Enterococcus faecium clinical isolate BM4524, resistant to vancomycin and susceptible to teicoplanin, harboured a chromosomal vanB cluster, including the vanSB / vanRB two-component system regulatory genes. Enterococcus faecium strain BM4525, isolated two weeks later from the same patient, was resistant to high levels of both glycopeptides. The ddl gene of BM4525 had a 2 bp insertion leading to an impaired d -alanine: d -alanine ligase. Sequencing of the vanB operon in BM4525 also revealed an 18 bp deletion in the vanSB gene designated vanSBΔ . The resulting six amino acid deletion partially overlapped the G2 ATP-binding domain of the VanS BΔ histidine kinase leading to constitutive expression of the resistance genes. Sequence analysis indicated that the deletion occurred between two tandemly arranged heptanucleotide direct repeats, separated by 11 base-pairs. The VanS B , VanS BΔ and VanR B proteins were overproduced in Escherichia coli and purified. In vitro autophosphorylation of the VanS B and VanS BΔ histidine kinases and phosphotransfer to the VanR B response regulator did not differ significantly. However, VanS BΔ was deficient in VanR B phosphatase activity leading to accumulation of phosphorylated VanR B . Increased glycopeptide resistance in E. faecium BM4525 was therefore a result of the lack of production of d -alanyl- d -alanine ending pentapeptide and to constitutive synthesis of d -alanyl- d -lactate terminating peptidoglycan precursors, following loss of d -alanine: d -alanine ligase and of VanS B phosphatase activity respectively. We suggest that the heptanucleotide direct repeat in vanSB may favour the appearance of high level constitutively expressed vancomycin resistance through a ‘slippage’ type of genetic rearrangement in VanB-type strains.
The C-terminal d-alanyl-d-alanine (d-Ala-d-Ala) residues of bacterial pentapeptide cell wall precursors are the target for glycopeptide antibiotics, such as vancomycin and teicoplanin, which inhibit the transglycosylation and transpeptidation reactions in peptidoglycan assembly (Reynolds, 1989). The d-Ala residues are incorporated into peptidoglycan precursors as a dipeptide synthesized by the Ddl d-Ala:d-Ala ligases.
Acquired resistance to glycopeptides in enterococci by synthesis of modified peptidoglycan precursors ending in the depsipeptide d-alanyl-d-lactate (d-Ala-d-Lac) instead of the dipeptide d-Ala-d-Ala (Arthur et al., 1996a) can be classified in three types, VanA, VanB and VanD, depending upon the level of resistance to vancomycin and susceptibility or resistance to teicoplanin (Arthur et al., 1996a; Depardieu et al., 2003). VanA-type strains display high-level inducible resistance to both vancomycin and teicoplanin whereas VanB-type strains have variable levels of inducible resistance to vancomycin only as teicoplanin is not an inducer (Quintiliani et al., 1993; Arthur et al., 1996a). VanD-type strains are characterized by constitutive resistance to moderate levels of the two glycopeptides (Depardieu et al., 2003).
The organization of the vanA, vanB and vanD operons is similar (Arthur et al., 1996a; Evers and Courvalin, 1996; Casadewall and Courvalin, 1999; Depardieu et al., 2003). Three proteins are required for glycopeptide resistance: a dehydrogenase (VanH, VanHB or VanHD) to reduce pyruvate to d-Lac, a ligase (VanA, VanB or VanD) to synthesize the depsipeptide d-Ala-d-Lac, and a d,d-dipeptidase (VanX, VanXB or VanXD) to hydrolyse the d-Ala-d-Ala dipeptide synthesized by the host Ddl ligase (Arthur et al., 1996a). In VanA- and VanB-type strains, a d,d-carboxypeptidase (VanY or VanYB) contributes to vancomycin resistance by hydrolyzing the C-terminal d-Ala residue of late peptidoglycan precursors, when elimination of d-Ala-d-Ala by VanX is incomplete (Arthur et al., 1998).
Expression of VanA- and VanB-type resistance is regulated by the VanS/VanR two-component signal transduction system, composed of a membrane-bound histidine kinase (VanS or VanSB) and a cytoplasmic response regulator (VanR or VanRB) that acts as a transcriptional activator (Arthur et al., 1992; Evers and Courvalin, 1996). Several members of the histidine protein kinase family are known to act as phosphatases, catalysing the dephosphorylation of the associated response regulator. Regulation by environmental signals may therefore take place by modulating either the kinase or the phosphatase activity of the sensor. Accumulation of the phosphorylated form of the response regulator is thus the end result of a series of reversible enzymatic reactions: autophosphorylation of the kinase, autophosphatase activity of the kinase, phosphotransfer to the regulator, backtransfer to the kinase, autophosphatase activity of the regulator and phosphoprotein phosphatase activity of the kinase towards the regulator (Hoch and Silhavy, 1995; Inouye and Dutta, 2003).
Interestingly, a comparison of the vanA and vanB gene clusters reveals that although the dehydrogenases, ligases and d,d-dipeptidases share high levels of amino acid sequence identity, the VanS/VanR and VanSB/VanRB two-component systems are only distantly related (Arthur et al., 1996a). The two response regulators are 22% identical whereas the histidine kinases have only 13% sequence identity, with unrelated amino-terminal sensing domains. VanS has been shown in vitro to display both kinase and phosphatase activity towards VanR (Wright et al., 1993) and a compelling body of genetic evidence indicates that this is likely to also be the case for VanSB, although it has not been demonstrated biochemically (Baptista et al., 1997; Arthur et al., 1999). Under inducing conditions, the sensor is thought to act primarily as a kinase, leading to phosphorylation of the response regulator and activation of the PR or PRB promoters. This allows transcription of the regulatory (vanRS, vanRBSB) genes and of the PH or PYB promoters for the resistance (vanHAX, vanHBBXB) genes (Arthur et al., 1997; 1999). Because VanS is required for negative regulation of the PH and PR promoters in the absence of glycopeptides, the phosphatase activity is thought to predominate under non-inducing conditions, preventing accumulation of the phosphorylated form of the response regulator (Arthur et al., 1997; 1999).
