INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

Over the last few years, Enterococcus has emerged as an important bacterial pathogen in nosocomial infections (13). The acquisition of specific mechanisms of resistance to different antibiotics, especially for the species Enterococcus faecium, has rendered infections with these microorganisms difficult to treat (8, 25); in just 10 years, antibiotic resistance has spread rapidly among enterococci and has become an important public health concern (11, 14). Macrolide-lincosamide-streptogramin (MLS) antibiotics constitute an alternative therapy for the treatment of insidious enterococcal infections. Three different mechanisms account for the acquired resistance to MLS antibiotics in gram-positive bacteria: modification of the drug target, inactivation of the drug, and active efflux of the antibiotic. In the first case, a single alteration of the 23S rRNA confers broad cross-resistance to macrolide-lincosamide-streptogramin B (MLS_B) antibiotics, whereas the inactivation mechanism confers resistance only to structurally related MLS antibiotics. Regarding the pump mechanisms, the mefA (4), mefE (34), msrA (29), and mreA (5) genes have been involved in the active efflux of macrolides in gram-positive bacteria. The mef and mreA genes have been associated with macrolide resistance, and the msrA gene has been associated with macrolide and streptogramin B resistance. Erythromycin resistance by erm methylases of the ermB-ermAM hybridization class has been described in Enterococcus isolates (3, 15, 19). However, even though some reports indicate the presence of a putative erythromycin efflux pump in this bacterial genus (21; H. Fraimow and C. Knob, Abstr. 97th Gen. Meet. Am. Soc. Microbiol., 1997, abstr. A-125, p. 22, 1997), little is known of the presence of such resistance determinants in enterococci.

The work described here was designed to study the presence of different erythromycin resistance genes in Enterococcus isolates of different species and with a variety of erythromycin susceptibility patterns. A novel intrinsic gene that encodes a putative ABC transporter was identified in all E. faecium isolates and presumably accounts for the higher macrolide MICs for this species in comparison with those for other enterococci (27).

(This work was presented in part at the 39th Interscience Conference on Antimicrobial Agents and Chemotherapy, San Francisco, Calif., 26 to 29 September 1999.)

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Bacterial isolates. Seventy-eight isolates of different Enterococcus species with a variety of erythromycin susceptibility patterns were included in this study (see Table 1): E. faecalis, n = 27; E. faecium, n = 23; E. durans, n = 8, E. avium, n = 6; E. hirae, n = 9; E. gallinarum, n = 3; and E. casseliflavus, n = 2. Sixty-three isolates were obtained from human clinical samples from the Hospital San Millán of Logroño, Spain; 2 E. faecium, 8 E. hirae, and 2 E. gallinarum isolates were of animal origin; and 1 E. hirae isolate (CECT 302), 1 E. gallinarum isolate (CECT 970), and 1 E. casseliflavus isolate (CECT 969) were from the Spanish Culture Type Collection. Species identification was based on the biochemical API 20 Strep system (BioMerieux, la Balme, France) and was also carried out according to the biochemical scheme of Facklam and Collins (10).

Susceptibility testing. Susceptibility testing was performed by the agar dilution method in Mueller-Hinton (MH) agar (Difco, Detroit, Mich.) by the standard method of the National Committee for Clinical Laboratory Standards (26). The antibiotics tested were erythromycin and spiramycin (Sigma Chemical Co.), azithromycin (Hoechst Marion Roussel, Romainville, France), and pristinamycin I (Rhône Poulenc Rorer, Paris, France).

Antibiotic inactivation test. A bioassay for the detection of the erythromycin inactivation mechanism was performed (12, 35). Enterococcus strains were incubated in brain heart infusion broth (Difco) with 40 µg of erythromycin per ml for 48 h. After centrifugation, 25 µl of the supernatant was deposited on sterile disks over MH agar plates previously seeded with Micrococcus luteus ATCC 9341. The plates were incubated for 24 h, and the zone sizes around the disks, which indicate the antibiotic remaining in the culture medium after incubation with cells, were measured. The zone sizes were compared with those produced when the antibiotic was incubated with either medium alone or with the erythromycin-susceptible E. faecium AR10 (erythromycin MIC, 0.5 µg/ml).

PCR analysis of erythromycin resistance genes. The presence of genes involved in MLS resistance with a methylation mechanism was determined by PCR amplification of known erm genes by using primers specific for ermA, ermB, and ermC (33) and for ermTR (17, 32). The presence of genes involved in antibiotic efflux systems was determined by PCR with gene-specific primers and conditions for the amplification of the mefA/E (33) and msrA (35) genes. The low-level erythromycin-resistant Enterococcus isolates that gave negative results in the PCR experiments described above were analyzed by using degenerate erm primers (2). Positive and negative controls were included in all experiments. Genomic DNA for PCR analysis was obtained with the Instagene matrix system (Bio-Rad) according to the manufacturer's instructions.

