Unité des Agents Antibactériens, Institut Pasteur, 75724 Paris, Cedex 15, France,1 Department of Biochemistry, University of Cambridge, Cambridge CB21QW, United Kingdom2
Received 29 May 2002/ Returned for modification 24 August 2002/ Accepted 20 September 2002
| ABSTRACT |
|---|
|
|
|---|
| INTRODUCTION |
|---|
|
|
|---|
The general organization of the vanD operon is similar to those present in VanA- and VanB-type strains (12, 21, 24). 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 hydrolyze the D-Ala-D-Ala dipeptide synthesized by the host D-Ala:D-Ala Ddl ligase and thereby limit synthesis of precursors containing the target for glycopeptides (3, 13, 20, 49). In VanA- and VanB-type strains, a penicillin-insensitive and Zn2+-dependent D,D-carboxypeptidase (VanY and 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 (3, 8, 9). Certain PBPs which function as D,D-carboxypeptidases preferentially cleave depsipeptide substrates (46), whereas the Zn2+-dependent VanY D,D-carboxypeptidase exhibits a higher catalytic efficiency for hydrolysis of substrates ending in D-Ala-D-Ala (3). The VanYD D,D-carboxypeptidase is distinct from VanY and VanYB since it displays substantial identity with some penicillin-binding proteins (17, 21). These catalytic-serine D,D-carboxypeptidases are susceptible to benzylpenicillin (48). VanZ, which confers teicoplanin resistance by an unknown mechanism, and VanW, with an unknown function, encoded by the vanA and vanB clusters, respectively, do not have counterparts in the vanD cluster (6, 24).
Synthesis of the resistance proteins is regulated at the transcriptional level by two-component regulatory systems (VanR-VanS and VanRB-VanSB) (11, 24). VanS is a putative membrane-associated sensor that controls the level of phosphorylation of VanR (55). Phosphorylation of the VanR and VanRB response regulators enhances the affinity of the proteins for the regulatory regions of the PR, PRB and PH, PYB promoters, and allows transcription of the regulatory (vanRS and vanRBSB) and resistance (vanHAX and vanHBBXB) genes, respectively (4, 5, 24, 28, 30). The VanR-VanS system activates the PH promoter for cotranscription of the vanH, vanA, and vanX genes in response to the presence of vancomycin or teicoplanin in the culture medium (8, 11). In contrast, the VanRB-VanSB system mediates activation of the PYB promoter only in the presence of vancomycin, and lack of induction by teicoplanin accounts for susceptibility of VanB-type strains to this antibiotic (8, 24). Low-level resistance to vancomycin in VanB strains results from a limited capacity to synthesize D-Ala-D-Lac and to hydrolyze D-Ala-D-Ala, leading to coproduction of D-Ala and D-Lac-ending peptidoglycan precursors (8).
Four VanD-type strains of Enterococcus faecium have been reported so far, and clinical isolates BM4339 and BM4416 (also designated N97-330) have been extensively studied (17, 22, 39, 43, 45). These two VanD-type strains are characterized by constitutively expressed resistance to moderate levels of vancomycin (MIC, 16 to 256 µg/ml) and teicoplanin (MIC, 2 to 64 µg/ml) despite the presence of the vanRD and vanSD genes expressed from the PRD promoter (22, 43, 45).
Strain 10/96A is resistant to vancomycin (MIC, 256 µg/ml) and to low levels of teicoplanin (MIC, 4 µg/ml) (23). A PCR product, obtained using the D-Ala:D-X ligase universal degenerate primers V1 and V2, was sequenced and revealed 83 to 85% identity with structural genes for VanD ligases (17, 21, 39). The vanD cluster of strain 10/96A was sequenced partially and found to contain two open reading frames (ORFs) encoding a dehydrogenase, VanHD, and a D-Ala:D-Lac ligase, VanD (23). The operon, which is not transferable, confers resistance by constitutive synthesis of peptidoglycan precursors ending in D-Ala-D-Lac, which represent the main components of cell wall cytoplasmic precursors. In contrast to the VanYD activities in strains BM4339 and BM4416, the VanYD D,D-carboxypeptidase activity in membrane fractions of strain 10/96A was not inhibited by penicillin G (23, 43, 45).
