Staphylococcus aureus Isolated From Retail Meat and Meat Products in China: Incidence, Antibiotic Resistance and Genetic Diversity

This study was to estimate the prevalence and characteristics of Staphylococcus aureus from 1,850 retail meat and meat products in China during July 2011 to June 2016. The samples were collected covering most provincial capitals in China, including 604 raw meat, 601 quick-frozen meat, and 645 ready-to-eat meat. Using the qualitative and quantitative methods, all 39 cities had S. aureus-positive samples, and S. aureus was detected in 35.0% (647/1,850) of the samples. The levels of S. aureus in retail meat showed that the MPN value of the majority of the positive samples ranged from 0.3 to 100 MPN/g. Twenty-four antibiotics were used to test all 868 S. aureus isolates for antibiotic susceptibility. Only 11 isolates (1.26%) were susceptible to all antibiotics, whereas most isolates (821/868, 94.6%) showed resistance or intermediary resistance to more than three or more antibiotics. Of these strains, 104 (12.0%) were resistant to more than 10 antibiotics. However, the most frequent resistance was observed to ampicillin (85.4%), followed by penicillin (84.6%), erythromycin (52.7%), tetracycline (49.3%), kanamycin (45.3%), telithromycin (30.1%), clindamycin (29.6%), streptomycin (21.1%), norfloxacin (20.4%), gentamicin (19.4%), fusidic acid (18.4%), ciprofloxacin (16.9%), chloramphenicol (13.1%), amoxycillin/clavulanic acid (11.0%), and others (<10%). 7.4% of isolates (62/868) were confirmed as methicillin-resistance S. aureus (MRSA). By molecular typing analysis, there were 164 spa types and 111 STs were identified, including 15 novel spa types and 65 newly STs by multilocus sequence typing (MLST) and spa typing. Despite the wide genetic diversity observed among the 868 isolates, a great proportion of the population belonged to finite number of major clones: ST1-t127 (93/868, 10.7%) and ST7-t091 (92/868, 10.6%), ST5-t002 (42/868, 4.8%), ST398-t034 (40/868, 4.6%), ST188-t034 (38/868, 4.4%), ST59-t437 (30/868, 3.5%), ST6-t701 (29/868, 3.3%), and ST9-t899 (27/868, 3.1%) in China. This study reflects S. aureus was readily detected in Chinese retail meat and meat products but the level were not very excessive. In this study, the high antibiotic resistance is alarming and raising public health concern. In additions, most of molecular types of isolates have been linked to human infections around the world, indicating that these types of S. aureus in China have a theoretical pathogenic potential.


INTRODUCTION
Recent years, many pathogens were responsible for food safety. Staphylococcus aureus is recognized as one of the major foodborne pathogens in fresh and ready-to-eat products and responsible for various infections around world (Diep et al., 2006). It could grows at temperature between 15 • C and 45 • C and at NaCl concentrations as high as 15% (Behling et al., 2010). This bacterium multiplies quickly at room temperature to produce toxins that cause illness. Naturally, the distribution of S. aureus was ubiquity in the world, but the most important infection source of S. aureus was food. Every year, S. aureus involved in about 241,000 illnesses of foodborne disease in USA (Scallan et al., 2011;Wu et al., 2018). In 2013, there were 12.5% of foodborne bacterial outbreaks are caused by S. aureus in China, which showed third most frequently pathogen after Vibrio parahaemolyticus (27.8%) and Salmonella (23.1%) (Wei-Wei et al., 2018).
In the last decades, the spread use of antibiotics in bacteria is increasing the emergence of multidrug resistant strains (MDR), which showed great challenges to public health. As a formidable adaptive capacity strain, S. aureus can adapt to varying environmental conditions and rapidly become resistant to virtually all antibiotics (Mccallum et al., 2010). Nowadays, more and more MDR S. aureus were reported in food poisoning outbreaks and isolated from food product in previous researched (Sauer et al., 2008;Huang et al., 2009;Gharsa et al., 2012;Papadopoulos et al., 2018). Moreover, in recent years, methicillinresistance S. aureus (MRSA) is attracting extensive attention. It usually showed multiple antimicrobial resistance and listed as one of 12 families of bacteria that pose the greatest threat to human health by WHO in 2017 (Govindaraj and Vanitha, 2018).
Nowadays, various of molecular subtyping approaches [e.g., pulsed-field gel electrophoresis (PFGE), multilocus sequence typing (MLST), and staphylococcal protein A (spa) typing] have been developed for the characterization of S. aureus. For epidemiology and evolutionary studies, as well as discriminatory power, it is important to produce unambiguous results that are readily comparable among different laboratories, and it is necessary to have a system for classifying the relationships among closely related strains to monitor changes and patterns in clonal lineages over time or space (O'Hara et al., 2016). Of these methods, MLST and spa typing showed highly clonal population structure identified for S. aureus in many previous study (Kanika et al., 2011;Fetsch et al., 2014;Basanisi et al., 2017). As a widely accepted method, MLST, which is based on DNA sequence and relied on analysis of relatively conserved genes that encode essential proteins, has proven very useful for epidemiology and evolutionary studies (Saunders and Holmes, 2007). spa typing is another efficient typing method for S. aureus, it based on sequencing of the polymorphic X region of the protein A gene (spa) (Hallin et al., 2009). It showed highly concordant with MLST. Some studies suggest that it is suitable for epidemiology and evolutionary investigations based on studies of European and international isolates (Koreen et al., 2004;Hallin et al., 2007;Strommenger et al., 2008). However, other studies question the use of a single locus method such as spa typing for epidemiologic investigations since recombination events might distort the underlying clonal relationships (O'Hara et al., 2016). Both two methods can assigned to MLST STs and spa types, which easily to compare with other laboratories in the world.
Of various food products surveyed, meat and meat products are widely known to be an important reservoir for S. aureus and involved in several outbreaks (Aydin et al., 2011;Hennekinne et al., 2012;Wang et al., 2013;Sallam et al., 2015). The investigate of S. aureus from meat can implement a system monitoring. In China, the qualitative and quantitative data of this bacterium in retail meat from different areas were limit. Therefore, the prevalence and levels of S. aureus from South China to North China was to investigate in retail meat and meat products in this study. After isolation and identification, antibiotic susceptibility test, as well as spa typing and MLST were used to determine the genetic background among the S. aureus isolates.

