genome sequence of S. aureus does not reveal an additional lysine-specific tRNA, suggesting that both the L-PG and protein synthesis pathways use the same lysyl-tRNA.

A detailed analysis of the mechanism of L-PG synthesis may help finding inhibitors of MprF, which would render a great number of bacterial pathogens susceptible to human host defenses and could help to combat multiply resistant, L-PG-producing bacteria such as S. aureus, Pseudomonas aeruginosa, or Enterococcus faecalis.


We thank Friedrich Götz, Hans-Georg Sahl, Ralph Jack, and Willem P. Nieuwenhuizen for support and helpful discussions; Rafik Oueslati, Gabriele Hornig, and Nicole Schatlowski for technical assistance, and Helen Hölscher for reading the manuscript. This work was supported by grants from the German Research Council (FOR 449/T2) and from the University of Tubingen to A.P.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
View Abstract