Cell surface attachment structures contribute to biofilm formation and xylem colonization by Erwinia amylovora

Abstract

Biofilm formation plays a critical role in the pathogenesis of Erwinia amylovora and the systemic invasion of plant hosts. The functional role of the exopolysaccharides amylovoran and levan in pathogenesis and biofilm formation has been evaluated. However, the role of biofilm formation, independent of exopolysaccharide production, in pathogenesis and movement within plants has not been studied previously. Evaluation of the role of attachment in E. amylovora biofilm formation and virulence was examined through the analysis of deletion mutants lacking genes encoding structures postulated to function in attachment to surfaces or in cellular aggregation. The genes and gene clusters studied were selected based on in silico analyses. Microscopic analyses and quantitative assays demonstrated that attachment structures such as fimbriae and pili are involved in the attachment of E. amylovora to surfaces and are necessary for the production of mature biofilms. A time course assay indicated that type I fimbriae function earlier in attachment, while type IV pilus structures appear to function later in attachment. Our results indicate that multiple attachment structures are needed for mature biofilm formation and full virulence and that biofilm formation facilitates entry and is necessary for the buildup of large populations of E. amylovora cells in xylem tissue.

Fig. 1.

Genetic maps of the individual genes and sets of genes deleted in this study. Individual genes are shown as block arrows, arrow direction corresponds to the direction of transcription, and gene length is drawn to scale in base pairs. The EAM gene designations are from the genome sequence of E. amylovora ATCC 49946 (EMBL accession no. FN666575). If multiple genes are shown in a set (e.g., ppdD, hofB, and hofC), the entire set of genes was deleted. In the Δfim, Δflg-3, and Δflg-4 mutants, 7, 21, and 13 genes, respectively, were deleted as shown.

Fig. 2.

Images of putative attachment structures of E. amylovora. (A) TEM imaging of a planktonic E. amylovora cell grown in broth culture and negatively stained. Peritrichous flagella are indicated by arrows. (B) TEM image of E. amylovora in planta. Putative attachment structures connect bacterial cells to host cells. (C) SEM image of E. amylovora cells attached to vascular tissue within a Gala apple tree. Imaged E. amylovora cells were found within a biofilm, with multiple appendages that protrude from the bacterial cell and attach to the host surface, as indicated by the arrows.

Fig. 3.

Biofilm formation in vitro by E. amylovora. (A) Quantification of biofilm formation on glass coverslips of deletion mutants of E. amylovora and complemented strains. All of the mutants, except the ΔfimD mutant, demonstrated significant deficiencies in biofilm formation compared to Ea1189 (P < 0.05, Student t test). Complementation restores the phenotype to one similar to that of the wild type. (B) Bright-field imaging of biofilm formation on glass coverslips. All of the deletion mutants tested exhibit a significant visual reduction in attachment to the glass surface after 48 h, except the ΔfimD mutant.

Fig. 4.

Flow cell imaging after 48 h of growth measuring aggregates as brightness of fluorescence, where higher intensity equals larger aggregates. The image shows aggregation and biofilm development over area (x and y) and intensity (z). Note that intensity mapping had the maximum value set at 250 and maps were developed with five color layers. Additional layers of color indicate greater signal intensity. (A) Ea1189 exhibiting a typical biofilm consisting of aggregates. (B) The Δfim mutant, a representative of biofilm-negative behavior, including very few to no aggregates but showing bacterial growth. (C) The ΔfimD mutant showing few aggregates throughout but not at levels as great as those of the wild type.

Fig. 5.

Bacterial populations in immature pear fruit over 3 days. A shift in the population growth of the ΔhofC and Δhof mutants and reductions in the Δfim, Δflg-3, and Δflg-4 mutants are evident. Significant alterations in population are indicated by asterisks, and significant differences (P < 0.05) were determined using the Student t test.

Fig. 6.

Measurement of necrosis progression in Gala apple tissue at 3 and 7 days postinoculation. Scissor cuts perpendicular to the mid-vein of a leaf were used for inoculations. The ΔhofC, ΔfimD, Δfim, and Δflg-4 mutants were significantly reduced in virulence compared to the wild-type Ea1189 strain, and the Δams mutant was nonpathogenic (significant differences [P < 0.05, t test] are indicated by asterisks) at 7 days postinoculation.

Fig. 7.

SEM images of E. amylovora Ea1189 within xylem tissue (A) and mesophyll tissue (B) of apple leaves. Distinct biofilms are present within the vascular tissue (black arrow), as indicated by the aggregates filling the vascular space, and smaller populations are present within the mesophyll tissue (white arrows). The biofilm formation-deficient Δfim mutant was visible within vascular tissue (C) and mesophyll tissue (D), which was typical of all of the deletion mutants studied, and exhibited larger populations within mesophyll tissue. Relatively few cells were able to gain entry into xylem tissue (black arrow).