We report the organization of the vanD gene cluster in E. faecium 10/96A and the regulation of expression of the resistance genes. We also show that a single mutation in the chromosomal ddl gene accounts for the lack of precursors terminating in D-Ala-D-Ala in this strain and that an insertion sequence in the vanSD gene is likely to be responsible for constitutive expression of resistance.
| MATERIALS AND METHODS |
|---|
|
|
|---|
|
Plasmid construction. The plasmids were constructed as follows (Fig. 1).
|
(ii) Plasmid pAT638. A strategy similar to that used for construction of pAT637 was followed to clone the vanYD gene of strain 10/96A. A sequence deduced from the alignment of the 3' end of the vanSD genes of BM4339 and BM4416 was used to design primer SD1 (5' GTTTTGAGGTTACATTGC). YD4 (5' GGTAATAGGGACTGTTCGGAT) contained 21 bases complementary to the sequence of the 3' end of the vanYD gene from strain 10/96A. These primers, used with total DNA from 10/96A as a template, yielded a product with the expected size of 1,060 bp that was cloned into pCR-Blunt generating plasmid pAT638.
(iii) Plasmid pAT639. To complete the sequence of the vanD operon of strain 10/96A, the portion upstream from vanYD was amplified using primers RDNH2 (5' ATGAATGAAAAAATCTTAGTGG) and SYD4 (5' TTACGATTTTCCTACGG) and total DNA as a template. Alignment of the vanRD genes from BM4339 and BM4416 was used to design the RDNH2 primer, complementary to a sequence conserved at the 5' end of these genes. When combined with primer SYD4, specific for the intergenic region upstream from the vanYD gene of 10/96A, RDNH2 led to amplification of a 3,725-bp fragment. This PCR fragment, with an unexpectedly large size, was cloned into pCR-Blunt, generating plasmid pAT639, and was sequenced.
(iv) Plasmids pAT635 and pAT633. For construction of pAT635(P2vanXDcat) and pAT633(P2vanYDcat), the vanXD and vanYD genes of BM4339 were amplified using primer pairs XD1NH2-XD1COOH and YD1NH2-YD1COOH, respectively, and BM4339 total DNA as a template. Oligodeoxynucleotides XD1NH2 (5' CAGTGAGCTCCGGTTTTACGCTTTCTG) and YD1NH2 (5' CAGTGAGCTCGCGAAAACATAAATCGC) harbored a SacI restriction site (underlined) and 17 bases complementary to the sequence upstream from vanXD or vanYD of BM4339, respectively. Oligodeoxynucleotides XD1COOH (5' AGTGTCTAGACTAGGCAATGCAAAAAT) and YD1COOH (5' AGTGTCTAGATTACTGGGCTTTGATTT) contained an XbaI restriction site (underlined), the stop codon (italicized), and 14 bases complementary to the 3' end sequence of vanXD or vanYD, respectively. The SacI and XbaI restriction sites allowed directional cloning of vanXD or vanYD upstream from the cat reporter gene of the shuttle vector pAT79 carrying the P2 promoter to generate pAT635(P2vanXDcat) and pAT633(P2vanYDcat). The 672-bp insert of plasmid pAT635(P2vanXDcat) corresponded to nucleotides 5057 to 5728 of the vanD operon of BM4339 and included the ribosome binding site (RBS), the initiation codon, the vanXD coding sequence, and the stop codon of the gene. The 1,153-bp insert of pAT633(P2vanYDcat) corresponded to nucleotides 1950 to 3102 of the vanD operon of BM4339 and consisted of the vanYD coding sequence with its RBS and initiation and stop codons.
