Occurrence and Patterns of Enterotoxin Genes, spa Types and Antimicrobial Resistance Patterns in Staphylococcus aureus in Food and Food Contact Surfaces in Singapore

Staphylococcus aureus contamination of food and food contact surfaces is a public health concern given its virulent and antimicrobial-resistant properties worldwide. In this study, a total of 181 MSSA isolates were analyzed for SE genes, antimicrobial resistance patterns, and spa types. Overall, 24.9% of isolates were positive for SE gene detection, with sea being the most prevalent classical SE (18.8%). The most predominant sample sources for SE gene contamination were hand swabs for sea (6/48), meat dishes for seb (3/14) and seafood dishes for sec (2/24). Antimicrobial resistance was also observed at relatively high frequencies for the clinically important antibiotics penicillin G and ampicillin (both 54.7%), followed by tetracycline (14.9%) and azithromycin (8.8%). In addition, characterization of spa types revealed spa type t5078 to be the most predominant (40.3%), with significant associations between spa types t127 and t5521 and the sea gene. This study offers insights into the enterotoxin gene and antimicrobial resistance profiles of S. aureus in cooked or ready-to-eat food to inform future surveillance and epidemiological studies.


Introduction
Staphylococcus aureus has been recognized as a ubiquitous pathogen responsible for Staphylococcal food poisoning (SFP), a gastrointestinal intoxication resulting from the ingestion of food contaminated by enterotoxigenic S. aureus [1]. While S. aureus does not form spores, their opportunistic nature encourages growth in a wide range of temperatures (7 to 48.5 • C) and pHs (4.2 to 9.3) [2]. These characteristics promote the growth and spread of S. aureus in many food products, especially meat and meat products, poultry and egg products, unpasteurized milk, and dairy products [3]. After contamination, improper storage conditions and poor hygiene practices accelerate the growth of S. aureus, allowing it to reach the cell density necessary for enterotoxin production. Hence, it is crucial that proper hygiene standards are adhered to during food processing and storage to minimize the spread and growth of the pathogen and its enterotoxins throughout the food processing chain.
Staphylococcal enterotoxins (SEs) produced by coagulase-positive staphylococci are the main causatives agents of SFP. SEs are resistant to heat, proteolytic enzymes, and other environmental conditions [4]. Due to their stable nature, the detection of SEs is a definite method for the confirmation of outbreaks and the enterotoxigenicity of strains. There are over 20 S. aureus enterotoxins identified. Based on serological classification, they are grouped as classical genes and non-classical genes (new SEs). Classical genes are the Amplification using multiplex PCR was conducted using the following parameters: initial denaturation of the strand at 98 • C for 30 s, followed by 30 cycles of denaturation at 98 • C for 10; annealing at 61 • C for 30 s; extension at 72 • C for 30 s; and final extension for 10 min at 72 • C. For amplification using singleplex PCR, the following parameters were used: initial denaturation of the strand at 98 • C for 30 s, followed by 35 cycles of denaturation at 98 • C for 10 s; annealing at 57 • C for 30 s; extension at 72 • C for 30 s; and final extension for 10 min at 72 • C. PCR-positive MSSA samples were confirmed with a latex agglutination test (PBP2) (Oxoid) and a cefoxitin disc (Oxoid) using the disc diffusion method following Clinical and Laboratory Standards Institute (CLSI) guidelines [18,19].
The amplified products were visualized using gel electrophoresis on a 1.5% agarose gel for multiplex PCR 1, singleplex PCR 1 and 2, and a 2% agarose gel for multiplex PCR 2. Detectable PCR bands were confirmed to contain the virulence genes.

