Plant pathogenic microbes have the remarkable ability to manipulate biochemical, physiological, and morphological processes in their host plants. These manipulations are achieved through a diverse array of effector molecules that can either promote infection or trigger defense responses. We describe a general functional genomics approach aimed at identifying extracellular effector proteins from plant pathogenic microorganisms by combining data mining of expressed sequence tags (ESTs) with virus-based high-throughput functional expression assays in plants. PexFinder, an algorithm for automated identification of extracellular proteins from EST data sets, was developed and applied to 2147 ESTs from the oomycete plant pathogen Phytophthora infestans. The program identified 261 ESTs (12.2%) corresponding to a set of 142 nonredundant Pex (Phytophthora extracellular protein) cDNAs. Of these, 78 (55%) Pex cDNAs were novel with no significant matches in public databases. Validation of PexFinder was performed using proteomic analysis of secreted protein of P. infestans. To identify which of the Pex cDNAs encode effector proteins that manipulate plant processes, high-throughput functional expression assays in plants were performed on 63 of the identified cDNAs using an Agrobacterium tumefaciens binary vector carrying the potato virus X (PVX) genome. This led to the discovery of two novel necrosis-inducing cDNAs, crn1 and crn2, encoding extracellular proteins that belong to a large and complex protein family in Phytophthora. Further characterization of the crn genes indicated that they are both expressed in P. infestans during colonization of the host plant tomato and that crn2 induced defense-response genes in tomato. Our results indicate that combining data mining using PexFinder with PVX-based functional assays can facilitate the discovery of novel pathogen effector proteins. In principle, this strategy can be applied to a variety of eukaryotic plant pathogens, including oomycetes, fungi, and nematodes.

To overview the gene content of the important pathogen Phytophthora infestans, large-scale cDNA and genomic sequencing was performed. A set of 75,757 high-quality expressed sequence tags (ESTs) from P. infestans was obtained from 20 cDNA libraries representing a broad range of growth conditions, stress responses, and developmental stages. These included libraries from P. infestans-potato and -tomato interactions, from which 963 pathogen ESTs were identified. To complement the ESTs, onefold coverage of the P. infestans genome was obtained and regions of coding potential identified. A unigene set of 18,256 sequences was derived from the EST and genomic data and characterized for potential functions, stage-specific patterns of expression, and codon bias. Cluster analysis of ESTs revealed major differences between the expressed gene content of mycelial and spore-related stages, and affinities between some growth conditions. Comparisons with databases of fungal pathogenicity genes revealed conserved elements of pathogenicity, such as class III pectate lyases, despite the considerable evolutionary distance between oomycetes and fungi. Thirty-seven genes encoding components of flagella also were identified. Several genes not anticipated to occur in oomycetes were detected, including chitin synthases, phosphagen kinases, and a bacterial-type FtsZ cell-division protein. The sequence data described are available in a searchable public database.

Detailed analysis of the inheritance of molecular markers was performed in the oomycete plant pathogen Phytophthora infestans. Linkage analysis in the sexual progeny of two Dutch field isolates (cross 71) resulted in a high-density map containing 508 markers on 13 major and 10 minor linkage groups. The map showed strong clustering of markers, particularly of markers originating from one parent, and dissimilarity between the parental isolates on linkage group III in the vicinity of the mating-type locus, indicating a chromosomal translocation. A second genetic map, constructed by linkage analysis in sexual progeny of two Mexican isolates (cross 68), contained 363 markers and is thus less dense than the cross 71 map. For some linkage groups the two independent linkage maps could be aligned, but sometimes markers appeared to be in a different order, or not linked at all, indicating chromosomal rearrangements between genotypes. Graphical genotyping showed that some progeny contained three copies of a homologous linkage group. This trisomy was found for several linkage groups in both crosses. Together, these analyses suggest a genome with a high degree of flexibility, which may have implications for evolution of new races and resistance development to crop protection agents.

No abstract

In Phytophthora infestans, a cluster of three dominant avirulence genes is located on the distal part of linkage group VIII. In a mapping population from a cross between two Dutch field isolates, probe M5.1, derived from an amplified fragment length polymorphism (AFLP) marker linked to the Avr3-Avr10-Avr11 cluster, hybridized only to DNA from the parent and F1 progeny that is avirulent on potato lines carrying the R3, R10, and R11 resistance gene. In the virulent parent and the virulent progeny, no M5.1 homologue was detected, demonstrating a deletion on that part of linkage group VIII. P. infestans is diploid, so the avirulent strains must be hemizygous for the region concerned. A similar situation was found in another mapping population from two Mexican strains. The deletion was also found to occur in many field isolates. In a large set of unique isolates collected in The Netherlands from 1980 to 1991, 37% had no M5.1 homologue and the deletion correlated strongly with gain of virulence on potato lines carrying R3, R10, and R11. Also, in some old isolates that belong to a single clonal lineage (US-1) and are thus highly homogenous, deletions at the M5.1 locus were detected, indicating that this region is unstable.