(v) Plasmids pAT634 and pAT636. To construct pAT636(P2vanXDcat) and pAT634(P2vanYDcat) from 10/96A, a strategy identical to that used for construction of pAT635(P2vanXDcat) and pAT633(P2vanYDcat) from BM4339 was followed. The vanXD and vanYD genes of strain 10/96A were amplified using primer pairs XD4NH2-XD4COOH and YD4NH2-YD4COOH, respectively, with 10/96A total DNA as a template. Primers XD4NH2 (5' CAGTGAGCTCAGGGTTTACGCTTTCTG) and YD4NH2 (5' CAGTGAGCTCGCGAAAAAATAAATCGC) harbored a SacI site (underlined) and 17 bases complementary to the sequence upstream from vanXD or vanYD of strain 10/96A, respectively. Primers XD4COOH (5' AGTGTCTAGACTAGGCAATGCAAAAAT) and YD4COOH (5' AGTGTCTAGATCACTGGGCCTTGATTT) contained an XbaI site (underlined), the stop codon (italicized), and 14 bases complementary to the 3' end sequence of vanXD or vanYD of strain 10/96A, respectively. The vanXD and vanYD PCR products were digested with SacI and XbaI and cloned under the control of the P2 promoter of the shuttle vector pAT79, leading to plasmids pAT636(P2vanXDcat) and pAT634(P2vanYDcat), respectively (11). The 672- and 1,153-bp inserts of pAT636(P2vanXDcat) and pAT634(P2vanYDcat) corresponded, respectively, to nucleotides 6875 to 7546 and nucleotides 3765 to 4917 of strain 10/96A and consisted of the vanXD or vanYD coding sequences of 10/96A with their RBS and initiation and stop codons.
(vi) Plasmid pAT640. The chromosomal ddl gene from E. faecium 10/96A with its RBS was amplified by PCR from total DNA with the previously described 4147-1 and 4147-2 oligodeoxynucleotides (26). Primers 4147-1 and 4147-2 contain, respectively, SacI and XbaI restriction sites that allow directional cloning of the ddl gene under the control of the constitutive P2 promoter and upstream from the cat reporter gene of the shuttle vector pAT79 (11). The 1,135-bp insert of the resulting pAT640(P2ddlG184Scat) plasmid contained the mutated ddl gene with the single G184S mutation and its own RBS.
The nucleotide sequences of the amplified fragments were redetermined.
Strain constructions. Plasmids pAT635(P2vanXDcat) and pAT633(P2vanYDcat) from BM4339 and plasmids pAT636(P2vanXDcat) and pAT634(P2vanYDcat) from 10/96A were introduced into E. faecalis JH2-2 by electrotransformation. E. faecium BM4512 was obtained by introduction of plasmid pAT640(P2ddlG184Scat) into E. faecium BM4339 by electrotransformation (Table 1). Transformants selected on spectinomycin, 60 µg/ml for JH2-2 or 120 µg/ml for BM4339, were screened for resistance to chloramphenicol. Plasmid DNA from chloramphenicol-resistant clones was digested with EcoRI plus HindIII and compared to the restriction profiles of pAT635(P2vanXDcat) and pAT633(P2vanYDcat) and those of pAT636(P2vanXDcat), pAT634(P2vanYDcat), and pAT640(P2ddlG184Scat) purified from E. coli Top10 to screen for DNA rearrangements.
Nucleotide sequencing. Plasmid DNA was extracted with the commercial Wizard Plus Minipreps DNA purification system (Promega, Madison, Wis.), labeled with a dye-labeled ddNTP Terminator cycle sequencing kit (Beckman Coulter UK Ltd., High Wycombe, United Kingdom), and the samples were sequenced and analyzed with a CEQ 2000 automated sequencer (Beckman).
Computer analysis of sequence data. Determination of the degrees of identity and similarity with known proteins was carried out using BLASTN, BLASTX, and BLASTP (2) and FASTA (42) from the Genetics Computer Group suite of programs.
Contour-clamped homogeneous electric field gel electrophoresis.
