mecA-related structure in methicillin-resistant coagulase-negative staphylococci from street food in Taiwan

Antibiotic-resistant patterns, a mecA homologue complex, and staphylococcal cassette chromosome mec (SCCmec) were analysed in samples of ready-to-eat (RTE) street food in Taiwan. RTE food samples (270) were collected in three densely populated Taiwanese cities between June and November 2014. Among 14 strains being identified as methicillin-resistant coagulase-negative staphylococci (MRCoNS), genetic diversities was determined by PFGE analysis. SCCmec types IV, V, VIII and TXG-24 were detected in 9, and mecASs (a mecA homologue) detected in 8. The mecASs gene complex from S. sciuri subsp. sciuri TXG-24 was found to be closely related to those found in both S. sciuri subsp. sciuri (ATCC29062) and S. sciuri subsp. rodentium (ATCC700061). SCCmecTXG24 carries a class A mec complex, a ccrA5B3-like gene complex, a heavy metal gene complex, and an IS1216 mobile element carrying tet(S). Matching identity to ccrA5 was 84.5% for ccrA in S. pseudintermedius KM241. Matching identify to ccrB3 was 92.1% for ccrB in S. pseudintermedius AI16. Similar ccrA and SCCmec boundary sequences suggest that SCCmec is easily transmitted to coagulase-negative staphylococci (CoNS). Based on MRCoNS strains identified in this research, Taiwanese RTE food products likely carry multiple antibiotic resistance genes that can be transmitted to hospitals and other clinical settings.

Scientific RepoRts | 7:42205 | DOI: 10.1038/srep42205 et al. have described S. sciuri as a reservoir for a methicillin-resistant gene 11 , and Ruzauskas et al. have reported the cross-sectional prevalence of methicillin-resistant S. haemolyticus in companion animals 7 .
It is generally accepted that RTE food products serve as reservoirs for antimicrobial-resistant bacteria, but transmission and resistance mechanisms in Taiwan require further investigation. For this project, we looked at proportions of methicillin-resistant coagulase-negative staphylococci (MRCoNS) found in samples of spring rolls, cold noodles, and fruit platters collected from RTE vendors in the densely inhabited cities of Kaohsiung, Taichung and Taipei, and attempted to determine their antibiotic resistance mechanisms.
Genetic analysis of the mecA Ss gene complex and SCCmec TXG24 . Gene analysis results indicate the presence of two mecA homologues (mecA and mecA Ss ) in S. sciuri subsp. sciuri TXG-24. Genomic structure analysis data for the mecA Ss region are shown in Fig. 1 shares a high degree of similarity with TXG-24, except for the upstream ugpQ of two hypothetical protein genes, the ABC transporter gene, and the amino acid/polyamine/organocation (APC) family transporter gene. The SCCmec element of S. sciuri subsp. sciuri TXG-24 has a complex genomic structure that contains a class A mec gene complex (IS431-mecA-mecR1-mecI), an IS1216 mobile element carrying tet(S), partial DNA recombinase with methyltransferase, a heavy metal-resistant gene complex, and a ccr gene complex (Fig. 2). The mec gene complex of SCCmec TXG24 is closely related to S. sciuri subsp. carnaticus GVGS2 (HG515014) and S. pseudintermedius KM241 (AM904731), except for two hypothetical protein genes and a truncated mecR2 gene. The SCCmec TXG24 region containing partial DNA recombinase with methyltransferase is highly similar to the comparative region of Streptococcus suis SC84 (FM252031.1), except for a truncated apt gene. Compared to S. capitis CR01 (KF049201), the heavy metal-resistant gene complex is associated with the absence of two cadmium-resistant genes (cadD and cadB). The proximal left boundary of SCCmec consists of the ccr gene complex, the putative helicase gene, and some hypothetical protein genes that are associated with comparative regions in S. sciuri subsp. carnaticus GVGS2 and S. pseudintermedius KM241.