Although VanB-type strains do not display teicoplanin resistance, mutations in the vanSB sensor gene were obtained in vitro, following selection on teicoplanin, that led to constitutive or teicoplanin-inducible expression of the resistance genes (Baptista et al., 1997) and in vivo in animal models (Aslangul et al., 1997). Derivatives of VanB-type strains resistant to teicoplanin have been isolated from two patients, following treatment with vancomycin (Hayden et al., 1993) or teicoplanin (Kawalec et al., 2001) but the isolates were not studied further.
We report the study of VanB-type Enterococcus faecium BM4524 resistant to vancomycin and susceptible to teicoplanin and of E. faecium BM4525, isolated from the same patient two weeks later, displaying constitutive resistance to high levels of both vancomycin and teicoplanin. We demonstrated that purified VanSB acts as both a histidine protein kinase and a VanRB phosphoprotein phosphatase in vitro. We also showed that increased glycopeptide resistance in BM4525 is due to the combination of a frameshift mutation leading to the loss of Ddl ligase activity and to constitutive synthesis of pentadepsipeptide precursors secondary to the loss of VanSB phosphatase activity following deletion of a six amino acid stretch which partially overlaps the conserved G2 ATP-binding domain.
Enterococcus faecium BM4524 and BM4525 harbour the vanB operon
Two vancomycin-resistant E. faecium strains, BM4524 and BM4525, were isolated from the same patient at a two week interval. Strain BM4524 was resistant to vancomycin only (Table 1) whereas BM4525 was resistant to very high levels of both vancomycin and teicoplanin (Table 1). The two strains had indistinguishable pulse field gel electrophoresis (PFGE) patterns after digestion of total DNA by SmaI or NotI (data not shown). To explain why the phenotypes of the two strains were different, the glycopeptide resistance genotype of BM4524 and BM4525 was determined by amplification with primers specific for vanA and vanB genes that are detected most frequently. Surprisingly, both strains contained the vanB gene cluster, a genotype which usually confers resistance to variable levels of vancomycin but not to teicoplanin (Quintiliani et al., 1993; Arthur et al., 1996a). To test the presence of other vancomycin resistance genes, PCR amplification was performed with DNA from strains BM4524 and BM4525 using primers specific for the vanD, vanC-1, vanC-2, vanC-3 and vanE genes. No products were obtained indicating that only the vanB resistance gene was present (data not shown).
Table 1. . MICs of glycopeptides and cytoplasmic peptidoglycan precursors synthesized.
a. MICs were determined by the method of Steers et al. (1959 ). Vm, vancomycin; Te, teicoplanin.
. Peptidoglycan synthesis was inhibited by adding ramoplanin to the cultures for 15 min.
Uninduced Vm (4 µg ml−1)
Uninduced Vm (4 µg ml−1)
Based on sequence differences, the vanB gene cluster can be divided into three subtypes: vanB-1, -2 and -3 (Patel et al., 1998; Dahl et al., 1999). The sequence of the vanB gene in strains BM4524 and BM4525 was identical to vanB-2 (Gold et al., 1993) and the 11 point mutations and 5 bp deletion in the vanB-2 operon were also present in the 309 bp region spanning the 175 bp vanSB-vanYB intergenic region and adjacent coding sequences in BM4524 and BM4525 (Dahl et al., 1999).
The vanB gene cluster was assigned to a chromosomal fragment of ∼ 600 kb in strains BM4524 and BM4525 by contour-clamped homogeneous electric field gel electrophoresis of total DNA digested with I-CeuI followed by successive hybridization with rrs and vanB probes (data not shown).
Organization of the vanB operons in BM4524 and BM4525
The amplification products obtained following PCR mapping of BM4524 and BM4525 genomic DNA with primers complementary to E. faecalis V583 prototype vanB locus (Evers and Courvalin, 1996) were of the expected size, indicating that the vanB gene cluster was present with an organization identical to that in strain V583, i.e. as two adjacent operons, one containing the regulatory genes, vanRBSB, the other the vanYBWHBBXB resistance genes (Fig. 1). No large insertions or deletions in the-non-coding regions were detected. However, no PCR products were obtained when one of the primers used for amplification was complementary to the region upstream from vanRB or downstream from vanXB (Fig. 1), indicating that the genomic environment of the vanB cluster in strains BM4524 and BM4525 differs from that in V583.
Transfer of VanB-type glycopeptide resistance is associated with the movement of large conjugative elements, such as Tn1549 (Garnier et al., 2000). PCR mapping of the regions flanking the vanB cluster in both strains, carried out with primers complementary to Tn1549, indicated that they were similar to this transposon except for a 1.418-kb insertion downstream from vanXB that displayed 56% identity with ORFB of IS150 from Escherichia coli (Fig. 1). This insertion sequence was recently described as being present in several vanB2-type Enterococcus faecium strains (Lee and Kim, 2003). However, attempts to transfer vancomycin resistance from strain BM4525 to E. faecalis JH2-2 or to E. faecium BM4107 recipients by filter mating were unsuccessful, as were all efforts to transform strains BM4524 or BM4525 with plasmid DNA by electroporation or conjugation (data not shown). The deduced amino acid sequences of VanRB, VanB, VanYB and VanXB were identical in strains BM4524 and BM4525 and displayed 98–100% identity with those of the corresponding proteins in Tn1549 (Fig. 1).
