Occurrence of subtilase cytotoxin and relation with other virulence factors in verocytotoxigenic Escherichia coli isolated from food and cattle in Argentina.

We investigated the presence of the gene of subtilase cytotoxin (SubAB), described in certain highly virulent verocytotoxigenic E. coli strains, in isolates from Argentina and its relation with other virulence factors. The gene subA was present in eae-negative strains mostly associated with saa, vt2 and ehxA genes.

Verocytotoxigenic Escherichia coli (VTEC) are a diverse group of E. coli strains characterized by the production of verotoxins (VT1 and/or VT2) which are regarded as their main virulence factors (6). VTEC are an important cause of gastrointestinal disease in humans (7, 12, 17) and life-threatening complications such as haemolytic uraemic syndrome (HUS).

More recently, it has been reported that some VTEC strains also produce another toxin called subtilase cytotoxin (SubAB). It was identified by Paton et al. (19) from an E. coli O113:H21 strain, which was responsible for an outbreak of HUS in South Australia in 1998, and since then has been detected in several other VTEC serotypes (5, 8, 14, 19, 20). Furthermore, Tozzoli et al. (25) found the first evidence that SubAB can also be produced by vt-negative E. coli isolated from cases of childhood diarrhoea. SubAB is encoded in the megaplasmid, and is the prototype of a new family of AB5 toxins comprising a single 35 kD A subunit which is a subtilase-like serine protease and a pentamer of B subunits, which mediates binding to glycolipid receptors on the target cell surface (19, 24).

SubAB was shown to be cytotoxic to Vero cells and lethal for mice, causing extensive microvascular thrombosis as well as necrosis in the brain, kidney, and liver (11, 19). The extreme cytotoxicity of this toxin for eukaryotic cells is due to a specific single-site cleavage of the essential endoplasmic reticulum chaperone BiP/GRP78 which is a master regulator of endoplasmic reticulum function (21). Its cleavage by subtilase cytotoxin represents a previously unknown trigger for cell death.

The cytotoxin SubAB has been described in certain highly virulent VTEC strains which are negative for the locus of enterocyte effacement (LEE), but the global distribution of SubAB-encoding VTEC strains is unknown. Furthermore, non-O157 VTEC including LEE negative strains predominate in Argentina, where HUS prevalence is the highest in the world.

Consequently, our aim was to investigate the presence of the subA gene in strains isolated from various sources in Argentina. The relation with other VTEC virulence factors and with the serotype was also evaluated.

A total of 95 strains were selected from a well characterized, previously described strain collection of our Laboratory. To assess potential associations among subA, other virulence factors and serotypes, we selected representative strains taking into account the presence of vt1, vt2, eae, saa, and ehxA genes determined in previous studies (15, 16, 22). In the present work, a previously described multiplex PCR, which detects subA, vt1 and vt2, was used for subA screening (20). Samples were obtained by boiling a dilution 1:25 of the bacterial culture for 10 min. Amplification products were visualized in 2% agarose gels, stained with ethidium bromide.

We found that 21 VTEC strains of diverse origins were positive for subA gene. They belonged to serotypes O2:H5, O20:H19, O39:H49, O79:H19, O88:H21, O113:H21, O141:H7, O141:H8, O178:H19. In O non-typable strains, subA was detected among those expresing H7, H8 and H19. To our knowledge this is the first time subA gene is found in VTEC serotypes O2:H5, O20:H19, O79:H19, O88:H21, O141:H7 and O141:H8. Within serotypes O20:H19, O113:H21, O141:H8, O178:H19 and ONT:H19 both subA-negative and subA– positive isolates were found (Table 1), in agreement with the reports of authors, such as Cergole-Novella et al. (1), Newton et al. (13) and Irino et al. (3).

Table 1

Presence of subA and other virulence genes in the selected Argentine VTEC strains.

