The Superior Adherence Phenotype of E. coli O104:H4 is Directly Mediated by the Aggregative Adherence Fimbriae Type I

Abstract

Whereas the O104:H4 enterohemorrhagic Escherichia coli (EHEC) outbreak strain from 2011 expresses aggregative adherence fimbriae of subtype I (AAF/I), its close relative, the O104:H4 enteroaggregative Escherichia coli (EAEC) strain 55989, encodes AAF of subtype III. Tight adherence mediated by AAF/I in combination with Shiga toxin 2 production has been suggested to result in the outbreak strain's exceptional pathogenicity. Furthermore, the O104:H4 outbreak strain adheres significantly better to cultured epithelial cells than archetypal EAEC strains expressing different AAF subtypes. To test whether AAF/I expression is associated with the different virulence phenotypes of the outbreak strain, we heterologously expressed AAF subtypes I, III, IV, and V in an AAF-negative EAEC 55989 mutant and compared AAF-mediated phenotypes, incl. autoaggregation, biofilm formation, as well as bacterial adherence to HEp-2 cells. We observed that the expression of all four AAF subtypes promoted bacterial autoaggregation, though with different kinetics. Disturbance of AAF interaction on the bacterial surface via addition of α-AAF antibodies impeded autoaggregation. Biofilm formation was enhanced upon heterologous expression of AAF variants and inversely correlated with the autoaggregation phenotype. Co-cultivation of bacteria expressing different AAF subtypes resulted in mixed bacterial aggregates. Interestingly, bacteria expressing AAF/I formed the largest bacterial clusters on HEp-2 cells, indicating a stronger host cell adherence similar to the EHEC O104:H4 outbreak strain. Our findings show that, compared to the closely related O104:H4 EAEC strain 55989, not only the acquisition of the Shiga toxin phage, but also the acquisition of the AAF/I subtype might have contributed to the increased EHEC O104:H4 pathogenicity.

Keywords: AAF; Enteroaggregative E. coli; adherence; autoaggregation; biofilm; enterohemorrhagic E. coli.

Figure 1.

Genetic organization of the aaf/I, III, IV and V operons and amino acid identities of the AggDCBA subunits. (a) Schematic representation of the genetic organization of aaf/I, III, IV and V operons (not to scale). (b) Amino acid identities of the AggDCBA subunits of AAF/I, IV and V in respect to the subunits of AAF/III in percent. Overall, AAF/V is more closely and AAF/IV more distantly related to AAF/III than AAF/I is to AAF/III

Figure 2.

Western blot analysis of EAEC strain 55989 agg3⁻ heterologously expressing AAF/I, III, IV and V. Shown is a Western blot analysis of the pBAD24-based expression of the aaf/I, III, IV and V determinants. Decreasing amounts of total protein (10 µg, 5 µg and 2.5 µg) of EAEC strain 55989 agg3⁻ carrying plasmid pAAF/I (a), pAAF/III (b), pAAF/IV (c) or pAAF/V (d) were used for semi-quantitative Western blot analysis using antibodies raised against each major fimbrial subunit. EAEC strain 55989 agg3⁻ carrying the empty vector pBAD24 was used as negative control (NC). Corresponding sizes of immunoreactive bands are given on the right (size marker; M)

Figure 3.

Electron microscopic images of EAEC strain 55989 agg3⁻ expressing different AAF variants. Shown are negatively stained EM images of EAEC strain 55989 agg3⁻ heterologously expressing AAF/I, III, IV and V (black arrows), as well as EAEC 55989 agg3⁻ containing the empty expression vector (NC)

Figure 4.

Impact of different AAF variants on bacterial autoaggregation. Shown is the relative decrease in OD₆₀₀ when compared to the OD₆₀₀ at the start of the experiment. Autoaggregation was tested for E. coli strain 55989 agg3⁻ carrying one of the plasmids pBAD24aafI, pBAD24aafIII, pBAD24aafIV and pBAD24aafV or pBAD24 (NC). Samples were taken in ten min intervals until reaching a plateau. (a) Representative results of an autoaggregation assay. Except for E. coli strain 55989 agg3⁻ expressing AAF/IV, the bacterial culture sedimented completely at t = 60 min. (b) Shown are average values for each time point and standard deviations of three biological replicates. Whereas the sedimentation of AAF/I-, AAF/III- and AAF/V-expressing cells differed significantly (p < 0.05) from the negative control already after 10 min, AAF/IV-expressing cells differed only after 30 min significantly from the negative control (the first time point at which the sedimentation speed is significantly different from the negative control is indicated with an * for each AAF variant). The sedimentation speed of AAF/IV-expressing cells was significantly different from AAF/I-, AAF/III- and AAF/V-expressing bacteria until t = 20 min (indicated with **)

Figure 5.

Autoaggregation of mixed EAEC 55989 agg3⁻ populations expressing different AAF variants. Shown are fluorescence microscopic images of aggregates formed by EAEC strain 55989 agg3⁻ heterologously expressing the same, different or no AAF operon. In order to be able to distinguish the bacteria expressing the different AAF subtypes, the bacteria carried an additional identical plasmid, except for the encoded fluorophore (YFP, yellow/CFP, cyan). The bacterial cells formed more or less homogeneous aggregates, but only upon expression of AAF

Figure 6.

Biofilm formation by EAEC strain 55989 agg3- expressing AAF/I, AAF/III, AAF/IV and AAF/V depends on the AAF subtype and temperature. Shown is the biofilm formation in LB medium at 20°C, 30°C and 37°C. For details see text

Figure 7.

Impact of AAF subtype expression on the aggregative adherence phenotype. Shown are bacterial aggregates formed by EAEC strain 55989 agg3⁻ heterologously expressing AAF/I, AAF/III, AAF/IV and AAF/V on HEp2 cells as indicated. NC: negative control (pBAD24); Medium: uninfected HEp2 cells; Scale bar: 10 µm; Upper panels: 10x magnification, Lower panels: 40x magnification. White arrows indicate rudimentary honey comb formation