Molecular Surveillance of MRSA in Raw Milk Provides Insight into MRSA Cross Species Evolution

ABSTRACT Methicillin-resistant Staphylococcus aureus (MRSA) in foods has been associated with severe infections in humans and animals worldwide. In the present study, the molecular characteristics of livestock-associated MRSA (LA-MRSA) and human-associated MRSA (hMRSA) isolates obtained in China, as well as MRSA isolates obtained from raw milk in 2018, were investigated. In total, 343 (20.38%; 343/1,683) S. aureus isolates were obtained from 1,683 raw milk samples from 100 dairy farms in 11 provinces across China. Among these, 49 (2.91%; 49/1,683) were mecA-positive MRSA. All LA-MRSA isolates were resistant to penicillin and highly resistant to erythromycin, sulfisoxazole, and clindamycin. Bioinformatic analysis the 49 genomes of LA-MRSA and 71 previously published hMRSA genomes isolated from Chinese individuals in 2018 indicated that blaZ, erm, ant(6)-Ia, aph(3′)-III, tet(K), cat, and aph(2″)-Ia were more prevalent in MRSA from raw milk (P < 0.05) compared to hMRSA. Additionally, hMRSA isolates were more significantly associated with ST5 (P < 0.01) compared to LA-MRSA; in contrast, ST338 was more prevalent among LA-MRSA isolates (P < 0.01). Likewise, the SCCmec type II was only detected in hMRSA isolates, whereas SCCmec type V and IV were more prevalent among LA-MRSA (P < 0.01). Furthermore, core-genome phylogenetic analysis showed the endemic characteristics of LA-MRSA in local provinces, as well as the close evolutionary relationships between MRSA from cattle and humans. Finally, homology analysis of mecA and blaZ genetic contexts revealed a high possibility of horizontal transmission of MRSA resistance genes among raw milk-associated and hMRSA strains, which increases the risk for public health. IMPORTANCE Methicillin-resistant Staphylococcus aureus (MRSA) is considered a public health concern as it is resistant to multiple antibiotics, thus being in zoonotic transmission of antibiotic resistance genes. MRSA causes serious public health issues and leads to hard-to-treat infections in humans and animals; therefore, it was meaningful to determine the prevalence of MRSA in raw milk samples and investigate phenotype and genotype of antimicrobial resistance and molecular characteristics in livestock-associated MRSA (LA-MRSA) and human-associated MRSA (hMRSA) in China, which could provide a theoretical basis for preventing and controlling the spread of MRSA between livestock and humans.


