INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Staphylococcal enterotoxins (SEs) are members of a subgroup of related protein exotoxins in the pyrogenic toxin (PT) family exhibiting a diverse number of shared (e.g., superantigenicity) and unique biological activities (7, 18, 42). SEs are distinguishable from other members of the PT family by their ability to induce emesis and diarrhea when ingested (17). Thus, they are causative agents in staphylococcal food poisoning (SFP), one of the main food-borne diseases having major economic and public health impacts (5, 6, 62). For Staphylococcus aureus, the production of SEs, such as SEA, SEB, SEC, SED, and SEE, is well documented (7, 8, 30, 40).

Unlike the situation for S. aureus, only limited and conflicting data are available regarding enterotoxin production in non-S. aureus staphylococcal species (10, 23, 25, 46, 52, 57, 63, 65). One of these species, the coagulase-positive S. intermedius, is the predominant staphylococcal species isolated from healthy and diseased dogs and has been isolated from a number of other carnivores, horses, and birds (16, 36, 54). In single cases, this species appears to be also responsible, as a zoonotic pathogen, for canine-inflicted human wound infections and invasive infections in immunocompromised patients (43, 45, 59, 64). Occasionally, the possession of enterotoxins has been reported for S. intermedius, and its involvement in an outbreak of SFP has been considered previously (4, 24, 37, 41). Previous reports regarding S. intermedius enterotoxins were restricted particularly to immunological procedures, which are generally limited in sensitivity and known for nonspecific results (14, 38, 53). DNA hybridization- and amplification-based approaches offer alternatives to detect staphylococcal isolates bearing the genetic sequences to produce enterotoxins irrespective of their toxin expression (56). Hence, there is an obvious need for a systematic verification of the enterotoxigenic potential of S. intermedius on a genomic level. For this purpose, a large panel of veterinary S. intermedius isolates and a control group of different animal-derived isolates of coagulase-negative staphylococci (CNS) were tested for the presence of SE-encoding genes by a multiplex PCR DNA-enzyme immunoassay (DNA-EIA).

(This study was presented in part at the 11th European Congress of Clinical Microbiology and Infectious Diseases, Istanbul, Turkey, 2001.)

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains. From September 1999 to August 2000, 1,669 consecutive animal specimens (canine, n = 590; equine, n = 896; feline, n = 150; other, n = 33) from the skin or mucosal surfaces of animals with skin or mucocutaneous infections were investigated (Table 1). The specimens were submitted by veterinary care centers in northwestern Germany to the Department of Microbiology and Infectious Diseases, School of Veterinary Medicine, Hannover, Germany. No duplicate isolates were included. In addition, 11 human S. intermedius isolates previously collected were studied (49).

Identification. Identification was based on growth characteristics on Columbia agar with 5% (vol/vol) sheep blood at 37°C and positive catalase production. The staphylococcal isolates were identified biochemically by the automated ID 32 Staph system (bioMérieux, Marcy l'Etoile, France). The presence of bound or free coagulase was confirmed by latex slide agglutination for clumping factor, protein A, and capsular polysaccharides (Pastorex Staph-Plus; Sanofi Diagnostics Pasteur, Marnes-la-Coquette, France). In addition, an in-house slide coagulase test using rabbit and human citrate plasma and a coagulase tube test using rabbit plasma (bioMérieux) were performed. Concerning the S. intermedius isolates, a possible misidentification regarding S. aureus was excluded by PCR for detection of S. aureus-specific nuc and coa genes; both PCR procedures were done with primer pairs as described previously (22, 33).

Media and chemicals. Brain heart infusion broth from Merck (Darmstadt, Germany) was used for cultivation of staphylococcal strains. Chemicals were purchased from Amersham Buchler (Braunschweig, Germany) or from Serva (Heidelberg, Germany). All PCR reagents and kits were purchased from Boehringer (Mannheim, Germany). The primers and 5'-biotinylated probes were prepared by MWG Biotech (Ebersberg, Germany).

DNA isolation procedures. Staphylococcal cells were collected from 0.5 ml of an overnight brain heart infusion broth culture. Cells were pelleted by centrifugation at 5,000 × g for 20 min, resuspended in 185 µl of TE buffer (20 mM Tris chloride, 2 mM EDTA [pH 8.0]) with 15 µl of recombinant lysostaphin (15 mg/ml) (Sigma, St. Louis, Mo.), and incubated at 37°C for 30 min. DNA was subsequently extracted with a QIAamp tissue kit (Qiagen, Hilden, Germany) according to the manufacturer's recommendations. Nucleic acid samples were eluted with distilled water and adjusted to a final concentration of 1 µg/ml according to A₂₆₀ values.

