High prevalence of Cfr-producing Staphylococcus species in retail meat in Guangzhou, China

Background The emergence and wide distribution of the transferable gene for linezolid resistance, cfr, in staphylococci of human and animal origins is of great concern as it poses a serious threat to the public health. In the present study, we investigated the emergence and presence of the multiresistance gene, cfr, in retail meat sourced from supermarkets and free markets of Guangzhou, China. Results A total of 118 pork and chicken samples, collected from Guangzhou markets, were screened by PCR for cfr. Twenty-two Staphylococcus isolates obtained from 12 pork and 10 chicken samples harbored cfr. The 22 cfr-positive staphylococci isolates, including Staphylococcus equorum (n = 8), Staphylococcus simulans (n = 7), Staphylococcus cohnii (n = 4), and Staphylococcus sciuri (n = 3), exhibited 17 major SmaI pulsed-field gel electrophoresis (PFGE) patterns. In 14 isolates, cfr was located on the plasmids. Sequence analysis revealed that the genetic structures (including ΔtnpA of Tn558, IS21-558, ΔtnpB, and tnpC of Tn558, orf138, fexA) of cfr in plasmid pHNTLD18 of a S. sciuri strain and in the plasmid pHNLKJC2 (including rep, Δpre/mob, cfr, pre/mob and partial ermC) of a S. equorum strain were identical or similar to the corresponding regions of some plasmids in staphylococcal species of animal and human origins. Conclusions To the best of our knowledge, this is the first study to report the presence of the multiresistance gene, cfr, in animal meat. A high occurrence of cfr was observed in the tested retail meat samples. Thus, it is important to monitor the presence of cfr in animal foods in China.


Background
Linezolid is considered to as the last treatment option for infections caused by methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant Enterococci and penicillin-resistant Streptococcus [1]. Mutations in the drug target site (23S rRNA or ribosomal proteins L3 and L4) are the most common mechanisms of linezolid resistance. Due to the low frequency of target mutation, the frequency of linezolid resistance is also relatively low [2]. However, emergence of the transferable linezolid resistance gene, cfr, in clinical isolates poses a challenge in linezolid treatment. cfr gene encodes an RNA methyltransferase, which modifies the adenine residue at position 2503 of the 23S rRNA gene and thereby confers resistance to phenicols, lincosamides, oxazolidinones, pleuromutilins, and streptogramin A antibiotics (the PhLOPS A phenotype) as well as decreases susceptibility to the 16-membered macrolides spiramycin and josamysin [3][4][5].
Since its first detection from a bovine Staphylococcus sciuri isolate in 1997 [6], cfr has been globally transmitted among different bacteria, such as Staphylococcus spp., Enterococcus spp., Macrococcus spp., Jeotgalicoccus spp., Streptococcus suis, Escherichia coli, Bacillus spp., Proteus vulgaris [7,8]. This gene is widely distributed in the isolates of both human and animal origin, especially in China [8]. A recent study has described this gene in farm environments [9]. However, there has been no study on the distribution of cfr in retail meat. In the present study, we investigated the presence and the genetic background of this multiresistance gene in retail meat samples sourced from supermarkets and free markets of Guangzhou, China.

