1Department of Plant Pathology, Kansas State University, Manhattan, KS 66506-5502, USA; 2Department of Plant Pathology, University of the Philippines, Los Baņos, College, Laguna 4031, Philippines; 3International Rice Research Institute, PO Box 933, Manila, Philippines; 4National Crop Experiment Station, RDA, Suweon 441-100, Korea; 5Centro Internacional de la Papa, Apartado 1558, Lima 12, Peru

Background and objectives
Host-plant resistance has been used extensively for disease control in many crop species; however, many resistance sources are not long-lasting as a result of rapid changes in the pathogen. Although many resistance genes have been found in rice germplasm, there is no easy way to predict the quality or durability of these resistance genes. We hypothesize that resistance genes imposing a high penalty for adaptation will probably be durable. By elucidating the molecular changes involved in pathogen adaptation and the associated fitness cost, it may be possible to develop a proactive approach to predict the durability of resistance genes available for deployment. Here we determine how the bacterial blight pathogen of rice, Xanthomonas oryzae pv. oryzae (Xoo), adapts to rice containing a resistance gene not previously deployed in the Philippines, and what the consequences of that adaptation are on pathogenic fitness.

Materials and methods
The genetic structure of Xoo populations was evaluated over a 3-year period in plots established in the Philippines at two sites separated by 60 km. Xoo populations were sampled from three near-isogenic rice cultivars containing the Xa4, Xa7 and Xa10 genes for resistance to Xoo and the recurrent parent IR24. The population structure was analysed using a PCR-based assay [1]. Pathogen race and aggressiveness were determined by comparison of lesion lengths after inoculation of a set of near-isogenic differentials. The avrXa7-containing fragment was detected by DNA blot analysis using avrXa10 as a probe [2].

Results and conclusions
The population of Xoo in the two sites comprised two lineages throughout the study. The dominant lineage was composed of two races, and only one of those, race 9, was virulent to and was detected on rice with Xa7. Distribution of race 9 strains in the plots containing the cultivar with Xa7 at both sites indicated that the virulent strains were not introduced, but were the result of selection within the plots. Disease incidence and severity were high on cultivars without Xa7 at both sites in all years. Despite the presence of race 9 strains in the first year, disease incidence and severity on the cultivar with Xa7 remained low over six cropping seasons. A laboratory strain mutated in avrXa7 showed a fitness penalty in terms of bacterial multiplication and lesion development in the plant. To determine if loss of avrXa7 function in field isolates had a negative impact on pathogen fitness, race 9 strains were compared for the presence of a fragment corresponding to (avrXa7) and for aggressiveness to Xa7. Three distinct groups (9a, 9b and 9c) were defined, indicating that virulence to Xa7 occurred through three independent events. These groups showed different levels of aggressiveness to Xa7 and to IR24. One group (9c), which is virulent to Xa7 and lacks the avrXa7-sized fragment, was not detected after 1994. We suggest that changes in the avrXa7 gene reduce the persistence of that bacterial population in the field, and predict that Xa7 would provide durable resistance to Xoo in the Philippines. Future studies will address the costs that other resistance genes impose on pathogen fitness, and the utility of this approach in designing and deploying rice cultivars.

1. Vera Cruz CM, Ardales EY, Skinner DZ et al., 1996. Phytopathology 86, 1352-1359.
2. Hopkins C, White F, Choi SH et al., 1992. Molecular Plant-Microbe Interactions 6, 451-459.