Clonal Distribution and Differential Occurrence of the Enterotoxin Gene Cluster, egc, in Carriage- versus Bacteremia-Associated Isolates of Staphylococcus aureus

Abstract

The Staphylococcus aureus enterotoxin gene cluster, egc, was detected in isolates from healthy individuals and in those from patients with bacteremia. The egc genes cooccur and are slightly enriched in strains from healthy carriers (present in 63.7% of carriage-associated isolates versus 52.9% of invasive isolates; P = 0.03). Multilocus sequence typing revealed that successful staphylococcal clones usually harbor the egc locus.

Staphylococcal enterotoxins are responsible for food-associated outbreaks of diarrhea among humans. The toxins generate visible pathological lesions in the stomach and the upper part of the small intestine (5). Many of the enterotoxins display superantigen characteristics and are obvious targets for antistaphylococcal therapies. Genes encoding several of the enterotoxins are physically clustered in the Staphylococcus aureus genome (9, 16). The locus encoding the enterotoxins SEG, SEI, SEM, SEN, and SEO is currently known as egc (enterotoxin gene cluster) (9). Although this cluster is highly prevalent among S. aureus strains in general, antibodies are rarely raised against egc enterotoxins, which is a unique feature of this group of enterotoxins (7). It is intriguing that the prevalence of egc genes in isolates of S. aureus is negatively correlated with the severity of infection (6). For SEA, the situation is precisely opposite: the toxin gene is significantly more often present in the invasive isolates (6, 14). The conclusion could be that one or more of the egc-encoded enterotoxins provide protection against severe sepsis. Additional research into this phenomenon is clearly warranted.

We here investigate whether strains isolated during Dutch S. aureus screening studies of carriage per se and S. aureus strains causing bacteremia (11, 19) can be differentiated on the basis of the absence or presence of the egc locus.

Three hundred ninety-one strains of S. aureus were included in the present study (Table 1). Most of the carriage isolates (n = 118) derived from the study by Wertheim et al. (19), in which patients were sampled at the time of hospitalization (MUP isolates). Another set of carriage strains (n = 86) derived from a group of elderly persons (over 50 years of age) in the open community (ERGO isolates) (11). As case strains, bacteremia-associated S. aureus strains (n = 180) from both these populations were included (11, 19). In addition, the group of invasive MUP isolates included strains isolated from pus (n = 7), and the nasal S. aureus carriage state for the bacteremic patients was also determined (12).

TABLE 1.

Presence of egc locus in carriage versus invasive isolates (n = 391)

Isolate	No. (%) of isolates
	Negative for SEM	Positive for SEM	Total
Carriage isolates
MUP	45 (38.1)	73 (61.9)	118 (100)
ERGO	29 (33.7)	57 (66.3)	86 (100)
Total	74 (36.3)	130 (63.7)a	204 (100)
Invasive isolates
MUP	47 (50.0)c	47 (50.0)d	94 (100)
ERGO	41 (44.1)	52 (55.9)	93 (100)e
Total	88 (47.1)	99 (52.9)b	187 (100)

^aBy Fisher's exact test (two-sided) done in comparison with total SEM⁺ value for invasive isolates; P = 0.0316.

^bBy Fisher's exact test (two-sided) done in comparison with total SEM⁺ value for carriage isolates; P = 0.0316.

^cForty-three blood culture isolates and 4 pus isolates.

^dForty-four blood culture isolates and 3 pus isolates.

^eAll 93 blood culture isolates.

Strains were grown on blood agar plates (Becton-Dickinson, Le Pont de Claix, France), and DNA was extracted using a bacterial DNA kit III and a Magnapure system (Roche Molecular Systems, Lelystad, The Netherlands). The SEM gene was amplified from 10 ng DNA using SEM-specific primers (40 cycles of 1 min at 94°C, 2 min at 55°C, and 3 min at 72°C) (6). PCR mixtures contained Supertaq DNA polymerase (Sphaero Q, Leiden, The Netherlands). PCR products were analyzed by agarose gel electrophoresis. The PCR products were of the expected lengths. PCRs specific for SEG, SEO, SEI, and SEN were performed for the MUP strains (n = 212) (6), as was multilocus sequence typing (11, 17, 18). Statistical analyses involved Fisher's exact test (two-sided), with a P value of <0.05 considered significant.