Spa Typing
The spa typing for the S. aureus isolates was performed using the following method [20][21][22][23]. The X region of the protein A gene was amplified using PCR with four primer sets: 1113f and 1514r; 1095f and 1517r; 1084f and 1618r; and 238f and 1717r. If no PCR amplification was detected with one of the primer sets, the other three sets were used for PCR amplification instead.
Amplification was conducted using the following parameters: For primers 1113f and 1514r, initial denaturation of the strand at 98 • C for 30 s is followed by 35 cycles of denaturation at 98 • C for 10 s, annealing at 61 • C for 30 s, extension at 72 • C for 30 s, and final extension for 10 min at 72 • C. For primers 1095f and 1517r, initial denaturation of the strand at 98 • C for 30 s is followed by 35 cycles of denaturation at 98 • C for 10 s, annealing at 45 • C for 30 s, extension at 72 • C for 30 s, and final extension for 10 min at 72 • C. For primers 1084f and 1618r, 238f and 1717r, initial denaturation of the strand at 98 • C for 30 s is followed by 35 cycles of denaturation at 98 • C for 10 s, annealing at 55 • C for 30 s, extension at 72 • C for 30 s, and final extension for 10 min at 72 • C. The amplified products were visualized using gel electrophoresis on 1.5% agarose gel. Detectable PCR bands were confirmed to contain the spa gene.
PCR products were purified using the QIAquick ® PCR Purification Kit (Qiagen, Hilden, Germany) and sequenced by capillary electrophoresis using BigDye Terminator chemistry (AIT Biotech, Singapore). Sequences were analyzed using BioNumerics v7.6 to determine spa types.

Statistical Analysis
All statistical analyses were performed using GraphPad Prism 8.0 (GraphPad Software, LLC, San Diego, CA, USA). A p-value < 0.05 was considered statistically significant. Nonrandom associations between categorical variables (spa type and SE gene) were determined using the Fisher's exact test. Cluster analysis was performed on BioNumerics v7.6 and a maximum distance of 2 was used to determine closely related spa types in the same cluster.

Occurrences and Distribution of SE Genes
The prevalence of S. aureus in food and food contact surfaces is 15.4% (237/1540 samples). All 237 S. aureus isolates were determined to be methicillin-susceptible S. aureus (MSSA). Of the 237 S. aureus isolates tested, 181 could be typed using spa sequencing. The remaining strains that could not be typed were excluded from the analysis. The frequency of isolates from food and food contact surfaces are shown in Table 3. The occurrence of each SE gene across all isolates is shown in Table 4 below.  The most common classical SE gene is the sea gene, while the most common nonclassical genes are seg and sei. In total, seg and sei genes had the highest occurrence (25/96) (26.0%) across all SE genes. sed and see genes were not detected in any of the isolates tested in this study.
The predominance of classical SE genes was categorized based on their food and food contact surface categories, as shown in Figure 1 below. The sea, seb and sec genes were predominantly found in hand swabs (6/48), meat dishes (3/17) and seafood dishes (2/26) respectively.

Distribution of Spa Types
The most predominant classical SE gene, sea, was observed to have the highest proportion in t5521, representing 73% (11/15) of all sea-positive isolates. The associations between spa type t5521 and the presence of the sea gene (p < 0.0001) and between spa type t127 and the sea gene (p = 0.0138) were determined to be statistically significant using Fisher's exact test.

Distribution of Spa Types
The most predominant classical SE gene, sea, was observed to have the highest proportion in t5521, representing 73% (11/15) of all sea-positive isolates. The associations between spa type t5521 and the presence of the sea gene (p < 0.0001) and between spa type t127 and the sea gene (p = 0.0138) were determined to be statistically significant using Fisher's exact test. The most predominant classical SE gene, sea, was observed to have the highest proportion in t5521, representing 73% (11/15) of all sea-positive isolates. The associations between spa type t5521 and the presence of the sea gene (p < 0.0001) and between spa type t127 and the sea gene (p = 0.0138) were determined to be statistically significant using Fisher's exact test.    A minimum-spanning tree was constructed to perform spa clustering analysis for all isolates (Figure 3). Clusters were arbitrarily assigned to clustering complexes spa CC01 to spa CC04, according to the four definitive clusters observed. Spa types were partitioned into complexes when the distance between connected nodes was ≤2.