E. faecium BM4525 produces a non-functional d-Ala:d-Ala ligase
In an attempt to explain the unusual high-level resistance to glycopeptides of strain BM4525, the chromosomal ddl genes from BM4524 and BM4525 were amplified by PCR and sequenced. The ddl gene of BM4525 had a 2 bp TA insertion at the 3′ end (position 962 with respect to the translational start site) resulting in a frameshift mutation and the creation of a TGA stop codon at position 991. This mutation would be expected to lead to the synthesis of a 330 amino acid truncated peptide instead of the putative full-length 358 amino acid Ddl. The determination of the nature and relative amounts of the peptidoglycan precursors produced by the two strains suggests that the d-Ala:d-Ala ligase is indeed inactive in strain BM4525, as the strain synthesized pentadepsipeptides exclusively (Table 1).
When E. faecium BM4524 was grown in the absence of vancomycin, UDP-MurNAc-pentapeptide was the unique precursor synthesized whereas, after incubation with vancomycin (4 µg ml−1), UDP-MurNAc-pentadepsipeptide was the main compound produced (Table 1). In contrast, E. faecium BM4525 only synthesized pentadepsipeptides, even in the absence of induction, indicating constitutive peptidoglycan precursor synthesis directed by the vanB operon. Synthesis of d-Ala-d-Lac ending precursors and almost total elimination of d-Ala-d-Ala terminating peptidoglycan precursors are necessary and sufficient for high-level vancomycin and teicoplanin resistance (Arthur et al., 1996b). Accordingly, increased glycopeptide resistance in strain BM4525 was associated with increased synthesis of UDP-MurNAc-pentadepsipeptide to the detriment of UDP-MurNAc-pentapeptide, an apparent loss of activity of the host d-Ala:d-Ala ligase and constitutive expression of the resistance genes.
d,d-peptidase activities in strains BM4524 and BM4525
Expression of the resistance genes was studied by analysis of the VanXBd,d-dipeptidase and VanYBd,d-carboxypeptidase activities, which were assayed by determining the amount of d-Ala released from hydrolysis of the dipeptide d-Ala-d-Ala and pentapeptide UDP-Mur-NAc-l-Ala-γ-d-Glu-l-Lys-d-Ala-d-Ala respectively (Table 2). Determination of d,d-dipeptidase specific activity provides a direct estimate of the capacity of a strain to hydrolyse d-Ala-d-Ala and also an indirect estimate for synthesis of d-Ala-d-Lac, because the resistance genes are co-regulated at the transcriptional level (Arthur et al., 1996b; 1997).
Table 2. .d , d -dipeptidase (VanX) and d , d -carboxypeptidase (VanY) activities in extracts from E. faecium BM4524 and BM4525.
The d,d-dipeptidase activity was measured in the 100 000 g cytosolic fraction of bacterial crude extracts following growth in the presence of various concentrations of vancomycin or teicoplanin as an inducer (Table 2). In BM4524, synthesis of d,d-dipeptidase was inducible by vancomycin only, in a concentration-dependent manner. The lack of induction by teicoplanin accounted for dissociated resistance to vancomycin only in this strain. In BM4525, the d,d-dipeptidase was constitutively synthesized at a moderate level, similar to that obtained for BM4524 after induction with the highest concentration of vancomycin tested. Similar results were obtained for d,d-carboxypeptidase activity in membrane preparations (Table 2), which was inducible by vancomycin only in BM4524 and constitutive in BM4525 and UDP-MurNAc-tetrapeptide was detected in both strains (Table 1).
Transcriptional analysis of the vanB gene cluster
As indicated above, production of depsipeptide peptidoglycan precursors was constitutive in strain BM4525. Expression of the vanB operon was examined by primer extension in E. faecium BM4524 and BM4525 during growth in the presence or absence of vancomycin. As shown in Fig. 2, the vanB operon transcription initiation site was located 48 base pairs upstream from the translation initiation codon of vanYB. The vanB operon was expressed in strain BM4524 only in the presence of vancomycin (Fig. 2, lanes 3 and 4) whereas in strain BM4525 the operon was expressed constitutively (Fig. 2, lanes 1 and 2) confirming that expression of glycopeptide resistance is inducible in BM4524 and constitutive in BM4525.
The position of the transcriptional start site of PYB in both strains was the same as in VanB-type E. faecalis V583 (Evers and Courvalin, 1996). The sequence of the upstream region revealed potential − 35 and − 10 regions sharing similarities with the consensus sequence of promoters recognized by the vegetative form of RNA polymerase, EσA (AGGGAA N17 CATAAT), although the − 35 sequence was poorly conserved, in agreement with the fact that the promoter is positively regulated (Fig. 2).
A deletion in the VanSB sensor of strain BM4525
VanB-type strains with an impaired host d-Ala:d-Ala ligase require vancomycin for growth, as d-Ala-d-Lac is only produced under inducing conditions. VanB-type mutants selected by glycopeptides for teicoplanin resistance have amino acid substitutions in the amino-terminal domain of VanSB or in the vicinity of histidine residue 233, corresponding to the putative autophosphorylation site, that confer, respectively, teicoplanin-inducible or constitutive expression of the resistance genes (Baptista et al., 1997). The vanSB genes of E. faecium BM4524 and BM4525 were amplified by PCR and sequenced. No mutations were found in vanSB of strain BM4524 relative to the V583 prototype (Evers and Courvalin, 1996). However, the nucleotide sequence of the vanSB gene from strain BM4525 revealed an 18 bp deletion removing codons 401–406 (RSRKSG) in VanSB, now referred to as VanSBΔ. VanSB is a Class I histidine kinase, with the typically conserved H, N, G1, F and G2 boxes (Dutta et al., 1999), and the six amino acid deletion partially overlapped the conserved G2 motif which is part of the ATP-binding domain (Fig. 3). Sequence analysis indicated that the deletion occurred between two tandem direct repeats of seven base pairs, separated by 11 base-pairs (Fig. 3). Production of the VanSBΔ truncated sensor was thus associated with constitutive expression of glycopeptide resistance in strain BM4525.