Sample Collection
From July 2011 to June 2016, a total of 1,850 retail meat samples were collected from supermarkets, fairs, and farmers' markets. The samples, including 604 raw meat (bacon/sausage, poultry, pork, mutton, and beef), 601 quick-frozen meat (frozen dumpling/steamed stuffed bun, frozen poultry, frozen pork, frozen mutton, and frozen beef), and 645 ready-to-eat (RTE) meat (roast chicken/duck, salt-baked chicken, stewed meat sausage, and ham), were obtained from 39 cities, which covered most of provincial capitals of China (Figure 1). All collected samples were tightly sealed with sterile plastic wrap and placed in a cold box at a temperature lower than 4 • C, then transported to an accredited laboratory and subjected to microbiological analysis within 24 h.

Isolation and Identification of S. aureus
The examination of S. aureus proceeded according to GB 4789.10-2010 of food microbiological (National Food Safety Standards of China) with slight modification. Briefly, 25 g of food sample was homogenized in 225 mL of tryptic soy broth with 10% sodium chloride (Huankai, Guangzhou, China). Subsequently, 1, 0.1, and 0.01-mL aliquots were transferred to tubes containing 9, 10, and 10 mL in triplicate with trypticase soy broth (Huankai) supplemented with 10% NaCl. Each tubes were incubated at 37 • C for 48 h, respectively. A loopful of enrichment broth culture (10 µL) was streaked onto chromogenic S. aureus agar plates (Huankai) and incubated at 37 • C for 24 h. One to four colonies with pink color were purified on nutrient agar medium. The purified colonies were analyzed via coagulase activity test by freeze-dried Rabbit Plasma (Huankai), and the API STAPH test strips (bio Merieux, Marcy-1'Etoile, France) was used. The MPN value was determined on the basis of the number of positive tube(s) in each of the three sets using the MPN table.

Antibiotic Susceptibility Testing
All S. aureus isolates were tests for antibiotic susceptibility by the disk diffusion method using Mueller-Hinton agar (Bauer et al., 1966) and commercially available discs (Oxoid, UK). The antimicrobial agents used were amoxycillin/clavulanic acid, ampicillin, cefepime, cefoxitin, penicillin G, ceftazidime, amikacin, gentamicin, kanamycin, streptomycin, chloramphenicol, clindamycin, erythromycin, telithromycin, ciprofloxacin, norfloxacin, tetracycline, linezolid, trimethoprim/sulphamethoxazole 1:19, rifampicin, quinupristin/dalfopristin, teicoplanin, nitrofurantoin, and fusidic acid. MICs for linezolid resistant isolates by disk diffusion were also confirmed by the agar dilution method on Mueller-Hinton agar. The results were scored according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI, 2015). Staphylococcus aureus ATCC25923 and Escherichia coli ATCC25922 were included as a control. The presence of the mecA/mecC gene was studied by PCR in all cefoxitin-resistant isolates (Perez-Roth et al., 2001;Stegger et al., 2012).

