Macrolide-lincosamide-streptogramin resistance phenotypes and genotypes of coagulase-positive Staphylococcus aureus and coagulase-negative staphylococcal isolates from bovine mastitis

Background There are limited data available on macrolide-lincosamide-streptogramin (MLS) resistance of Staphylococcus aureus (S. aureus) and coagulase-negative staphylococci (CoNS) from bovine milk in China. To address this knowledge gap, MLS resistance was determined in 121 S. aureus and 97 CoNS isolates. Minimum inhibitory concentrations (MICs) of MLS antibiotics were determined by an agar dilution method, while differentiation of MLS phenotypes was performed by a double-disc diffusion test. MLS resistance genotypes were determined by PCR for corresponding resistance genes. Results Forty (33.1 %) S. aureus and 65 (67.0 %) CoNS were resistant to erythromycin, whereas all 218 isolates were susceptible to quinupristin/dalfopristin. Among 40 erythromycin-resistant (ER-R) S. aureus and 65 ER-R CoNS isolates, 38 S. aureus and 40 CoNS isolates exhibited the inducible MLS (iMLS) resistance phenotype and 2 S. aureus and 20 CoNS isolates expressed the constitutive MLS resistance (cMLS) phenotype. At the same time, 5 CoNS isolates exhibited resistance to erythromycin but susceptibility to clindamycin (the MS phenotype). An inactivating enzyme gene lnu(A), methylase genes erm(C) and erm(B), efflux genes msr(A)/msr(B), a phosphotransferase gene mph(C), an esterase gene ere(A) and the streptogramin resistance determinant vga(A) were detected individually or in combinations. Among them, genes lnu(A), erm(C) and mph(C) predominated. The ereA gene was detected for the first time in staphylococci of bovine milk origin. Resistance genes also existed in erythromycin-susceptible isolates. Conclusions Our study demonstrated a high level of resistance to MLS antibiotics in staphylococci from bovine mastitic milk, especially with a high rate of the iMLS phenotype in S. aureus isolates. These data suggest that MLS antibiotics should be used judiciously to treat or prevent bovine mastitis caused by staphylococci.


Background
Bovine mastitis is the most costly disease for the dairy industry worldwide. Although a wide variety of pathogens have been isolated as causative agents of this disease, Staphylococcus aureus (S. aureus) is considered as one of the most important pathogens due to its resistance to certain antibiotics and its propensity to recur chronically. Recently, coagulase-negative staphylococci (CoNS) have been considered as opportunistic pathogens that cause bovine mastitis in many countries and could be therefore described as emerging mastitis pathogens [1,2]. Increasing attention has been paid to CoNS in both subclinical and clinical mastitis cases throughout the world [3,4]. Macrolidelincosamide-streptogramin (MLS) antibiotics, including erythromycin, clindamycin and spiramycin, are frequently used for treatment of bovine mastitis [5,6]. Thus, results from an in vitro susceptibility testing are an important tool to guide a veterinarian in selecting the most efficacious antimicrobial agent(s) for therapeutic and prophylactic intervention.
Three mechanisms are mainly responsible for acquiring resistance to MLS antibiotics in staphylococci: (1) target site modifications by methylation or mutation; (2) active efflux of antibiotics; or (3) inactivation of antibiotics. The first mechanism includes target site modifications by a methylase encoded by one or more of the erm genes, methylating 23S rRNA and thereby altering binding sites for MLS antibiotics [7]. Phenotypically, this resistance appears either inducible (resistant to 14-and 15-membered macrolides and susceptible to 16-membered macrolides, lincosamides and streptogramin B) or constitutive (resistant to all forms of these antibiotics) [8]. The second mechanism involves a macrolide efflux pump encoded by msr(A) and/or msr(B) genes. This pump protein belongs to the ABC transporter family and exports 14-membered macrolides and streptogramin B antibiotics from bacterial cells, while lincosamide and streptogramin A antibiotics remain unaffected (the MS phenotype) [9]. The third mechanism encompasses several enzymes. A lincosamide nucleotidyltransferase encoded by the lnu(A) gene confers resistance only to lincosamides and has been detected in CoNS isolates from bovine mastitis [10]. Esterases encoded by ere(A)/(B) genes hydrolyze the lactone ring of the macrocyclic nucleus [11]. Furthermore, vga(A)/(B) genes have been characterized as a determinant of streptogramin A resistance [11]. Finally, the macrolide phosphotransferase C encoded by the mph(C) gene inactivates some macrolide antibiotics and has been detected in CoNS isolated from bovine subclinical mastitis [12].
The reported resistance of S. aureus and CoNS isolated from bovine mastitis to MLS antibiotics in different countries was generally low [12,13]. Meanwhile, there was a paucity of data regarding MLS-resistance phenotypes and genotypes of S. aureus and CoNS isolated from bovine mastitis in China, except one study [5]. The objective of this study was to determine the MLS resistance phenotypes and genotypes of 121 S. aureus and 97 CoNS isolates from mastitic milk from dairy farms of the Shaanxi province in Northwestern China.

