Genotypes, Exotoxin Gene Content, and Antimicrobial Resistance of Staphylococcus aureus Strains Recovered from Foods and Food Handlers

Abstract

Staphylococcal food poisoning, one of the most common food-borne diseases, results from ingestion of one or more staphylococcal enterotoxins (SEs) produced by Staphylococcus aureus in foods. In the present study, 64 S. aureus isolates recovered from foods and food handlers, associated or not associated with food-poisoning outbreaks in Spain, were investigated. They were assigned to 31 strains by spa typing, multilocus sequence typing (MLST), exotoxin gene content, and antimicrobial resistance. The strains belonged to 10 clonal complexes (CCs): CC5 (29.0%), CC30 (25.8%), CC45 (16.1%), CC8, CC15 (two strains each), CC1, CC22, CC25, CC59, and CC121 (one strain each). They contained hemolysin genes (90.3%); lukED (77.4%); exfoliatin genes eta, etd (6.5% each), and etb (3.2%); tst (25.8%); and the following enterotoxin or enterotoxin-like genes or clusters: sea (38.7%), seb (12.9%), sec (16.1%), sed-selj with or without ser (22.9%), selk-selq (6.5%), seh, sell, selp (9.7% each), egc1 (32.3%), and egc2 (48.4%). The number of se and sel genes ranged from zero to 12. All isolates carrying tst, and most isolates with genes encoding classical enterotoxins (SEA, SEB, SEC, and SED), expressed the corresponding toxin(s). Two CC5 isolates from hamburgers (spa type t002, sequence type 5 [ST5]; spa type t2173, ST5) were methicillin resistant and harbored staphylococcal cassette chromosome mec (SCCmec) IVd. Six (19.4%) were mupirocin resistant, and one (spa type t120, ST15) from a food handler carried mupA (MIC, 1,250 μg/ml). Resistance to ampicillin (blaZ) (61.3%), erythromycin (ermA-ermC or ermC) (25.8%), clindamycin (msrA-msrB or msrB) (16.1%), tetracycline (tetK) (3.2%), and amikacin-gentamicin-kanamycin-tobramycin (aphA with aacA plus aphD or aadD) (6.5%) was also observed. The presence of S. aureus strains with an important repertoire of virulence and resistance determinants in the food chain represents a potential health hazard for consumers and merits further observation.

INTRODUCTION

Staphylococcal food poisoning (SFP) is a common intoxication that results from the consumption of foods containing sufficient amounts of one or more preformed enterotoxins (3, 29). Symptoms have a rapid onset and comprise nausea and vomiting, with or without diarrhea (43). The illness is usually self-limiting but occasionally can be severe enough to warrant hospitalization (33). The real incidence of SFP is underestimated, mainly due to misdiagnosis and because sporadic episodes and minor outbreaks are not reported. Despite this, 29 outbreaks were reported in the European Union in 2008; 414 persons were affected, and 26 required hospitalization (17). With regard to Spain, 431 outbreaks have been recorded during the period from 1994 to 2008 (10, 17, 21, 32). Four of these outbreaks occurred in Asturias (a Northern Spanish region), at three different restaurants in 2002 (31) and at a nursing home for the elderly in 2006, and the associated isolates were examined in this work.

Staphylococcus aureus, the casual agent of SFP, is a common commensal of the skin and mucosal membranes. Food handlers carrying enterotoxin-producing strains in their noses or hands are regarded as the main source of food contamination, via manual contact or through respiratory secretions (41). S. aureus enterotoxins (SEs) belong to the broad family of pyrogenic toxin superantigens and have emetic activity (6, 18). Classical SEs (SEA, SEB, SEC, SED, and SEE) have been discovered in studies of S. aureus strains involved in SFP outbreaks and are classified into distinct serological types (15). Related toxins that lack emetic activity or have not been tested for it are designated staphylococcal enterotoxin-like toxins (SEls) (30). All known se and sel genes are located on mobile genetic elements, including the νSaβ genomic island, which carries the enterotoxin gene clusters known as egc1 (seg, seln, sei, selm, selo) (νSaβ type I), egc2 (seg, seln, selu, sei, selm, selo) (νSaβ type III), and variants thereof (5, 13, 42); S. aureus pathogenicity islands (SaPIs) (etd, tst, seb, sec, sell, selq, selk) (34); prophages (lukPV, eta, sea, see, selp, selk, selq) (19); and plasmids (etb, seb, sed, selj, ser, ses, set) (8, 36).