Analysis of insertion sequence element carrying the tet(S) tetracycline-resistant gene.
The tet(S)-carrying IS1216 mobile element was found at the 3′ end of ΔmecR2 (Fig. 2). According to our sequence analysis, orf25-orf26-orf27-tet(S) had a high degree of similarity with both the Lactococcus lactis subsp. lactis pK214 plasmid (GenBank accession number X92946) and Streptococcus dysgalactiae subsp. equisimilis NTUH_1743 (EF682209) (Fig. 3). Comparisons of IS1216 regions revealed exceptionally high degrees of shared identity (99.4% and 99.6%) with the L. lactis sp. lactis pK214 plasmid, but much lower degrees of shared identity (69.1% and 76.5%) with S. dysgalactiae subsp. equisimilis NTUH_1743 due to a truncated gene. The ΔtnpA gene was only found downstream of orf25 in L. lactis sp. lactis pK214.
ccr gene phylogenetic trees. SCCmec is a genetic element that encodes methicillin resistance and that carries a unique site-specific recombinase (the ccr gene) in charge of SCCmec element integration and excision 6,12 . For the present study, we identified a ccr gene complex in S. sciuri subsp. sciuri TXG-24. Lengths of ccrA and ccrB were 1350 and 1629 bp, respectively. Phylogenetic trees for the ccrA and ccrB sequences (23 each) are shown in Fig. 4a and b. ccrA matching identity was 84.5% to ccrA5 in S. pseudintermedius KM241 (GenBank accession number AM904731). ccrB matching identity was 92.1% to ccrB3 in S. pseudintermedius AI16 (LN864705.1). SCCmec TXG24 boundaries. To investigate SCCmec TXG24 boundaries, we aligned the left and right boundaries of SCCmec types I-VII with the SCCmec element of S. sciuri subsp. carnaticus GVGS2 (Fig. 5). SCCmec TXG24 integration occurred at almost the same nucleotide position at the 3′ end of the orfX gene as the SCCmec complex of S. sciuri subsp. carnaticus GVGS2 and S. pseudintermedius KM241, with both sharing identical direct repeats (DR) at their left and right boundaries. However, nucleotide positions in the other SCCmec types were different from that of SCCmec TXG24 , and the inverted repeats (IR) of each SCCmec type were variant.

Discussion
In their study of five types of RTE food products in Taiwan 15.4% of other meat samples, and 13.6% of all vegetable samples. A separate study conducted in southern Taiwan found a 9.5% incidence of S. aureus contamination in RTE food products purchased from warehouse stores, 12.7% from traditional markets, and 19.0% from supermarkets 3 . The two research teams reported the presence of different pathogens in RTE food, but did not address antimicrobial susceptibility or resistance pattern tendencies. For the present study, we isolated 14 MRCoNS strains that were resistant to at least one antibiotic, and identified the dominant sources as spring rolls filled with salad ingredients and stewed ground pork wrapped in thin pastry dough, both prepared by glove-wearing vendors ( Table 1). The fillings and pastry cracks are likely bacteria reservoirs 13 . The second most common source was cold noodles mixed with some kind of sauce, with bacterial proliferation likely due to the relatively higher pH value of the sauce or improper storage temperature 2 . Bacterial  contamination of fruit platters (the third most common source) was likely due to the improper cleaning of knives. Regardless of actual cause or transmission route, the data indicate that RTE food contamination is a likely avenue for transmitting antibiotic-resistant genes and food-borne diseases [14][15][16] .
Determining genetic relationships in bacterial isolates is an important task for monitoring the spread of bacteria. In one study conducted in Turkey, genetic diversity data for 154 multi-drug-resistant strains of S. aureus found in 1,070 RTE food samples suggested multiple routes for various isolates 17 . In the present study, only two S. saprophyticus isolates (TPE-21 and TPE-32, both from Taipei city) shared the same pulsotype, indicating genetic diversity in our RTE food samples (Supplementary Fig. S1).
Staphylococcal enterotoxin (SE) contaminated food have been reported in foodborne illness 18  These studies and our finding revealed that SEs production of staphylococcal isolates may be associated with food poisoning 19,20 . Our study found that the dominant SE genes were selk (5/14, 35.7%) and seln (5/14, 35.7%), followed by sec (3/14, 21.4%) ( Table 1). Further studies are warranted to determine the importance of SEs-producing CoNS in RTE food.
To date, the mecA gene has been found in multiple homologues, including mecA1 (mecA1, mecA Ss and mecA Sv ) 21,22 , mecA Sf 22 , and mecC (formerly mecA LGA251 ) 23 . Although mecA Ss (from S. sciuri) and mecA Sv (from S. vitulinus) share 80% and 91% identities with mecA, respectively, neither gene is associated with oxacillin resistance 22 . mecA Sf (from S. fleurettii), which belongs to the class A mec complex, shares 99% identity with the mecA gene, suggesting that S. fleurettii may be the ancestor of the SCCmec element in MRSA 22 . We found a  close relationship between the TXG-24 mecA Ss gene complex and a comparative region of S. sciuri subsp. sciuri ATCC29062 (GenBank accession number AB547234.1) that is not associated with oxacillin resistance (Fig. 1).
The CoNS-acquired mecA gene, which has been the focus of multiple studies, is a likely reservoir for transmitting antibiotic-resistant genes [7][8][9][10][11] . Of the 14 MRCoNS strains that we tested, the most prevalent was S. sciuri-a widespread Staphylococcus species among animals and humans. Reported in a wide range of food products, this bacteria has been described as a reservoir for the methicillin-resistant gene 11,24,25 . S. sciuri was first described by Kloos et al. and originally isolated from both human and animal skins 26 . According to one study, ccr genetic diversity in methicillin-susceptible S. sciuri may be useful for capturing the mecA gene and assembling the SCCmec element 27 . Two research teams have shown that S. saprophyticus, S. haemolyticus and S. lentus in RTE foods are likely routes for antibiotic-resistant gene transmission in Poland 28,29 . Specifically, Podkowik et al. reported that 40% (17/42) of the CoNS strains they examined were resistant to 4 or more antibiotics, especially 15 isolates (36%) harbouring the mecA gene 28 . Chajecka-Wierzchowska et al. found that 56.9% (33/58) of the CoNS strains they tested were resistant to at least one antibiotic, with 24 isolates (41.3%) harbouring the mecA gene 29 . In Taiwan, SCCmec types IV and V have been described as prevalent in community-associated MRSA; these same SCCmec types were also found in the MRCoNS strains we analysed for the present study (Table 2) 30,31 . Combined, these data indicate that MRCoNS strains can serve as reservoirs for transmitting the SCCmec element to and from MRSA.
Many ccr gene complexes have been identified in MRSA 23,32,33 . Multiple ccr variants have been found in CoNS strains-for example, ccrA5B3 in S. pseudintermedius KM241 and both ccrA5B13 and ccrA5B9 in S. sciuri 27,34 . Different compositions of the ccr gene complex may be due to dissimilarities in recognised insertion sites 35,36 . We found similar ccrA boundaries sequences in S. sciuri subsp. sciuri TXG-24, S. pseudintermedius KM241, and S. sciuri subsp. carnaticus GVGS2, suggesting that SCCmec is easily transmitted across these and perhaps other Staphylococcus species (Figs 4a and 5).
In summary, we found that CoNS strains in contaminated RTE food samples collected in three Taiwanese cities were resistant to multiple types of antibiotics; it is likely that the associated antibiotic-resistant genes can be easily transmitted to other food products, to the homes of consumers, and to hospitals and other clinics. Since S. sciuri carries diverse ccr genes that are globally distributed, further research is called for to determine or refute its role as a reservoir for antibiotic-resistant gene transmission. For each sample, 10 g were homogenised using a stomacher sample blender, and enriched in brain-heart infusion broth (BD Biosciences) overnight at 37 °C. Single loopfuls of each bacterial suspension were plated on mannitol salt agar. Single colonies were placed on Muller-Hinton agar with 2% NaCl and 4 μ g/ml oxacillin. Bacterial identification was performed by dnaJ gene sequencing as previously described 37 . Antimicrobial susceptibility testing. Antimicrobial susceptibility testing was performed using standard agar dilution methods according to Clinical and Laboratory Standards Institute guidelines 38 . Minimum inhibitory concentration (MIC) was defined as the lowest concentration of antibiotic preventing bacterial growth after 16-20 h of incubation at 37 °C. The following antimicrobial agents were tested: erythromycin, gentamicin, levofloxacin, oxacillin, tetracycline and vancomycin.