Detection of MLS resistance phenotypes
In order to differentiate different types of resistance phenotype for erythromycin-resistant (ER-R) isolates, a doubledisk diffusion test (D test) was performed with erythromycin (15 μg/disc) and clindamycin (2 μg/disc), following the procedure recommended by CLSI [17]. Staphylococcal isolates showing resistance to erythromycin (zone size ≤13 mm) but being sensitive to clindamycin (zone size ≥21 mm) and producing a D-shaped zone of inhibition around clindamycin with flattening towards erythromycin disc was defined as having an inducible type of MLS resistance (D + , iMLS). In addition, resistance to erythromycin (zone size ≤13 mm) as well as to clindamycin (zone size ≤14 mm) indicated a constitutive type of MLS resistance (cMLS). Staphylococcal isolates showing resistance to erythromycin (zone size ≤13 mm) while being sensitive to clindamycin (zone size ≥21 mm) with no blunting zone were classified as the MS phenotype.

Detection of MLS resistance genotypes
Staphylococcal isolates were incubated in the Brain Heart Infusion broth (Oxoid) at 37°C for 16-18 h. Then, bacteria were harvested by centrifugation. Plasmid and chromosome DNA of bacterial isolates were extracted using a commercial DNAout kit (Tiandz Inc., Beijing, China) as described previously [14]. The screening of MLS resistance determinants including methylase genes erm(A), erm(B) and erm(C); phosphotransferase genes mph(A) and mph(C); lincosamide nucleotidyltransferase genes lnu(A) and lnu(B); erythromycin esterase genes ere(A) and ere(B); streptogramin resistance genes vga(A), vga(B), vgb(A) and vgb(B), and the macrolide efflux determinants msr(A)/msr(B) was performed by PCR using the specific primers as described in previous studies [11,12,[18][19][20]. PCR products were randomly selected and sequenced to ensure specificity and accuracy. Sequence comparisons were performed using the Basic Local Alignment Search Tool (BLAST) program (http://www.ncbi.nlm.nih.gov/BLAST/).
Among 97 CoNS isolates, 65 isolates exhibited MLS resistance phenotypes. Among them, 40 showed the iMLS phenotype and 20 expressed the cMLS phenotype, while 5 exhibited the MS phenotype. The MICs of the antimicrobial agents tested are summarized in Table 2. Eighteen isolates with cMLS phenotypes exhibited a high-level of resistance to erythromycin, clindamycin, azithromycin, spiramycin and tylosin with MIC values of ≥256 μg/mL. Furthermore, 1 S. haemolyticus with the iMLS phenotype exhibited MICs ≥256 μg/mL for erythromycin and azithromycin while MICs for spiramycin and tylosin were 64 μg/mL and 128 μg/mL, respectively. In addition, 39 CoNS isolates with the iMLS phenotype showed a complete cross-resistance to erythromycin and azithromycin with MICs of ≥256 μg/mL. However, MIC values of 16-membered macrolides tylosin (2-8 μg/mL) and spiramycin (2-16 μg/mL) were in the susceptible ranges.
The iMLS phenotype rate of ER-R S. aureus (38/40) and ER-R CoNS (40/65) isolates was much higher in this study than previous studies, underlining the importance of routine screening of bovine S. aureus and CoNS isolates for inducible resistance phenotypes. Wang et al. [5] reported that the inducible MLS resistance phenotype was detected in 38 out of 72 S. aureus isolates from cows with clinical mastitis in Inner Mongolia of China. In another study, only 3 isolates with the iMLS phenotype were found out of 22 ER-R CoNS in Germany [12]. The reason for the higher rate of the iMLS phenotype in our study is not clear.     Table 3). The msr(A)/msr(B) genes were found in 6 isolates, which were all positive for erm(C) or erm(B) genes and displayed iMLS phenotypes (Table 4) The simultaneous presence of two or more MLS antibiotic resistance genes was also detected ( Table 4). The simultaneous presence of two or more macrolide resistance genes in the same S. aureus or CoNS isolate is well-known and has been reported previously for S. aureus or CoNS isolates from bovine mastitis [5,12,21].