The changing epidemiology of S. aureus over recent decades, including the emergence of the novel methicillin-resistant S. aureus (MRSA) ST398 clone linked to food production animals (livestock-associated MRSA), highlights the necessity of monitoring the S. aureus clones that are circulating through the food chain (27). S. aureus contributes substantially to the virulence and resistance gene pool of Gram-positive bacteria, and molecular analysis of the circulating strains could help us understand the exchange of such genes between bacterial species and clones (1). In this study, we examined a collection of S. aureus strains recovered in the Spanish Region of Asturias from manually handled foods and food handlers, associated or not associated with SFP outbreaks recorded in the region. The objectives were (i) to establish the genetic backgrounds of the food-related isolates and to compare them with isolates that have been circulating in human carriers or causing disease in hospitals in Asturias, (ii) to investigate the distribution of exotoxin genes in food-related isolates and the production of tst and classical enterotoxins, and (iii) to evaluate the resistance properties of these isolates, paying particular attention to the presence of MRSA.

MATERIALS AND METHODS

Bacterial strains.

The S. aureus isolates analyzed in this study (64 isolates) were collected from food handlers (14 isolates) and manually handled foods (50 isolates) in Asturias (see Table 1). Thirty-five isolates recovered at the Public Health Laboratory (LSP) of Asturias from foods and employees during three poisoning outbreaks (outbreaks 1 to 3) that occurred at different restaurants were included. These isolates had been partially characterized in a previous study (31). New isolates (29 isolates) were collected from foodstuffs associated with a fourth outbreak, which took place at a nursing home for the elderly; from a variety of foods in the Security Food Program of the Agency for Environmental Health and Consumption of Asturias; and from healthy carriers attending a training course for food handlers. The new isolates were biochemically confirmed as S. aureus (API Staph system; bioMérieux) and were tested for hemolytic (sheep blood agar; Oxoid, Madrid, Spain), thermonuclease (DNase test agar; Oxoid), and coagulase (Slidex Staph Plus; bioMérieux, Marcy-l'Etoile, France) activities. Production of classical enterotoxins was determined by reverse passive latex agglutination by using the commercial SET-RPLA kit (Oxoid) according to the manufacturer's recommendations. All isolates were also screened for toxic shock syndrome toxin 1 (TSST-1) by using the TST-RPLA kit (Oxoid).

Table 1.

