Contribution of vraSR and graSR Point Mutations to Vancomycin Resistance in Vancomycin-Intermediate Staphylococcus aureus

The major human pathogen Staphylococcus aureus poses a significant public health threat and is becoming increasingly resistant to currently available antibiotics, including vancomycin, which is the frontline drug for methicillin-resistant S. aureus infections (2, 10). Vancomycin-intermediate Staphylococcus aureus (VISA) is considered to have emerged from vancomycin-susceptible Staphylococcus aureus (VSSA) through multiple genetic alterations (4, 9). We have recently reported that a mutated response regulator gene, graR, of VISA Mu50 could convert the hetero-VISA Mu3 into a VISA strain when it is overexpressed in Mu3 (4, 16). Here, we report a follow-up study with the genetic analysis and engineering of a vancomycin-susceptible clinical S. aureus strain, Mu50Ω, related to Mu50.

Mu50Ω was isolated from the patient from whom strain Mu50 had been isolated one and a half years before. The patient was hospitalized because of a relapsed surgical site infection, and the vancomycin-susceptible strain was identified among VISA isolates indistinguishable from Mu50. Despite its vancomycin susceptibility, the strain possessed the same pulsed-field gel electrophoresis pattern as Mu50; thus, we named it Mu50Ω, suspecting its close relationship with Mu50 (5, 13).

To understand the genetic basis behind Mu50Ω's vancomycin susceptibility, we determined whole-genome sequence of Mu50Ω. Genome sequencing was performed using the chromosome-walking method with Mu50 genome sequence (accession no. BA000017) as a scaffold for primer design as well as orientation of the contiguous PCR fragments of Mu50Ω. A 2,878,428-bp-long whole-genome sequence of Mu50Ω with the same general features of Mu50 genome (14), including G+C content, rRNA operon, tRNA, transfer-messenger RNA, IS431, Tn554, SCCmec, bacteriophage, and genomic island, was successfully determined. The resulting sequence was then used for genome comparison with Mu50, and loci containing nucleotide differences between Mu50Ω and Mu50 chromosomes including single nucleotide polymorphisms (SNPs) or deletion/insertion regions were identified. All nucleotide differences were confirmed by resequencing the corresponding regions of Mu50Ω and Mu50 genomic DNAs. With this, we finally confirmed that there were only a total of 10 SNPs and 3 deletions/insertions which would cause, in either Mu50 or Mu50Ω, a single-amino-acid replacement in the products of four genes, a premature termination of the products of two genes, and a total or partial deletion of four genes (Table 1). These differences were considered as candidates responsible for differences in vancomycin susceptibility between VISA Mu50 and VSSA Mu50Ω. In particular, differences in the vraSR (12) and graSR (4) two-component regulator systems, which are reported to be highly related to vancomycin resistance (4, 7, 11, 12, 15, 17), were evident. The vraS of Mu50Ω differed from that of Mu50 for the presence of a premature stop codon (Fig. 1A). Besides, both Mu50 vraS and Mu50Ω vraS had a T→A transition at nucleotide position 14, generating an Ile5→Asn5 substitution, compared to the vraS genes of VSSA N315 and 12 other VSSA strains for which genomes have been published (Table 1 and Fig. 1A). Concerning graSR, graR of Mu50Ω lacked the mutation identified in Mu50 (Table 1 and Fig. 1A) (4, 16).

