Spontaneous deletion of the methicillin resistance determinant, mecA, partially compensates for the fitness cost associated with high-level vancomycin resistance in Staphylococcus aureus

Abstract

Treatment of infections caused by Staphylococcus aureus is often confounded by the bacterium's ability to develop resistance to chemotherapeutic agents. Methicillin-resistant S. aureus (MRSA) arises through the acquisition of staphylococcal chromosomal cassette mec (SCCmec), a genomic island containing the methicillin resistance determinant, mecA. In contrast, resistance to vancomycin can result from exposure to the drug, a mechanism that is not dependent upon a gene acquisition event. Here we describe three MRSA strains that became resistant to vancomycin during passage in the presence of increasing concentrations of the drug. In each case two derivative strains were isolated, one that had lost mecA and one that retained mecA during passage. Strain 5836VR lost mecA by the site-specific chromosomal excision of SCCmec, while the other two strains (strains 3130VR and VP32) deleted portions of their SCCmec elements in a manner that appeared to involve IS431. Conversion to vancomycin resistance caused a decrease in the growth rate that was partially compensated for by the deletion of mecA. In mixed-culture competition experiments, vancomycin-resistant strains that lacked mecA readily outcompeted their mecA-containing counterparts, suggesting that the loss of mecA during conversion to vancomycin resistance was advantageous to the organism.

FIG. 1.

Detection of mecA. (A) PCR for the detection of mecA from the strains indicated along the top. (B) SmaI-digested total cellular DNA separated by PFGE. Black arrows show bands present in the parental strains and the vancomycin-resistant, mecA-positive strains that shift (white arrowheads) in the vancomycin-resistant, mecA-negative strains, indicating a loss of DNA. Molecular weights (in kilobase pairs) are indicated on the left.

FIG. 2.

Schematic of SCCmec elements and the deletions in vancomycin-resistant derivatives. Depicted are the SCCmec elements of the parental strains as well as the remnants of SCCmec present in the vancomycin-resistant, methicillin-susceptible derivatives. The region deleted from the parent to the derivative is indicated by dotted lines. Selected open reading frames (orfs) are shown as block arrows, identified in the key at the bottom. All parental strains contained SCCmec type II. The deletions in strains 3130VR and VP32 mapped precisely to the copy of IS431 flanking PUB110, while 5836VR deleted SCCmec in a manner consistent with CcrAB-mediated recombination. All deletions included the mec operon (mecA, mecRI, and mecI) as well as ccrAB.

FIG. 3.

Site-specific SCCmec excision in strains 3130 and 5827. CcrAB-mediated excision of SCCmec was monitored by use of a PCR-based technique. PCR was performed with total cellular DNA isolated from strains 3130 and 5827 containing ccrAB on a multicopy plasmid. (A) Schematic of the region of the staphylococcal chromosome where SCCmec is integrated (top), the staphylococcal chromosome where SCCmec has excised (middle), and the excised, circular SCCmec element (bottom). Selected open reading frames are shown as block arrows for orientation; vertical lines indicate SCCmec attachment sites; and small triangles represent the primers used to amplify the right SCCmec attachment site (triangles 1), the left SCCmec attachment site (triangles 2), the chromosomal junction from which SCCmec excised (triangles 3), and the excised, circular SCCmec element (triangles 4). (B and C) Agarose gel electrophoresis of PCR amplification products from strain 3130 and 5827, respectively. The numbers used to designate the lanes correspond to the primer sets depicted by triangles in panel A.

FIG. 4.

Growth analysis. The doubling time is shown for each strain listed. Doubling times with error bars (standard deviations) were calculated from the exponential growth phases of seven independent growth curves prepared for each strain. Statistical analysis for comparison of the results for each derivative strain to those for its parent strain was performed by a paired t test.

FIG. 5.

Mixed-culture competition experiments. Equal numbers of two different strains were mixed in liquid culture and passaged for up to 5 days. The bacteria were enumerated by plating them on differential selective medium. Differential selectable markers were introduced on plasmids pRN5543 (chloramphenicol) and pCN36 (tetracycline). Each strain pair was competed twice, with the plasmids switched to correct for the possibility of differential fitness costs of each plasmid.

FIG. 6.

Competition of 3130V32/pRN5543 and 3130V32ΔmecA. To determine if the fitness differential evident with the deletion of portions of SCCmec was due to the loss of mecA, 3130V32/pRN5543 and 3130V32ΔmecA were analyzed in mixed-culture competition experiments. Strain 3130V32ΔmecA is strain 3130V32 with mecA insertionally inactivated by tetM. Equal numbers of 3130V32/pRN5543 and 3130V32ΔmecA cells were mixed in liquid culture and passaged for 8 days. The bacteria were enumerated by plating them on differential selective medium (3130V32/pRN5543 on chloramphenicol and 3130V32ΔmecA on tetracycline).

FIG. 7.

Competition of 450MHomo, 450MHomoex, and 450MHomoΔmecA. Equal numbers of two different strains were mixed in liquid culture and passaged for up to 5 days. The bacteria were enumerated by plating them on differential selective medium (450MHomo on oxacillin, and 450MHomoΔmecA on tetracycline; 450MHomoex could not be directly selected for, and so the total colony count on antibiotic-free medium is shown).