Features of food-related S. aureus isolates analyzed in this work

Source (no. of isolates)a	Strainb	Toxin(s)c	Virulence profile	Resistance profiled	CCe	spa type	MLSTf
O1, FH1 (2)	1	NA	hla hlb hld hlg lukED egc1 splF agrIII	AMP (blaZ::Tn552)	CC30	t021	ST30
O1, FH2	2	NA	hla hld hlgv lukED splF agrII	MUP (mupA)	CC15	t120	ST15
O1, cooked lamb (5)	3	SEC	hla hld hlg lukED sec sel egc1 ear agrI	AMP (blaZ::Tn552)	CC45	t050	ST45
O1, shellfish salad (5)	3	SEC	hla hld hlg lukED sec sel egc1 ear agrI	AMP (blaZ::Tn552)	CC45	t050	ND
O1, paella (4)	3	SEC	hla hld hlg lukED sec sel egc1 ear agrI	AMP (blaZ::Tn552)	CC45	t050	ND
O1, cream cake (2)	3	SEC	hla hld hlg lukED sec sel egc1 ear agrI	AMP (blaZ::Tn552)	CC45	t050	ND
O2, FH3	4	NA	hla hld hlg hlgv eta agrIII	AMP [ERY CLII] MUP (blaZ::Tn552 [ermA ermC])	CC1	t177	ST3
O2, FH4	5	TSST-SEA	hla hld hlg hlgv eta lukED tst sea egc2 splF agrIII	AMP [ERY CLII] MUP (blaZ::Tn552 ermC)	CC30	t012	ND
O2, FH5	6	TSST	lukED tst seh egc2 splF agrIII	AMP [ERY CLII] MUP (blaZ::Tn552 ermC)	CC30	t136	ST34
O2, cream cake	7	NA	hla hlb hld hlg hlgv lukED seh egc2 agrIV	AMP ERY (blaZ [msrA-msrB])	CC5	t616	ST1619
O2, vegetables (2)	8	SEA	hla hlb hld hlgv sea bsaB splF agrIV	AMP [ERY CLII] TET (blaZ::Tn552 ermC tetK)	CC5	t701	ST6
	9	SEC	hla hlb hld hlg hlgv lukED sec egc1 ear agrI	AMP (blaZ)	CC22	t790	ST217
O2, beef	9	SEC	hla hlb hld hlg hlgv lukED sec egc1 ear agrI	AMP (blaZ)	CC22	t790	ST217
O3, FH6	8	SEA	hla hlb hld hlgv sea bsaB splF agrIV	AMP [ERY CLII] TET (blaZ::Tn552 ermC tetK)	CC5	t701	ND
O3, FH7	10	NA	lukED sek seq ear bsaB splF agrIV	AMP (blaZ)	CC8	t008	ST8
O3, FH8	11	SEB	hla hld hlgv etd lukED seb egc1 ear bsaB splF agrIV	AMP (blaZ::Tn552)	CC25	t7125	ST26
O3, FH9	8	SEA	hla hlb hld hlgv sea bsaB splF agrIV	AMP [ERY CLII] TET (blaZ::Tn552 ermC tetK)	CC5	t701	ND
O3, stuffed crab	8	SEA	hla hlb hld hlgv sea bsaB splF agrIV	AMP [ERY CLII] TET (blaZ::Tn552 ermC tetK)	CC5	t701	ND
O3, shellfish salad (2)	12	NA	hla hld hlgv lukED egc1 splF agrII		CC5	t1560	ST146
	13	NA	hla hlb hld hlgv lukED egc1 splF agrII	AMP (blaZ)	CC5	t1560	ND
O3, cream cake (2)	14	NA	hla hlb hld hlgv lukED egc1 splF agrII		CC5	t1560	ND
	13	NA	hla hlb hld hlgv lukED egc1 splF agrII	AMP (blaZ)	CC5	t1560	ND
O4, Russian salad	15	TSST-SEA	hld hlg tst sea egc2 agrIII	AMP (blaZ::Tn552)	CC30	t012	ND
Cake	16	(−)-(−)-SED	hla hlb hld hlg hlgv lukED etd sea seb sed sej sep ser egc2 ear agrII	ERY (msrB)	CC5	t002	ND
Cake	17	(−)-(−)-SED	hla hlb hld hlg hlgv lukED sea seb sed sej sep egc2 agrIV	AMP (blaZ)	CC8	t008	ND
Cake	18	SED	hla hld hlgv lukED sed sej sep egc1 ear splF agrII	ERY (msrA-msrB linA/linA′)	CC5	t002	ST5
Stuffed eggs	19	(−)-SEC	hla hlb hld hlg hlgv lukED sea sec sel egc2 ear agrI	AMP [ERY CLII] (blaZ::Tn552 [ermA-ermC])	CC45	t015	ST45
Sponge cake (5)	20	SEA-SEC	hld hlg lukED sea sec sel egc2 ear agrI	AMP (blaZ::Tn552)	CC45	t015	ND
Sponge cake (5)	21	TSST-SEA	hld hlg lukED tst sea egc2 agrIII	AMP (blaZ::Tn552)	CC30	t840	ST30
Swiss roll (5)	22	(−)-SEB-SEC	hla hlb hld hlg hlgv sea seb sec sek seq egc2 ear agrIV		CC59	t216	ST59
Cheese	23	NA	agrIII		CC30	t012	ND
Hamburger	24	TSST-(−)-SED	hla hlb hld hlg hlgv lukED tst sea sed sej ser egc2 ear agrI		CC45	t015	ND
Hamburger	25	TSST-SEA-SED	hla hlb hld hlg hlgv lukED tst sea sed sej ser egc1 agrIII	AMP (blaZ::Tn552)	CC45	t604	ST546
Hamburger	26	SED	hla hlb hld hlgv lukED sed sej ser egc1 splF agrII	[AMP OXA] [AMK GEN KAN TOB] MUP ([blaZ SCCmec IVd] [aacA aphD-aphA])	CC5	t002	ST5
Hamburger	27	SEA-SED	hla hlb hld hlg hlgv lukED sea sed sej ser egc2 splF agrII	[AMP OXA] CLI [AMK GEN KAN TOB] MUP ([blaZ SCCmec IVd] unk [aadD-aphA])	CC5	t2173	ST5
TC, FH10	28	TSST	hla hlb hld hlg lukED tst egc2 splF agrIII		CC30	t166	ST34
TC, FH11	29	NA	hla hld hlgv lukED ear agrII		CC15	t3474	ST15
TC, FH12	30	NA	hla hlb hld hlgv etb egc2 agrIV		CC121	t272	ST51
TC, FH13	31	TSST	hla hlb hld hlg tst seh egc2 splF agrIII		CC30	t166	ND

