Genetic Basis of Resistance to Fusidic Acid in Staphylococci^▿

Bacterial strains and culture.

Table 1 lists staphylococcal strains used and generated in the present study. S. lugdunensis strains were isolated from patients in Sweden and were identified on the basis of typical morphology and/or smell, resistance to desferrioxamine, and ornithine decarboxylase and pyrrolidonyl arylamidase activity (1). Escherichia coli XL-10 Gold and XL-1 Blue (Stratagene, Amsterdam, The Netherlands) were used as cloning hosts.

Unless otherwise stated, strains were cultured in Luria-Bertani broth with aeration or on Luria-Bertani agar at 37°C. Susceptibility testing and determination of bacterial doubling times were performed as described previously (9).

DNA manipulation.

Recombinant DNA methods were standard (16). Routine PCR amplification was performed with Platinum Pfx (Invitrogen, Paisley, United Kingdom), while the FailSafe PCR System (Epicenter, Madison, WI) was used for long PCR. Southern hybridization for detection of fusB (14) was performed on EcoRI-digested DNA at 60°C. Detection of fusC was performed in an identical fashion but used a probe generated by PCR with the oligonucleotide primers fusCU and fusCL (Table 2).

Recombinant DNA constructs were routinely created by using plasmid pCU1 (2). Constructs for tetracycline-regulated gene expression were generated by using plasmid pAJ96, a derivative of pALC2073 (3) that carries a transcriptional terminator downstream of the cloning site. For the former, PCR amplicons were introduced into the BamHI restriction site of pCU1 by virtue of engineered BamHI restriction sites at the 5′ end of oligonucleotide primers (Table 2), while engineered KpnI/SacI restriction sites were used for ligation into pAJ96. Constructs were propagated in E. coli before recovery and introduction into S. aureus RN4220 by electroporation. Tetracycline (100 ng/ml) was used to induce expression from the xyl/tetO promoter on pAJ96.

GenBank accession numbers.

The fusA gene of S. intermedius (NCTC 11048) has been assigned GenBank accession number AY776250, and that from S. lugdunensis was assigned GenBank accession number DQ866810.

FusB homologues mediate resistance to fusidic acid in S. aureus and S. saprophyticus.

A whole-genome sequence analysis of S. aureus strain MSSA476 was recently reported (8). This strain is resistant to fusidic acid and carries a gene (SAS0043) encoding a homologue (YP_042173) of the FusB protein. The gene and the fusidic acid resistance phenotype may be associated (8), although this has not been confirmed. We also identified a staphylococcal gene (SSP2165) encoding a FusB homologue (YP_302255) in the genome of S. saprophyticus ATCC 15305 (10). Although most staphylococcal species are inherently susceptible to fusidic acid, S. saprophyticus is intrinsically resistant to the antibiotic (6). These two staphylococcal FusB homologues exhibit 44% (YP_042173) and 47% (YP_302255) identity to FusB and 41% identity to each other (Fig. 1).

To establish whether these FusB-like proteins represent functional homologues of FusB (i.e., confer resistance to fusidic acid), the genes encoding them were PCR amplified, along with their upstream expression signals, using oligonucleotide primers SAS0043U/L and SSP2165U/L for SAS0043 and SSP2165, respectively (Table 2) and introduced into S. aureus RN4220 on plasmid pCU1. Both genes conferred resistance to fusidic acid (Table 1). Consequently, these proteins appear to form part of a FusB protein family that confer resistance to fusidic acid in staphylococci. Based on the precedent of ≤80% amino acid identity to represent the dividing line between one resistance determinant and a related one (12), we have designated these proteins FusC (YP_042173 from MSSA476) and FusD (YP_302255 from ATCC 15305), and their corresponding genes were designated fusC and fusD.

To establish the relative abilities of the FusB, FusC, and FusD proteins to confer resistance in S. aureus, constructs were generated enabling identical expression of the three genes. Thus, fusC and fusD were PCR amplified with oligonucleotide primers MSSAkpn/MSSAsac and SAPkpn/SAPsac (Table 2), resulting in upstream regions and ribosome binding sites identical to those of fusB. The fusB gene was amplified by using primers fusBkpn/fusBsac (Table 2). All three genes were expressed under identical induction conditions from the xyl/tetO promoter on plasmid pAJ96 in S. aureus RN4220. Under these conditions, FusB demonstrated a greater ability to protect S. aureus from fusidic acid compared to FusC and FusD (Table 1).

