Virulence properties of Escherichia coli O111:H12 strains

Abstract

Seventeen Escherichia coli O111:H12 strains isolated from the feces of children with acute diarrhea were studied with regard to their adherence properties and other virulence characteristics. All strains showed an aggregative adherence pattern to HEp-2 cells and agglutinated bovine and sheep red cells in the presence of mannose. These strains did not have gene sequences homologous to the aggregative adherence fimbria I gene and did not react with any of the DNA probes used to detect other virulence genes in enteropathogens. With one exception, the O111:H12 strains did not induce fluid accumulation in the rabbit ileal loop assay, although 16 of the strains had the enteroaggregative E. coli heat-stable enterotoxin 1 (EAST1) gene sequences. A 60–70 MDa plasmid was present in 16 of the strains studied. We conclude that the O111:H12 serotype, one of the first E. coli identified in infantile diarrhea, belongs to the enteroaggregative E. coli category but the genes encoding its adherence phenotype are distinct from those previously described.

Enteroaggregative Escherichia coli, Adherence pattern, Virulence properties, Mannose-resistant hemagglutination activity, EAST1

1 Introduction

Since its description in 1950 [1]Escherichia coli O111:H12 has been found in association with outbreaks and sporadic cases of diarrhea in children in several countries [2, 3]. On the basis of these epidemiological studies, this serotype is considered a diarrheogenic E. coli[4, 5], but its virulence mechanisms have not been established.

Cravioto et al. [6] in their pioneering work on the adhesion of enteropathogenic E. coli (EPEC) to HEp-2 cells found that the four O111:H12 strains included in the study had the ability to adhere to those cells, but they made no distinction among the different adherence patterns of E. coli. Scotland et al. [7] studied eight strains of E. coli O111:H12 and showed that six were non-adherent and two exhibited the diffuse adherence (DA) pattern to HEp-2 cells. In subsequent work, Scotland et al. [8] described two strains of the same serotype as enteroaggregative (EAggEC). Recently, Giammanco et al. [9] isolated three E. coli O111:H12 strains, two displaying the DA and one displaying the aggregative adherence (AA) pattern.

Campos et al. [10] in an extensive study on the clonal structure and virulence properties of E. coli O111 serogroup showed that serotypes O111:H12, O111:H4 and O111:HND (later, this serotype was found to be H10) belonged to a single clone, designated ET1. While all O111 serotypes reacted with one or more probes that detect virulence genes in enteropathogens, the O111:H12 strains did not react with any of the probes. Furthermore, these strains showed a variable pattern of adherence to HEp-2 cells, but for the majority of the strains no pattern could be determined. The purpose of this study was to investigate the adherence properties and other virulence characteristics of E. coli O111:H12 strains recently isolated in São Paulo and in Rio de Janeiro.

2 Materials and methods

2.1 Bacterial strains

Seventeen E. coli O111:H12 strains isolated in São Paulo and in Rio de Janeiro were included in this study. The five strains from São Paulo were isolated between 1990 and 1993 and the 12 strains from Rio de Janeiro, during 1987–1988. The 17 strains were isolated from the feces of children under 2 years of age, with acute diarrhea attending outpatients services, both in São Paulo and Rio de Janeiro. All the strains were kept in nutrient agar slants at room temperature once they were isolated or received at our laboratory. Before starting this study, the O and H antigens of the strains were confirmed by the methods recommended by Ewing [11]. The 17 strains and the strains described previously as ET1 [10] were examined by ‘enterobacterial repetitive intergenic consensus sequence (ERIC)-PCR’ and ribotyping (data not shown) and proved to belong to the same clone.

2.2 HEp-2 cell adherence assay

The ability to adhere to HEp-2 cells was determined as described by Cravioto et al. [6]. When no adherence pattern was seen after 3 h of incubation, the test was repeated with a 6-h incubation period.

2.3 Hemagglutination assay

Bacterial strains were examined for hemagglutination of bovine and sheep red blood cells as described by Old [12] in the presence and absence of 1%d-mannose.