Identification of cfr-positive Staphylococcus isolates
Of the 118 retail meat samples tested, a total of 22 cfr-positive Staphylococcus isolates were detected in 12 pork samples and 10 chicken samples. The 22 cfr-positive staphylococcal isolates included Staphylococcus equorum (n = 8), Staphylococcus simulans (n = 7), Staphylococcus cohnii (n = 4), and Staphylococcus sciuri (n = 3). In addition, one cfr-positive Macrococcus caseolyticus isolate was obtained from a chicken sample. In total, 15.8% and 26.2% pork and chicken samples carried cfr-positive isolates, respectively.
Clonal analysis of cfr-positive staphylococci and location of cfr Pulsed-field gel electrophoresis (PFGE) of 22 cfr-positive staphylococci revealed 17 major SmaI PFGE patterns (Table 1). Eight S. equorum isolates showed five different PFGE patterns, with two chicken strains from the same market presenting indistinguishable patterns. Six distinct PFGE patterns were identified for the seven S. simulans isolates, with only two pork isolates from different markets presenting similar PFGE patterns. For the four S. cohnii isolates, three PFGE patterns were identified, with two pork isolates from the same market presenting identical patterns. Each of the three S. sciuri isolates exhibited distinct PFGE patterns. In summary, most of the cfr-positive staphylococcal isolates were genetically distinct, but a clonal transfer of cfr-positive staphylococcal isolates had occurred either in the same or among different markets.
Results of Southern blotting indicated that 14 isolates harbored cfr in their plasmid DNA ( Table 1). The remaining eight isolates appeared to carry cfr in their genomic DNA; however, this assumption needs to be further confirmed by S1-PFGE. Only one cfr-carrying plasmid (designated as pHNLKJC2) that originated from a chicken isolate, TLKJC2, was transformed into Staphylococcus aureus RN4220. The transformant was confirmed by polymerase chain reaction (PCR) for cfr; it showed the same PFGE pattern as that of Staphylococcus aureus RN4220.
Analysis of the genetic environment of cfr in the plasmid pHNTLD18 and pHNLKJC2 Southern blotting confirmed that, in Staphylococcus equorum TLD18, cfr was located on a plasmid designed as pHNTLD18. An approximately 5.7-kb EcoRI fragment containing cfr was cloned and sequenced. A Tn558 variant was identified on the plasmid pHNTLD18, in which parts of the Tn558-associated transposase genes tnpA and tnpB were replaced by a cfr-carrying segment and the insertion sequence IS21-558 ( Figure 1A). Another resistance gene, fexA, encoding an exporter that mediates the active efflux of phenicols, was found to be located downstream of Tn558.
The sequences 1,926-bp upstream and 2,659-bp downstream of cfr on the plasmid pHNLKJC2 were obtained by primer walking. Basic local alignment search tool (BLAST) analysis of these sequences revealed a 3′truncated segment of the gene pre/mob upstream of cfr. Further upstream, an incomplete rep gene was detected. Analysis of the region downstream of cfr revealed the presence of a complete pre/mob gene. Immediately downstream of the pre/mob gene, an incomplete macrolide-lincosamide-streptogramin B (MLS B ) resistance gene ermC was detected ( Figure 1B).

Discussion
Lack of previous studies on the distribution of the multiresistance gene cfr among staphylococci in retail meat led us to screen 118 meat samples for the same. In our analysis, cfr was detected in 22 samples. The detection rate was 18.6%, which is higher than the detection rates of food animal samples in China [10,11]. The low fitness cost of cfr acquisition observed in staphylococcal isolates may account for the persistence of this multiresistance gene in retail meat even in the absence of an antimicrobial selection pressure [12]. The high detection rate found in this study suggested that cfr may be widely disseminated among staphylococci in the meats sold in China, increasing the possibility of this gene entering the food chain.
In this study, S. equorum (n = 8) was the predominant species among the 22 cfr-carrying isolates obtained from animal food sources. To the best of our knowledge, this is the first report of cfr in S. equorum. S. equorum was originally isolated from the skin of horses and was later found to be the predominant species of staphylococci in sausages and cheese samples [13][14][15][16][17]. S. equorum is used as one of the starter cultures in the preparation of smear-ripened cheese and cured meats such as sausages [15,16]. Since S. equorum present in retail meats has rare chances of coming in contact with antimicrobial agents, the origin and high prevalence of cfr in Staphylococcus equorum is intriguing.
The cfr-carrying segment (including rep, Δpre/mob, cfr, pre/mob and partial ermC) on the plasmid pHNLKJC2 from the chicken meat strain S. sciuri TLKJC2, was found to be similar to the corresponding plasmid regions from different staphylococcal species such as the plasmid pSS-03 (accession number JQ219851) from a bovine S. cohnii strain and the plasmid pMSA16 (accession number JQ246438) from a bovine MRSA ST9 strain in China ( Figure 1B) [10,18]. In addition, this cfr-carrying segment also showed high nucleotide sequence identity (98%) to the corresponding region of plasmid pSCFS1 (accession number AJ579365) from a bovine S. sciuri in Germany [19]. The cfr-carrying segment (including ΔtnpA of Tn558, IS21-558; ΔtnpB; and tnpC of Tn558, orf138, fexA) on the plasmid pHNTLD18 from the pork strain S. equorum TLD18 was identical to the corresponding segment of the  [21], and the plasmid pSS-02 (accession number JF834910) obtained from a porcine S. saprophyticus strain in China( Figure 1A) [10]. These results indicated that the horizontal transfer mediated by mobile genetic elements such as plasmids and insertion sequences may contribute to the spread of cfr and suggested that it is possible to transfer cfr via mobile genetic elements from staphylococcal isolates of animal origin to the bacterial strains in the human body through meat consumption, posing a serious threat to the public health. The MICs of the cfr-positive staphylococci indicated multiresistance phenotype in these strains other than the PhLOPS A phenotype, suggesting limited therapeutic options to control these cfr-carrying staphylococci. Most of the cfr-positive staphylococcal isolates showed low-level linezolid resistance with MIC values ranging from 4 to 16 mg/L; this result is in agreement with previously reported linezolid MICs among cfr-carrying staphylococci from farm animals and humans [10,11,22]. In addition, five of the cfr-positive isolates had linezolid MIC values of 2 mg/L, which is the same as the typical linezolid MIC 90 value and not consistent with MIC value shifts observed for isogenic cfr-negative/positive staphylococcal strain pairs [23]. This finding indicated that cfr, although present, is not functional or is partially silenced in some way [24]. Although Staphylococcus strains isolated from meat samples showed low-level of linezolid resistance in the present study, emergence of the multiresistance gene cfr in meat poses a potentially significant threat to the public health, considering that the cfr-mediated linezolid resistance can rapidly and widely spread among different bacterial species.