^aO1 to O4, outbreaks 1 to 4; FH, food handler; TC, training course. The order in which the strains are listed corresponds to the year of isolation. Thirty-six isolates from foods (26 isolates) and food handlers (10 isolates) were recovered during four poisoning outbreaks. Three of these outbreaks (O1 to O3) occurred in different restaurants during 2002 (31), and the fourth (O4) occurred in a nursing home for the elderly in 2006. Twenty-four isolates were collected from non-outbreak-associated foods during 1997, 2006, and 2007, and four isolates were recovered from the nasal cavities of healthy carriers attending a training course for food handlers in 2008. The number of isolates is given only when there were more than one.

^bThe 64 food-borne isolates were assigned to 31 strains.

^cProduction of “classic” staphylococcal enterotoxins (SEA, SEB, SEC, and SED) and TSST-1 was determined by reverse passive latex agglutination in this or a previous study (31). NA, not applicable; (−), negative.

^dFor the determination of resistance phenotypes and genotypes, the antimicrobial agents tested were ampicillin-penicillin (AMP), oxacillin-methicillin (OXA), gentamicin (GEN), amikacin (AMK), kanamycin (KAN), tobramycin (TOB), tetracycline (TET), erythromycin (ERY), clindamycin (CLI) (CLI^I, inducible), chloramphenicol, ciprofloxacin, moxifloxacin, rifampin, linezolid, vancomycin, tigecycline, mupirocin (MUP), trimethoprim, and trimethoprim-sulfamethoxazole. Antimicrobials in brackets belong to the same class. unk, unknown; all screened genes involved in clindamycin resistance tested negative by PCR amplification.

^eCC, clonal complex according to MLST and supported by spa typing.

^fND, not determined.

Antimicrobial susceptibility testing.

All isolates were tested for antimicrobial susceptibility by the disk diffusion method using Mueller-Hinton agar and commercially available discs (Oxoid). The antimicrobial agents used were ampicillin, penicillin, oxacillin, methicillin, gentamicin, amikacin, kanamycin, tobramycin, tetracycline, erythromycin, clindamycin, chloramphenicol, ciprofloxacin, moxifloxacin, rifampin, linezolid, vancomycin, tigecycline, mupirocin, trimethoprim, and trimethoprim-sulfamethoxazole. MICs for mupirocin (GlaxoSmithKline, United Kingdom) and vancomycin (Laboratorios Normon SA, Madrid, Spain) were determined by the agar dilution method on Mueller-Hinton agar (Oxoid, Madrid, Spain), with concentrations ranging from 0 to 1,500 μg/ml and from 0 to 4 μg/ml, respectively. The results were scored according to the guidelines of the Clinical and Laboratory Standards Institute (12) or were categorized as low (MIC, 8 to 64 mg/liter) or high (MIC, ≥512 μg/ml) resistance in the case of mupirocin (37). S. aureus NCTC 8325 was included as a control. To estimate the rate of inducible lincosamide resistance, the double-disc diffusion test was performed as reported elsewhere (40).

PCR amplification.

Genomic DNA was purified using the phenol-chloroform and proteinase K (50 μg/ml; Roche Diagnostics, Spain) method (39), preceded by a lysis step with lysostaphin (0.02 μg/ml; Sigma, St. Louis, MO). Staphylococcal cassette chromosome mec (SCCmec) typing of MRSA strains was carried out as described by Zhang et al. (46). In addition, all isolates were tested by simplex PCR amplifications (see Tables S1 and S2 in the supplemental material) of genes that confer resistance to ampicillin-penicillin (blaZ), methicillin-oxacillin (mecA and SCCmec), macrolides (msrA, msrB), lincosamides (linA/linA′), macrolides-lincosamides-streptogramin B (ermA, ermB, ermC), tetracyclines [tet(K), tet(L), tet(M), tet(O)], aminoglycosides (aacA plus aphD, aadD, aphA), phenicols (cat::pC194, cat::pC221, cat::pC223, fexA), and mupirocin (mupA). The isolates were also tested by simplex PCR for virulence determinants encoding hemolysins (hla, hlb, hld, hlg, hlg variant), leukotoxins (lukED, lukM, lukPV), exfoliatins (eta, etb, etd), toxic shock syndrome toxin (tst), SEs (sea, seb, sec, sed, see, seg, seh, sei, ser, ses, set), and SEls (selj, selk, sell, selm, seln, selo, selp, selq, selu) and for the type of accessory gene regulatory (agr) system (agrI, agrII, agrIII, agrIV). Markers for the νSaβ genomic islands (splF, bsaB) and SaPIs (ear) were also screened. The sea to see, selj, and seg to sei genes have been tested previously in isolates associated with outbreaks 1 to 3 (31), but all other genes were screened in the present work. The presence or absence of different types of egc clusters, νSaβ genomic islands, SaPIs, prophages, and plasmids was inferred on the basis of positive PCR results for relevant gene combinations (3).