2.4 Rabbit ileal loop assay

Eight strains were tested for fluid accumulation in the rabbit ileal loop assay according to the procedure described by Trabulsi [13]. Ligated ileal loops of approximately 10 cm long (seven loops per animal; separated by 5 cm long inter-loops) were inoculated with 1 ml of culture supernatants. The animals were killed 24 h after inoculation and the respective ileal loops were observed for fluid accumulation. A LT-I producing enterotoxigenic E. coli (strain 40T) and E. coli K12 C600 were included as controls. Each strain was tested in three rabbits.

2.5 DNA hybridization

The E. coli strains were tested by colony blot hybridization, by using specific DNA probes as described previously [10, 14] for: AA (aggregative adherence); bfpA (bundle-forming pilus); DA (diffuse adherence); eaeA (E. coli attaching and effacing); EAF (EPEC adherence factor); EHEC (E. coli enterohemorrhagic factor); Inv (E. coli invasiveness); LT-I (E. coli heat-labile enterotoxin type I); SLT-I (Shiga-like toxin I); SLT-II (Shiga-like toxin II); STh (heat-stable enterotoxin from E. coli of human origin); and STp (heat-stable enterotoxin from E. coli of porcine origin). The prototype wild-type strains from which DNA probes are derived were used as positive controls of each probe. E. coli K12 C600/pBR322 was used as negative control in all hybridization tests.

2.6 Detection of AAF/I and EAST1 gene sequences by PCR

PCR for detection of the EAggEC heat-stable enterotoxin 1 (EAST1) gene sequences was performed as described previously [15], using the primers 5′-CCATCAACACAGTATATCCGA and 5′-GGTCGCGAGTGACGGCTTTGT, which generate a 111-bp product. For the detection of the aggregative adherence fimbria I (AAF/I) gene sequences, the primers were used: 5′-GCGTTAGAAAGACCTCCAATA and 5′-GCCGGATCCTTAAAAATTAATTCCGGC, which generate a 432-bp product [16]. E. coli K12 C600 and EAggEC strain 17-2 were used as negative and positive controls, respectively.

2.7 Plasmid profiles

Plasmid DNAs were extracted by the alkaline lysis procedure [17], electrophoresed in 0.7% agarose gel with molecular-mass size markers, stained with ethidium bromide and detected under UV illumination.

3 Results and discussion

The results of all tests are summarized in Table 1. The negative results obtained with the DNA probes show that the E. coli O111:H12 strains lack the genes usually found in the other categories of diarrheogenic E. coli, including EAggEC. According to this, the O111:H12 strains may be defined as EAggEC by their adherence pattern to HEp-2 cells (Fig. 1), but the genes involved in this phenotype are different from those that have been identified so far. We have not looked for the gene responsible for the AAF/II fimbriae mentioned by Nataro et al. [18], because molecular probes are not available yet. Nevertheless, the gene for this fimbria is present in the EAggEC plasmid [19], which is not harbored by our strains. Regarding fimbriae, it should be mentioned that Yakubu et al. [20] recently identified a mannose-resistant fibrillar hemagglutinin that mediated the agglutination of sheep and bovine cells by two E. coli O111:H12 strains. These strains did not show an AA pattern to HEp-2 cells, but the authors did only a 30-min adherence test, which probably would not detect it. So, we suspect that our strains bear this kind of fimbria, since they also agglutinate both types of red blood cells.

1

Virulence characteristics of 17 E. coli O111:H12 strains

Strain number	Adherence pattern		MRHA		Rabbit ileal loop	DNA probesa	PCR for		60–70 MDa plasmid
3 h	6 h	bovine	sheep	AAF/I	EAST1
235	AA	NT	+	+	−	−	−	+	+
236	IN	AA	+	+	−	−	−	+	+
250	IN	AA	+	+	−	−	−	+	+
253	IN	AA	+	+	NT	−	−	−	+
276	IN	AA	+	+	+	−	−	+	−b
23-I	IN	AA	+	+	NT	−	−	+	+
114-II	IN	AA	+	+	NT	−	−	+	+
132-I	IN	AA	+	+	−	−	−	+	+
178-II	IN	AA	+	+	NT	−	−	+	+
182-I	IN	AA	+	+	NT	−	−	+	+
189-II	IN	AA	+	+	NT	−	−	+	+
197-I	IN	AA	+	+	−	−	−	+	+
226-I	IN	AA	+	+	−	−	−	+	+
270-V	IN	AA	+	+	NT	−	−	+	+
278-II	IN	AA	+	+	NT	−	−	+	+
289-I	IN	AA	+	+	NT	−	−	+	+
370-IV	IN	AA	+	+	−	−	−	+	+