Overall Occurrence of SE Genes
In this present study, 24.9% (45/181) of the isolates were detected for the presence of at least one SE gene. The incidence of S. aureus detected with at least one or more SE genes in this present study (24.9%) was relatively lower than that reported in other countries, such as Korea (48.0%), China (54.4%), and Italy (55.5%) [24][25][26], which was expected. One of the most common foods associated with SFP is milk and dairy products [3]. In Singapore, only heat-treated milk is permitted to be sold for direct human consumption [27], hence there is a lower risk of S. aureus contamination as compared to raw milk, which could be a possible reason for the lower incidence compared to other countries. Another possible reason is that food handlers in Singapore are required to undergo a compulsory food safety course to equip them with the basic hygiene knowledge required for handling food, resulting in a lower incidence of S. aureus. The lower incidence of S. aureus with SE genes also correlates with the occurrence of SFP outbreaks, as Singapore has no known reported SFP outbreaks compared to other Asian countries [6][7][8].
Among the ten SE genes tested, the classical SE gene sea (18.8%, 18/96) and the nonclassical SE genes sej (26.0%, 25/96) and sei (26.0%, 25/96) were detected at the highest frequencies. For the classical SE gene sea, observation was similar to studies conducted in other countries, such as Taiwan (29.9%) and Iran (25.5%), where the sea gene was most prevalent among the classical genes [5,28]. The sea gene was most commonly isolated in cases of SFP and was frequently isolated in SFP outbreaks in Japan and the United States [3,8]. Enterotoxins sea and seb are known to cause approximately 90% of staphylococcal food poisoning worldwide [29]. The presence of enterotoxin genes in these isolates suggests the isolates' potential to produce toxins under favorable conditions and cause staphylococcal food poisoning (SFP) if allowed to grow in large numbers in food.
In contrast, several other studies by Hait et al. [30] and Tang et al. [31] have found non-classical genes to be the most predominant genes detected in the isolates investigated. Non-classical genes are new types of genes that have lower expression than classical genes. A non-classical sei gene has been detected in food poisoning-associated S. aureus isolates in Switzerland, the United Kingdom, and Japan [32][33][34]. However, despite the presence of sei genes in these isolates, it remains an open question whether these isolates have produced enterotoxins in sufficient amounts in food to cause SFP. The detection of enterotoxin sei in food should be explored further to make an accurate association between the sei gene and its ability to cause food poisoning.