Typing techniques.

All isolates were subjected to staphylococcal protein A (spa) typing, and at least one representative of each spa type was also analyzed by multilocus sequence typing (MLST) (16, 20). Genomic DNA, extracted as indicated above, was used in PCR amplifications for the spa (spa typing), arcC, aroE, gplF, gmk, pta, tpi, and yqiL (MLST) genes. PCR products generated by both techniques were purified with the GFX PCR DNA gel band purification kit (GE Healthcare, Madrid, Spain) and were sequenced at Macrogen Inc. (Seoul, South Korea). The Ridom Spa Server (http://spaserver.ridom.de/) and the MLST website (http://www.mlst.net) were used to assign spa types and sequence types (ST), respectively. Isolates were grouped into a single clonal complex (CC) when at least five of the seven housekeeping genes included in the MLST scheme were identical (16).

Strain assignment and frequency determination.

Isolates with the same spa type, ST, agr type, virulence profile, and resistance profile were considered a single strain. The percentages reported in Results and Discussion were calculated with respect to the total number of strains (31 strains) (see Table 1).

RESULTS

Exotoxin gene profiles and association with mobile genetic elements.

As shown in Table 1, the majority of the strains included in this study contained hemolysin genes (90.3%) (two to five hl genes) and the lukED leukotoxin gene (77.4%); fewer strains carried tst (25.8%), eta, etd (6.5% each), or etb (3.2%); and none tested positive for lukPV or lukM. One or more genes for classical SEs—sea (38.7%), seb (12.9%), sec (16.1%), and sed (22.6%)—were detected in 54.8% of the strains. Other enterotoxin genes, including selj (22.6%), selk-selq (6.5%), seh, sell, selp (9.7% each), and ser (16.1%), were also represented, while see, ses, and set were absent. Most strains (80.7%) harbored egc1 (32.3%) or egc2 (48.4%). In all, the number of enterotoxin genes ranged from zero (four strains) to 12 (one strain), with an average of 6.

Markers for νSaβ (splF, bsaB) and SaPIs (ear) were found in 35.5%, 48.4%, and 9.7% of the strains, respectively (Table 1). Different combinations of lukED, egc genes, splF, and bsaB support the occurrence of νSaβ type I (splF-lukED-egc1) (six strains), type II (bsaB-splF-lukED) (one strain), and type III (splF-egc2) (one strain). Additional combinations, such as splF-bsaB-lukED-egc1 (one strain), lukED-egc2-splF (four strains), lukED-egc1 (three strains), lukED-egc2 (seven strains), splF-lukED (one strain), splF-bsaB (one strain), lukED alone (one strain), and egc2 alone (three strains), suggest deletions in known types of νSaβ or the existence of new variants. On the other hand, detection of ear together with certain se or sel genes is consistent with the presence of SaPI3 (ear-seb-selk-selq) (one strain), SaPI5 (ear-selk-selq) (one strain), or SaPImw2 (ear-sec-sell) (three strains). In other cases, the presence of se or sel genes could be presumptively associated with prophages, which are known to carry sea (12 strains) and selp (3 strains), or with plasmids, where sed or selj with or without ser have been located (7 strains).

Expression of tst and classical se genes.

In all tst-, sec-, and sed-positive strains (eight, five, and seven, respectively), the toxins encoded could be detected immunologically, while 58.3% and 50% of the strains carrying sea or seb, respectively, expressed the corresponding toxin. In four strains, production of two classical enterotoxins (SEA plus SEC, SEA plus SED, and SEB plus SEC) was detected, and TSST-1 was expressed together with SEA, SED, and SEA plus SED.

