The Oryza sativa-Magnaporthe grisea pathosystem is of economic importance but it is also a model for the plant-fungus interaction studies. A gene-for-gene system was described for this pathosystem. However, the number of specific interactions characterized at the molecular level is still restricted. After cloning the avirulence gene ACE1 of M. grisea (Böhnert et al. 2004, Plant Cell), we undertook the positional cloning of the corresponding resistance gene Pi33. Pi33 was localised on chromosome 8 of rice in two independent crosses (Berruyer et al. 2003, TAG). A fine mapping allowed us to place the gene between two markers distant of 240 kb. In this area, a cluster of 9 candidate genes was identified by analysis of the sequence of Nipponbare (which is deprived of Pi33). This analysis was confirmed by the sequencing of two BAC clones covering the same region in IR64 (which carries a resistance allele of Pi33). To identify the gene among the candidates, several strategies were developed. Silencing experiments by RNAi were started. Partial sequences of the different candidates were compared between varieties with or without Pi33. Polymorphism of flanking markers was also characterized. Finally, expression of the candidates was measured by RT-PCR in varieties with or without Pi33.