		Presence or absence of indicated gene
Serotype	Source (n° of isolates)	subA	Vt1	Vt2	eae	ehxA	saa
02:H5	S (1)	+	–	+	–	–	+
02:H25	F (1)	–	–	+	–	–	–
05:H-	C (1)	–	+	–	+	+	–
08:16	B (2); F (1)	–	+	–	–	–	+
015:21	F (1)	–	–	+	–	–	–
O20:H19	F (1); H (1); S (1)	+	+	+	–	+	+
O20:H19	G (1)	+	–	+	–	+	+
O20:H19	C (1)	–	+	+	–	+	+
O20:H19	C (1)	–	+	+	–	–	–
O20:H19	H (1)	–	–	+	–	–	–
O22:H8	H (2)	–	+	+	–	+	+
O22:H8	B (1)	–	–	+	–	–	–
O25:H19	F (1)	–	–	+	–	+	–
O26:H11	C (5)	–	+	–	+	+	–
O26:H11	C (2)	–	–	+	+	+	–
O39:H49	S (1)	+	+	+	–	+	+
O39:H49	S (4)	+	–	+	–	+	+
O74:H28	S (1)	–	–	+	–	+	+
O79:H19	S (1)	+	–	+	–	+	+
O88:H21	H (1)	+	+	+	–	+	+
O91:H21	F (2); G (1); H (1)	–	–	+	–	+	+
O103:H-	C (1)	–	+	+	+	+	–
O113:H21	B (1); G (1); V (1)	+	–	+	–	+	+
O113:H21	F (2); H (1)	–	–	+	–	–	–
O116:H21	B (1); G (1)	–	–	+	–	+	+
O120:H19	F (1)	–	–	+	–	+	+
O141:H7	S (1)	+	+	+	–	+	+
O141:H8	G (1)	+	–	+	–	+	+
O141:H8	G (1)	–	+	+	–	+	+
O145:H-	F (5); S (1)	–	–	+	+	+	–
O145:H-	F (4)	–	+	–	+	+	–
O145:H-	F (1)	–	+	–	+	–	–
O146:H21	F (1)	–	–	+	+	–	–
O157:H7	F (4); H (1)	–	–	+	+	+	–
O174:H21	B (2); C (1); F (4); S (2)	–	–	+	–	–	–
O174:H21	S (1)	–	+	+	–	+	+
O174:H21	F (1)	–	+	–	–	–	–
O175:H8	F (2)	–	–	+	–	–	–
O177:H-	F (1)	–	–	+	+	+	–
O178:H19	B (1)	+	–	+	–	+	+
O178:H19	B (1); F (2)	–	–	+	–	–	–
ONT:H7	B (1)	+	–	+	–	+	+
ONT:H8	H (1)	+	+	+	–	+	+
ONT:H19	B (1)	+	–	+	–	+	+
ONT:H19	B (3)	–	–	+	–	+	+
ONT:H19	B (1)	–	+	+	–	+	+
ONT:H21	V (1)	–	–	+	–	+	+
ONT:H21	F (1)	–	+	+	–	+	+
ONT:HNT	H (1)	–	–	+	–	+	+

In addition to SubAB, VTEC can present other megaplasmid-encoded virulence factors such as an enterohaemolysin (Ehx), considered as an indicator of megaplasmid presence, and the STEC autoagglutinating adhesin (Saa), only present in LEE-negative VTEC strains (2, 4, 9, 18, 23, 26). In our study, most of the subA-positive strains were also ehxA-positive with the exception of a strain belonging to O2:H5 serotype, positive for subA and negative for ehxA. We found two combinations among subtilase-positive strains: subA+/ehxA+/saa+ (20 strains) and subA+/ehxA-/saa+ (only one strain). Although subtilase gene was always detected in saa-positive strains, strains positive for saa were not necessarily subA-positive. Paton & Paton (20) and Karama et al. (5) also found some subtilase negative strains among saa-positive VTEC.

We have previously reported several saa variants among VTEC strains in our collection (10). Subtilase gene was found in strains carrying saa variants 1, 2, 3, 4 or 5, and therefore there was no evident association between saa variants and the presence/absence of subA.

As we expected, subA was found in eae-negative strains (which were screened for the presence of all known eae variants). These results are in accordance with those of Paton & Paton (20), Osek (14), Khaitan et al. (8), Cergole-Novella et al. (1), Karama et al. (5) and Irino et al. (3). However, Newton et al. (13) reported one VTEC strain harboring both subA and eae genes.

All subA-positive VTEC isolates carried vt2 either alone or with vt1 gene. Of 21 VTEC strains that were tested positive for subA, 14 strains had the gene encoding the vt2 toxin only and 7 strains were positive for both vt1 and vt2. Paton & Paton (20) demonstrated a strong association between the presence of subA and VTEC carrying the vt2 gene only, although they found a few VTEC strains that were subA-positive vt2-negative. Osek (14), Karama et al. (5) and Irino et al. (3) also found one vt1-positive vt2-negative VTEC isolate which was subA-positive.

The association of subA with vt2 and ehxA, in addition to the described potential of SubAB to augment clinical manifestations of VTEC infection or to cause disease in its own right, highlights the risk of some LEE-negative VTEC.

In summary, our results reveal the presence of subtilase cytotoxin gene in Argentine VTEC isolates belonging to several non-O157:H7 serotypes, and the existence of different combinations of megaplasmid encoded factors, confirming once again the great genetic variability of these plasmids.