Antimicrobial drug resistance.

Nine of the 31 strains (29.0%) were susceptible to all compounds tested, and none were resistant to chloramphenicol, ciprofloxacin, moxifloxacin, rifampin, linezolid, vancomycin, tigecycline, trimethoprim, or trimethoprim-sulfamethoxazole. The remaining strains were resistant to three or more classes (here considered multiresistant) (19.4%), two classes (6.5%), or one class (45.2%) of antimicrobials. The frequencies of resistance to individual agents were 61.3% for ampicillin-penicillin, 6.5% for oxacillin-methicillin, 25.8% for erythromycin, 19.4% for clindamycin, 3.2% for tetracycline, 6.5% for amikacin-gentamicin-kanamycin-tobramycin, and 19.4% for mupirocin. The mupirocin MICs were 32 μg/ml (three strains from food handlers associated with outbreak 2), 64 μg/ml (two strains from hamburgers), and 1,250 μg/ml (one strain from a food handler in outbreak 1). As shown in Table 1, all strains resistant to ampicillin-penicillin carried the blaZ gene, which encodes a β-lactamase, and in many of these strains (63.2%), the gene was associated with transposon Tn552, as demonstrated by PCR amplification of a fragment spanning from blaZ to the transposase gene of Tn552. The two oxacillin-methicillin-resistant strains (MRSA), both from hamburgers, carried SCCmec IVd. The genes implicated in erythromycin resistance together with inducible resistance to clindamycin were ermC (three strains) or ermA and ermC (two strains), while resistance to erythromycin only was associated with the presence of either msrB or msrA and msrB (one and two strains, respectively). The aphA gene was found in the two aminoglycoside-resistant strains together with aacA plus aphD or aadD, and the plasmid-associated tetK gene was responsible for tetracycline resistance (one strain). The strain with high-level resistance to mupirocin (MIC, 1,250 μg/ml) carried mupA, while the low-level-resistant strains lacked this gene.

Clones and clonal complexes.

The strains were assigned to 21 spa types and 15 STs and were distributed among 10 CCs (Tables 1 and 2). The most frequent CCs were CC5 (29.0%), CC30 (25.8%), and CC45 (16.1%). CC8, CC15 (two strains each), CC1, CC22, CC25, CC59, and CC121 (one strain each) were also represented, but the emerging CC398 was not detected. The four groups of agr systems—agrI (16.1%), agrII (29.0%), agrIII (32.3%), and agrIV (22.6%)—were found in the food-related strains (Table 1), and a strong correlation between different CCs and certain agr groups was observed: CC1-agrIII, CC5-agrII, CC15-agrII, CC22-agrI, CC30-agrIII, CC45-agrI, CC45-agrIII, and CC121-agrIV (38, 45). The agrIV group, previously reported in CC5 and CC25 in Asturias (2, 4), was found here in CC5, CC8, CC25, and CC59 isolates. The two MRSA strains from hamburgers, acquired at a local supermarket, had different spa types (t002 and t2173), but both belonged to ST5 (CC5) and had the agrII system.

Table 2.

Clonal complexes, sequence types, and spa types found in food-related S. aureus strains

CC (no. [%] of strains)	ST (no. of strainsa)b	spa type (no. of strainsa)	Origin (no. of strainsa)c
CC5 (9 [29.0])	ST5 (3), ST6, ST146, ST1619	t002 (3), t616, t701 (1), t1560 (3), t2173	O2-F (2), O3-F/FH (3), F (4)
CC30 (8 [25.8])	ST30 (2), ST34 (2)	t012 (3), t021, t136, t166 (2), t840	O1-FH, O2-FH (2), O4-F, F (4)
CC45 (5 [16.1])	ST45 (2), ST546	t015 (3), t050, t604	O1-F, F (4)
CC8 (2 [6.5])	ST8	t008 (2)	O3-FH, F
CC15 (2 [6.5])	ST15 (2)	t120, t3474	O1-FH, TC-FH
CC1 (1 [3.2])	ST3	t177	O2-FH
CC22 (1 [3.2])	ST217	t790	O2-F
CC25 (1 [3.2])	ST26	t7125	O3-FH
CC59 (1 [3.2])	ST59	t216	F
CC121 (1 [3.2])	ST51	t272	TC-FH

^aGiven only when there was more than one strain.