To prove the significance of the above mutations in vancomycin susceptibility, we engineered the VSSA Mu50Ω strain stepwise, with individual genetic traits of N315 or Mu50. Briefly, this was carried out by removing the stop codon from the Mu50Ω vraS and subsequently adding the Mu50-specific nonsynonymous mutation into its graR gene, using the plasmid pKOR1-mediated substitution technique as described previously (1, 16). A total of three Mu50Ω mutants were constructed: (i) Mu50Ω-vraSn, which had replacement of Mu50Ω vraS with N315 vraS; (ii) Mu50Ω-vraSm, which had replacement of Mu50Ω vraS with Mu50 vraS; and (iii) Mu50Ω-vraSm-graRm, which had replacement of Mu50Ω vraS and graR with Mu50 vraS and Mu50 graR (Fig. 1A). With each step of the above construction, we observed a stepwise increase of vancomycin resistance, from VSSA to the VISA phenotype. MIC determination demonstrated that the resulting constructs had remarkable increases in vancomycin MIC, from 0.5 mg/liter for Mu50Ω to 3.5 mg/liter for Mu50Ω-vraSn, 4.5 mg/liter for Mu50Ω-vraSm, and 6 mg/liter for Mu50Ω-vraSm-graRm (the latter value is equivalent to that of Mu50). Consistent with the increased vancomycin MICs, population analysis showed that Mu50Ω-vraSn and Mu50Ω-vraSm had a significant increase in vancomycin-resistant subpopulations compared to the parent strain Mu50Ω, and Mu50Ω-vraSm-graRm had almost the same vancomycin resistance profile as Mu50 (Fig. 1B), exhibiting the conversion of VSSA Mu50Ω into a VISA strain. Toward other antibiotics targeting cell wall synthesis, including teicoplanin, bacitracin, and β-lactams, the mutants also showed significant decreased susceptibility, whereas no changes were observed in the susceptibility to other antibiotic groups—e.g., aminoglycosides, quinolones, and tetracyclines (Table 2) —indicating that the genome engineering of vraS and graR mutations did not affect Mu50Ω's susceptibility profile toward non-cell-wall-associated antibiotics. Altogether, it is interesting to note that (i) loss of vancomycin resistance in Mu50Ω is due to the truncated translation of VraS protein; (ii) full translation of VraS protein seems to be indispensable for the resistance to cell-wall-associated antibiotics; (iii) restoration of vraS function alone did not raise vancomycin resistance of Mu50Ω to the level of VISA Mu50; and (iv) a subsequent point mutation in graR is needed for the enhancement of vancomycin resistance in a Mu50Ω clone with a mutated vraS to reach that of Mu50.

This study clearly revealed that direct engineering of the Mu50Ω genome with Mu50 vraS partially restores its vancomycin resistance, and with a further Mu50 graR substitution on the mutant's chromosome, a vancomycin resistance level similar to that of Mu50 is achieved. This reconfirmed our view that activation of VraSR and GraSR two-component regulatory systems in VSSA will lead to a VISA phenotype through the activation of the genes under their regulation (13, 16).

The two-component regulator systems vraSR and graSR were identified as glycopeptide resistance-associated genes in previous studies (4, 12) and are known to positively modulate the regulation of cell wall biosynthesis (11, 12, 16). Cell wall thickening is the molecular mechanism of vancomycin resistance in VISA, which is associated with the peptidoglycan-clogging mechanism that prevents the passage of vancomycin through a thickened peptidoglycan layer (3, 5, 6, 8). VraSR is known to be important in the uptake of materials needed for cell wall synthesis (12), and the activation of vraSR, which upregulates the expression of enzymes involved in the peptidoglycan synthesis pathway such as PBP2, PBP1A/1B, MurZ, etc., could occur by a single point mutation in vraS (Y. Katayama et al., submitted for publication). Overexpression of mutated graR, designated graRm (“m” stands for Mu50), in a vraSR-activated strain, hetero-VISA Mu3, raised vancomycin MIC and cell wall thickness significantly, while introduction of the intact graR into Mu3 did not increase vancomycin MIC and cell wall thickness appreciably (16). No significant increase was observed in vancomycin MIC and cell wall thickness when graRm was introduced into N315 that has an intact vraS (16). Although detailed studies on the physiological function and activation mechanism of vraSR and graSR are necessary, the data presented in this study clearly indicate that the combination effect of graSR and vraSR mutations found in VISA Mu50 contributes to vancomycin resistance. Nevertheless, whether the role of vraSR and graSR found in the Mu50 cell lineage could be applied to other clinical VISA strains in VISA phenotype conversion remains to be investigated.