Correlation between the MIC values of MLS resistance phenotypes and phenotypes
The possible relationship between MLS resistance phenotypes and genotypes was also explored. Among the 40 ER-R CoPSA isolates, 8 isolates with iMLS phenotypes were sensitive to 16-membered macrolides spiramycin and 32 isolates with iMLS phenotypes were sensitive to tylosin. Those isolates were all erm(B) and/or erm(C) positive. As for the 65 ER-R CoNS isolates, 4 erm(B) and/or erm(C) positive isolates with the MS phenotype and 28 erm(B) and/or erm(C) positive isolates with the iMLS phenotype were sensitive to 16-membered macrolides spiramycin and tylosin, respectively. Furthermore, 1 S. warneri with the MS phenotype and 9 CoNS isolates with the iMLS phenotype were also sensitive to 16-membered macrolides spiramycin and tylosin. Those 10 isolates were negative for erm genes but positive for other MLS resistance genes, such as msr(A)/(B), mph(C), ere(A), lnu(A) or vga(A). In general, erm-carrying ER-R S. aureus and CoNS isolates with iMLS or MS phenotypes possessed a high degree of resistance to erythromycin, azithromycin and clindamycin (inducible), while having a low rate of resistance to 16-membered macrolides tylosin and/or spiramycin. It has been reported that the lactone rings of 16-and 14-membered macrolides adopt distinctly diverse conformations, thereby enabling the former compounds to avoid steric hindrance with the nucleotide A2058 mutation in   [5] e MIC breakpoints of this antibiotic were not available E. coli [22]. Such a mechanism may be also responsible for differential sensitivity to 16-and 14-membered macrolides in staphylococci. In addition, differential effects of 14 and 15-membered macrolides versus 16-membered macrolides on expression of erm genes could contribute to our results. Expression of erm genes can be either inducible or constitutive. Inducible erm genes expression is controlled at a post-transcriptional level, which involves a structure upstream from the erm gene composing of a leader peptide and a series of inverted repeats. Formation of different mRNA secondary structures in this regulatory region in the presence or absence of an inducer allows or prevents the translation of the erm gene transcripts [8]. Only 14and 15-membered macrolide can induce erm expression, while 16-membered macrolides, lincosamides, or streptogramins are not able to induce erm gene expression [8]. However, why 10 CoNS without the erm gene were also sensitive to 16-membered macrolides spiramycin and tylosin in this study is unclear and will be further studied. In addition, previous studies have shown that erm gene expression can quickly and irreversibly switch from inducible expression to constitutive expression under selective pressure due to the structural alterations (sequence deletions of varying length, duplications and mutations),   [23][24][25]. Therefore, different conformational rearrangements in the mRNA structure or structural alterations (deletions, duplications or mutations) in the upstream regulatory region of erm genes could be one of the plausible reasons of our isolates and such resistance mechanism will be further studied.

Characterization of MLS resistance genotypes in erythromycin-susceptible isolates
Among 81 erythromycin-susceptible (ER-S) CoPSA isolates, 79 isolates were positive for lnu(A), 69 for erm(C), 47 for mph(C), 66 for erm(B), 64 for msr(A)/(B), 20 for ere(A) and 10 for vga(A) genes (Table 5), but all were susceptible to the corresponding antibiotics (erythromycin, azithromycin, spiramycin, tylosin or clindamycin) in the antibiotic susceptibility testing, due to unknown reasons. As for 32 ER-S CoNS isolates, 7, 9, 20, 11, 7 and 9 CoNS isolates harbored erm(B), erm(C), msr(A)/(B), mph(C), ere(A), vga(A) genes, respectively (Table 5). Furthermore, the lincosamide nucleotidyltransferase gene, lnu(A), was detected in all ER-R S. aureus, ER-R CoNS isolates, ER-S CoNS and 79 ER-S CoPSA isolates (Table 3; Table 5). The presence of lnu(A) among staphylococcal isolates from bovine mastitis has been reported [5,10,12]. The lnu(A) gene is mainly carried by small rolling-circle plasmids and it mediates only a low-level of resistance to the lincosamide pirlimycin [10]. The ere(A) gene was detected for the first time in staphylococci of bovine milk origin. Our results are in agreement with previous studies which detected erm(C), lnu(A), mph(C) or erm(A) genes in susceptible S. aureus or CoNS isolates [12,26,27]. When Martineau et al. [26] subcultured 4 erythromycin susceptible strains harboring the erm(C) gene with increasing concentration of the antibiotic, they found that those susceptible strains all become resistant. Thus, we need to be vigilant when we use MLS antibiotics on dairy farms.

Conclusions
In summary, a very high rate of iMLS (95 %, 38/40) phenotype of ER-R S. aureus and MLS resistance phenotype (67 %, 65/97) of CoNS isolates from milk of mastitic cows was found in this study in comparison with previous studies, presumably due to extensive use of MLS antibiotics in dairy cows in our region. Our results suggest that MLS antibiotics should be used judiciously for therapeutic and prophylactic intervention of staphylococci infection.