Occurrence of SE Genes according to Food and Food Contact Surface Category
Due to the widespread occurrence of classical genes among SFP outbreaks [5], this study will focus on the comparison of the occurrence of classical SE genes across food and food contact surfaces. Of the food and food contact surfaces, hand swabs had the highest incidence of the sea gene (12.5%, 6/48). This is similar to studies conducted in Brazil and Japan, where high occurrences of the sea gene were detected in hand swab samples [35,36]. Without proper hygiene practices, such as wearing gloves during food preparation, S. aureus can be transmitted from human skin to food. This suggests that food handlers without proper hygiene care may increase the risk of contamination in food, as they act as vectors for the spread of enterotoxigenic S. aureus to food [2,35,37], which increases the risk of consumers consuming food contaminated with enterotoxigenic S. aureus. Contamination by food handlers contributes significantly to food poisoning outbreaks. In the United States, 42% of outbreaks between 1975 and 1988 were attributed to contamination by food handlers [2].
Of the food and food contact surfaces, meat dishes had the highest incidence of enterotoxin seb (21.4%, 3/14). Previous studies have reported few or no detections of the seb gene in S. aureus isolates in retail meat samples [38,39], which is interesting to note as the seb gene is directly associated with human contamination [40]. Meat dishes collected in this study could be more susceptible to human contamination, as the dishes, including chicken rice and duck rice, involve post-cooking manipulation, such as cutting and shifting the meat from chopping board to plate. Potential contamination sources include cutting boards, knives, or improper hygiene practices by food handlers [24]. Similar to the sea gene, the seb gene has remarkable stability against heat and proteolytic digestion [29,41,42]. Contamination of food with the seb gene in suitable numbers could result in severe food poisoning as well [19]. The occurrence of the sec gene among S. aureus isolates was highest in seafood dishes (7.7%, 2/26). Similar findings were reported in a study where 12.5% (1/8) of fish products were contaminated with the sec gene [26].
The presence of enterotoxin genes in S. aureus isolates is not necessarily a definitive indication of protein expression in these genes, as these genes may be non-functional or silent due to point mutations [43]. In addition, the level of enterotoxin production is dependent on other factors, including pH, water activity, temperature, and other parameters [34]. Knowledge on the occurrence of enterotoxin genes in this study, therefore, does not reflect the true enterotoxigenic potential of the S. aureus isolates. This limitation calls for greater research into the expression of genes in isolates retrieved from food to inform exposure and quantitative microbiological risk assessment (QMRA). Nonetheless, the presence of S. aureus strains with multiple enterotoxin genes still presents a threat to public health with respect to the consumption of contaminated food and contamination by food handlers.
Spa type t5078 has been linked to MSSA isolates isolated from patients in different countries. In Singapore, spa type t5078 was discovered in a MSSA isolate that was detected on an infected indwelling graft in a patient suffering from chronic renal failure [47]. In Taiwan, spa type t5078 was isolated from blood samples from patients, which were then discovered to be MSSA isolates [48]. According to Tunsjø et al. [49], S. aureus shares similar virulence genes, pathogenicity islands, and bacteriophages with S. argenteus. This is consistent with a study by Aung et al. [50], where 50% (12/24) of S. argenteus isolates were classified into spa type t5078 and other spa types with similar repeat profiles to t5078. Spa type t084 was found to be the most predominant spa type among MSSA isolates in a children's hospital in Poland and in the United States, with reports of invasive infections and being present in healthcare-associated and community-onset infections [51,52]. In another study, spa type t084 was also one of the predominant spa types among MSSA isolates among healthcare workers and patients [53]. The third predominant spa type, t5521 (6.1%, 11/181), was not actively studied in many countries. In a study conducted by Uhlemann et al. [54], t5521 was identified as one of eight new spa types isolated from S. aureus isolates from patients in Martinique. However, as t5521 is a relatively new spa type, no further extensive research was conducted.
Spa type t127 was also associated with an MSSA outbreak caused by ice-cream in Germany, with a high concomitance with the sea, in concordance with the results of this study (p = 0.0138) [55]. This links spa type t127 to potential food poisoning events. However, the statistically significant association (p < 0.0001) between spa type t5521 and sea in this present study has not been reported in other studies to the best of our knowledge. Further research is recommended to validate the statistical associations, which can aid in surveillance and epidemiological analysis of S. aureus infections and SFP outbreaks [56].
One limitation of spa typing in this study was the high proportion of non-typable spa types (23.6%), either due to the low quality of tandem repeats or no sequence generated. Future studies could consider using WGS to evaluate the reliability of spa typing by PCR. Spa-typing has been effective in distinguishing S. aureus from various sources, which will be relevant and useful for the epidemiological determination of food sources in outbreak investigations. Although the S. aureus strains in this study were isolated from surveillance and risk assessment studies and not from outbreak investigations, studying the genetic patterns of S. aureus isolates in food and food contact surfaces will be useful to understand the molecular epidemiology of these isolates, which will be useful in cases of improper hygiene practices or food handling during food production and storage.