Overproduction and purification of VanSB and VanSBΔ
To determine if the six amino acid deletion in VanSBΔ affected the autophosphorylation, phosphotransfer or phosphatase activity of the histidine kinase, the VanSB, VanSBΔ and VanRB proteins were overproduced and purified. The vanRB coding sequence was cloned in plasmid pET28/16, creating a translational fusion adding six histidine residues to the carboxy-terminus of VanRB and placing the gene under the control of an inducible T7 bacteriophage promoter.
Histidine kinases of the EnvZ family have two amino-terminal transmembrane domains. In VanSB, these are located between residues 10–30 and 136–153, as determined using the algorithm of Claros and von Heijne (1994). To overproduce the isolated cytoplasmic histidine kinase domains of VanSB (residues M159 to L447) and VanSBΔ (residues M159 to L441), the corresponding DNA fragments were cloned in pET28/16, adding six histidine residues to the carboxy-terminus of the proteins. SDS-PAGE analysis showed overproduced bands in the soluble fraction of crude extracts from cells carrying plasmids pAT816(pET28ΩvanSB), pAT817(pET28ΩvanSBΔ) and pAT818(pET28ΩvanRB), which were absent in the control extract from cells carrying the pET28/16 vector (Fig. 4, lanes 3, 5, 7 and 2 respectively). The overproduced bands corresponding to VanSB and VanSBΔ displayed the expected apparent molecular masses (33.3 kDa and 32.7 kDa, respectively; Fig. 4, lanes 3 and 5) whereas the recombinant VanRB protein migrated slightly higher than the expected size (approximately 31 kDa instead of the deduced molecular mass of 26.2 kDa; Fig. 4, lane 7). Such an anomalous electrophoretic migration has previously been reported for other response regulators such as DegU or YycF (Dahl et al., 1991; Howell et al., 2003).
The His-tagged VanRB, VanSB and VanSBΔ proteins were purified in a single step using a Ni-NTA agarose column yielding 14, 12 and 12 mg of purified protein per litre of cell culture respectively. SDS-PAGE analysis revealed a purity greater than 95% (Fig. 4, lanes 4, 6, and 8).
Autophosphorylation of VanSB and VanSBΔ and phosphotransfer to VanRB
Autophosphorylation of VanSB and VanSBΔ was performed by incubation of the purified proteins with [γ-32P]-ATP followed by SDS-PAGE and autoradiography (Fig. 5A and B, lanes 1). Autophosphorylation of the two kinases was equally efficient (Fig. 5A and B, lanes 1) and took place in a time-dependent manner, reaching an equilibrium after approximately 60 min (data not shown). Phosphate transfer was inhibited by EDTA (data not shown). These results indicate that the six amino acid deletion in VanSBΔ, which partially overlaps the G2 ATP binding domain, did not affect autophosphorylation of the histidine kinase as compared to intact VanSB.
Purified VanRB was incubated with purified [32P]-VanSB or [32P]-VanSBΔ and phosphotransfer to VanRB was analysed following SDS-PAGE and autoradiography (Fig. 5A and B, lanes 2–7). Rapid phosphotransfer to VanRB was observed with both histidine kinases although it occurred slightly more efficiently from VanSB than from VanSBΔ (Fig. 5A and B, lanes 2–7). This suggests that constitutive glycopeptide resistance of strain BM4525 may be due to a difference in phosphatase activity between VanSB and VanSBΔ rather than a difference in kinase activity.
Phosphatase activity of VanSB and VanSBΔ
Numerous response regulators can be phosphorylated in vitro using low molecular weight phospho-donors such as acetyl phosphate (Lukat et al., 1992). In order to generate large amounts of radioactively labelled phosphorylated VanRB in the absence of its cognate histidine kinase, the purified response regulator was incubated in the presence of acetyl-[32P]-phosphate. Phosphorylation of VanRB by acetyl-[32P]-phosphate was slow, reaching equilibrium in approximately 60 min (data not shown).
The phosphatase activities of VanSB and VanSBΔ were compared by determining hydrolysis of phosphorylated VanRB over time in the presence or absence of the histidine kinases. Isolated [32P]-VanRB was extremely stable in vitro(Fig. 6A) with a calculated t1/2 = 150 min (Fig. 7). Addition of purified VanSB strongly accelerated decomposition of [32P]-VanRB (Fig. 6B), by approximately 11-fold (t1/2 = 14 min) (Fig. 7) whereas addition of VanSBΔ only had a 2.5-fold effect on [32P]-VanRB hydrolysis (t1/2 = 60 min; Figs 6C and 7). Dephosphorylation of VanRB by VanSB or VanSBΔ was unaffected by the addition of ATP at a final concentration of 1 mM (data not shown). No backtransfer of phosphate was observed from VanRB to the VanSB or VanSBΔ kinases, suggesting that only hydrolysis to Pi occurs (Fig. 6B and C). These results indicate that the six amino acid deletion overlapping the G2 domain is essentially associated with the loss of phosphatase activity of VanSBΔ, leading to constitutive expression of the resistance genes in strain BM4525.
Although the ubiquitous two-component systems (TCSs) constitute one of the largest known families of transcriptional regulators in bacteria, VanS/VanR is the only system controlling expression of antibiotic resistance genes (Arthur et al., 1996a; 1999; Arthur and Quintiliani, 2001). In addition to their kinase activity, many of the histidine kinases also act as phosphoprotein phosphatases, accelerating the dephosphorylation of their cognate response regulators. Whereas the phosphorylated form of the histidine kinases is relatively stable, the intrinsic stability of phosphorylated response regulators varies widely, with isolated proteins displaying half-lives ranging from 23 s for CheY (Ganguli et al., 1995) to 80 min for DegU (Dahl et al., 1992) and 180 min for Spo0F (Zapf et al., 1998). The time-scale response of the signal transduction system can thus be modulated through the combined kinase/phosphatase activities of the histidine kinases controlling the amount of the phosphorylated form of the response regulator present in the cell in response to environmental signals. In many cases, regulation occurs essentially through the control of the phosphatase activity of the kinase such as EnvZ (Igo et al., 1989) or NtrB (Keener and Kustu, 1988; Kamberov et al., 1994) of E. coli, and FixL of Rhizobium meliloti (Lois et al., 1993). In the sporulation phosphorelay of Bacillus subtilis, additional specific phosphatases target either the Spo0F or Spo0A response regulators to prevent the bacteria from sporulating under inappropriate conditions (Perego et al., 1996; Perego, 2001). Phosphatase activity is particularly critical in the case of response regulators whose phosphorylated form is highly stable, to ensure that the protein is not permanently activated.
