Phenotypic and genotypic characterization of antibiotic resistance of methicillin-resistant Staphylococcus aureus isolated from hospital food

Background Pathogenic biotypes of the Methicillin-resistant Staphylococcus aureus (MRSA) strains are considered to be one of the major cause of food-borne diseases in hospitals. The present investigation was done to study the pattern of antibiotic resistance and prevalence of antibiotic resistance genes of different biotypes of the MRSA strains isolated from various types of hospital food samples. Methods Four-hundred and eighty-five raw and cooked hospital food samples were cultured and MRSA strains were identified using the oxacillin and cefoxitin disk diffusion tests and mecA-based PCR amplification. Isolated strains were subjected to biotyping and their antibiotic resistance patterns were analyzed using the disk diffusion and PCR methods. Results Prevalence of S. aureus and MRSA were 9.69 and 7.62%, respectively. Meat and chicken barbecues had the highest prevalence of MRSA. Prevalence of bovine, ovine, poultry and human-based biotypes in the MRSA strains were 8.10, 8.10, 32.43 and 48.64%, respectively. All of the MRSA strains recovered from soup, salad and rice samples were related to human-based biotypes. MRSA strains harbored the highest prevalence of resistance against penicillin (100%), ceftaroline (100%), tetracycline (100%), erythromycin (89.18%) and trimethoprim-sulfamethoxazole (83.78%). TetK (72.97%), ermA (72.97%), msrA (64.86%) and aacA-D (62.16%) were the most commonly detected antibiotic resistance genes. Conclusions Pattern of antibiotic resistance and also distribution of antibiotic resistance genes were related to the biotype of MRSA strains. Presence of multi-drug resistance and also simultaneous presence of several antibiotic resistance genes in some MRSA isolates showed an important public health issue Further researches are required to found additional epidemiological aspects of the MRSA strains in hospital food samples.


Background
Consumption of contaminated food is one of the most common cause of outbreak of food-borne diseases in hospitals [1]. Based on the general weakness of the hospitalized patients and also the possibility of occurrence of suppression in their immune system, hospital foodstuffs should have a high quality and safety [1].
The most cases of food-borne outbreaks in hospitals are occurred due to the consumption of food contaminated with Staphylococcus aureus (S. aureus) [2,3]. Staphylococcus aureus is commonly found in nose and respiratory system and on the skin [2,3]. It is responsible for the occurrence of nosocomial and community-acquired infections, food-borne diseases and food poisoning [2,3]. Occurrence of different types of gastrointestinal diseases which are known by vomiting, nausea, abdominal cramps, weakness and diarrhea and also toxic shock syndrome are attributed to the S. aureus strains [2,3].
One of the most interesting questions about the contamination of foods with S. aureus concerns the source of these contamination organisms [4][5][6]. When foods or some of their ingredients are of animal origin it can be of importance to determine whether the strains of S. aureus isolated originate from animals or from humans [4][5][6]. S. aureus can be disseminated in the host's environment. Presence of several biotypes in the S. aureus strains (bovine, ovine, poultry and human) with different microbiological characters has increased the importance of the issue [4][5][6]. Identification of S. aureus biotypes is a practical approach to determine the exact routes of food contamination and also find their microbiological and epidemiological similarities and differences [4][5][6]. Biotyping of S. aureus strains isolated from hospital food samples is essential to trace their origin and their public health significance, to investigate the relationship of the strains, and to determine their diversity within and between samples [4][5][6].
Food-borne S. aureus strains are usually resistant against several types of antibiotics [7][8][9]. Nowadays, methicillinresistant S. aureus (MRSA) has become a serious problem in hospitals [7][8][9]. Documented data revealed that about 50-70% of the S. aureus strains isolated from the hospital environment were MRSA [7][8][9]. MRSA strains are responsible for about 100,000 cases of infections with around 20% mortality rate each year in the United States [7]. High pathogenicity of MRSA strains [9], its high resistance to several types of antibiotics [9] and its foodborne aspects [9] have increased the importance of isolation of MRSA in hospital food samples. Staphylococcal food poisoning is an intoxication that results from the consumption of foods containing sufficient amounts of one (or more) preformed enterotoxin [2,3]. Therefore, the risk of MRSA contaminated food might be due to the important factors like cross-contamination [2,3,7].
MRSA strains of both animal and human origins are believed to serve as important reservoirs of antimicrobial resistance genes which can transfer and integrate into the MRSA genome leading to the emergence of new and potentially more resistant strains [10][11][12][13]. Documented data revealed that presence of certain antibiotic resistance genes is responsible for occurrence of severe antibiotic resistance [10][11][12][13]. Reports showed that high presence of mecA, aacA-D, tetK and tetM, ermA and ermC, msrA, linA and vatA, vatB and vatC antibiotic resistance genes in the S. aureus strains isolated of foodstuffs caused severe occurrence of resistance against methicillin, aminoglycosides, tetracyclines, macrolide-lincosamide-streptogramin B, macrolides, lincosamides and streptogramin A groups of antibiotics, respectively [10][11][12][13].
Reports of methicillin-resistant strains are challenging due to the large proportion of methicillin-resistant strains and increasing numbers of isolates reinforcing the need to revise their importance to food safety [10][11][12][13][14]. Therefore, screening of these elements is important for public health and despite the importance of such a screen, limited data are available for MRSA at the species level among the hospital food samples.
MRSA strains have been tested in hospital food samples to assess microbiological safety, sanitation conditions during processing, and storage quality of products. High pathogenicity of MRSA strains and general weakness of hospitalized patients make it necessary to assess the presence of MRSA strains in hospital food samples. The current research was done to study the prevalence rate and antimicrobial resistance properties of the MRSA biotypes isolated from various types of raw and cooked hospital food samples in Iran.

