Phylogenetic backgrounds and virulence profiles of atypical enteropathogenic Escherichia coli strains from a case-control study using multilocus sequence typing and DNA microarray analysis

Abstract

Atypical enteropathogenetic Escherichia coli (EPEC) strains are frequently detected in children with diarrhea but are also a common finding in healthy children. The aim of this study was to compare the phylogenetic ancestry and virulence characteristics of atypical (eae positive, stx and bfpA negative) EPEC strains from Norwegian children with (n = 37) or without (n = 19) diarrhea and to search for an association between phylogenetic ancestry and diarrhea. The strains were classified in phylogenetic groups by phylogenetic marker genes and in sequence types (STs) by multilocus sequence typing. Phylogenetic ancestry was compared to virulence characteristics based on DNA microarray analysis. Serotyping and pulsed-field gel electrophoresis (PFGE) were also performed. All four phylogenetic groups, 26 different STs, and 20 different clonal groups were represented among the 56 atypical EPEC strains. The strains were separated into three clusters by overall virulence gene profile; one large cluster with A, B1, and D strains and two clusters with group B2 strains. There was considerable heterogeneity in the PFGE profiles and serotypes, and almost half of the strains were O nontypeable. The efa1/lifA gene, previously shown to be statistically linked with diarrhea in this strain collection (J. E. Afset et al., J. Clin. Microbiol. 44:3703-3711, 2006), was present in 8 of 26 STs. The two phylogenetic groups B1 and D were weakly associated with diarrhea (P = 0.06 and P = 0.09, respectively). In contrast, group B2 was isolated most frequently from healthy controls (P = 0.05). In conclusion, the atypical EPEC strains were heterogeneous both phylogenetically and by virulence profile. Phylogenetic ancestry was less useful as a predictor of diarrhea than were specific virulence genes.

FIG. 1.

Phylogenetic relationships between 56 atypical EPEC strains from Norwegian children. A rooted phylogenetic tree was constructed by the neighbor-joining algorithm based on the Kimura two-parameter model of nucleotide substitution. The ST with the number of isolates (in brackets) is given at each branch tip. Bootstrap values greater than 50% based on 1,000 replications are shown at the internal nodes. Clonally related STs are shown by shaded boxes. The classification in phylogenetic groups, distribution of LEE gene variants, the presence of OI-122, lpfA, and EHEC-related genes, and the sources of the strains are shown for each ST. For combinations of LEE gene variants, see Table 3.

FIG. 2.

Phylogenetic relationships between 56 atypical EPEC strains from Norwegian children are presented as a phylogenetic splits network based on the neighbor-net algorithm using a p distance matrix. STs are indicated at the branch tips. The sources of the strains are listed in parentheses (patients/controls). STs belonging to the same phylogenetic group are enclosed by an ellipse.

FIG. 3.

Principal component analysis of the distribution of 94 putative virulence genes or gene variants in 56 atypical EPEC strains isolated from Norwegian children. Scores (A) and loadings (B) for the first two principal components; the phylogenetic group of each strain is indicated by the design of the marker, while the color shows whether the strain was isolated from children with (red) or without (blue) diarrhea. The loadings show genes previously shown to be statistically linked with diarrhea in red (2).

FIG. 4.

PFGE cluster analysis compared to phylogenetic ancestry, virulence profile, phylogenetic group, and source of atypical EPEC strains from Norwegian children with or without diarrhea. Clusters identified by the cluster cutoff method (see the text) are visualized by dense lines in the dendrogram. Two strains (Trh7 and Trh10) were not typeable by the PFGE method used. The STs 35 and 526 share six of seven alleles but have not been assigned to a ST complex. See Table 1 for more information.