Statistical Analysis
For statistical analysis, the bacterial numbers were converted to base-10 logarithms. MPN values < 0.3 MPN/g were set to 0.15, and MPN values > 110 MPN/g were assigned the maximum value (Motes et al., 1998). The data of prevalence and levels of S. aureus-positive samples sorted by different food type and sampling sites were compared by using the analysis of variance (SPSS v21.0). Differences between the mean values were significant when P < 0.05.

Prevalence and Levels of S. aureus in Retail Meat in China
In this survey, a total of 1,850 samples were collected in 39 Chinese cities. Overall, there were 647 samples showed coagulase positive and confirmed for S. aureus by API STAPH test. The average incidence of S. aureus was 35.0% (647/1,850), the MPN value of the majority of the positive samples ranged from 0.3 to 110 MPN/g, which showed the geometric mean was 10.35 MPN/g. Out of the 647 positive S. aureus samples, 33 (5.1%) exceeded 100 MPN/g, whereas 319 (49.3%) were below 1 MPN/g.
The analyzed food products were classified into three categories (raw meat, quick-frozen meat, and ready-to-eat meat), and the values of S. aureus contamination in each sample were determined ( Table 1). Among the analyzed categories, raw meat was the most frequently contaminated with S. aureus, with a prevalence that reached 51.0%. The prevalence of S. aureus in raw poultry, raw mutton, raw beef, raw pork and bacon/sausage was 67.9, 54.5, 50.4, and 18.6%, respectively. The quick-frozen meat showed 43.4% of contamination for S. aureus in this survey, including 60.9% of positive samples in quick-frozen poultry, 50.0% in quick-frozen pork, 31.4% in quick-frozen beef, 30.9% in quick-frozen mutton, and 29.4% in quick-frozen dumpling. The prevalence of S. aureus in RTE meat was lower than that in other samples, which detected 12.2% of RTE samples were positive for S. aureus. S. aureus were detected in 25.0% (3/25) of RTE beef, 12.7% (22/173) of RTE pork, and 11.8% (54/456) of RTE poultry, whereas RTE mutton was free of S. aureus. Combining the MPN values, raw meat showed the highest levels of S. aureus (14.05 MPN/g), followed by RTE meat (9.11 MPN/g) and quick-frozen meat (6.37 MPN/g). Most of the S. aureuspositive samples exhibited less than 10 MPN/g, such as RTE poultry (9.89 MPN/g), quick-frozen poultry (7.37 MPN/g), pork (6.67 MPN/g), RTE pork (6.24 MPN/g), quick-frozen dumpling (4.99 MPN/g), quick-frozen mutton (1.96 MPN/g), quick-frozen pork (1.14 MPN/g), and bacon/sausage (0.59 MPN/g).
The distribution of S. aureus among different sampling sites is shown in Table 2. However, all of the 39 cities had S. aureuspositive samples, ranging from 16.7% in Shenzhen to 48.8% in Beihai. The prevalence of S. aureus exceeded 40% in one third of sampling sites. The highest contamination level in Beihai was 20.94 MPN/g, which showed 47.6% (10/21) of the positive samples were exceeded 10 MPN/g, whereas the lowest level in Changsha was only 0.43 MPN/g, which showed 80.0% (8/10) of the positive samples were less than 1 MPN/g. A total of 1,201 samples were collected from South China, with 35.3% prevalence, and 26 positive samples (26/424, 6.1%) had S. aureus densities exceeding the detectable level of 100 MPN/g. In North China, S. aureus was detected in 223 out of 649 samples (34.4%), of which only 7 samples (3.1%) of these samples had levels exceeding 100 MPN/g, whereas 127 samples (57.0%) yielded <1 MPN/g. Thus, there is no significantly difference (p > 0.05, χ 2 test) between South China and North China.
A phylogenetic tree based on the 7 concatenated MLST sequences shows the relatedness between the strains. The STs of S. aureus isolates were further analyzed based on the sampling sites and food sources (Figures 2A,B). All STs were fell into three different cladogram (designated as A, B, and C) which originate from CC7. Most novel were single-locus variant. Cladogram A contained CC6, CC5, CC1, CC188, ST97, ST20, ST25 and some relevant single-locus variant; Cladogram B included CC398, CC59, ST522, ST30 and some relevant single-locus variant; Cladogram C contained ST72, ST8, ST630, and ST692. It could be found that three different cladogram were evolved in directions, but distribute in different cities and different food sources, which showed high genetic diversity of S. aureus in China. The quickfrozen meat and raw meat were the major types in most of the CCs, whereas RTE meat was dominant for CC72. Interestingly, CC72 was only identified in S. aureus isolated from South China (Guangzhou, Macao, Hong Kong and Haikou).

