We thank the following for kindly providing strains of Fusarium oxysporum f. sp.

Complementary auxotrophic nitrate-nonutilizing (nit) mutants were used to investigate vegetative compatibility within 27 strains of Verticillium dahliae isolated from several hosts originating from Africa, Asia, Europe and the United States. Using about 500 nit mutants generated from these strains, three vegetative compatibility groups, 1, 2 and 4, were identified. Simultaneously, virulence of each strain was assessed on cultivars of Gossypium hirsutum, G. barbadense and G. arboreum, based upon Foliar Alteration Index (FAI) and Browning Index (BI) estimation. The strains in VCG1 were of both the cottondefoliating pathotype and race 3 (on cotton) but were non pathogenic on tomato; those in VCG2 and VCG4 were of the nondefoliating pathotype and belonged to different races on cotton and on tomato. Hyaline mutants deriving from parental wild-type strain showed differences in pathogenicity but were always assigned to the parental VCG. A relationship was established between VCGs and the taxonomic position of host plants. Data from nit pairings indicated that the sub-populations of V. dahliae (VCGs) may not be completely isolated genetically.

The first two authors contributed equally to this work. SummaryLipid peroxidation, often associated with hypersensitive cell death, may be initiated either by active oxygen species (AOS) or lipoxygenases (LOX). Here we report a detailed analysis of this oxidative process in both incompatible and compatible interactions between the cotton cultivar Reba B50 and Xanthomonas campestris pv. malvacearum (Xcm). The hypersensitive reaction (HR) was characterized by a massive production of polyunsaturated fatty acid (PUFA) hydroperoxides together with typical tissue dehydration. Among these, isomers peroxidized on carbon 9, largely predominant, were chiral, showing an excess in the S enantiomer. The HR process was accompanied by an increase in 9S-LOX activity and preceded by transcription of a LOX gene (GhKLox1). These results showed that: (i) AOS produced during the oxidative burst were not involved in PUFA peroxidation during HR; and (ii) as previously described in elicited leaves of tobacco, the massive enzymatic lipid peroxidation was closely associated with hypersensitive cell death. During disease development in this cotton cultivar, the 9-lipoxygenation of PUFAs was late, weak, preceded by a faint accumulation of GhKLox1 transcripts, and associated with chlorosis but not with necrosis. Consequently, the main difference between incompatible and compatible interactions was in the precocity and intensity of the oxidative process, rather than in its nature. These data provide the evidence for a correlation between lipid peroxidation and hypersensitive cell death induced by pathogens.