AA: aggregative adherence pattern; IN: indefinite adherence; MRHA: mannose-resistant hemagglutination activity; NT: not tested; AAF/I: aggregative adherence fimbria I; EAST1: EAggEC heat-stable enterotoxin 1. ^aAll strains were negative with probes for AA, bfpA, DA, eaeA, EAF, EHEC, Inv, SLT-I, SLT-II, LT-I, STh, and STp (see Section 2). ^bThis strain showed a plasmid of 85 MDa.

1

Adherence of E. coli O111:H12 to HEp-2 cells. A: Indefinite pattern after a 3-h test. B: Aggregative pattern after a 6-h test.

While the O111:H12 strains used in this study behaved uniformly with regard to their ability to adhere to HEp-2 cells and to agglutinate sheep and bovine red cells, other authors have described strains of this serotype as non-adherent [7], diffusely adherent [7, 9] and aggregative [8, 9]. Probably these differences are due to the time of incubation of the adhesion test and to the age of the bacterial culture examined. As we have shown, the AA pattern of our strains appeared in its typical form only after 6 h of incubation of the adhesion test used and in our experience (data not shown), old O111:H12 strains behave variably with regard to their adherence pattern. At present it is not readily apparent why our strains exhibit their AA pattern only after 6 h of incubation.

At least three toxins have been described in EAggEC, but none of them has been firmly implicated in diarrhea [21]. Of these toxins, the best characterized is EAST1, the gene of which we detected in 16 of our 17 strains. The similar frequency of this toxin in some diarrheogenic E. coli suggests that it is associated with diarrhea [22]. Our effort to detect the production of other toxins using the ileal loop assay was fruitless, since only one strain (no. 276) provoked fluid accumulation. Probably this was due to the production of one of the toxins active in ileal loops [23]. Interestingly, this strain differed from the others by the presence of a 85 MDa plasmid. It is also clear from our findings that EAST1 was either not expressed or alone does not cause fluid secretion in ileal loops, since all strains tested had the EAST1 gene sequences.

The presence of a 60–70 MDa plasmid in all but one strain of the 17 investigated suggests that both AA and mannose-resistant hemagglutination activity (MRHA) were mediated by this plasmid. In the study reported by Nataro et al. [24], the AA pattern of the strain was encoded by a plasmid of a similar size. Yamamoto et al. [25] have also shown that in one strain (O127:H2) a 71-MDa plasmid was responsible for the AA pattern.

EAggEC differ significantly in their pathogenic potential, as shown by Nataro et al. in volunteer studies [18]. However, several epidemiological studies have shown that these bacteria are associated with persistent diarrhea [21]. Certainly this is not the case with O111:H12 strains, since strains of this serotype are most often associated with acute diarrhea [3–5]. The 17 strains examined in this study were all isolated from children with acute diarrhea, of which three required rehydration and one died.

It is generally accepted that EPEC were the first E. coli implicated in diarrhea [2, 4, 5]. However, it is interesting to mention that E. coli O111:H12, an EAggEC, and E. coli O111:H2 and E. coli O55:H6, two typical EPEC serotypes [10, 14], were described at the same time as agents of diarrhea [1].

Acknowledgements

This work was supported by Grants 62.02366/92-2 (PADCT/CNPq) and 92/4890-2 (FAPESP). We thank Dr. Adriana H. Regua-Mangia (Escola Nacional de Saúde Pública, FIOCRUZ, Rio de Janeiro, RJ, Brazil) for supplying the bacterial strains isolated in Rio de Janeiro and Dr. Thomas S. Whittam (The Pennsylvania State University, University Park, PA, USA) for comments on the manuscript.