DISCUSSION
Recent years, although many studies about the prevalence of S. aureus in retail food have been reported in China (Wang et al., 2012Zhang et al., 2013;Chao et al., 2015), the investigation of qualitative and quantitative in national level are lacking. Therefore, a more comprehensively investigation for the incidence of S. aureus isolated from retail meat foods is needed. In this study, samples collected covered all provincial capitals of China, and the food types were also more diversification. Moreover, in conjunction with resistance properties and molecular types, it is better to understand the genetic background of these strains.
In this study, we also identified 62 MRSA isolates by the disk diffusion method and mecA/C detection. It showed 7.14% of MRSA in S. aureus isolates from retail meat. Most of the MRSA strains were resist to β-lactams and other antibiotics which showed 77.4% (48/62) were resistant to more than 10 antibiotics. This result as similar with the results of research by Wang et al. (2017) and Hao et al. (2015) in China. In fact, the prevalence of MRSA varies greatly by geographical location in retail meats. For example, MRSA was present in 1.9% of 3,520 retail meats in the USA (Ge et al., 2017), 0.5% (13/2,810) in Korea (Kim et al., 2015), 1.6% (5/318) in Spain, 13% in Denmark (19/145) (Tang et al., 2017), 11.9% in the Netherlands (264/2,217), and 24.8% in Canada (655/2,640) (Narvaez et al., 2016). Sample size, geographic area and collection period may responsible for the differences observed (Ge et al., 2017).
The genetic types of 868 S. aureus isolates acquired from raw meat and meat products in China were characterized by MLST and spa typing. Comparing the STs and spa types, it seems that spa typing had a superior distinguishability than MLST. In this study, more than one spa type was detected in the major of STs. This result is consistent with those of previous studies by Hata et al. (2010) and Song et al. (2015). In our study, despite the wide genetic diversity observed among the 868 S. aureus meat isolates, a great proportion of the population belonged to finite number of major clones: ST1, ST7, ST5, ST398, ST188, ST6, ST59, ST88, ST15, and ST9 (74.3%, 645/868) were the dominant STs in S. aureus from retail meat. In concordant with MLST types, t127, t091, t002, t189, t034, t701, t437, t899, t796, t084, t3092, t085, t164, and t1376 (63.7%, 553/868) were the dominant spa types. According to the S. aureus MLST database (https://pubmlst.org/ bigsdbdb=pubmlst_saureus_isolates&page=profiles), there were 448 strains of ST1, 54 strains of ST7, 3,993 strains of ST5, 1,412 strains of ST398, 183 strains of ST188, 82 strains of ST6, 172 strains of ST59, 108 strains of ST88, 485 strains of ST15, and 50 strains of ST9 were demonstrating, the origins of them covered US, Canada, Poland, The Netherlands, Denmark, Australia, Bulgaria, Italy, Malaysia, Ivory Coast, France, Japan, Switzerland, and many other countries around world. Most of them isolated from the blood, pus, nasal swab, or skin swab. Beside, these types of S. aureus isolates have been relevant to a variety of clinical infections (Grundmann et al., 2010;Neocleous et al., 2010;Lozano et al., 2011;Valentindomelier et al., 2011;He et al., 2013;Suhaili et al., 2016). it indicates that these types of S. aureus strains have a theoretical pathogenic potential.
In Denmark, Tang et al. (2017) reported that the most frequent MRSA spa type and MSSA spa type were both CC398-t034. In USA, 25.9% of MRSA were identified as ST398-t034 (Hanson FIGURE 2 | Minimum spanning tree of MLST data for 868 S. aureus isolates. Each circle represents one ST, the size of which is related to the number of strains within this ST, while the branch numbers account for the number of allele differences between connected STs. The colors in the circles in (A) represent the sampling locations, and the colors in the circles in (B) represent the food sources. The gray zones surrounding some STs indicate that these genotypes belong to the same clonal complex. et al., 2011). In Germany, Feßler et al. (2011) also found 28 of 32 food related MRSA were CC398. In fact, CC398 strains are usually found in retail chicken, turkey, pork, and beef (Lozano et al., 2009;Huber et al., 2010;Argudín et al., 2011;Verhegghe et al., 2013), and have been perceived as a livestockassociated pathogens. Interestingly, we only found 7 isolates that were MRSA ST398, whereas MSSA ST398 was a common clone (52 isolates) in retail meat, and meat products from China in this