Samples
From June 2015 to June 2016, a total of 485 various types of raw and cooked hospital food samples including raw meat (n = 38), raw chicken (n = 37), raw fish (n = 9), meat barbecue (n = 31), chicken barbecue (n = 82), grilled fish (n = 19), soup (n = 94), salad (n = 56) and cooked rice (n = 119) were randomly collected from the big hospitals of the Isfahan province, Iran. Samples were immediately transferred to the Food Hygiene Research Center of the Islamic Azad University of Shahrekord in cooler with ice-packs. All food samples showed normal physical characters including odor, color and consolidation.

Isolation and identification of S. aureus
Each sample was aseptically weighed in an analytical balance and 25 g were transferred into a sterile plastic bag. Then, 225 mL of buffered peptone water (Merck, Germany) was added and homogenized in a Stomacher Bagmixer 400 W (Interscience, Saint-Nom, France) for 2 min. Five milliliter aliquot of the enriched homogenate was transferred into 50 mL Trypticase Soy Broth (TSB, Merck, Germany) supplemented with 10% NaCl and 1% sodium pyruvate. After incubation at 35°C for 18 h, a loopful of the culture was plated onto Baird-Parker agar supplemented with egg yolk tellurite emulsion (Merck, Germany) and incubated overnight at 37°C. Black shiny colonies surrounded by 2 to 5-mm clear zones were further identified on the basis of Gram staining, hemolytic activity on sheep blood agar (Merck, Germany), catalase activity, coagulated test (rabbit plasma), oxidase test, glucose O/F test, resistance to bacitracin (0.04 U), mannitol fermentation on Mannitol salt agar (Merck, Germany), urease activity, nitrate reduction, phosphatase, deoxyribonuclease (DNase, Merck, Germany) test, voges-proskaver (Merck, Germany) test and carbohydrate (xylose, sucrose, trehalose and maltose, fructose, lactose, mannose) fermentation tests [11].

Identification of Methicillin-resistant Staphylococcus aureus strains
Cefoxitin (30 μg) and oxacillin (1 μg) susceptibility tests were used to distinguish the MRSA strains from S. aureus isolates of hospital food samples. All tests were performed using the guidelines of the Clinical and Laboratory Standards Institute (CLSI) [15].
MRSA isolates were identified another time using the PCR-based amplification of mecA gene. MRSA strains were sub-cultured on TSB media (Merck, Germany) and further incubated for 48 h at 37°C. Genomic DNA was extracted from bacterial colonies using the DNA extraction kit (Fermentas, Germany) according to manufacturer's instruction.