Genomic DNA embedded in agarose plugs was digested for 3 h at 37°C with 0.01 U of I-CeuI, an intron-encoded endonuclease specific for rRNA genes (34). Fragments were separated on a 0.8% agarose gel using a CHEF-DRIII system (Bio-Rad Laboratories, Hercules, Calif.) under the following conditions: total migration, 24 h; initial pulse, 60 s; final pulse, 120 s; voltage, 6 V/cm; included angle, 120°; and temperature, 14°C. The DNA fragments were transferred to a nitrocellulose membrane and hybridized successively under stringent conditions at 68°C to an
-32P-labeled 16S rRNA (rrs) probe obtained by amplification of an internal portion of the rrs gene with primers RWO1 and DG74 (27) and to a vanD probe obtained by PCR with primers D4-1 and D4-2 and 10/96 total DNA as a template (Fig. 1). The amplification product used to generate the probe was labeled with [
-32P]dATP (3,000 Ci/mmol; Amersham Pharmacia Biotech) by Megaprime using a commercially available kit (Amersham).
Analysis of peptidoglycan precursors. Extraction and analysis of peptidoglycan precursors was performed as described previously (37, 49). Enterococci were grown in BHI medium without or with vancomycin (4 µg/ml) to the mid-exponential phase (A600 = 1). Ramoplanin was added to inhibit peptidoglycan synthesis, and incubation was continued for 15 min to cause accumulation of peptidoglycan precursors. The bacteria were harvested, and the cytoplasmic precursors were extracted with 8% trichloroacetic acid (15 min, at 4°C), desalted, and analyzed 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 (VanX) and D,D-carboxypeptidase (VanY) activities. The enzymatic activities in the supernatant and in the resuspended pellet fraction were assayed as described previously (8, 48). Strains were grown until the optical density at 600 nm reached 0.7 in the absence or presence of vancomycin at various concentrations (1, 8, and 64 µg/ml) for induction of 10/96A or with spectinomycin (60 µg/ml) to counterselect loss of derivatives of pAT79. Bacteria were then lysed by treatment with lysozyme (2 mg/ml) at 37°C, followed by sonication, and the membrane fraction was pelleted (100,000 x g, 45 min). The supernatant (S100) and resuspended pellet (C100) were collected and assayed for D,D-peptidase (VanX or VanY) activities by measuring the D-Ala released from substrate hydrolysis (D-Ala-D-Ala, 6.56 mM, or UDP-MurNAc-L-Ala-D-Glu-L-Lys-D-Ala-D-Ala, 5 mM) through coupled indicator reactions using D-amino acid oxidase and horseradish peroxidase (8, 48). 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.
Preparation of membrane fragments and binding of benzyl[14C]penicillin. The membrane fragments were prepared and labeling was carried out as described elsewhere (48). Briefly, a culture at an optical density at 600 nm of 1.0 was centrifuged, the pellet was washed in 50 mM Tris HCl (pH 7.2) and resuspended, and osmotic lysis was achieved in the presence of lysozyme (400 µg/ml) and muramidase (70 µg/ml) after incubation at 37°C. DNase (25 µg/ml) and MgCl2 (5 mM) were added, and after 3 min at 37°C, the suspension was cooled to 4°C, centrifuged, and washed. The membrane fraction was resuspended in 50 mM Tris HCl (pH 7.2) and incubated at 37°C with benzyl[14C]penicillin (1 µg/ml). After addition of unlabeled penicillin G (3 mg/ml) and sample buffer (New England Biolabs), the membrane proteins were solubilized by heating at 98°C. The labeled membrane proteins were run on a 12% polyacrylamide gel. The gel was stained, dried, and set up for phosphorimaging overnight to detect PBPs and determine the amount of penicilloyl-protein complex. Autoradiography was carried out for 5 weeks to reveal minor PBPs.
Nucleotide sequence accession number. The 7,546-bp fragment containing the vanD gene cluster of strain 10/96A was submitted to GenBank and assigned accession no. AY082011.