Phosphatase activity plays a crucial role in controlling VanA-type vancomycin resistance. Indeed, as a member of the EnvZ/OmpR family of TCSs, phosphorylated VanR displays considerably high stability (t1/2 = 10 h), and VanS has been shown in vitro to act as a phospho-VanR phosphatase, catalysing dephosphorylation of the response regulator (Wright et al., 1993). It has been suggested that the Met 55 residue of VanR plays an important role in its low autophosphatase activity, based on studies of the Spo0F response regulator (Zapf et al., 1998). Genetic evidence indicates that VanS acts primarily as a phosphatase under non-inducing conditions, and as a kinase in the presence of glycopeptides, leading to phosphorylation of the response regulator and activation of the resistance genes (Arthur et al., 1997). This is also thought to be the case for the distantly related VanSB kinase, as mutations in vanSB conferring constitutive expression of glycopeptide resistance led to substitutions of conserved residues known to play a critical role in the phosphatase activity of the two related kinases EnvZ and PhoR (Aiba et al., 1989; Yamada et al., 1989; Baptista et al., 1997).
We have examined the molecular basis for inducible or high level constitutive vancomycin resistance in two VanB-type clinical isolates from the same patient. E. faecium BM4524 was resistant to vancomycin only, whereas derivative BM4525 was resistant to high levels of both vancomycin and teicoplanin (Table 1). A comparative analysis of the two strains indicated that the chromosomal ddl gene in BM4525 was disrupted by a 2 bp insertion leading to production of an impaired d-Ala:d-Ala ligase and thus to the exclusive synthesis of peptidoglycan precursors terminating in d-Lac (Table 1). Enterococci containing a vancomycin resistance cluster and having an impaired d-Ala:d-Ala ligase can only grow in the presence of vancomycin as this antibiotic is required for induction of the resistance genes and the strains rely entirely on synthesis of peptidoglycan precursors containing d-Ala-d-Lac for growth (Sahm et al., 1989; Baptista et al., 1997; Van Bambeke et al., 1999; Arthur et al., 1999). We show here that the vanB resistance genes (vanYBWHBBXB) are constitutively expressed in strain BM4525 (Fig. 2 and Table 1), thus bypassing the requirement for glycopeptides.
Sequence comparisons of the vanB operons in the two strains revealed a single mutation in the vanSB gene of strain BM4525 consisting in an 18 bp deletion (AGAAAAAGTGGGCGAAGT) removing amino acids 402–407 in VanSBΔ and partially overlapping the conserved G2 motif which is part of the ATP-binding domain (Fig. 3). The deletion occurred between two tandemly arranged heptanucleotide direct repeats, separated by 11 base-pairs. An in vitro approach was adopted in order to determine how this deletion might lead to constitutive expression of the resistance genes in strain BM4525.
The VanRB response regulator and the VanSB and VanSBΔ kinases (cytoplasmic domains) were overproduced in E. coli and purified (Fig. 4). Both kinases were autophosphorylated when incubated with [γ-32P]-ATP (Fig. 5) and were able to phosphorylate VanRB, although VanSBΔ was slightly less efficient than VanSB in phosphotransfer to the response regulator (Fig. 5). VanRB catalysed its own phosphorylation in the absence of kinase, using acetyl-phosphate as a phosphodonor (Fig. 6), as shown for VanR (Wright et al., 1993) and other response regulators (Lukat et al., 1992). In the absence of VanSB or VanSBΔ, phospho-VanRB was highly stable, with a half-life of 150 min (Fig. 7).
A recent structure- and mathematical modelling-based comparative analysis of TCSs led to the prediction that VanSB should act as a so-called ‘bifunctional sensor’, i.e. one endowed with both kinase and phosphatase activities (Alves and Savageau, 2003). We show that this is indeed the case, as the addition of VanSB strongly accelerated hydrolysis of [32P]-VanRB (t1/2 = 14 min), indicating that VanSB also acts as a phospho-VanRB phosphatase. The addition of ATP had no effect on VanSB phosphatase activity, as also reported for VanS, in contrast to the prototype EnvZ histidine kinase (Igo et al., 1989; Wright et al., 1993). Addition of VanSBΔ had a much less pronounced effect on phospho-VanRB hydrolysis, leading to a t1/2 = 60 min, which indicates a significant loss in phosphatase activity following the deletion of amino acids 402–407. This six amino acid stretch partially overlaps the conserved G2 motif which is part of the ATP-binding domain. This is in agreement with a recent report indicating that the G2 box of EnvZ and its surrounding residues play an important role in modulating phosphatase activity (Zhu and Inouye, 2002). Interestingly, an identical 18 bp deletion in vanSB was previously obtained in a vancomycin- and teicoplanin-dependent mutant with an impaired d-Ala:d-Ala ligase (strain BM4391) following growth in vitro in the presence of teicoplanin (Van Bambeke et al., 1999). This strain had two other point mutations in the vanSB gene in comparison to strain BM4525, which may explain the difference in phenotype (Van Bambeke et al., 1999).