Conclusions
To the best of our knowledge, this is the first study to report a surprisingly high occurrence of cfr in retail meat samples in Chinese markets. Animal meat harboring bacteria containing the transmissible cfr would be a serious threat to the public health as these bacteria may act as reservoirs for spreading cfr to bacteria that infect humans, particularly in environments with a large microbial community. Recently, cfr was detected in human isolates in China [20,25]. Thus, more attention needs to be paid to the possibility that cfr can find its way through the food chain to commensal or pathogenic bacteria of humans. Considering that a limited number of meat samples were used and that to from only one city in China, the results of the present study regarding dissemination of cfr among staphylococcal species from animal food sources in China is not conclusive. Thus, continuing the surveillance of cfr gene in meat distributed in China is critical to limit its dissemination, which could potentially threaten the human health.

Methods
Sample collection, identification of species, and cfr detection In February 2012, 72 pork samples and 46 chicken samples were collected from five free markets and one supermarket in Guangzhou. The meat samples were incubated in Luria-Bertani (LB) broth for enrichment. Then, the cultured broth was streaked onto selective media plates of Baird-Parker agar supplemented with 10 mg/L florfenicol. One isolate per sample was selected for further analysis. Whole-cell DNA was prepared according to a previously described protocol [26]. The presence of cfr was screened by PCR with previously described primers [5]. Species identification of the cfr-carrying strains was performed by the API-Staph System (bioMérieux, France) and further confirmed by 16S rRNA sequencing [27].

Molecular typing and transformation
PFGE of all cfr-positive Staphylococcus isolates was performed by using the CHEF Mapper System (Bio-Rad Laboratories, Hercules, CA), according to the previously described protocol [10]. All the plugs of genomic DNA were digested with SmaI (TaKaRa Biotechnology, Dalian, China). The PFGE patterns were interpreted according to the criteria described by Tenover et al. [28]. The location of cfr was determined by Southern blotting.
Cfr-carrying plasmids of the isolates were extracted by using the QIAGEN Plasmid DNA Midi Kit (Qiagen, Hilden, Germany) and then transferred into S. aureus RN4220 by electrotransformation, as described previously [29]. The transformants were selected on media supplemented with 10 mg/L florfenicol, screened for cfr by PCR amplification, and confirmed by PFGE.

Antimicrobial susceptibility testing
The MIC values of all cfr-positive original Staphylococcus isolates and transformants were determined by the broth microdilution method, according to the recommendations specified in CLSI documents M100-S22 [30]. The results were interpreted according to Eucast breakpoints (http://www.eucast.org/clinical_breakpoints/). Isolates with an MIC of ≥16 mg/L were tentatively considered to be florfenicol-resistant [26]. The reference strain S. aureus ATCC 29213 was used for quality control.

Cloning and sequencing of the regions flanking cfr
The regions flanking cfr in the transformant obtained from the isolate TLKJC2 were determined by PCR mapping. The plasmid DNA of the isolate TLD18 was extracted and digested with EcoRI. The digested fragments were cloned into the pUC18 vector, and the recombinant plasmid (designated as pUC18-cfr) was introduced into Escherichia coli DH5α with subsequent selection for the transformant (designated as E. coli DH5α-pUC18-cfr) on media supplemented with 10 mg/L florfenicol. The approximately 5.7-kb segment in pUC18-cfr, including cfr and its flanking regions, was sequenced by primer walking. The DNA sequences were compared to those deposited in GenBank using the BLAST program (http://www.ncbi. nlm.nih.gov/BLAST).