Multiplex PCR DNA-EIA. The multiplex PCR DNA-EIA to detect simultaneously the SE-encoding genes (sea, seb, sec_1-3, sed, and see) was performed as previously described (9). For carryover prevention, the uracil DNA glycosylase system (Boehringer) was used. The amplification was performed in an automated thermocycler with a hot bonnet (Hybaid, Teddington, United Kingdom). For hybridization of amplified DNA from the multiplex PCR, a generic DNA-EIA (GEN-ETI-K DEIA; Sorin, Saluggia, Italy) was used according to the manufacturer's recommendations. The strips were incubated with an optimized concentration of the biotinylated probes (20 pg/µl for each probe). Reactions were judged as positive if the signal reached or exceeded the cutoff value of 0.150 absorbance units above the mean value of determinations of toxin-negative reference strains. Positive controls included DNA from toxin-producing S. aureus strains (SEA, KN543; SEB, ATCC 14458; SEC, ATCC 19095; SED, ATCC 23235; and SEE, ATCC 27664). In addition to standard PCR controls for contamination events, S. aureus Cowan 1 and S. epidermidis DSM 20044 served as negative controls.

Sequencing of the sec fragment. A 271-bp sec PCR amplification fragment of the S. intermedius strain K1144 was purified by gel filtration (Centri-Sep columns; Princeton Separations, Adelphia, N.J.) followed by direct sequence determination. After thermal cycling with fluorescent dye-labeled terminators using the primers SEC-3 and SEC-4 (9), the nucleotide sequence was determined on an automated DNA sequencer (ABI PRISM 310; Applied Biosystems Division, Perkin-Elmer Corp., Foster City, Calif.). The sequences were analyzed by alignment with known sequences of sec genes of S. aureus (20, 21, 26) and S. intermedius (32) by using OMIGA 2.0 software (Oxford Molecular, Oxford, United Kingdom).

Detection of toxin production. Culture filtrates of the S. intermedius isolates were tested for the presence of SEA to SEE by an EIA (Ridascreen Set A, B, C, D, E; R-Biopharm, Darmstadt, Germany) and, except for SEE, by semiquantitative reversed passive latex agglutination (RPLA; Unipath, Basingstoke, Hampshire, United Kingdom) assay. For EIA, a cutoff value of 0.150 absorbance units above the mean value of determinations of negative controls was defined as a positive reaction. The RPLA assay was interpreted visually as recommended by the manufacturer.

Statistical analysis. The chi-square test was used to analyze differences between frequencies of isolation of S. intermedius from canine specimens and those for other veterinary specimens. Fisher's exact test (one tailed) was used to assess differences regarding the presence of enterotoxin genes between canine S. intermedius isolates and other animal isolates.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Testing of S. aureus reference strains. The multiplex PCR with primer pairs specific for sea, seb, sec, sed, and see genes successfully amplified fragments from the DNAs of S. aureus reference strains (Fig. 1). PCR amplification was confirmed by respective hybridization reactions in a DNA-EIA system for all reference strains used. Furthermore, 100% correlation between the results of the enterotoxin analysis and the detection of the corresponding gene was observed, as shown for SEC in Table 3. The sensitivity of the multiplex PCR-EIAs reached the expected range as previously described (9). None of the primer pairs and the respective hybridization probes reacted with any of the negative control strains which also tested negative in both immunoassays.

View larger version (21K): [in this window] [in a new window]   FIG. 1.   Agarose gel electrophoresis pattern showing PCR-amplified products in multiplex PCR for the SE genes for S. aureus control strains and animal-derived sec-positive S. intermedius isolates. Lanes M1 and M2, DNA molecular size markers (1-kb/100-bp DNA ladder); lane 1, positive control with multiplex PCR of all enterotoxin genes tested (sea, seb, sec, sed, and see); lane 2, SEC-positive S. aureus ATCC 19095 with amplification product of sec gene; lane 3, negative control strain S. aureus Cowan 1; lane 4, S. epidermidis negative control strain DSM 20044; lanes 5 to 11, representatives of sec-PCR-positive S. intermedius isolates. Sizes are marked in base pairs on the left.

Isolation of S. intermedius. The frequency of isolation of S. intermedius from various specimens of different animals is shown in Table 1. Overall, 281 (16.8%) S. intermedius isolates were collected over 1 year from 1,669 consecutive animal specimens; they consisted of 247 (41.9%) S. intermedius isolates from 590 canine specimens and 34 (3.2%) isolates from 1,079 specimens of other animals (P < 0.0001). Most of the canine S. intermedius isolates (n = 189) were derived from skin specimens.