The deletions in strains BM4525 and BM4391 occurred between the two tandemly arranged heptanucleotide direct repeats which appear to be specific to the vanSB gene. The fact that this spontaneous deletion was observed in a clinical isolate and during growth in vitro suggests that it may occur relatively frequently under selective pressure. This is reminiscent of the expression of Tn5-encoded streptomycin resistance, where mutants expressing the str gene were obtained in E. coli following a deletion between tandem repeats. As for vanB, the same deletion was identified in clinical isolates as well as in strains grown in vitro in the laboratory (Mazodier et al., 1986; Genilloud et al., 1988), and the deletion was suggested to affect a post-translational event. Similarly, deletion of one of two tandemly repeated heptanucleotides accounts for streptothricin resistance in a Campylobacter coli animal isolate, a genetic rearrangement that was easily reproduced in vitro (Jacob et al., 1994).
The structure of the tandemly arranged heptanucleotide direct repeats in the vanSB gene suggests that the deletion may result from a RecA-independent recombination event, according to the replication ‘slippage’ model, due to slipped misalignment of the nascent DNA strand during replication (Bzymek and Lovett, 2001). This type of structure is also reminiscent of the simple sequence contingency loci of Haemophilus influenzae and Neisseria meningitidis involved in phase variation, where the small tandem DNA repeats have evolved as a means of rapidly generating large amounts of genetic variation, thus increasing fitness and allowing the bacteria to survive unanticipated challenges (Bayliss et al., 2001). It is therefore tempting to speculate that the repeated sequence in vanSB may also act as a simple sequence contingency locus, allowing bacteria to constitutively express high-level vancomycin resistance when challenged. Mutations in vanSB that lead to constitutive expression of the resistance genes are readily selected in vancomycin-dependent bacteria, not only by teicoplanin but also in media devoid of antibiotics, as such mutations will allow growth in the absence of the inducer. Thus, vancomycin may indirectly select for constitutive teicoplanin resistance in two steps. This may explain the emergence of teicoplanin resistance in a patient treated with vancomycin and may also account for constitutive expression of vanB-related gene clusters in clinical isolates of E. faecalis that harbour null mutations in the ddld-Ala:d-Ala ligase gene.
Strains, plasmids and growth conditions
Bacterial strains and plasmids used in this study are listed in Table 3. Enterococcus faecium BM4524, isolated from the blood of a patient, was resistant to vancomycin and susceptible to teicoplanin (Table 1). Enterococcus faecium BM4525, isolated two weeks later from the bile of the same patient, was resistant to high levels of both vancomycin and teicoplanin (Table 1). Enterococcus faecalis JH2-2 is a derivative of strain JH2 that is resistant to fusidic acid and rifampicin (Jacob and Hobbs, 1974). Escherichia coli strain Top10 (Invitrogen) was used as a host for recombinant pCR-Blunt plasmids with kanamycin selection (50 µg ml−1) and E. coli strain BL21 λDE3 (Studier and Moffatt, 1986) for protein overexpression and purification.
597 bp PCR fragment (vanSB′, vanYB′) of BM4525 cloned into pCR-Blunt
1232 bp PCR fragment (vanB) of BM4525 cloned into pCR2.1
3050 bp PCR fragment (ORFA, ORFB, orf7, orf8, xis) of BM4525 cloned into pCR-Blunt
1042 bp PCR fragment (vanSBΔ) of BM4525 cloned into pCR-Blunt
690 bp PCR fragment (vanRB) of BM4525 cloned into pCR-Blunt
884 bp NcoI-XhoI PCR fragment (vanSB) of 268–10 cloned into pET28/16
866 bp NcoI-XhoI PCR fragment (vanSBΔ) of BM4525 cloned into pET28/16
663 bp NcoI-XhoI PCR fragment (vanRB) of BM4525 cloned into pET28/16
1215 bp PCR fragment (vanRB′) of BM4525 cloned into pCR-Blunt
Strains were grown in brain heart infusion (BHI) broth or agar (Difco Laboratories) at 37°C. The MICs of glycopeptides were determined by the method of Steers et al. (1959), with 105 CFU per spot on BHI agar after 24 h of incubation.
Recombinant DNA techniques
Plasmid DNA isolation, digestion with restriction endonucleases, amplification of DNA by PCR with Pfu DNA polymerase (Stratagene), ligation of DNA fragments with T4 DNA ligase (Amersham Pharmacia Biotech) and transformation of E. coli Top10 with recombinant plasmid DNA were performed by standard methods (Ausubel et al., 1987). Total DNA from enterococci was prepared as previously described (Le Bouguénec et al., 1990).
The plasmids were constructed as follows (Fig. 1). The vanSB′-vanYB′ intergenic region, the vanB2, vanRB and vanSB genes were amplified individually using BM4525 total DNA as template, and primer pairs VB38-VB27, VB7-VB8, O1-O2 and P9-O4, respectively, designed from the E. faecalis V583 vanB cluster (Fig. 1). The PCR fragments were cloned into the pCR-Blunt vector, leading to plasmids pAT811(vanSB′-vanYB′), pAT815(vanRB) and pAT814(vanSB), or into plasmid pCR2.1, leading to plasmid pAT812(vanB2) and were sequenced.
To complete the 3′ sequence of the vanB2 operon in strain BM4525, the region downstream from vanXB was amplified using primers MR1 (5′-CATCTTTTGTAATACAT-3′) and MR7 (5′-TCCTCAATCGTCAGCGT-3′) designed from the vanB cluster of strain 268–10 (Garnier et al., 2000) and BM4525 total DNA as template (Fig. 1). The 3050 bp PCR fragment was cloned into pCR-Blunt, leading to plasmid pAT813(pCR-BluntΩISEnfa3, orf7, orf8, xis) and sequenced.