General Antimicrobial Resistance Patterns
In this study, resistance to beta-lactams, specifically penicillin G and ampicillin, was observed at the highest frequency (54.7%, 99/181). The results are in agreement with other reports regarding the resistance of S. aureus detected in food to penicillin G in the United States (67.4%), Kuwait (82.0%), China (83.7%), and Western Algeria (60.8%) [57][58][59][60]. Notably, ampicillin and penicillin G resistance occurred at the same frequency (54.7%), similar to the results observed in bovine milk samples in China (91.4%) [61] and MSSA isolates in Trinidad and Tobago (11%) [62]. Penicillin resistance through beta-lactamase is conferred by the blaZ gene, which can be chromosomal or plasmid-encoded [63,64]. Furthermore, the spread of Penicillin G resistance occurs with the spread of resistant strains of S. aureus, where food could act as a vector [63,65]. However, as many clinically relevant S. aureus strains do possess beta-lactamase functions [66], penicillin is unlikely to be used for treatment of SFP, and thus the high resistance to both penicillin G and ampicillin in most sample categories identified in this study could be inherent. However, this study showed limited resistance to amoxycillin/clavulanic acid (0.6%), perhaps owing to the beta-lactamase inhibition activity of clavulanic acid [67]. Therefore, it could be postulated that beta-lactams are resistant. S. aureus isolates in this study were likely due to the presence of beta-lactamase activity.
The findings from this current study also showed that tetracycline resistance was high (27/181, 14.9%). Other studies in the United States (56.4%) and China (24.4%) have reported varied resistance to tetracycline [57,68,69]. The varied resistance to tetracycline in different countries could be explained using the varying usage of tetracycline in animal feeds, and the treatment of bacterial infections in plants, agriculture, and human medicine [70]. While tetracycline resistance in this study is high, compared to other antimicrobial agents tested, the frequency is still considerably low compared to other countries and thus should not be a cause for concern.
To date, there is a limited understanding of the transmission of antimicrobial-resistant S. aureus through food and food-contact surfaces. Food provides a conducive environment for the growth of bacteria. In addition, food chains are important in the spread of antimicrobial resistance between food and the environment [71,72]. These suggest that ready-to-eat food and food contact surfaces can be potential environmental sources for the colonization and circulation of antimicrobial-resistant S. aureus in the community [37,71]. Antimicrobial resistant S. aureus will not be a food safety concern if enterotoxin genes are not expressed and allowed to grow in sufficient numbers in food. However, the consumption of food contaminated with enterotoxigenic S. aureus with antimicrobial resistance could pose a serious food safety and public health risk [59]. In addition, antimicrobial-resistant S. aureus in food could contribute to a larger part of the environmental resistome. Hence, it is crucial to monitor the antimicrobial resistance and enterotoxigenicity of MSSA in retail food to understand epidemiological changes and develop strategies to prevent the contamination of the pathogen in food.

Antimicrobial Resistance Patterns according to Food and Food Contact Surface Category
The results indicated that tetracycline resistant S. aureus was high in bread products (3/5, 60%). This was reported in China as well (23.3%) [59]. Studies conducted in other countries have shown the possibility of associating antimicrobial resistance with a particular type of food, such as in Iran, where chloramphenicol resistance was identified in food products made from poultry meat, which correlated to the use of chloramphenicol to treat infections in poultry [73]. Due to the limited availability of an equal number of isolates across different sample categories, this study did not have the chance to show that a particular food or food contact surface category was at increased risk of acting as a vehicle for antimicrobial transmission. More data and larger sample sizes are required to calculate risk ratios and draw conclusions about whether an association between antimicrobial resistance and food or contact surfaces is causal in nature.

Conclusions
In conclusion, this study analyzed the patterns of SE genes, spa types, and antimicrobial resistance of S. aureus in food and food contact surface samples. This study revealed the occurrence of antimicrobial-resistant or enterotoxigenic S. aureus in food and food contact surface samples, suggesting that food or food contact surfaces can be potential vehicles for spreading S. aureus. Hence, there is a need for constant monitoring of food hygiene. In addition, findings from this study offer epidemiological insights to inform future surveillance and quantitative microbiological risk assessment.