Detection of enterotoxin genes in S. intermedius and CNS. A total of 292 S. intermedius isolates were tested for the possession of SE-encoding genes. Thirty-three (11.3%) of these isolates were found to be sec positive by PCR amplification (Tables 2 and 3). Figure 1 shows a representative agarose gel electrophoresis pattern of sec-positive S. intermedius isolates. All positive results were confirmed by the sec-specific DNA-EIA hybridization procedure (Table 3). Except for one equine isolate, all animal S. intermedius isolates with positive sec gene amplification were derived from canine specimens (31 of 247 [12.6%]) (P = 0.07). In addition, 1 of 11 human S. intermedius isolates was sec positive (Table 2). None of the S. intermedius isolates tested showed positive amplification results for the genes encoding enterotoxins SEA, SEB, SED, and SEE.

Sequencing results. Representative of the sec-positive amplification results of canine isolates, a specific 271-bp amplification fragment of S. intermedius strain K1144 was sequenced directly in both directions (Fig. 2). Subsequently, the nucleotide sequence of this fragment was aligned with published nucleotide sequences of different SEC subtypes. For the sequenced fragment, 100% identity to the sequence of the canine type of enterotoxin C (SEC_canine) was found (32). Sequencing of the sec-positive isolates N940276, derived from human bronchoalveolar lavage fluid, and K822, derived from an equine skin specimen, likewise showed identity with the sec_canine nucleotide sequence (Fig. 2).

View larger version (26K): [in this window] [in a new window]   FIG. 2.   Alignment of sequences coding for SEC including reference sequences of S. aureus subtypes, S. intermedius canine subtype, and clinical sec-positive isolates derived from canine, equine, and human specimens. The EMBL accession numbers of the sec nucleotide sequences used for the alignment are as follows: sec1, X05815; sec3, X51661; seccanine, U91526. For sec2, a sequence was used as reported by Bohach and Schlievert (21). The sequences of the three S. aureus subtypes were merged, since there were no differences in the aligned region as showed here. Dots indicate identity.

Enterotoxin production analysis. Overall, there was a complete agreement between enterotoxin gene detection and enterotoxin production for 289 (99.0%) of the 292 S. intermedius isolates. Except for three S. intermedius isolates, the amplification of sec-specific sequences was confirmed for 30 (90.9%) isolates on a protein level by semiquantitative detection of corresponding toxin production using two different immunological tests (Tables 2 and 3). Two of three isolates with noncorresponding results (equine isolate K822 and canine isolate E865) were negative in both EIA and RPLA, whereas the third isolate (canine isolate E826) showed a weak SEC production value near the cutoff in the EIA but was negative in RPLA (Table 3).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Whereas for S. aureus the possession of enterotoxins is known to be a cause of food poisoning and septic shock-like illness (7, 19, 47), only scant and contradictory information is available concerning the production of enterotoxins by S. intermedius and other non-S. aureus staphylococci and their pathogenic role. Since some studies showed evidence for the production of exotoxins such as -toxin and enterotoxins A and C by S. intermedius (4, 24, 31), an enterotoxigenic potential of this species may be assumed. Furthermore, a unique molecular variant of type C SEs was described elsewhere as SEC_canine for pyoderma isolates (32). Also, Hirooka et al. described enterotoxigenic canine S. intermedius isolates (37). In addition, an outbreak of food poisoning with a butter blend due to S. intermedius has been reported previously (41). However, valid molecular epidemiology data concerning S. intermedius enterotoxin genes are still lacking. Thus, the aim of the study reported here was to investigate systematically the enterotoxigenic potential of S. intermedius by using a multiplex PCR-DNA-EIA compared with the results of two different immunoassays.

In this study, S. intermedius was isolated from more than 50% of canine cutaneous specimens (41.9% of all specimens) tested. This is in agreement with previous findings of carrier rates of 40 to nearly 100% for specimens of different cutaneous and mucocutaneous origin of healthy and diseased dogs (2, 3, 11, 27). The isolation of S. intermedius in other animal hosts compared to dogs was significantly lower (3.2%), which is in agreement with the results of previous observations (29, 35, 48).

Positive PCR amplification and DNA-EIA hybridization results were detected in 31 (12.6%) of 247 canine S. intermedius isolates, all possessing sec gene sequences, whereas no sequences of other enterotoxin genes studied were found. Regarding S. intermedius isolates of other animals tested, only one sec-positive isolate was identified from an equine specimen. Canine isolates tend to have a higher prevalence of the sec gene but not to a statistically significant extent. Thus, in contrast to equine, feline, and other animal hosts, a substantial portion of the canine population tested harbors enterotoxigenic S. intermedius. The quite frequent detection of the sec gene in S. intermedius isolates studied here (overall, 11.3%) was comparable with the SEC prevalence in S. aureus of about 5 to 15% (1, 44, 50).

