Interaction of the GraRS two-component system with the VraFG ABC transporter to support vancomycin-intermediate resistance in Staphylococcus aureus

Abstract

Current treatment for serious infections caused by methicillin-resistant Staphylococcus aureus relies heavily upon the glycopeptide antibiotic vancomycin. Unfortunately, this practice has led to an intermediate resistance phenotype that is particularly difficult to treat in invasive staphylococcal diseases, such as septicemia and its metastatic complications, including endocarditis. Although the vancomycin-intermediate resistance phenotype has been linked to abnormal cell wall structures and autolytic rates, the corresponding genetic changes have not been fully elucidated. Previously, whole-genome array studies listed numerous genes that are overexpressed in vancomycin-intermediate sensitive strains, including graRS (SACOL0716 to -0717), encoding a two-component regulatory system (TCRS), as well as the adjacent vraFG (SACOL0718 to -0720), encoding an ATP-binding cassette (ABC) transporter; but the exact contribution of these genes to increased vancomycin resistance has not been defined. In this study, we showed that isogenic strains with mutations in genes encoding the GraRS TCRS and the VraFG ABC transporter are hypersensitive to vancomycin as well as polymyxin B. Moreover, GraRS regulates the expression of the adjacent VraFG pump, reminiscent of gram-positive bacteriocin-immunity regulons. Mutations of graRS and vraFG also led to increased autolytic rates and a more negative net surface charge, which may explain, in part, to their increased sensitivity to cationic antimicrobial peptides. Taken together, these data reveal an important genetic mediator to the vancomycin-intermediate S. aureus phenotype and may hold clues to the selective pressures on staphylococci upon exposure to selective cationic peptide antibiotics used in clinical practice.

FIG. 1.

Hypersensitivity of a COL graR mutant to VAN. (A) Growth curves of wild-type COL and four response regulator mutants in CSMHB medium without VAN. (B) Cells grown with 1 μg/ml VAN. Note that only the COL graR mutant exhibited a growth defect. (C) All strains carry a pEPSA5-based plasmid (see Materials and Methods). Cells grown with VAN are indicated by (V). Note that only the noncomplemented COL graR mutant retains a growth defect in the presence of VAN. Similar results were obtained with mutants in the RN6390 background (not shown).

FIG. 2.

Regulation of the graRS locus. (A) Diagram of the 5,800-bp chromosomal locus including the graRS and vraFG genes. Shown are the SA open reading frame designation of the COL genome, secondary gene names, proposed functions as given in Entrez Gene (27), gene sizes, translated protein sizes, and amino acid changes between proteins of the COL and Mu50 strains. (B, C, and D) Shown in the upper panels are autoradiographs from the Northern blot assay. To the left of the images are RNA size standards (designated in kb), while to the right of the images are arrows indicating likely cross-reactivity of the labeled probe to rRNA bands (B and C). The lower panels contain the ethidium bromide-stained RNA gels of 23S and 16S rRNA prior to Northern blotting to indicate comparable loading between strains. WTR, wild-type revertant. (B) Northern analysis using the indicated Mu50-derived RNA that was probed with a 400-bp ³²P-labeled DNA fragment internal to the graS gene. Note the presence of a transcript, albeit expectedly smaller (∼2.2 kb), in the Mu50 graR mutant strain (gray arrow). (C) Northern analysis using the indicated Mu50-derived RNA that was probed with a 400-bp ³²P-labeled DNA fragment internal to the vraG gene. Note the greatly reduced level of vraFG transcript in the Mu50 graR mutant strain. (D) Northern analysis using the indicated COL-derived RNA that was probed with a 400-bp ³²P-labeled DNA fragment internal to the vraG gene. RNA was isolated from wild-type COL (lanes 1 to 3), the graR mutant (lanes 4 to 6), the graR mutant with graR expressed from pEPSA5 (lanes 7 to 9), and the graR mutant with empty vector pEPSA5 (lanes 10 to 12). Cells were treated with an antibiotic (Ab) 10 minutes prior to RNA isolation: V, VAN (20 μg/ml); I, imipenem (20 μg/ml); -, no treatment.

FIG. 3.

Autolysis of Mu50 wild-type and mutant strains. The graph reveals the reduction of bacterial density over time in the presence of buffer containing 0.05% Triton X-100 at 30°C (see Materials and Methods). The mean percent loss of cell density at the indicated times are overlaid with standard deviations from three independent experiments. *, statistical significance of the indicated strain with respect to the Mu50 parental wild-type strain at the 3-hour time point, as determined by the paired Student t test (P < 0.01). WTR, wild-type revertant.

FIG. 4.

Binding of cytochrome c to whole S. aureus cells. The graph shows percent binding of cytochrome c after 10 min of incubation with S. aureus at room temperature. Detailed below the graph are the parental strain (Mu50 or COL) and genotypes of various genetic constructs. Data represent the means and standard deviations from three independent experiments. *, statistical significance of the indicated strain compared to the respective control strain (lane 1 versus lanes 2 to 5, lane 6 versus lane 7, lane 8 versus lane 9, and lane 10 versus lanes 11 to 13), as determined by the paired Student t test (P < 0.01). WTR, wild-type revertant.