DISCUSSION
S. aureus, a major pathogen that threatens public health, can acquire resistance to antibiotics and produce a series of virulence factors. The clinical usage of methicillin led to the emergence of MRSA, and this evolution continued as a result from human contact or food consumption, and MRSA has been increasingly found in companion and food animals (7). In the present study, raw milk samples from 100 dairy farms in 11 Chinese provinces were collected to investigate the prevalence and resistance of MRSA isolates obtained from cattle. In addition, the molecular characteristics of MRSA isolates from cattle origin were analyzed along with hMRSA isolates collected in China in 2018 available in the NCBI database. Moreover, 343 S. aureus isolates (20.38%; 343/1,683) were obtained from 1,683 raw milk samples. Interestingly, the isolation rate of S. aureus in the present study was significantly lower (P , 0.01) than that related in previous studies conducted in the Philippines  (41.94%, 39/93) (19) and Greece (29%, 162/559) (20). The isolation rate of S. aureus (52.67%; 79/150) in samples from Shanghai was similar to that from Shihezi (58.1%, 54/93) (21). Furthermore, 49 mecA-positive MRSA isolates (2.91%; 49/1,683) were confirmed, although no significant difference (P . 0.05) was observed considering the findings reported in a study conducted in Italy (4%; 3/75) (22). Moreover, the percentage rate of MRSA recovery among S. aureus isolates in the present study was 14.29% (49/343), although no significant difference (P . 0.05) was observed with the findings reported by Badua et al. (23.08%; 9/39), but was higher (P , 0.01) than 0.99% (5/162) among milk samples from dairy goats in Greece (20).
Moreover, antimicrobial susceptibility profile of MRSA isolates was determined by broth microdilution method. MIC values indicated that MRSA isolates from cattle were resistant to penicillin (100%; 49/49) and highly resistant to erythromycin (89.80%, 44/ 49), sulfisoxazole (85.71%, 42/49), and clindamycin (81.63%, 40/49), which were similar to the findings reported in a previous study conducted in the Philippines (19), but was lower than the rates reported in a study conducted in Great Britain (23).
Genotyping of MRSA isolates was based on MLST, SCCmec, and spa typing. MLST, combined with SCCmec and spa typing, constitutes a useful tool for exploring the origin and evolution of S. aureus, and also provides a general naming system for S. aureus (27). ST239 and ST5 are hospital-associated MRSA (HA-MRSA) lineages that are prevalent in China and other Asian countries, whereas community-associated MRSA (CA-MRSA) lineages such as ST59, ST338, ST30, ST72, and ST8 are widespread in different geographical regions (7,28). In the study, minimum spanning tree analysis confirmed that ST5 (42.25%; 30/71) was the dominant lineage among hMRSA isolates; however, the ST338 variant (40.82%; 20/49) was the predominant type in raw milk samples. Interestingly, ST59 coexisted in MRSA from human (n = 11) and cattle (n = 13) origin. In addition, Bayesian analysis showed that hMRSA and LA-MRSA isolates shared certain clusters in the core genome phylogenetic analysis. Combined with SNPs detection, the results indicated a high possibility of horizontal transmission of MRSA isolates could occur between raw milk-and humanderived strains, which increases the risks to public safety. Animal-to-human transmission can occur through three main routes, i.e., direct contact, environmental contamination, or handling of products from infected animals; however, MRSA-contaminated raw milk is one of the most direct and dangerous routes for oral transmission to humans.
In total, 13 SCCmec types were identified in cattle and human MRSA isolates. Among these, SCCmec type V (33.33%; 40/120), II (30.83%; 37/120), and IV (28.33%; 34/120) were predominant among MRSA isolates, whereas SCCmec type IV was the most predominant type and accounted for nearly half of all MRSA isolates reported in a previous study (19,29). In contrast, STs within SCCmec types II, III, and IV were similar to those reported in another study (29). Among spa types, 22 spa types were identified in total, among which predominated t437 (34.17%; 41/120) and t2460 (20.83%; 25/120), which was similar to the findings reported in a study with HA-MRSA isolates which showed that spa types t437 and t2460 are highly epidemic in China.
Moreover, the analysis of the genetic context of mecA in MRSA isolates (difference in SNPs prevalence , 500) indicated that the homology of mecA was similar within the genus (either human or cattle origin) and highly similar within different species, thus demonstrating that LA-MRSA found in contaminated raw milk increases the diffusion risk of MRSA between veterinary and human spheres, which was consistent with the findings in a previous study with MRSA in pork (30). Moreover, molecular analysis revealed that blaZ was found within 16 different genetic contexts, all of which contained the structure blaI-blaR1-blaZ. In particular, type I (n = 43) was the most common genetic backbone, and incomplete insertion sequences DIS431mec and DIS1272were found both upstream and downstream blaI-blaR1-blaZ, respectively, which was similar to genetic contexts reported in previous studies conducted in Switzerland (31,32). In addition, the percentage of sequence homology in the blaZ genetic context among different sources was greater than 90%, which further strengthens the hypothesis of horizontal transfer of blaZ between LA-MRSA and hMRSA isolates. In contrast, lsa(E) (n = 3) was detected only in ST9-LA-MRSA, commonly associated with lnu(B), and insertion sequences DIS431R and DISVlu1 were found upstream and downstream the lsa(E)-lnu(B) structure. This genetic backbone had 100% sequence identity with the corresponding sequences of S. aureus strain 2868B2 available in the NCBI database (GenBank accession no. NZ_CP060141.1) (33). Thus, the information obtained from the characterization of marker genes and the genetic context can be used epidemiologically to elucidate outbreaks, identify possible sources of colonization (livestock or humans), and differentiate between community-and hospital-acquired strains. Moreover, it is important to identify and limit the emergence of MRSA strains belonging to the ST9-t3446-SCCmec type IX genotype and/or carrying acquired drug resistance genes mecA, blaZ, aac (69)-aph (20), aadD, ant (6)-Ia, lsa (E), dfrG, tet (M), fexA, and lnu (B) as reported elsewhere (34).
Herein, a large-scale investigation of antimicrobial resistance in MRSA isolates obtained from raw milk samples from dairy farms across China was conducted, which enable comparing the molecular characteristics of LA-MRSA and hMRSA. All MRSA isolates evaluated in the present study were resistant to penicillin and highly resistant to erythromycin, sulfisoxazole, and clindamycin. Except for mecA, erm was the most prevalent resistance gene found in MRSA isolates from cattle and human origins. Moreover, the number of ARGs carried by LA-MRSA was greater than that carried by hMRSA isolates; in particular, blaZ, erm, ant(6)-Ia, aph (39)-III, tet(K), cat, and aph(20)-Ia were significantly more prevalent in raw milk isolates than in the genomes of MRSA of human origin. In addition, the enterotoxin gene sel was predominant in milk, whereas hMRSA isolates commonly harbored sec. Furthermore, ST5 and ST338 were the most prevalent lineages among MRSA isolates of human and cattle origins, respectively. Additionally, the relationship of MRSA strains origin raw milk samples from the same province was closer in the core-genome phylogenetic tree. With the homology analysis of marked genetic environments of mecA and blaZ, all the results demonstrated that there was a highly possibility of horizontal transmission of MRSA isolates between foodborne raw milk and human-derived strains, which increased the risk to public safety.

MATERIALS AND METHODS
Sample collection and selection of bacterial isolates. A total of 1,683 raw milk samples were aseptically collected from healthy dairy cows of scalable farms located in different Chinese provinces. All samples (collected as triple duplicates) were stored momentarily in an insulated icebox and transported to the laboratory immediately after collection. Then, 1 mL of raw milk sample was mixed with 9 mL of 7.5% sodium chloride broth, and the mixture was incubated overnight at 37°C for enrichment. An aliquot of the bacterial suspension was then transferred uniformly onto CHROMagar S. aureus isolation medium (CHROMagar, Paris, France) followed by incubation for 24 h at 37°C. One colony was then selected and inoculated in 1 mL of brain heart infusion (BHI) (Land Bridge, Beijing, China) for incubation for 6 h at 37°C under shaking. The identification of selected colonies as S. aureus was conducted using a Vitek 2 Compact GP ID Card (bioMérieux, Marcy-l' Etoile, France).
Molecular identification of MRSA isolates. DNA of identified S. aureus isolates was extracted from pre-enrichment BHI using the TIANamp Bacteria DNA Kit (Tiangen, Beijing, China) according to the manufacturer's instructions. mecA-positive MRSA isolates were confirmed using multiplex PCR (PCR) targeting the 16S