A collection of Pseudomonas syringae and viridiflava isolates was established between 1993 and 2002 from diseased organs sampled from 36 pear, plum and cherry orchards in Belgium. Among the 356 isolates investigated in this study, phytotoxin, siderophore and classical microbiology tests, as well as the genetical methods REP-, ERIC-and BOX-(collectively, rep-) and IS50-PCR, enabled identification to be made of 280 isolates as P. syringae pv. syringae (Pss), 41 isolates as P. syringae pv. morsprunorum (Psm) race 1, 12 isolates as Psm race 2, three isolates as P. viridiflava and 20 isolates as unclassified P. syringae. The rep-PCR methods, particularly BOX-PCR, proved to be useful for identifying the Psm race 1 and Psm race 2 isolates. The latter race was frequent on sour cherry in Belgium. Combined genetic results confirmed homogeneities in the pvs avii, and morsprunorum race 1 and race 2 and high diversity in the pv. syringae. In the pv. syringae, homogeneous genetic groups consistently found on the same hosts (pear, cherry or plum) were observed. Pathogenicity on lilac was sometimes variable among Pss isolates from the same genetic group; also, some Psm race 2 and unclassified P. syringae isolates were pathogenic to lilac. In the BOX analyses, four patterns included 100% of the toxic lipodepsipeptide (TLP)-producing Pss isolates pathogenic to lilac. Many TLP-producing Pss isolates non-pathogenic to lilac and the TLP-non-producing Pss isolates were classified differently. Pseudomonas syringae isolates that differed from known fruit pathogens were observed in pear, sour cherry and plum orchards in Belgium.

The pathogenicity of 99 Belgian Pseudomonas syringae strains representative of the genetic diversity encountered in Belgian fruit orchards was evaluated by using 17 pathogenicity tests conducted on pear, cherry, plum, lilac, sugar beet and wheat. The P. syringae pv. morsprunorum strains were pathogenic to stone fruit species but the race 1 strains possessing the cfl gene involved in coronatine production were pathogenic in more tests than those lacking the gene. Also, sweet cherry twigs were a better material to detect pathogenic strains of race 1 and sour cherry twigs of race 2, which accorded with race 2 presence in sour cherry orchards in Belgium. Three groups were defined in the pv. syringae based on pathogenicity. One group pathogenic in 71.1% of the tests and to lilac included toxic lipodesipeptideproducing (TLP+) strains. The second group pathogenic in 26.8% of the tests and non-pathogenic to lilac included TLP+ strains. The thirth group pathogenic in 9.1% of the tests and almost specifically pathogenic to pear included TLP− strains. The three groups were genetically heterogeneous. Although strain-host relationships were noted within the pv. syringae, aptata and atrofaciens when considering the strain origins, such relationships were not found in the pathogenicity tests, suggesting that pathogenicity tests could probably not reproduce all the aspects of the host-pathogen interactions. None of the pathogenicity tests was able to provide all the information provided by the complete study. A test on pear buds indicated that strains different from the pv. syringae were pathogenic to pear.