To obtain the 5′ sequence upstream from vanRB, the region upstream from vanRB was amplified with BM4525 DNA as template and the specific primer RB13 (5′-TGACGTTGAAA GAGGCAG-3′) designed from the vanB cluster of strain 268–10 (Garnier et al., 2000) used in combination with RB5 (5′-GTTCATGCCCGTTCATA-3′), complementary to a previously sequenced portion of the 5′ end of the vanRB gene from BM4525 (Fig. 1). The 1215 bp PCR fragment was cloned into pCR-Blunt, leading to plasmid pAT819 (pCR-BluntΩPRB,vanRB′ ) and sequenced.
Plasmid pET28/16 (Chastanet et al., 2003) was used for protein overexpression and purification. For construction of pAT816(pET28ΩvanSB) and pAT817(pET28ΩvanSBΔ), a portion of the vanSB gene of strain 268–10 and the vanSBΔ gene of strain BM4525, which had a deletion of 18 bp overlapping the G2 ATP binding domain, were amplified using 268–10 and BM4525 total DNA as template, respectively, and primer pair SB1-SB2. Oligodeoxynucleotide SB1 (5′-GGTGGTCTC CCATGACAACGCCGATCAAAG-3′) contained a BsaI restriction site (underlined) designed to generate a cohesive end compatible with NcoI, and 19 bases complementary to codons 159–165 of vanSB or vanSBΔ. Oligodeoxynucleotide SB2 (5′-CTCCTCGAGTGTTGATGTGAGCGGTAAATC-3′) contained a XhoI site (underlined) in place of the TAA stop codon and the CTA leucine codon and 21 bases complementary to codons 440–446 of vanSB or to codons 434–440 of vanSBΔ. The 884 bp and 866 bp pCR products from vanSB and vanSBΔ, respectively, were digested by BsaI and XhoI and cloned between the NcoI and XhoI restriction sites of plasmid pET28/16 to generate plasmids pAT816(pET28ΩvanSB) and pAT817(pET28ΩvanSBΔ).
To construct pAT818(pET28ΩvanRB) from BM4525, a similar strategy was used. Plasmid pAT818(pET28ΩvanRB) was constructed by cloning a 657 bp BsaI-XhoI fragment corresponding to the vanRB coding sequence, amplified using oligonucleotides RB3 and RB4 and BM4525 total DNA as template, in the NcoI and XhoI sites of pET28/16. Oligodeoxynucleotide RB3 (5′-GGTGGTCTCCCATGTC GATACGAATTCTACT-3′) contained a BsaI restriction site (underlined), and 19 bases complementary to codons 1–7 of vanRB of BM4525. Oligodeoxynucleotide RB4 (5′-CTCCTC GAGTAATGATTCCTCCAATCGGT-3′) contained a XhoI site (underlined) replacing the TAA stop codon and 20 bases complementary to codons 213–219 of vanRB.
Plasmid DNA was extracted with the commercial Wizard Plus Miniprep DNA purification system (Promega) and sequencing was performed by the dideoxynucleotide chain termination method with [α-35S]-dATP (Amersham Pharmacia Biotech) and the T7 Sequenase version 2.0 DNA sequencing kit (Amersham). The PCR fragments were purified with a PCR purification kit (Qiagen), labelled with a dye-labelled ddNTP Terminator Cycle Sequencing Kit (Beckman Coulter UK) and the samples were sequenced and analysed with a CEQ 2000 automated sequencer (Beckman).
Database comparisons and sequence analysis
Sequence analysis was performed with the blastn, blastx and blastp (Altschul et al., 1997) and fasta (Pearson and Lipman, 1988) programs, using the GCG sequence analysis software package (version 10.1, Genetics Computer Group, Madison, Wisconsin) and transmembrane domains were predicted using von Heijne's algorithm and the toppredii program (Claros and von Heijne, 1994).
Extraction and analysis of peptidoglycan precursors was performed as described (Reynolds et al., 1994). Enterococci were grown in BHI medium without or with vancomycin (4 µg ml−1) to the mid-exponential phase (OD600 = 1). Ramoplanin was added to inhibit peptidoglycan synthesis and incubation was continued for 15 min to allow accumulation of peptidoglycan precursors. Cells were harvested and the cytoplasmic precursors were extracted with 8% trichloroacetic acid (15 min at 4°C), desalted and analysed by high-performance liquid chromatography. Results were expressed as the percentages of total late peptidoglycan precursors represented by UDP-MurNAc-tripeptide, UDP-MurNAc-tetrapeptide, UDP-MurNAc-pentapeptide and UDP-MurNAc-pentadepsipeptide that were determined from the integrated peak areas.
d,d-dipeptidase (VanXB) and d,d-carboxypeptidase (VanYB) activities
The enzymatic activities in the cytosolic and resuspended pellet fractions were assayed as described previously (Arthur et al., 1996a). Strains were grown in the absence or presence of vancomycin or teicoplanin at various concentrations (1, 8 and 64 µg ml−1) until the optical density at 600 nm reached 0.7. Bacteria were then lysed by treatment with lysozyme (2 mg ml−1) at 37°C, followed by sonication and the membrane fraction was pelleted (100 000 g, 45 min). The cytosolic fraction (S100) and resuspended pellet (C100) were collected and assayed for d,d-peptidase (VanXB or VanYB) activities by measuring the d-Ala released from substrate hydrolysis (d-Ala-d-Ala or l-Ala-γ-d-Glu-l-Lys-d-Ala-d-Ala) through coupled indicator reactions using d-amino acid oxidase and horse radish peroxidase. Specific activity was defined as the number of nanomoles of product formed at 37°C per minute per milligram of protein contained in the extracts.
Extraction of total RNA
Enterococcal strains were grown in BHI (100 ml) at 37°C to an optical density at 600 nm of 0.8, in the presence or absence of vancomycin (8 µg ml−1). Bacteria were pelleted and frozen immediately. RNA was extracted essentially as described previously (Glatron and Rapoport, 1972; Chastanet et al., 2001) and resuspended in 50 µl of water treated by DEPC. RNA concentrations were determined by measuring absorbance at 260 nm.