Because the sec primers and the sec-specific probe of the multiplex PCR-EIA anneal in homologous regions of the different S. aureus sec gene variants (sec₁, sec₂, and sec₃) as well as of the S. intermedius sec_canine gene (except for the 1-bp difference in SEC-5B probe), the assay used here covered the sec gene variants of both species. Concerning slight differences in the nucleotide sequences of the sec subtypes, sequencing was performed to subtype the amplified sec gene fragments. All sec amplicons sequenced were represented by sec_canine type (32). This applied not only to the sec-positive canine isolates but also to the equine and human isolates.

A comparison of the molecular detection of the enterotoxin genes was performed between the results obtained from two immunological assays for analysis of enterotoxin production in vitro. Overall, there was a complete agreement for 289 (99.0%) of the 292 S. intermedius isolates regarding all enterotoxins tested. Concordant positive results for SEC production and the respective toxin gene amplification-hybridization were obtained from 30 (90.9%) of 33 sec-positive isolates, thus resulting in three sec gene-possessing S. intermedius isolates without, or with only weak, enterotoxin production in vitro. Regarding S. aureus, it is known that a number of enterotoxigenic strains with positive reactions in biological assays (monkey-feeding assay) tested negative in immunological assays as described previously for the optimum sensitivity plate, a modified Ouchterlony method (13, 28, 58). With the advent of the EIA methods, it was shown that several of these strains produced only very small amounts of enterotoxin (10 to 15 ng) (58). Thus, negative or very weak results of immunoassays might be due to expression that is nonexistent or too weak to be detected even by the more sensitive modern immunological assays, or they could be explained by variations in antigenic composition of the toxin protein.

Because to date the amount of enterotoxin required to cause illness in humans is not known (12, 13), the molecular-based detection of staphylococci that produce only minimal enterotoxin is of particular importance. In principle, differences between toxin gene expression in food products or artificial culture and that in the natural habitat have to be considered (40). In addition, enterotoxin production in food is influenced by many factors that do not influence DNA-based methods (34).

The positive result for SEA production but negative result by sea-specific PCR in one S. intermedius isolate might be due to nonspecific results, since immunological methods can be influenced by various factors (14, 38, 53). In addition, cross-reactions with newly described SEs (15, 39, 51, 55, 66) or still-unknown enterotoxins sharing a close antigenic relatedness could cause positive results in immunological assays but remain unrecognized in molecular tests. Recently, Jarraud et al. (39) reported on an enterotoxin gene nursery (egc) of S. aureus which was probably generated from an ancestral gene through gene duplication and variation, suggesting further, still-unknown enterotoxins in addition to previously identified members of the expanded PT family.

In general, the numbers and types of microorganisms present in a finished food product are influenced strongly by, among other things, infected or colonized persons practicing poor personal hygiene (40). Although S. intermedius is adapted primarily to carnivore hosts, its entry into human foods is not precluded. Both invasive and noninvasive zoonotic transmissions from dogs to humans have been documented elsewhere (43, 59-61). As supported by the sec_canine-positive human S. intermedius isolate reported here, enterotoxigenic S. intermedius strains may contaminate food products during handling and processing via animal-human contacts. Once enterotoxigenic S. intermedius strains are in foods, their growth may be expected to lead to the production of enterotoxins. Thus, such strains may play a role as a causative agent of SFP, as had already been demonstrated by a food-related outbreak in the western United States (41). Given the completely negative results of enterotoxin testing of the CNS control group, the occurrence of SEs in animal CNS isolates may be at least very rare.

In summary, this is the first study in which a large number of S. intermedius isolates was tested systematically for staphylococcal toxin genes. It has been demonstrated on both molecular and immunological levels that the coagulase-positive species S. intermedius possesses a substantial enterotoxigenic potential, particularly for SEC. Hence, in addition to the role of S. aureus in the etiology of SFP, we may assume a causative role particularly for canine S. intermedius isolates via canine-human contact. Therefore, S. intermedius should not be disregarded when screening food products for contamination with enterotoxin-producing pathogens or studying outbreaks of SFP. Furthermore, the application of a PCR-based method offers an alternative method for the rapid and specific detection of enterotoxigenic S. intermedius isolates, even for isolates with no or only very weak production of enterotoxins in vitro.