DNA isolation and Southern blot hybridization. Plasmid and genomic DNAs from Enterococcus isolates were extracted by alkaline lysis as described previously (31) by lysozyme treatment for 1 h. Southern blotting was performed with total DNAs and plasmid DNAs from erythromycin-resistant and erythromycin-susceptible E. faecium isolates by using an msrA-like PCR fragment (obtained from erythromycin-resistant isolate E. faecium E134) as the probe. This probe was labeled with digoxigenin according to the manufacturer's instructions (Boehringer Mannheim). The homology of ermTR PCR amplicons with the ermTR gene was also analyzed by Southern blotting. An ermTR fragment from a group G Streptococcus strain, strain S211, which contained the ermTR gene, was labeled with digoxigenin and was used as a probe. Hybridizations were carried out at 50°C, and in all cases, positive and negative controls were included.

DNA sequencing. The amplicons were obtained with msrA-specific primers (35) by PCR analysis of genomic DNAs of the following strains: E. faecium AR10 (erythromycin MIC, 0.5 µg/ml), E. faecium E136 (erythromycin MIC, 32 µg/ml) and E. faecium E134 (erythromycin MIC, >128 µg/ml). The amplicons were then purified and sequenced. The amplicon obtained with ermA-specific primers (33) by PCR analysis of genomic DNA of E. faecalis E307 was also purified and sequenced. Automatic sequencing (ABI 310 gene sequencer; Perkin-Elmer) was carried out by using the same primers used for the PCRs. Analysis of the sequences was performed with the aid of Wisconsin Package (version 9.1; Genetics Computer Group, Madison, Wis.).

Nucleotide sequence accession number. The nucleotide sequence of the E. faecium msrC gene has been assigned GenBank accession no. AJ243209.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

The MICs of erythromycin, azithromycin, spiramycin, and pristinamycin I were determined for 78 strains belonging to seven different Enterococcus species. Two groups of strains (highly resistant and susceptible) could be distinguished among all Enterococcus species, depending on the MICs obtained. One E. faecium strain (strain E136) with a low level of resistance to the macrolides was also found. All the Enterococcus isolates included in this study were classified according to their species and their erythromycin susceptibility patterns, and a variety of erythromycin resistance mechanisms was investigated by PCR (Table 1). These groups proved to be homogeneous: all strains that belonged to the same group had the same antibiotic resistance determinants.

The inactivation tests described above were performed with 19 strains of the seven different enterococcal species included in this study with a variety of erythromycin susceptibility patterns. No significant differences in zone sizes were observed between susceptible and resistant strains, nor were significant differences in zone sizes observed when they were compared with the zone sizes for erythromycin incubated under the same conditions but without the presence of bacterial cells. These results indicate that the strains that were analyzed do not express a detectable mechanism of erythromycin inactivation under these conditions.

Presence of erm methylase genes. The Enterococcus isolates were analyzed for the presence of erm methylase genes by PCR by using specific conditions for detection of the erm genes characterized in gram-positive bacteria (see Materials and Methods). When PCR analysis was carried out with specific primers for the amplification of the ermB gene, a band with the expected molecular size (639 bp) was obtained for 39 of the 40 highly erythromycin-resistant Enterococcus isolates (MICs, >128 µg/ml), independently of the species involved (12 E. faecium, 14 E. faecalis, 2 E. durans, 2 E. avium, 8 E. hirae, and 1 E. gallinarum isolates) (Table 1). For all these strains, the MICs of azithromycin, spiramycin, and pristinamycin I were always 64 µg/ml. The remaining highly resistant E. faecium isolate was positive by PCR with the ermA-specific primer and negative by PCR with ermB- and ermC-specific primers. This ermA amplicon was sequenced and was found to have 100% homology with the ermA gene described for Staphylococcus aureus (24); this gene has been frequently associated with macrolide resistance in S. aureus and coagulase-negative staphylococci (20). To our knowledge, this is the first description of the ermA gene in enterococci. No PCR fragment of the expected size was obtained from any of the enterococcal isolates for which erythromycin MICs were 32 µg/ml (10 E. faecium, 13 E. faecalis, 6 E. durans, 4 E. avium, 1 E. hirae, 2 E. gallinarum, and 2 E. casseliflavus isolates) with either ermA-, ermB-, or ermC-specific primers. The same results were obtained when PCRs were carried out with degenerate erm-specific primers. These results indicate that erm genes could be present only in highly macrolide-resistant strains of Enterococcus.