Sample collection and microbiological analysis.
Pulsed-field gel electrophoresis (PFGE). PFGE typing of SmaI-digested DNA (New England BioLabs, Ipswich, MA) was performed as previously described 39   Identification of SCCmec TXG24 and the mecA Ss gene complex. Genomic DNA from S. sciuri subsp.
sciuri TXG-24 was extracted manually. Total DNA was subjected to quality control using agarose gel electrophoresis and quantified by Qubit (Invitrogen, Thermo Fisher Scientific, Waltham, MA). The S. sciuri subsp. sciuri TXG-24 genome was sequenced using massively parallel sequencing Illumina (San Diego, CA). Two DNA libraries were constructed: a paired-end library with a 500 bp insert, and a mate-pair library with a 5 kb insert. Both libraries were sequenced with the HiSeq2500 ultra-high-throughput sequencing system (Illumina, San Diego, CA) (PE125 strategy). Library construction and sequencing was performed at Beijing Novogene Bioinformatics Technology Co., Ltd. An in-house quality control program was used for both paired-end and mate-pair reads.
Illumina PCR adapter reads and low quality reads were filtered and assembled with SOAPdenovo 43,44 to generate scaffolds. All reads were used for subsequent gap closures. SCCmec TXG24 and mecA Ss gene complex nucleotide sequences from S. sciuri subsp. sciuri TXG-24 were added to GenBank (accession numbers KX774481 and KX774480, respectively).
Phylogenetic tree analysis. . Phylogenetic trees were analysed by MEGA7 using the neighbour-joining method; tree topologies were estimated using bootstrap analyses with 2,000 replicates to achieve confidence intervals as indicated on each tree node 45 . Identities shown after each ccr gene were aligned and calculated using DNAman (Lynnon Biosoft, Quebec).
SCCmec type determination and mecA Ss gene detection. SCCmec types were determined by mec and ccr gene complexes as described in our previous study 39 . SCCmec TXG24 was determined by the class A mec complex and ccr gene (ccrA5B3) (Supplementary Table S1), and mecAs was determined by mecAs-F and mecAs-R (Supplementary Table S1).