study. Currently, infections caused by MSSA ST398 have been reported in humans and shown to cause infections more frequently than MRSA ST398 (Valentindomelier et al., 2011;Mediavilla et al., 2012;David et al., 2013;Li et al., 2015). In addition, the staphylococcal cassette chromosome mec (SCCmec) element types were diverse in MRSA ST398 . In Price et al. (2012) collected 89 CC398 S. aureus isolates (including MSSA and MRSA) from animals and humans spanning 19 countries and four continents by whole-genome sequence typing and found that livestock-associated MRSA CC398 originated as MSSA ST398 in humans. The lineage appears to have undergone a rapid radiation in conjunction with the jump from humans to livestock, where it subsequently acquired tetracycline and methicillin resistance (Price et al., 2012). Therefore, when antibiotic selection in connection with food animal production, it maybe raise the potential chance of MSSA CC398 to acquire the SCCmec cassette , however, further studies should be elucidate the possible rules for regarding MSSA ST398 as well.
In the present study, the majority types of MSSA was ST1-t127 (93/868, 10.7%) and ST7-t091 (92/868, 10.6%). In fact, ST7-t091 and ST1-t127 isolates have been reported as the fourth and sixth most prevalent clone, comprising both MSSA and MRSA strains, isolated from human invasive infections in 26 European countries (Grundmann et al., 2010). In Franco et al. (2011) demonstrated that t127/ST1 isolates can be assigned to two genetically different clusters (porcine and human) and hypothesized that ST1-t127 strains could represent another lineage of livestock-associated pathogens. On the other hand, CC59-t437 (41.9%, 26/62) was the predominant types in retail meat MRSA in China, followed by ST9-t899 (27.4%, 17/62). CC59-t437 is the predominant Asian community-associated MRSA (CA-MRSA) lineage, ranging from 35.8 to 76.7% with CA-MRSA in China . In Yang et al. (2017) were collected S. aureus strains in Beijing Children's hospital from respiratory tract, skin, and soft tissue, sterile sites in 104 children cases and found 61.7% of CA-MRSA was ST59-SCCmec IV-t437. Except China, ST59 was also reported in Vietnam, Japan, Australia, and other counties for CA-MRSA infection (Tang et al., 2007;Coombs et al., 2010;Higuchi et al., 2010). Therefore, further research will be proceed and found the reason of why ST59-t437 was the predominant types of MRSA isolates in retail meat in our study China. Besides, ST9-t899 was consider that the predominant S. aureus and MRSA genotype in pigs and related workers in Asia (Chuang and Huang, 2015). According to the previous studies, there were many studies have been reported the prevalence of this type of MRSA in Taiwan, Hong Kong, Malaysia, Thailand, and other countries from livestock (Neela et al., 2009;Graveland et al., 2011;Larsen et al., 2012;Lo et al., 2012). Thus, the majority types are different than those in studies in other counties. However, as well as livestock-associated lineages and clinical S. aureus clones, it revealed that these strains were widespread from retail meat and meat products in China. The pathogenic potential of these strains from retail meat food in China should not be ignored.

CONCLUSIONS
In conclusion, we report a wide scale and systematic investigation of S. aureus from retail meat and meat products in China, it supplements the nationwide qualitative and quantitative data of the prevalence and levels of S. aureus. The contamination of S. aureus was common in retail meat in China, but the levels of S. aureus were not very excessive. The prevalence rate of S. aureus in raw meat and quick-frozen meat was significantly higher than that in RTE meat. Most S. aureus isolates exhibited resistance to a variety of antimicrobials. By molecular typing analysis showed that these isolates had a high genetic diversity. The majority of these types have been linked to human infections worldwide, indicating that S. aureus strains of these types in retail meat related S. aureus in China have at least a theoretical pathogenic potential. However, the frequently types in our study are different than those in studies in other counties. Thus, further studies may need to elucidate the potential origins of these strains in China.

AUTHOR CONTRIBUTIONS
QW, JZ, SW, and TL conceived and designed the experiments. JH and FZ performed the experiments. SW, HW, and HZ analyzed the data. XY, LX, MC, YD, and SZ contributed reagents, materials, analysis tools. SW, JW, and QW contributed to the writing of the manuscript.