| RESULTS |
|---|
|
|
|---|
|
|
Characterization of the van genes in E. faecium 10/96A and of the deduced proteins. Plasmids pAT637(vanD' vanXD), pAT638(vanSD' vanYD'') and pAT639 (vanRD, SD', ORFA, ORFB, vanSD') were obtained by cloning the 632-, 1,060-, and 3,725-bp PCR fragments obtained from total DNA of 10/96A into the pCR-Blunt vector (Table 1 and Fig. 1). Sequencing of both strands of the inserts in these plasmids revealed the presence of the structural genes for the D,D-peptidases (VanXD and VanYD) and also for the two-component regulatory system (VanRD-VanSD) which was interrupted by an insertion sequence composed of two ORFs in the same orientation (Fig. 3). The genes for the VanHD dehydrogenase and for the VanD D-Ala:D-Lac ligase have been characterized previously (23). The complete organization of the 7,546-bp vanD gene cluster composed of eight genes, of which six were found with the same organization as those of the vanD operons in E. faecium BM4339 and BM4416, is shown in Fig. 1 and 3.
|
|
|
|
The C-terminal portion of VanSD in strain 10/96A contained the five blocks of conserved amino acids characteristic of transmitter modules in histidine protein kinases (Fig. 5). Histidine residue 140 of VanSD from strain 10/96A was aligned with histidine residues 166 of VanSD of BM4339 and BM4416, which are the putative sites of autophosphorylation of sensors (Fig. 5). The hydropathy profile of the N-terminal putative sensor domain of VanSD from strains BM4339 and BM4416 revealed the presence of two stretches of hydrophobic amino acids similar to those in VanS, VanSB, and EnvZ, suggesting a similar topology for these enzymes (data not shown). The ISEfa4 copy in vanSD of strain 10/96A not only removed one of the potential membrane spanning regions of VanSD, but the remaining larger portion of the protein would not have been produced. Lack of VanSD may lead to a high steady-state level of phosphorylated VanR and could thus account for the constitutive expression of the vanD operon in strain 10/96A.
|
As shown previously, four and five I-CeuI fragments from BM4339 and BM4416, respectively, hybridized with the rrs probe, and a 330-kb fragment from BM4339 or a 380-kb fragment from BM4416 cohybridized with a vanHDDXD probe (43). The comparative analysis indicated that strain 10/96A was distinct from these two isolates
D,D-Peptidase activities in strain 10/96A.
D,D-Dipeptidase and D,D-carboxypeptidase activities in E. faecium 10/96A were assayed by determining the amount of D-Ala released from hydrolysis of the dipeptide D-Ala-D-Ala and of the pentapeptide UDP-Mur-NAc-L-Ala-
-D-Glu-L-Lys-D-Ala-D-Ala, respectively (Table 3). The D,D-dipeptidase activity was measured in the supernatant of the lysed bacteria (after centrifugation at 100,000 x g) that had been grown in the presence of various concentrations of vancomycin (1, 8, and 64 µg/ml) as an inducer. As in BM4339, weak D,D-dipeptidase activity (VanXD) was found in the cytoplasmic extracts from induced or uninduced 10/96A (Table 3) (45). Since this strain has an impaired D-Ala:D-Ala ligase, it does not require an active VanX type D,D-dipeptidase for resistance.
The level of D,D-carboxypeptidase activity in the cytoplasmic fraction of 10/96A was low. However, the membrane preparation of this strain contained substantial activity, slightly weaker than that of membrane extracts of BM4339 and BM4416, in which the activity was inhibited by low concentrations of benzylpenicillin (Table 3) (48). In 10/96A the D,D-carboxypeptidase activity was not significantly induced by vancomycin, nor was it inhibited by benzylpenicillin (Table 3).
Comparison of D,D-peptidase activities from BM4339 and 10/96A in E. faecalis JH2-2. Strains BM4339 and 10/96A do not produce D-Ala-D-Ala-containing peptidoglycan precursors following mutations in the chromosomal ddl gene (Table 2 and Fig. 2). Consequently, as mentioned previously no D,D-dipeptidase activity is required for glycopeptide resistance in this genetic background.
