Distribution and phylogeny of immunoglobulin-binding protein G in Shiga toxin-producing Escherichia coli and its association with adherence phenotypes

Abstract

eibG in Shiga toxin-producing Escherichia coli (STEC) O91 encodes a protein (EibG) which binds human immunoglobulins G and A and contributes to bacterial chain-like adherence to human epithelial cells. We investigated the prevalence of eibG among STEC, the phylogeny of eibG, and eibG allelic variations and their impact on the adherence phenotype. eibG was found in 15.0% of 240 eae-negative STEC strains but in none of 157 eae-positive STEC strains. The 36 eibG-positive STEC strains belonged to 14 serotypes and to eight multilocus sequence types (STs), with serotype O91:H14/H(-) and ST33 being the most common. Sequences of the complete eibG gene (1,527 bp in size) from eibG-positive STEC resulted in 21 different alleles with 88.11% to 100% identity to the previously reported eibG sequence; they clustered into three eibG subtypes (eibG-alpha, eibG-beta, and eibG-gamma). Strains expressing EibG-alpha and EibG-beta displayed a mostly typical chain-like adherence pattern (CLAP), with formation of long chains on both human and bovine intestinal epithelial cells, whereas strains with EibG-gamma adhered in short chains, a pattern we termed atypical CLAP. The same adherence phenotypes were displayed by E. coli BL21(DE3) clones containing the respective eibG-alpha, eibG-beta, and eibG-gamma subtypes. We propose two possible evolutionary scenarios for eibG in STEC: a clonal development of eibG in strains with the same phylogenetic background or horizontal transfer of eibG between phylogenetically unrelated STEC strains.

FIG. 1.

Expression of EibG in STEC O91, as demonstrated by Western blotting of STEC O91 cell lysates with HRP-conjugated human IgG Fc fragment. In lanes 1 to 6, the following strains (STs are in parentheses) are shown: lane 1, O91:H14 (ST33); lane 2, Ont:H⁻ [wzy_O91+, fliC_H14] (ST33); lane 3, O91:H⁻ [fliC_H14] (ST33); lane 4, O91:H21 (ST442) (negative control); lane 5, O91:Hnt [fliC_H14] (ST33); and lane 6, OR:Hnt [wzy_O91+, fliC_H14] (ST33); lane M, molecular mass marker.

FIG. 2.

Unrooted neighbor-joining tree of all 21 unique eibG alleles based on complete (1,527 bp) eibG sequences, illustrating their phylogenetic relationships and grouping them into three subtypes (eibG-α, eibG-β, and eibG-γ). Bootstrap values (values > 50% are shown) using 1,000 replicates are displayed on the corresponding branches. In parentheses, the related STs for each eibG allele are given.

FIG. 3.

Protein alignment of the 21 EibG alleles predicted from their nucleotide sequences in comparison to the reference sequence (EibG 001).

FIG. 4.

Tanglegram based on the topology of neighbor-joining trees of the concatenated MLST gene sequences (3,423 bp) and of complete eibG gene sequences (1,527 bp). In parentheses, the allelic profiles of each ST and the corresponding serogroups (OR [Orough], autoagglutinable strains) are given. The tanglegram illustrates the association between the genomic background (represented by MLST data) and the different eibG alleles showing either clonal diversification of eibG within the same genomic background (e.g., eibG 018 and 019 within ST690, serogroup O91) or possible horizontal gene transfer of eibG between different genomic backgrounds (e.g., eibG 015 in ST33, serogroup O91, and in ST442, serogroup O146).

FIG. 5.

Minimum spanning tree based on the MLST allelic profiles, portraying the clonal distribution of the eibG-positive STEC strains in comparison to the HUSEC collection. Each dot represents a given sequence type (ST), and the size of the circle is proportional to the number of strains analyzed. Connecting lines of increasing length illustrate, and numbers on these lines indicate, the number of alleles that are different between two STs.

FIG. 6.

Patterns of adherence of STEC strains with different EibG subtypes to HCT-8 human intestinal epithelial cells. In panels A to H, the following strains are shown (serotype, ST, and eibG subtype are in parentheses): 1809/00 (O91:H⁻ [H14], ST33, eibG-α) (A), 4789/97-1 (O146:H21, ST442, eibG-α) (B), 4831/97 (OR:H45, ST656, eibG-α) (C), 99-02787 (OR:H10, ST745, eibG-α) (D), 06-03233 (O152:H⁻ [H14], ST13, eibG-β) (E), 0519/99 (OR:Hnt, ST753, eibG-γ) (F), 0520/99 (Ont:H30, ST753, eibG-γ) (G), and 1745/98 (O91:H21, eibG-negative control) (H). Bar = 10 μm.

FIG. 7.

Patterns of adherence of STEC strains with different EibG subtypes to FDK-R 971 bovine intestinal epithelial cells. In panels A to E, the following strains are shown (serotype, ST, and eibG subtype are in parentheses): 7140/96 (O91:H⁻ [H14], ST33, eibG-α) (A), 4789/97-1 (O146:H21, ST442, eibG-α) (B), 06-03233 (O152:H⁻ [H14], ST13, eibG-β) (C), 0520/99 (Ont:H30, ST753, eibG-γ) (D), and 1745/98 (O91:H21, eibG-negative control) (E). Bar = 20 μm.

FIG. 8.

Expression of different EibG protein subtypes in STEC, demonstrated by fluorescence staining of strains harboring eibG-α, eibG-β, and eibG-γ with Alexa 488-labeled human IgG Fc fragment. In panels A to C, the following strains are shown (serotype, ST, and eibG subtype are in parentheses): 99-02787 (OR:H10, ST745, eibG-α) (A), 06-03233 (O152:H⁻ [H14], ST13, eibG-β) (B),and 0520/99 (Ont:H30, ST753, eibG-γ) (C). Panel D shows the eibG-negative STEC O91:H21 strain 1745/98 (negative control), and panel E shows strain 99-02787 (OR:H10, ST745, eibG-α), which was preincubated with nonlabeled human IgG Fc fragment before the fluorescence staining. Bar = 10 μm.

FIG. 9.

Adherence phenotypes of clones harboring different eibG subtypes on HCT-8 cells. In panels A to E, the following clones are shown (eibG subtypes and their origins in parentheses): clone B-1-10 (eibG-α from STEC O91:H14, ST33) (A), clone B-10-9 (eibG-α from STEC O146:H21, ST442) (B), clone B-20-1 (eibG-β from STEC O152:H⁻ [H14], ST13) (C), clone B-18-2 (eibG-γ from STEC Ont:H30, ST753) (D), and E. coli BL21(DE3) harboring pGEM-T Easy vector (eibG-negative control) (E). Bar = 20 μm.