Presence of erythromycin efflux mef genes. The presence of erythromycin-efflux genes in Enterococcus isolates was analyzed by PCR with primers specific for the mefA and mefE genes. mef efflux pump genes have been detected in S. pyogenes (4), S. pneumoniae (34), and Streptococcus agalactiae (1), as well as in Micrococcus luteus, Corynebacterium jeikeiium, Corynebacterium spp., and viridans group streptococci (21). Previous reports have indicated that mefE might have an important role in erythromycin resistance in E. faecium; according to studies carried out in the United States (Fraimow and Knob, Abstr. 97th Gen. Meet. Am. Soc. Microbiol. 1997), 42% of the resistant strains have been reported to carry this determinant. In the same way, very recently, Luna et al. (21) have reported on the presence of mef genes in Enterococcus spp. However, we were unable to detect amplification with any of the enterococcal isolates when the mefA/E-specific primers were used in PCR analysis. This result may reflect a different geographical distribution of mef genes in E. faecium. In one of our samples (E. faecium E136), a faint band of 450 bp, which was larger than expected (348 bp), was obtained. E. faecium E136 was the only isolate with a low-level erythromycin resistance phenotype (MIC, 32 µg/ml) (Table 1). The homology of the 450-bp amplicon obtained from E. faecium E136 was analyzed by Southern blotting with a mefA-specific probe (obtained from S. pyogenes S2). The amplicon gave a positive signal upon hybridization at 50°C (data not shown). The positive results by both PCR analysis and Southern hybridization indicated the presence of a DNA sequence related to the mef sequence in this isolate (A. Portillo, A. Alonso, F. Ruiz-Larrea, M. Zarazaga, J. L. Martinez, and C. Torres, Abstr. 38th Intersci. Conf. Antimicrob. Agents Chemother., abstr. C-122, 1998). However, the different size of the PCR fragment indicated that it was not the same mef gene so far described.

msrC gene. PCR analysis for determination of the presence of msrA produced an unexpected result. The msrA gene, first identified in Staphylococcus epidermidis (29), confers resistance by an efflux system after induction with erythromycin, and an msrA-related gene, msrB, has been described in Staphylococcus xylosus (23). All our E. faecium isolates gave a PCR amplification product with the expected molecular size with msrA-specific primers (Fig. 1), irrespective of their MLS resistance phenotypes. Nevertheless, no msrA gene was found in any of the other Enterococcus species by the same PCR protocol. Similar results were obtained by hybridization at 50°C by using the msrA-like PCR fragment from E. faecium E134 as a probe; positive results were obtained for E. faecium strains, and negative results were obtained for the other enterococcal species. The msrA-like amplicons obtained from susceptible strain E. faecium AR10 (erythromycin MIC, 0.5 µg/ml), low-level-resistant strain E. faecium E136 (MIC, 32 µg/ml), and highly resistant strain E. faecium E134 (MIC, >128 µg/ml) were sequenced. No differences were observed among the sequences of the three fragments. However, even though the amplicons were of the expected molecular size, the sequence obtained was different from that of msrA. When compared with msrA, this novel gene showed an identity of 62% at the DNA level and a similarity of 72% at the amino acid level, with an overlap of 135 amino acids. This new gene was named msrC (GenBank accession no. AJ243209). The homology was extremely high for the regions that contained the nucleotide binding motifs and the signature sequence included in the ABC transporter domain described for other MLS efflux determinants from gram-positive bacteria (erythromycin, tylosin, carbomycin, pristinamycin, virginiamycin) (Fig. 2). This analysis strongly suggests that msrC also belongs to this efflux pump gene family. Figure 2 shows the two nucleotide-binding motifs of the msrC gene: motif A, which corresponds to a P loop, and motif B, which, together with the signature pattern for this class of ABC transporters, is located between the A and the B motifs of the ATP-binding site (30).

View larger version (62K): [in this window] [in a new window]   FIG. 1.   PCR amplification with msrA-specific primers. Lanes: 1, E. faecium E134 (erythromycin MIC, >128 µg/ml); 2, E. faecium E136 (MIC, 32 µg/ml); 3, E. faecium AR10 (MIC, 0.5 µg/ml); 4, E. faecalis E121 (MIC, >128 µg/ml); 5, E. durans AR23 (MIC, >128 µg/ml); 6, E. avium E402 (MIC, >128 µg/ml); 7, E. hirae P9 (MIC, >128 µg/ml); 8, E. gallinarum AR45 (MIC, >128 µg/ml); 9, E. casseliflavus C85 (MIC, 4 µg/ml). The DNA band of 405 bp is marked with an arrow.

View larger version (85K): [in this window] [in a new window]   FIG. 2.   Comparison of MsrC amino acid sequence with other sequences of ABC transporter. MsrC, AJ243209 efflux pump in E. faecium; MsrA, P23212 erythromycin resistance protein in S. epidermidis; MsrB, M81802 erythromycin resistance protein in S. xylosus; VgA, JC1204 putative ATP-binding protein involved in resistance to virginiamycin in S. aureus; VgaB, AAB95639 pristinamycin resistance protein in S. aureus; CarA, AAC32027 carbomycin resistance protein in Streptomyces thermotolerans; TlrC, P25256 tylosin resistance protein in Streptomyces fradiae; YDIF, O05519 ABC transporter (ATP-binding protein) in Bacillus subtilis. Nucleotide binding motifs and the ABC transporter signature (PS00211) are indicated by shaded areas.