Induction of ermAMR from a Clinical Strain of Enterococcus faecalis by 16-Membered-Ring Macrolide Antibiotics

Abstract

We cloned the MLS_B resistance determinant by PCR from a clinical isolate of Enterococcus faecalis 373, which is induced more strongly by a 16-membered-ring macrolide, tylosin, than by erythromycin. To elucidate the molecular basis of resistance of E. faecalis 373, we analyzed the cloned gene, designated ermAMR, by site-directed mutagenesis and reporter gene assay. Our results showed that an arginine-to-cysteine change in the seventh codon of the putative leader peptide endowed tylosin with resistance inducibility and that TAAA duplication enabled the control region to express the downstream methylase gene at a drastically increased level.

Macrolide-lincosamide-streptogramin B (MLS_B) antibiotics are chemically distinct but have similar effects on ribosomes. They bind to the peptidyltransferase region of 50S ribosomes, resulting in the inhibition of protein synthesis (17). Resistance to MLS_B antibiotics due to methylation of specific sites of 23S rRNA by N⁶-methyltransferase encoded by the erm gene family is widespread among clinical strains (8). In the case of ermC from Staphylococcus aureus, resistance is induced by the 14-membered-ring macrolide erythromycin and lincosamide celesticetin and is regulated by a translational attenuation mechanism, reviewed in references 20 and 21. Although the molecular basis of resistance induction has been well elucidated, the molecular mechanisms for the distinction between resistance inducers and noninducers are still ambiguous, except that some amino acid residues in the leader peptide of ermC are closely related to the resistance inducibility of inducer antibiotics (9, 10). Although 16-membered-ring macrolides are generally considered noninducers, except in the cases of Streptomyces and the selected mutant strain (5, 6, 16), we previously reported an unexpected MLS_B resistance phenotype in which the 16-membered-ring macrolide tylosin induced MLS_B resistance more strongly than the 14-membered-ring macrolide erythromycin in Enterococcus faecalis 373 (13).

To elucidate the molecular basis of MLS_B resistance of E. faecalis 373, we cloned a resistance determinant by PCR and analyzed it by site-directed mutagenesis and reporter gene assay. E. faecalis 373 genomic DNA was digested with BclI and hybridized with four ³⁵S-labeled probes specific for the coding regions of ermC (20), ermA (12), ermAM (4), and ermK (7). Two DNA fragments of the digested genomic DNA showed strong homology with the ermAM probe (data not shown). To clone the resistance determinant, PCR was performed with cycling at 97°C for 30 s, 70°C for 2 min, and 75°C for 2 min for 30 cycles by using 2 U of Vent DNA polymerase (New England Biolabs, Beverly, Mass.), 1.5 g of E. faecalis 373 genomic DNA, and PCR primers TN1 (5′-TTTTTTGGGGTCCCGAGCGCCTACGAGGAA) and TN2 (5′-GGCGCTAGGGACCTCTTTAGCTCCTTGGAAGCT), which were deduced from the ermAM sequence in Tn917 (15). The PCR-amplified 1.5-kb fragment, designated ermAMR, was cloned into the vector pBS42 (2) via PCR cloning vector pKF3 (Takara Shuzo, Otsu, Japan), yielding plasmid pEF42 (Fig. 1). The nucleotide sequence of the 1.5-kb PCR product of E. faecalis 373 was determined by the chain termination method of Sanger et al. (14). The completely sequenced leader region of ermAMR aligned with the same region of ermAM in Tn917 and is presented in Fig. 2.Sequence comparison revealed two mutations in the leader peptide of ermAMR: substitution C441T, which converts arginine into cysteine, and the duplication of TAAA at the T504 site. This second mutation generates a translation stop codon, thereby shortening the length of the putative leader peptide by nine amino acids.

FIG. 1.

Schematic representation of mutant plasmids used in this study. The ermAMR fragment, generated by PCR, was first cloned to the SmaI site of pKF3, and a PCR fragment with sticky ends was isolated by BamHI/BglII digestion of plasmid pKF3-ermAMR (not shown). The isolated DNA fragment was transferred to the BamHI site of pBS42, resulting in pEF42. To construct the ermAMR-lacZ fusion plasmid, the leader region fragment of pEF42 was generated by PCR, treated with EcoRI/BamHI, and fused to EcoRI/BamHI-treated pMM156, resulting in pEZ1. Reporter constructs with a mutated leader region (M1, TAAA deletion; M2, Cys→Arg substitution) were constructed in the same way for pEF42-M1, pEF42-M2, and pEF42-M3.

FIG. 2.

Alignment of ermAMR with ermAM in Tn917. C441T (Arg→Cys) substitution and TAAA duplication were discovered in the leader region.