Biotyping of the MRSA strains
Biotyping of the MRSA strains was done according to the method described by Devriese (1984) [16] with minor modifications. MRSA isolates were examined for production of staphylokinase and β-hemolysin, coagulation of bovine plasma and ability to growth on crystal violet agar. Staphylokinase production of the MRSA isolates were studied by incubating on bovine fibrin plates with or without dog serum as a plasminogen source. Staphylokinase production is determined by observation of lear zones on the bovine fibrin plates with dog serum. MRSA isolates were also cultured on 5% sheep blood agar media for determination of production of Beta-haemolysin. Development of broad discolored zones with sharp edges clearing at 4°C is a marker for production of Beta-haemolysin. Coagulation of bovine plasma is studied by adding 0.1 mL of an overnight culture of the MRSA isolates in Brain Heart Infusion broth (Merck, Germany) to tubes with diluted (1:3) bovine plasma. Occurrence of big clots within 6 h were considered as positive reaction. MRSA strains were also streaked into the crystal violet agar media and appearance of A (crystal violet spots with a bright or pale yellow color and yellow spots with blue margins), C (blue or violet growth spots with or without an orange tint) and E (white spots or white growth with a blue hue) types were studied.

Statistical analysis
Statistical analysis was done using the SPSS 21.0 statistical software (SPSS Inc., Chicago, IL, USA). Chi-square test and Fisher's exact two-tailed test were used to assess any significant relationship between prevalence of MRSA strains, their biotypes, antibiotic resistance genes and antibiotic resistance pattern. P value <0.05 was considered as statistical significant level.

Results
In this study, the prevalence of S. aureus strains in various types of raw and cooked hospital food samples were investigated and the results are shown in Table 2. Forty-seven out of 485 hospital food samples (9.69%) were positive for S. aureus. Furthermore, the prevalence of S. aureus in raw food samples with animal origin, cooked food samples with animal origin and cooked food samples without animal origin were 23.80, 9.09 and 4.08%, respectively. Chicken meat samples had the highest prevalence of S. aureus (27.02%) among all studied raw food samples. Meat barbecue samples had the highest prevalence of S. aureus (26.31%) among all studied cooked food samples with animal origin. Salad samples had the highest prevalence of S. aureus (7.41%) among all studied cooked food samples without animal origin. None of the raw and cooked fish samples were S. aureus positive. Statistically significant difference was seen in the prevalence of S. aureus between different types of food samples (P < 0.05). Moreover, a statistically significant difference was found in the prevalence of S. aureus between raw and cooked food samples.
Thirty-seven out of 47 S. aureus isolates (78.72%) were recognized as MRSA. The prevalence of MRSA strains in raw foods with animal origin, cooked foods with animal origin and cooked foods without animal origin were 16.66, 9.09 and 4.08%, respectively. All of the S. aureus isolates of meat barbecue, chicken barbecue, soup and salad samples had complete resistance against methicillin. Rice had the lowest prevalence of MRSA strains (20%). Statistically significant difference was seen in the prevalence of MRSA between different Table 1 Target genes, oligonucleotide primers and PCR conditions used for detection of antibiotic resistance genes in the MRSA strains isolated from various types of hospital food samples types of food samples (P < 0.05). Statistically significant difference was also found in the prevalence of S. aureus between raw and cooked food samples (P < 0.05). Table 3 represents the prevalence of different biotypes among the MRSA strains isolated from various types of hospital food samples. The prevalence of bovine, ovine, poultry and human-based biotypes in the MRSA strains isolated from various types of hospital food samples were 8.10, 8.10, 32.43 and 48.64%, respectively. The biotypes of the 2.70% of MRSA strains were determined as unknown. All of the MRSA strains recovered from soup, salad and rice samples were related to humanbased biotypes. Statistically significant difference was seen in the prevalence of different biotypes between different types of food samples (P < 0.05).