Carriage of antibiotic resistance determinants can impose a fitness cost in bacteria. Since constructs enabling equivalent expression of all three resistance genes had been constructed, it was possible to examine whether expression of fusB, fusC, or fusD was associated with a fitness cost in S. aureus. However, strains expressing these genes did not exhibit doubling times significantly different from RN4220 carrying pAJ96 (data not shown).

Resistance to fusidic acid mediated by FusB homologues in clinical strains of S. aureus.

A proportion of fusidic acid-resistant S. aureus clinical strains carry neither fusB nor mutations in fusA (14). Such strains might be resistant to fusidic acid through expression of FusB homologues. Relevant strains (CS strains; Table 1) were screened for the presence of genes encoding FusB homologues by Southern hybridization. All strains tested positive for fusC (Fig. 2). To establish whether these strains harbored fusC or polymorphic variants cross-hybridizing with fusC, the fusC genes from these strains were PCR amplified and sequenced (primers SAS0043U and SAS0043L; Table 2). Since these primers correspond to the start and end of the fusC gene, this analysis could not provide DNA sequence information for the gene termini. However, a DNA sequence for >95% of the gene was obtained. The fusC genes encoded protein products identical (in the sequenced portion) to that encoded by fusC from MSSA476. However, fusC from strain CS979 carried a silent mutation (S₁₆₂ was encoded by TCC rather than TCT).

Resistance to fusidic acid in S. intermedius and S. lugdunensis.

Resistant strains of S. intermedius (7) and S. lugdunensis were ∼64-fold less susceptible to fusidic acid than wild-type, sensitive strains of the same species (Table 1). These resistant strains were screened initially for nucleotide substitutions in fusA. Since the DNA sequence of fusA from these species was not available, oligonucleotide primers for PCR amplification and sequencing of this locus were designed against regions of the upstream (rpsG) and downstream genes (tufA) that are conserved between S. aureus and Bacillus subtilis. Thus, the primers rpsU and tufL (Table 2) were used to PCR amplify the entire fusA gene from both fusidic acid-sensitive and resistant strains, which were then sequenced in toto.

The fusA genes of S. lugdunensis and S. intermedius encode proteins with 95 and 92% identity, respectively, to EF-G from S. aureus N315. No nucleotide polymorphisms were identified in the fusA genes of the fusidic acid-resistant S. lugdunensis strains relative to the sensitive strains. However, six coding polymorphisms in fusA from S. intermedius CL1 and CL2 were identified compared to S. intermedius NCTC 11048 (NCTC 11048 residue shown first); N₁₈₇T, V₁₉₇T, D₂₃₁E, N₂₉₄D, S₃₆₈T, and T₄₈₈S. To examine whether these polymorphisms were responsible for the observed fusidic acid-resistance phenotype in CL1/CL2, the fusA genes of CL1, CL2, NCTC 11048 (negative control) and S. aureus CS957 (positive control; Table 1) were expressed in trans from pAJ96 in S. aureus RN4220. Expression of fusA from CS957 conferred resistance to fusidic acid, although expression of the S. intermedius fusA genes did not (data not shown).

Subsequently, representative fusidic acid-sensitive and -resistant strains of S. lugdunensis and S. intermedius were tested for the presence of fusB and fusC by Southern hybridization. Fusidic acid-resistant S. intermedius carried fusC (data not shown). The sequences of the fusC genes from S. intermedius CL1 and CL2 were identical to those found in MSSA476, with the exception of the same silent mutation identified in strain CS979.

Fusidic acid-resistant S. lugdunensis strains carried fusB (Fig. 2), a finding that was confirmed by PCR amplification and sequencing. We recently determined that fusB in the epidemic European fusidic acid-resistant impetigo clone (EEFIC) of S. aureus is located on a genomic island called SaRI_fusB, integrated into the chromosome downstream of the groEL gene (14a). Since fusB is present on a similar-sized EcoRI fragment in fusidic acid-resistant S. lugdunensis and in representatives of the EEFIC such as S. aureus 74136 (Fig. 2), we postulated that fusB may reside at the same chromosomal locus in S. lugdunensis. Southern hybridization established that the fusB-positive S. lugdunensis strains carried this resistance gene on the chromosome, since no hybridization of the fusB probe was detected with plasmid preparations (data not shown). Furthermore, using oligonucleotide primers specific for fusB (fusU; Table 2) and conserved regions of staphylococcal groEL (groL; Table 2), a PCR product (∼15 kb) was generated from S. aureus 74136 and the fusidic acid-resistant S. lugdunensis strains but not from fusidic acid-sensitive strains or from those that carry plasmid-borne fusB (Fig. 2).