We reasoned that these mutations are probably responsible for the unexpected resistance phenotype of E. faecalis 373. Thus, to address the effect of each mutation on the resistance phenotypes, the mutations were reversed, one by one and in combination, by site-directed mutagenesis. The site-directed mutagenesis was performed with the Altered Sites II system (Promega, Madison, Wis.), and each mutation was confirmed by sequence analysis. The specificity of induction of MLS_B resistance by various antibiotics was tested in Bacillus subtilis BR151 (trpC2 lys-3 metB10) (23) harboring pEF42, pEF42-M1 (TAAA deletion), pEF42-M2 (T441C), or pEF42-M3 (TAAA deletion and T441C), as described by Weaver and Pattee (data not shown) (19). To quantify the effect of each mutation on methylase expression, we constructed ermAMR-lacZ fusion plasmids in which the ermAMR leader region or the mutant leader regions were translationally fused to Escherichia coli β-galactosidase. Construction of an ermAMR-lacZ fusion plasmid was performed and is presented in Fig. 1. The leader region of ermAMR was PCR amplified with primers LP1 (5′-GCGAAT TCTTTTTTGGGGTCCCGAGCGCCTACGAGGAA) and LP2 (5′-CGTAAACGGGATCCGTTTCTTTTAAATTC). The 730-bp fragment was digested with BamHI and EcoRI and ligated to pMM156 harboring E. coli β-galactosidase (3). B. subtilis BR151 (carrying reporter construct pEZ1, pEZ2, pEZ3, or pEZ4) was tested for induction of β-galactosidase by erythromycin and tylosin. Cultures of B. subtilis BR151 containing the ermAMR-lacZ fusion plasmid and mutated ermAMR-lacZ fusion plasmids were grown separately at 35°C in SPII medium (1) to early log phase. The optimum induction concentration for each antibiotic was determined by measuring β-galactosidase activity at 90 min as a function of inducer concentration (data not shown). Cultures were induced for various times by the addition of erythromycin or tylosin at 0.2 μg/ml, the optimum induction concentration in B. subtilis BR151. β-Galactosidase assays were performed as described previously (11), except that bacteria were lysed by incubation with lysozyme (4 mg/ml for 30 min at 37°C) and the volume of the solutions used was reduced to facilitate the use of microtiter dishes. The results are presented in Fig. 3. In the case of plasmid pEZ4, which has the same leader sequence as ermAM, erythromycin induced an approximately 5.8-fold increase in β-galactosidase activity over a 120-min period, while tylosin failed to induce β-galactosidase expression, consistent with induction experiments reported previously (4). However, induction specificity was dramatically changed in plasmid pEZ2, which has the C441T mutation that changes arginine, the seventh codon of the putative leader peptide, to cysteine. Tylosin induced β-galactosidase activity approximately 5.3-fold, but erythromycin did not induce activity of plasmid pEZ2. The β-galactosidase activity of plasmid pEZ3, which has a TAAA duplication, increased approximately 3.8-fold at basal state and approximately 4.9-fold at induced state compared with those of pEZ4 when induced by erythromycin. Although the level of β-galactosidase expression was dramatically elevated in plasmid pEZ3, the induction specificity pattern of macrolide antibiotics was the same as that for pEZ4. In pEZ3, erythromycin induced gene expression more strongly than tylosin did, that is, a 7.5-fold increase by erythromycin and a 4.9-fold increase by tylosin. We reasoned that the TAAA duplication elevates the level of methylase expression, resulting in apparent constitutive resistance, as in the case of a previous report (22). In addition, in plasmid pEZ1, which has the original control region from E. faecalis 373, tylosin-induced β-galactosidase expression increased approximately 2.9-fold at basal state and approximately 6.3-fold at induced state compared with those of pEZ2. Although the difference between the induction efficiencies of the two antibiotics was not as large as in pEZ1, the induction specificities of the antibiotics were the same in pEZ2. Tylosin induced gene expression 5.2-fold, and erythromycin induced gene expression 3.7-fold. From the above-mentioned results, we conclude that the arginine-to-cysteine change in the seventh codon of the putative leader peptide endowed tylosin with resistance inducibility and that TAAA duplication enabled the control region to express the downstream methylase gene at a drastically increased level.

FIG. 3.

Induction of β-galactosidase activity in the ermAMR leader region (and mutant leader region) reporter constructs. B. subtilis BR151, harboring each construct, was tested for induction by erythromycin (•) and tylosin (○) as a function of induction time.

Sixteen-membered-ring macrolides have been notable for their inability to induce MLS_B resistance determinants in eubacteria. Although induction of ermSV in Streptomyces viridochromogenes by the 16-membered-ring macrolide tylosin was reported by Kamimiya and Weisblum (5) and induction of ermSF by tylosin was also described previously (6), it has been reported that none of the erm genes in eubacteria can be induced by 16-membered-ring macrolides except in the case of an artificially selected mutant strain of S. aureus (16). Our present data show that induction of erm expression by tylosin is not restricted to Streptomyces and can be observed in clinical eubacterial strains. The striking feature of the ermAMR reporter system in this study is that tylosin induces β-galactosidase activity more strongly than erythromycin. The phenomenon is more prominent in the pEZ2 reporter system, in which the TAAA duplication was deleted by site-directed mutagenesis. The only difference between ermAM and the pEZ2 reporter systems is the seventh amino acid of the putative leader peptide, i.e., arginine in ermAM and cysteine in pEZ2. In ermC, the induction specificity of different inducer antibiotics can be altered by mutations in the leader peptide (9, 10), so it is possible that the single-amino-acid modification in the leader peptide of ermAMR is responsible for the change in induction specificity. The change of amino acid in the leader peptide may result in conformational changes of nascent peptides on the translating ribosome (18). These changes give rise to the selective induction of erm expression by tylosin. Duplication of the TAAA sequence in the leader peptide elevated the gene expression dramatically, as shown in pEZ1 and pEZ3 in Fig. 3. TAAA duplication is located within one of the stem regions of the secondary structure of the leader region. However, using computational analysis, we could not observe critical conformational changes of the secondary structure, which would expose the SD2 region, resulting in constitutive expression. It is possible that the sequence duplication affects the level of gene expression by improving the efficiency of translational attenuation, which controls the downstream methylase gene.

Further research will be needed to address the role of the sequence duplication. Our results show that even a single amino acid change in the putative leader peptide sequence can alter the induction specificity of different antibiotics. Further research on the control of ermAMR will help elucidate how the leader region selects inducer antibiotics.

Nucleotide sequence accession number.

The nucleotide sequence of ermAMR has been assigned GenBank accession no. U86375.