^bNot all strains were tested by MLST.

^cO1 to O4, food poisoning outbreaks 1 to 4; F, foods; FH, food handlers; TC, training course.

DISCUSSION

The presence of staphylococcal species in foods is well documented, but information on the S. aureus lineages circulating through the food chain is still limited. In the present study, 64 food-related isolates were assigned to 31 strains by molecular typing techniques, in conjunction with virulence and resistance properties. These strains belonged to 10 CCs, and all except 1 (CC22) were also represented in isolates recovered from hospitals and healthy carriers in Asturias (2, 4; M. A. Argudín, M. C. Mendoza, and M. R. Rodicio, unpublished data). In addition, the most frequent CCs among food-related isolates (CC5, CC30, and CC45) were also the most frequent among isolates from healthy carriers of the region, and the three included potential outbreak strains. Strain 3, which was collected from different foods served at the outbreak 1 restaurant, belonged to spa type t050 and ST45 (CC45), carried the sec gene, and expressed the toxin. This strain was not found in two food handlers working at the same restaurant, although both were S. aureus carriers. Strain 8 (t701, ST6; CC5; positive for SEA; resistant to several antimicrobials) was recovered from stuffed crab and from two food handlers associated with outbreak 3; it was the probable cause of the outbreak, as already proposed by Martín et al. (31). Strain 8 was also present in vegetables from the outbreak 2 restaurant, but a total of six strains (strains 4 to 9), belonging to four CCs, were recovered from foods and food handlers at this restaurant. In addition, all but one of these strains carried at least one enterotoxin gene, and most produced SEA or SEC. With regard to outbreak 4, a single isolate (strain 15) is available. It was recovered from Russian salad served at a nursing home for the elderly, was positive for sea and SEA, and belonged to spa type t012 (ST30; CC30).

SEA, followed by SED, is the enterotoxin most frequently associated with SFP, although outbreaks caused by SEB, SEC, SEE, and SEH have also been reported (3, 29). In the food-related strains characterized in this study, the sea, seb, sec, sed, and seh genes, but not see, were detected, and SEA, SEB, SEC, and SED, for which commercial detection systems are available, were expressed in most strains carrying the genes. The relationship between other SEs or SEls and SFP is poorly understood, partly due to the unavailability of commercial kits for detection. In the present study, genes belonging to the egc cluster were very common, appearing in more than 80% of the strains, either alone or in combination with other enterotoxin genes. A high incidence of egc genes in food-borne S. aureus strains has also been reported by other authors (7, 9, 22, 28, 35). Accordingly, the pathogenic role of the egc enterotoxins may be underestimated.

As in S. aureus strains from other sources, the frequency of resistance to ampicillin-penicillin was high (61.3%) in the food-related isolates tested in this study, and the responsible gene, blaZ, was often associated with Tn552 (23). The frequency of resistance to other antimicrobial agents, including oxacillin (methicillin) (6.5%), erythromycin (25.8%), clindamycin (19.4%), and gentamicin-tobramycin (6.5%), was lower than that reported in Spanish hospitals (2, 4, 14). This is in agreement with data from Kérouanton et al. (25), showing that strains from SFP outbreaks were more susceptible to antimicrobials than human clinical isolates. However, the frequency of resistance to mupirocin (19.4%) was higher than that reported for nosocomial S. aureus strains in Spain (8.9%) (14), although only strain 2, from a food handler, harbored the mupA gene, which confers high-level resistance. Mupirocin is the antibiotic of choice for the eradication of S. aureus carriage, and resistance has been implicated in failure of the treatment (37). The two MRSA strains detected (strains 26 and 27) belonged to CC5 and carried SCCmec IVd. This subtype of SCCmec is uncommon in Spain and was associated only with CC8 (44). Both strains were recovered from hamburgers acquired at a local supermarket and produced SED (strain 26) or SEA plus SED (strain 27). The presence of these strains in hamburgers is noteworthy, because MRSA, a highly relevant nosocomial and community-acquired pathogen (11), has also been implicated in SFP outbreaks (24, 25, 26).

In conclusion, the molecular epidemiology study conducted shows the presence of different S. aureus strains containing important resistance and virulence determinants in foods and food handlers. These strains encode not only classical and novel enterotoxin genes but also major virulence factors, such as exfoliative toxins and TSST-1. The presence of these isolates in the food chain represents a potential health hazard for consumers and deserves further attention.