Application of amplified fragment length polymorphism fingerprinting for taxonomy and identification of the soft rot bacteria Erwinia carotovora and Erwinia chrysanthemi

Abstract

The soft rot bacteria Erwinia carotovora and Erwinia chrysanthemi are important pathogens of potato and other crops. However, the taxonomy of these pathogens, particularly at subspecies level, is unclear. An investigation using amplified fragment length polymorphism (AFLP) fingerprinting was undertaken to determine the taxonomic relationships within this group based on their genetic relatedness. Following cluster analysis on the similarity matrices derived from the AFLP gels, four clusters (clusters 1 to 4) resulted. Cluster 1 contained Erwinia carotovora subsp. carotovora (subclusters 1a and 1b) and Erwinia carotovora subsp. odorifera (subcluster 1c) strains, while cluster 2 contained Erwinia carotovora subsp. atroseptica (subcluster 2a) and Erwinia carotovora subsp. betavasculorum (subcluster 2b) strains. Clusters 3 and 4 contained Erwinia carotovora subsp. wasabiae and E. chrysanthemi strains, respectively. While E. carotovora subsp. carotovora and E. chrysanthemi showed a high level of molecular diversity (23 to 38% mean similarity), E. carotovora subsp. odorifera, E. carotovora subsp. betavasculorum, E. carotovora subsp. atroseptica, and E. carotovora subsp. wasabiae showed considerably less (56 to 76% mean similarity), which may reflect their limited geographical distributions and/or host ranges. The species- and subspecies-specific banding profiles generated from the AFLPs allowed rapid identification of unknown isolates and the potential for future development of diagnostics. AFLP fingerprinting was also found to be more differentiating than other techniques for typing the soft rot erwinias and was applicable to all strain types, including different serogroups.

FIG. 1.

Dendrogram derived from the UPGMA linkage of correlation coefficients between AFLP profiles from soft rot erwinias, and other erwinias and enterobacterial species. The levels of linkage representing the Pearson product-moment correlation coefficient (r) are expressed as percentages for convenience. The banding profiles against each branch represent normalized and background-subtracted digitized gel strips processed in GelCompar.

FIG. 2.

Cluster analysis of E. chrysanthemi and E. carotovora subspecies together with suggested species and subspecies clusters (clusters 1 to 4). ¥ indicates a previously uncharacterized strain. The dendrogram is derived from the UPGMA linkage of correlation coefficients between AFLP profiles. The levels of linkage representing the Pearson product-moment correlation coefficient (r) are expressed as percentages for convenience. The banding profiles against each branch represent normalized and background-subtracted digitized gel strips processed in GelCompar. The arrow indicates E. carotovora-specific AFLP bands. Ech, E. chrysanthemi; Ecc, E. carotovora subsp. carotovora; Eco, E. carotovora subsp. odorifera; Eca, E. carotovora subsp. atroseptica; Ecb, E. carotovora subsp. betavasculorum; Ecw, E. carotovora subsp. wasabiae.

FIG. 3.

Dendrogram derived from the UPGMA linkage of correlation coefficients between AFLP profiles from strains of E. carotovora subsp. atroseptica. The levels of linkage representing the Pearson product-moment correlation coefficient (r) are expressed as percentages for convenience. The banding profiles against each branch represent normalized and background-subtracted digitized gel strips processed in GelCompar.