Discussion
S. aureus is considered as one of the most common causes of nosocomial infections, as well as the cause of most cases of food poisoning in hospitals [4,20]. Hospital meals are an indispensable portion of patient care. Safe and complete meals can encourage patients to eat well and giving them the nutrients they need to recover from surgery or illness. Foodstuff contamination with S. aureus may occur directly from infected food-producing animals or may result from poor hygiene during production processes, or the retail and storage of food, since humans may also harbor microorganisms. The current research is the first report of the biotyping and molecular characterization of antibiotic resistance in the MRSA strains isolated from various types of raw and cooked hospital food samples. Findings obtained from this research revealed that the prevalence of S. aureus in different types of hospital food samples was 9.69%. The prevalence rate of the S. aureus in hospital food samples of our research was higher than that of Spain (6.10%) [21] and Iran (6.42%) [22], Portugal (11.10%) [23] and Brazil (50%) [24].
Investigations conducted in the U.S. as well as numerous other countries, including Canada, Taiwan, China, Denmark, South Korea, Austria, France, Belgium, Italy, and The Netherlands, have isolated MRSA mainly from different types of foods [4,20]. Costa et al. (2015) [25] revealed that 28.10% of hospital food samples harbored MRSA strains. They showed that the prevalence of MRSA strains in beef, chicken, pork and fish samples were 23.30, 23.30, 37.50 and 30%, respectively which was higher than that found in our study.
Biotyping is a simple method used to trace the origin of S. aureus strains isolated from food samples. The results of our study showed that all of the MRSA strains isolated from rice, salad and soup samples were derived from humans. Furthermore, 48.64% of MRSA isolates of hospital food samples had human origin. Generally, the results revealed the role of infected humans in the dissemination and also transmission of MRSA strains to hospital food samples. The role of food handlers in transmission of MRSA strains into the food samples has also been reported by Castro [26]. A study which was conducted by Kitai et al. (2005) [27] supported the high prevalence of poultry-based biotypes found in our investigation (71.42%). They reported that about 80% of all S. aureus isolates of foodstuffs belonged to the poultry-based biotypes, while prevalence of human-based biotypes was 22.10%. Normanno et al. (2007) [28] revealed that the prevalence of human, ovine, not-host-specific, bovine and poultrybased biotypes of the S. aureus isolates of Italian food samples were 50. 40, 23.20, 17.60, 7.20 and 1.60%, respectively.
S. aureus causes food intoxication and doesn't lead to food infection [2,3]. Therefore, the risk of MRSA contaminated hospital food might be due to the crosscontamination. High prevalence of MRSA strains in cooked food samples may be due to the crosscontamination of cooked foods through food handlers and kitchen equipment.
Most of the tetracycline-resistant MRSA strains harbored tetK and tetM genes. Prevalence of aacA-D gene was high among gentamicin, amikacin and kanamycin-resistant MRSA strains. Prevalence of msrA, ermA and ermC and linA were also significant among the macrolide, erythromycin and clindamycin-resistant MRSA strains. Therefore, the pattern of the antibiotic resistance of the MRSA strains of hospital food samples was confirmed by the PCR amplification of the specific antibiotic resistance genes. MRSA strains of our study had considerable prevalence of resistance against clindamycin (48.64%). The most imperative mechanism involving resistance against clindamycin is modulated by methylase enzyme which often encoded by ermA and ermC genes [38]. Prevalence of ermA and ermC antibiotic resistance genes among the MRSA strains of our research were 72.97 and 27.02%, respectively. Majority of our isolates carried two tetracyclines, two erythromycins, one macrolide and several streptogramin resistance determinants reveals a great diffusion of these types of resistance. TetK, ermA, msrA and aacA-D which encode resistance against tetracycline, erythromycin, macrolides and aminoglycosides were the most commonly detected antibiotic resistance genes in the MRSA strains isolated of hospital food samples. The literature survey did not indicate any report on the prevalence of vatA, vatB, vatC, msrA, ermA, ermC, linA, aacA-D, tetK and tetM genes among the MRSA strains of hospital food samples. Kumar et al. (2010) [39] reported that the most commonly identified antibiotic resistance genes among the S. aureus isolates of food samples were linA (51.60%), msrB (46.10%), tetK (34.40%), tetM (34.40%), msrA (26.60%) and aacA-D (26.60%). Karataş et al. (2017) [40] revealed the higher prevalence of ermA than ermc antibiotic resistance genes among the clindamycin, erythromycin and telithromycin-resistant and also higher prevalence of tetM than tetK antibiotic resistance genes among the tetracycline-resistant MRSA strains which were similar to our findings. Our results were also similar with those of the previous research which was conducted by Simeoni et al. (2008) [41]. They reported that the prevalence of tetM, tetO, tetK, ermA, ermB, ermC, aac, blaZ and mecA antibiotic resistance genes amongst the S. aureus strains isolated from meat samples were 100, 0, 91.66, 16.66, 33.33, 58.33, 0, 100 and 58.33%, respectively. High prevalence of tetK and tetM antibiotic resistance genes in the MRSA isolates can be clarified by their usual genetic locations. Presence of tetK gene on small multicopy plasmids and tetM on conjugative transposons contributes to the spread of these determinants [42]. Some of the MRSA strains harbored ermC gene. This gene is often located on small multicopy plasmids which are present in many different staphylococcal species [42]. The ermA gene is usually carried by transposons which could explain its high prevalence amongst the MRSA strains. Resistance to aminoglycosides (43.24 to 81.08%) which encode by the aacA-D gene (62.16%) is more prevalent amongst the human-based biotypes. It is because of this gene is usually more diffused in staphylococci of human origin [42]. Johler et al. (2011) [42] reported that prevalence of ermC, tetK and tetM antibiotic resistance genes among the S. aureus strains isolated from cases of food poisoning, milk and pork were 25, 4.87 and 0%, 50, 0 and 12.82 and 0%, 12.19, and 53.84%, respectively. Podkowik et al. (2012) [43] reported that the prevalence of tetracycline resistance genes (tetO, tetK and tetM) and erythromycin resistance methylase gene (ermA, ermB and ermC) among the S. aureus strains isolated from ready to eat meat products were 44 and 60%, respectively.