To test whether the vanXD and vanYD genes from BM4339 and 10/96A encode functional enzymes, the genes and their RBS were cloned under the control of the constitutive P2 promoter, leading to plasmids pAT635(P2vanXDcat) and pAT633(P2vanYDcat) from BM4339 and pAT636(P2vanXDcat) and pAT634(P2vanYDcat) from 10/96A, which were all electrotransformed into E. faecalis JH2-2 (Fig. 1). Although the deduced sequences of the two VanXD proteins do not contain mutations in the conserved residues known to be involved in zinc binding and catalysis (Fig. 4), only very weak hydrolysis of D-Ala-D-Ala was detected in cytoplasmic extracts from E. faecalis JH2-2 harboring pAT635(P2vanXDcat) and pAT636(P2vanXDcat) (Table 4). These results are in agreement with those obtained with crude extracts of BM4339 and 10/96A (Table 3).
No D,D-carboxypeptidase activity was detected in extracts from membrane or cytoplasmic fractions from JH2-2/pAT634(P2vanYDcat) harboring vanYD of 10/96A, whereas some activity was present in JH2-2/pAT633(P2vanYDcat) harboring vanYD of BM4339 (Table 4). Compared with the other VanD-type strains, E. faecium 10/96A produced almost exclusively UDP-MurNAc-pentadepsipeptide (95%), whereas UDP-MurNAc-tetrapeptide (2%) and UDP-MurNAc-tripeptide were present in insignificant amounts (Table 2). Despite the presence of the three conserved motifs in the same ORF corresponding to the cytoplasmic portion (Fig. 4), the frameshift mutation in the vanYD gene of strain 10/96A accounted for the lack of D,D-carboxypeptidase activity (Table 4). These results suggest that the truncated VanYD from 10/96A was not active due to loss of the domain containing the active site.
Analysis of PBPs from E. faecium 10/96A. D,D-Carboxypeptidases from BM4339 and BM4416 belong to the PBP family of catalytic serine enzymes but are susceptible to benzylpenicillin (43, 45, 48). We have studied the binding of benzyl[14C]penicillin to membrane preparations of E. faecium 10/96A in comparison with the other VanD-type strains, BM4339 and BM4416 (Fig. 6). A PBP which migrated as a doublet on sodium dodecyl sulfate gel with an apparent molecular mass of 40 to 42 kDa was present in strains BM4339 and BM4416 but absent from membranes of E. faecium 10/96A (Fig. 6). This result supported the hypothesis that the defect in the vanYD gene in the latter strain results in a lack of inhibition by benzylpenicillin of the D,D-carboxypeptidase present, which is presumably encoded by a different gene.
|
| DISCUSSION |
|---|
|
|
|---|
Compared with the VanA- and VanB-type strains, VanD-type strains BM4339, BM4416 (also designated N97-330), and 10/96A have negligible D,D-dipeptidase activity, encoded by vanXD alleles, despite the presence of critical residues implicated in the binding of Zn2+ and in catalysis (Tables 3 and 4) (36, 43, 45). Lack of such an activity should result in a glycopeptide-susceptible phenotype, since bacteria are unable to remove peptidoglycan precursors ending in D-Ala-D-Ala, the target for glycopeptides. However, the chromosomal ddl gene is disrupted by a 5-bp insertion in BM4339; by insertion of an IS19 (also called ISEfm1) element in BM4416; and by a single mutation, G184S, next to the serine involved in the binding of D-Ala1 in strain 10/96A (Fig. 2) (17, 21, 43). Both insertions and the mutation lead to production of quasi exclusively peptidoglycan precursors terminating in D-lactate (Table 2). Enterococci containing a vancomycin resistance cluster but with an impaired D-Ala:D-Ala ligase can only grow in the presence of vancomycin if these strains are inducible for vancomycin resistance. Such strains rely entirely for growth on synthesis of peptidoglycan precursors containing D-Ala-D-Lac instead of D-Ala-D-Ala (7, 15, 50, 53). In the VanD-type strains studied, there were no qualitative differences between the peptidoglycan precursors produced by uninduced or induced cells, indicating that the vanD clusters were expressed constitutively, thus bypassing the requirement for glycopeptides (Table 2).