Table 2 shows that a positive score in the SEM PCR is strongly associated with positive scores for the SEI, SEO, SEN, and SEG enterotoxin genes. Overall, when positive in the SEM PCR, 96.7% (116/120) of these strains were positive for all of the toxins associated with egc. When negative in the SEM PCR, 85/92 (92.4%) of these strains were negative for all of the other toxin genes. When carriage strains were compared to all invasive isolates (blood and pus derived), it appeared that the carriage isolates harbored the egc locus significantly more often (130/204 versus 99/187; P = 0.0316) (Table 1). Although a clear biological rationale is currently lacking, this observation suggests that the presence of egc is associated with noninvasiveness and a lower disease-invoking potential, as suggested earlier (6). When the pus isolates were excluded from the analysis and the bacteremia-associated strains were compared separately to the carriage strains, the significance was maintained. However, the possible protective effect of egc does not involve changes in mortality once bacteremia has developed (49.4% [39/79] egc positive in bacteremia with no S. aureus-related mortality versus 53.3% [8/15] with S. aureus-related mortality; P = 1.0). Again, a biological explanation is lacking, and it has to be emphasized that numbers of cases are low.

TABLE 2.

Association of PCR results for SEM with those for other enterotoxins of egc (validated by 212 isolates)

Enterotoxin	No. (%) of isolates
	Positive	Negative
SEM	120	92
SEI	120 (100)	92 (100)
SEO	119 (99.2)	86 (93.5)
SEN	119 (99.2)	90 (97.8)
SEG	117 (97.5)	92 (100)
GIMNO	116 (96.7)	85 (92.4)

Clonality is an important feature of the enterotoxin gene content of strains. Some clones seem to be missing the egc element (clonal complexes [CCs] 1, 7, 8, and 15 [Table 3 ]). Strains from CCs 5, 22, 25, 30, and 45 are positive in at least 92.5% of cases. Apparently, the egc element is associated with specific staphylococcal lineages, some of which are successful in the Rotterdam region (11).

TABLE 3.

Presence of the egc locus in different genetic clusters of S. aureus (n = 212)

CC or characterization by multilocus sequence typing	No. (%) of isolates
	Negative for SEM	Positive for SEM	Total
1	8 (88.9)	1 (11.1)	9 (100)
5	0 (0.0)	12 (100)	12 (100)
7	12 (92.3)	1 (7.7)	13 (100)
8	19 (100)	0 (0.0)	19 (100)
9	1 (33.3)	2 (66.7)	3 (100)
12	1 (100)	0 (0.0)	1 (100)
15	32 (100)	0 (0.0)	32 (100)
22	0 (0.0)	2 (100)	2 (100)
25	0 (0.0)	10 (100)	10 (100)
30	4 (7.5)	49 (92.5)	52 (100)
45	1 (2.7)	36 (97.3)	37 (100)
Singleton	12 (63.2)	7 (36.8)	19 (100)
New sequence type	2 (100)	0 (0.0)	2 (100)
Total	92 (43.4)	120 (56.6)	212 (100)

It is interesting that the egc enterotoxins may be associated with staphylococcal toxic shock syndrome and scarlet fever (8) and that some of these toxins are enriched in genital isolates (1). Furthermore, Becker et al. showed that seg and sei genes were found in strict combination in 53.0% of the invasive and 57.1% of the colonizing strains of S. aureus, whereas the sed and sej genes were found significantly more often in blood isolates (P = 0.037) (2). Ferry et al. showed that the presence of the egc locus was lower in S. aureus strains isolated from patients with septic shock than in those isolated from septic patients without shock, in strains isolated from patients with suppurative disease, and in carriage strains (6). The levels of antibodies against SEG and SEI are elevated in women; this may reflect mucosal adaptation of strains and a higher contact frequency for women (13). However, some of these studies were performed with limited numbers of strains, and sometimes controversial data are published. The association of SEG, SEH, SEI, SEO, and SEM with food poisoning, for instance, is still ill-defined (3, 4).

We here describe the presence of egc in a large collection of carriage and invasive isolates of S. aureus deriving from the same geographical region. egc is (i) slightly enriched among carriage strains, (ii) not associated with mortality in patients suffering from staphylococcal bacteremia, and (iii) coupled to certain clonal lineages. There are surprising gaps in the capacities of different human sera to neutralize these superantigens (7). Furthermore, egc toxins seem to be produced in lesser quantities and have a lower immunogenic response than other enterotoxins (6), which may also explain why strains producing these egc enterotoxins are tolerated in the nose. The function of the enterotoxins and their immune tolerance need to be studied in detail. Also, the precise enterotoxin gene content of the egc locus is still unknown. For instance, the SEU toxin gene was detected only recently (10). The host range of staphylococci with different toxin gene repertoires needs additional investigation as well (15). We conclude that the egc locus may enhance the carriage potential of an S. aureus strain, which might be explained by the fact that most of the successful clones uniformly contain the egc locus.