Primer extension analysis
The synthetic oligodeoxynucleotide PYB (5′-TGAACGA AATAATGAACGCACATAGAAAAGCCCGCTTTTC-3′) was 5′ end labelled with [γ-33P]-ATP (Amersham Pharmacia Biotech) and T4 polynucleotide kinase (Biolabs). The enzyme was inactivated for 15 min at 65°C. Labelled primer (1 pmol) was heated at 95°C for 2 min and then incubated in the presence of 25 µg of total RNA at 70°C for 10 min. The annealing of labelled primer with RNA was carried out at 50°C for 2 min in a mixture of 50 µl final volume composed of 1×‘first strand buffer’ (Invitrogen), a solution of 0.5 mM of all four dNTPS (Amersham), 10 mM DTT (Invitrogen) and 40 U RNasin (Amersham). The extension was performed with 40 U of avian myeloblastosis virus reverse transcriptase (Boehringer) for 45 min at 50°C. After phenol-chloroform extraction, the reaction products were precipitated with ethanol, redissolved in 10 µl of sterile water and 10 µl of loading buffer were added (Amersham). After heat denaturation, 4 µl were loaded onto 6% polyacrylamide-urea sequencing gels for electrophoresis. Sequencing reactions using primer PYB and plasmid pAT811(pCR-BluntΩvanSB′,vanYB′ ) DNA as a template were run in parallel to localize the transcription start site.
Overproduction and purification of VanSB, VanSBΔ and VanRB
Plasmids pAT816(pET28ΩvanSB), pAT817(pET28ΩvanSBΔ) and pAT818(pET28ΩvanRB) were introduced into an E. coli BL21 λDE3 strain, in which the T7 RNA polymerase gene is under the control of the inducible lacUV5 promoter (Studier and Moffatt, 1986), that also carries the pRep4 plasmid allowing co-expression of the GroESL chaperonin to optimize recombinant protein solubility (Amrein et al., 1995). The transformants were grown in LB medium at 30°C until mid-exponential phase (OD600 = 0.8). IPTG (1 mM) was added to induce recombinant protein production, and incubation was pursued for 5 h. The VanSB, VanSBΔ and VanRB proteins were then purified by IMAC (immobilized metal affinity chromatography) and concentrated in buffer A (50 mM NaH2PO4 pH 8, 300 mM NaCl, 50% glycerol) as previously described (Chastanet et al., 2003). Electrophoresis on 15% acrylamide gels was performed as described (Laemmli, 1970) and protein concentrations were determined using the Bio-Rad protein assay (Bradford, 1976).
Autophosphorylation of VanSB and VanSBΔ
Autophosphorylation of VanSB or VanSBΔ (1.2 nmol) was performed in a final volume of 100 µl in buffer B (final concentrations: 50 mM Tris-HCl, 50 mM KCl and 1 mM MgCl2, pH 7.5). The reaction was initiated by the addition of 5 µl of ATP (1 mM final) containing 200 µCi of [γ-32P]ATP and incubated at room temperature for 1 h. ATP was removed using Sephadex G-50 spin columns equilibrated with buffer B, and the phosphorylated protein was collected in a final volume of 500 µl.
Phosphorylation of VanRB by VanSB and VanSBΔ
Phosphotransfer to purified VanRB was carried out in buffer B. The reaction was initiated by the addition of 10 µl of the purified autophosphorylation reaction mixture of VanSB or VanSBΔ described above to a 15 µl reaction mixture containing 2.1 nmol of VanRB. After incubation for various periods of times (0.5, 1, 5, 10, 20 and 30 min) at room temperature, the reactions were quenched by the addition of 5 µl of β-mercaptoethanol stop solution (Sigma) followed by electrophoresis on 15% SDS-polyacrylamide gels and autoradiography.
Phosphorylation of VanRB by acetyl-[32P]-phosphate
VanRB(5.1 nmol) was incubated in 100 µl of buffer C (50 mM Tris-HCl, pH 7.8, 20 mM MgCl2, 0.1 mM dithiothreitol) containing 178 pmol (3.3 µCi) of acetyl-[32P]-phosphate (Amersham Pharmacia Biotech, custom synthesis) at room temperature for 30 and 60 min Excess acetyl-[32P]-phosphate was removed using Sephadex G-50 spin columns equilibrated with buffer C, and the protein was collected in a final volume of 100 µl.
Hydrolysis of phospho-VanRB by VanSB and VanSBΔ
The VanRB response regulator was labelled with acetyl-[32P]-phosphate for 1 h at room temperature as described above. The VanSB or VanSBΔ histidine kinases (approximately 1.6 nmol) were added to 3 nmol of phosphorylated VanRB and incubation was pursued for various periods of times (0, 0.5, 1, 5, 10, 20 and 30 min) in a 66 µl reaction mixture in Buffer C. Aliquots (10 µl) were withdrawn at designated time-points, and the reactions were quenched with β-mercaptoethanol stop solution followed by electrophoresis on 15% SDS-polyacrylamide gels and autoradiography. Radioactive gels were exposed to storage phosphor screens and scanned with a Molecular Dynamics Storm 860 optical scanner. Quantification of phosphorylated products was performed using the ImageQuant 5.1 software package (Molecular Dynamics).
We thank J. Bell for the gift of E. faecium BM4524 and BM4525, P. Reynolds for advice in determination of peptidoglycan precursors and S. Dubrac for helpful discussions and assistance in quantification of VanRB hydrolysis. This work was supported in part by a Bristol-Myers Squibb unrestricted biomedical research grant in infectious diseases and by the ‘Programme de Recherche Fondamentale en Microbiologie, Maladies Infectieuses et Parasitaires’ from the Ministère de l’Education Nationale de la Recherche et de la Technologie.