In VanA- and VanB-type strains, VanS or VanSB sensors act as a kinase in the presence of glycopeptides. VanS and VanSB also negatively control promoters PH, PYB and PR, PRB that mediate transcription of the resistance (vanHAX and vanHBBXB) and regulatory (vanRS and vanRBSB) genes, respectively, in the absence of glycopeptides (4, 5). Under noninducing conditions, the wild-type sensors are therefore considered to act as phosphatases preventing accumulation of the phosphorylated form of the response regulators. According to this model, a constitutive phenotype is associated with loss of the phosphatase activity of the kinase and expression of the resistance genes remains unaltered under noninducing or inducing conditions (4, 5, 15). In VanA-type strains, in the absence of VanS, dephosphorylation of VanR phosphorylated by an heterologous kinase is extremely slow compared to that of the related response regulators, leading to a high level of phospho-VanR and thus to constitutive high-level transcription of the resistance genes (5, 55). In VanB-type strains, constitutive expression of glycopeptide resistance is most probably due to an impaired dephosphorylation of VanRB by VanSB, since substitutions affecting homologous residues in related sensor kinases result in a defect of the phosphatase, but not of the kinase activity, of the proteins (1, 57). Constitutive expression of the vanB cluster is due to amino acid substitutions at two specific positions on either side of histidine 233, which corresponds to the putative autophosphorylation site of VanSB (15).
Alignment of the deduced amino acid sequences of the VanSD sensors from E. faecium BM4339, BM4416, and 10/96A revealed a mutation at position 173 in the sensor of BM4339, leading to a Pro-to-Ser substitution (Fig. 5). This substitution is in a critical region, since it alters a residue close to histidine 166, corresponding to the putative autophosphorylation site of VanSD, an observation which could account for the constitutive expression of the vanD cluster in BM4339 (Fig. 5). Previous comparison of the vanSD genes from BM4339 and N-97-330 (so called BM4416) showed that the latter strain had suffered a 1-bp deletion at position 670 (BM4339 numbering), which results in a frameshift mutation leading presumably to the synthesis of a 233-amino-acid truncated and nonfunctional sensor instead of a protein containing 381 amino acids as in BM4339 (17). Strain 10/96A contains a different type of mutational event which bypasses the requirement of glycopeptide for constitutive expression of the resistance genes. Insertion sequence ISEfa4 was found 45 bp downstream from the start site of the vanSD gene of strain 10/96A and it possessed the characteristics of the IS605-family (Fig. 3). IS605, detected in H. pylori, is unusual in that it contains, in opposite orientation, homologs of genes for the putative transposases of two other unrelated insertion sequence elements, IS200 from H. pylori and IS1341 from the thermophilic bacterium PS3 (32). ISEfa4 is characterized by (i) the absence of terminal inverted repeats, (ii) lack of duplication of target sequences, (iii) inserting with its left end next to 5'-TTTAAC, and (iv) two ORFs encoding putative transposases but in the same orientation.
To our knowledge, E. faecium 10/96A is only the third glycopeptide-resistant Enterococcus in which an insertion has been identified within a van gene but is the first in VanD-type strains. Disruption of vanY by IS1476 and insertion of IS1216V located towards the 3' end of vanS have been reported in VanA-type strains (35; A. L. Darini, M. F. Palepou, D. James, and N. Woodford, Letter, Antimicrob Agents Chemother. 43:995-996, 1999). The latter insertion leads to the loss of 11 amino acids from the C terminus of the VanS sensor and their possible replacement by 10 amino acids resulting from read-through of the inserted IS1216V (Darini et al., letter). According to the authors, this change would not affect the function of the VanS sensor, because the critical residues remain intact. The disruption of vanY leads to a decrease of its activity but has no phenotypic consequence, since in VanA-type strains, VanY is not necessary for vancomycin resistance (3, 9, 35). In the case of ISEfa4, the insertion led to the production of a truncated VanSD, allowing strain 10/96A to grow in the absence of glycopeptide in the medium.