Conclusions
The present investigation is the first report of the biotyping and study the antimicrobial resistance properties of MRSA strains isolated from raw and cooked hospital food samples. In this study, a total prevalence of 7.62% of MRSA as well as high prevalence of human and poultry-based biotypes was seen in hospital food samples. Considerable prevalence of resistance against penicillin, ceftaroline, tetracycline, erythromycin, trimethoprimsulfamethoxazole, doxycycline and gentamicin antibiotic agents and high distribution of tetK, ermA, msrA and aacA-D antibiotic resistance genes may pose a potential public health threat. Our findings exhibited that the pattern of antibiotic resistance and also distribution of antibiotic resistance genes were dependent on the biotype of MRSA strains. Human-based biotypes had a higher prevalence of resistance against human-based antibiotics and their corresponding resistance genes. Raw food samples with animal origin had the lower prevalence of human-based biotypes, while cooked food samples had a higher prevalence of human-based biotypes. Presence of multi-drug resistance and also simultaneous presence of several antibiotic resistance genes in some MRSA isolates must be considered as serious health hazard. Cooking food thoroughly is important, but preventing cross-contamination is the most effective ways to prevent occurrence of MRSA in hospital food. Further researches are required to understand higher epidemiological aspects of the MRSA strains in hospital food samples. Availability of data and materials All data generated or analyzed throughout this research are included in this published article.
Authors' contributions FSD designed the study and carried out the culture-based identification, biotyping and PCR genetic alignment. AAB and HG supported the study and carried out the samples collection, disk diffusion and statistical analysis. ER and AM carried out the writing and drafting of the manuscript. All authors read and approved the final manuscript.

Consent for publication
There was no consent for publication.