1The Sainsbury Laboratory, John lnnes Centre, Norwich Research Park, Colney, Norwich NR4 7UH, UK; 2Department of Plant and Microbial Biology, University of California, Berkeley, CA 94720-3102, USA; 3Plant Pathology and Weed Science Department, Horticulture Research International-Wellesbourne, Warwickshire CV35 9EF, UK

Background and objectives
In most cases plants defend themselves against pathogens through the recognition of a pathogen Avirulence (Avr) gene product by a plant resistance (R) gene product. The model plant Arabidopsis is a more tractable system for molecular genetic research than crop species. Therefore its interaction with pathogens is being studied to unravel disease resistance pathways. Many R loci (RPP loci) have been identified in Arabidopsis that trigger isolate-specific resistance to the oomycete pathogen Peronospora parasitica. The RPP5 gene shares similar structural motifs with the tobacco N gene and the flax L6 gene, which are recognized by a viral and rust pathogen, respectively [1]. This suggests that similar R-gene-mediated pathways operate in different plant species to different pathogen types. Two other Arabidopsis genes, NDR1 [2] and EDS1 [3] that encode anticipated resistance signalling components, are required for resistance specified by several RPP genes and other R genes specifying resistance to bacterial pathogens. To dissect genetically resistance signalling pathways in Arabidopsis, we analysed the dependence on EDS1 and NDR1 by a wider spectrum of R genes.

Materials and methods
Several independently isolated recessive mutant eds1 alleles and one recessive mutant mdr1 allele were used in this study. The mutations were crossed into three Arabidopsis backgrounds containing different R genes. Combinations of mutant and R loci were selected using PCR-based markers.

Results and conclusions
The extent of asexual sporulation of five different P. parasitica isolates on eds seedlings showed that RPP2, RPP4, RPPS, RPP10, RPP14 and RPP21 have an absolute requirement for EDS1. In contrast, resistance conferred by RPP8 to the Emco 5 isolate had no requirement for EDS1. Bacterial growth curves showed that EDS is fully required for RPS4 function, though similar experiments showed no requirement for EDS1 by RPM1, RPS2 and RPS5. Analysis of the requirement of the same spectrum of R genes for NDR1 showed that all of the R genes that were strongly EDS1-dependent had only a very weak requirement for NDR1 function. Conversely, strongly NDR1-dependent R genes operate independently of EDS1.

We conclude that the different requirements for EDS1 and NDR1 reflect the operation of two distinct R-gene-mediated signalling pathways that are not determined by pathogen type. Analysis of the different predicted R proteins by ourselves and other groups suggests that EDS1 or NDR1 dependence is more a feature of R protein type.

1. Parker JE, Coleman MJ, Szabo V et al., 1997. Plant Cell 9, 879-894.
2. Century KS, Shapiro AD, Repetti PP et al., 1997. Science 278, 1963-1965.
3. Parker JE, Holub EB, Frost LN et al.,1996. Plant Cell 8, 2033-2046.