The frameshift mutation in the vanYD gene of strain 10/96A results in a truncated polypeptide of 118 amino acids lacking the active site of a D,D-carboxypeptidase as indicated by the lack of activity of the protein after cloning the complete gene in E. faecalis JH2-2 (Table 4). The mutation disrupted the reading frame of the VanYD protein of 10/96A, which would otherwise have contained the active site motifs of a PBP (Fig. 4). These motifs are present in VanYD of BM4339 and BM4416, the proteins bind benzylpenicillin, and the D,D-carboxypeptidase activity is inhibited by benzylpenicillin (Fig. 4) (48). The truncated product of the vanYD gene of 10/96A did not bind penicillin (Fig. 6), nor was there any D,D-carboxypeptidase activity in the cytoplasmic fraction, implying that reinitiation of the C-terminal portion of the protein was unlikely to have occurred, particularly as no potential RBS was identified upstream from possible start sites. Surprisingly, substantial D,D-carboxypeptidase activity was detected in the membrane fraction (Table 3). This activity was not inhibited by benzylpenicillin, was not induced by vancomycin, and was presumed to be catalyzed by a totally different protein. Further investigation will indicate whether strain 10/96A has acquired a gene, not present in the vanD gene cluster, which encodes a VanY- or VanYB type protein.
Consideration of the peptidoglycan precursors of the three VanD strains supports this hypothesis. When peptidoglycan synthesis was blocked by ramoplanin in BM4339 and BM4416, UDP-MurNAc-tetrapeptide was present (21 and 24%, respectively) in addition to UDP-MurNAc-pentadepsipeptide (Table 2). As the D-Ala:D-Ala ligase of both these strains is inactive, little if any UDP-MurNAc-pentapeptide would have been present; consequently, tetrapeptide is likely to have resulted from removal of D-lactate from UDP-MurNAc-pentadepsipeptide. PBPs that function as D,D-carboxypeptidases hydrolyze esters in addition to peptides (46); therefore, it is presumed that VanYD of BM4339 and BM4416, which are PBPs and whose activity was inhibited by low concentrations of benzylpenicillin, will hydrolyze UDP-MurNAc-pentadepsipeptide with the production of tetrapeptide. The peptidoglycan precursors accumulated in strain 10/96A were almost exclusively UDP-MurNAc-pentadepsipeptide (95%), and only 2% UDP-MurNAc-tetrapeptide was present. The D,D-carboxypeptidase activity present in VanA- and VanB-type strains and catalyzed by VanY
1-45 preferentially hydrolyzes peptidoglycan precursors terminating in acyl-D-Ala-D-Ala (3). If the D,D-carboxypeptidase in membranes of 10/96A has the same specificity as VanY or VanYB, it would account for the lack of UDP-MurNAc-tetrapeptide in the precursors accumulated in this strain, as virtually no UDP-MurNAc-pentapeptide was available as substrate (Table 2). The D,D-carboxypeptidase activity is not required for resistance, but it is possible that the gene encoding this enzyme was acquired prior to the mutation, leading to a defective D-Ala:D-Ala ligase.
We have shown previously that, in VanA-type strains, increased transcription of the vanHAX operon is associated with increased incorporation of D-Ala-D-Lac into peptidoglycan precursors, to the detriment of D-Ala-D-Ala, and with a gradual increase in vancomycin resistance levels (8). More-complete elimination of D-Ala-D-Ala-containing precursors is required for teicoplanin resistance (8). An unusual feature of the VanD-type strains is their susceptibility to teicoplanin (MIC = 4 µg/ml), despite constitutive production of peptidoglycan precursors that terminate essentially only in D-Ala-D-Lac. The small amount of pentapeptide could have been synthesized by the VanD ligase. Teicoplanin susceptibility has been associated with mutations in the VanS sensor of some VanA-type strains (29).
The nucleotide divergence between the vanD alleles (BM4339, BM4416, and 10/96A) and the geographical dispersion of the isolates (Canada, United States, and Brazil) lead to the hypothesis that the VanD-type strains represent independent introductions in enterococci of gene clusters from undefined donor species.
| ACKNOWLEDGMENTS |
|---|
This work was supported in part by a Bristol-Myers Squibb unrestricted biomedical research grant in infectious diseases.
| FOOTNOTES |
|---|
Present address: Unité des Agents Antibactériens, Institut Pasteur, 75724 Paris Cedex 15. ![]()
| REFERENCES |
|---|
|
|
|---|