Second Africa Rice Congress, Bamako, Mali, 22–26 March 2010: Innovation and Partnerships to Realize Africa’s Rice Potential 5.3.1 Rice yellow mottle virus diversification impact on the genetic control of RYMV

Rice yellow mottle is the most important virus disease of rice in Africa. The disease is a major constraint due to its wide geographical distribution and the extent of yield losses it induces. The role of the causal agent Rice yellow mottle virus (RYMV) in virus–host interactions was studied in relation to the current knowledge of rice genetics. Most cultivated rice varieties are susceptible to the virus, but a monogenic and recessive high resistance has been found in two varieties of Oryza sativa and a few varieties of O. glaberrima. The high resistance RYMV1 encodes an eukaryotic translation initiation factor eIF(iso)4G and at least five alleles have been identified. However, studies on virus diversity indicated the occurrence of virus pathotypes capable of overcoming the high resistance gene at rates up to 40%. Such pathotypes could also be generated experimentally. At molecular level, RYMV pathogenicity was associated with mutations in the viral protein genome-linked (VPg), which interacts with the eIF(iso)4G factor. Patterns of the breakdown of alleles rymv1-2 in O. sativa and rymv1-3 in O. glaberrima cv. Tog5681 differed greatly and were modulated by virus adaptation features found in the VPg. The specificity of virus–host interactions between RYMV and rice suggests that the deployment of resistant varieties should take into account a good knowledge of virus populations. Introduction Rice yellow mottle virus (RYMV) was observed for the first time in Kenya in the late 1960s (Bakker, 1970). The disease is now present in all or almost all African rice-producing countries (Kouassi et al., 2005). It has never been reported outside of Africa. Since the beginning of the 1990s, RYMV has become a major constraint to rice production, mainly in west Africa, with losses that can reach 25–100% depending on the rice varieties (Abo et al., 1998). Cultivated rice belongs to two species Oryza sativa and O. glaberrima. Screening of many varieties of these two species as well as wild species of rice led to the identification of resistance sources. Types of resistance have been characterized in few rice varieties. There is, on the one hand, a partial resistance marked by a delay in the appearance of symptoms and in virus accumulation (Ioanidou et al., 2000; Fargette et al., 2002). It has polygenic determinism and has been identified in varieties of O. sativa subsp. japonica, e.g. Azucena. The second, high resistance is characterized by an absence of symptoms, very low virus accumulation and a blockage of virus movement (Ndjiondjop et al., 1999). It is monogenic and recessive. Progress in rice genetics has revealed that the high resistance gene RYMV1 has at least five alleles: rymv1-1, present in varieties susceptible to the virus; rymv1-2 identified in Gigante and Bekarosaka varieties of O. sativa subsp. indica (Ndjiondjop et al., 1999; Rakotomalala et al., 2008); all the other alleles have been identified in O. glaberrima: Tog5681 (rymv1-3), Tog5672 (rymv1-4) and Tog5674 (rymv1-5). The gene RYMV1 codes protein eIF(iso)4G that is the factor of the host interacting with the virus (Albar et al., 2006). There are ongoing efforts to introgress the resistance conferred by the RYMV1 gene into susceptible but yielding varieties. RYMV has a great molecular diversity (Traoré et al., 2005). There are more than six strains with distinct geographical distributions. Strains S1-AO and S1-AC are present in savannahs of west and central Africa, respectively. In west Africa, there is also a Sahelian strain (Sa) and a forest strain (S2/S3). Three other strains (S4, S5 and S6) have been identified in east Africa and Madagascar. Virus pathogenic diversity studies revealed the presence of RYMV isolates capable of avoiding the resistance conferred by the RYMV1 gene (Konate et al., 1997; Fargette et al., 2002; Traoré et al., 2006). The high occurrence of such isolates (40%) is an important parameter that can compromise the sustainability of resistance. The severity factor has been identified as being the virus protein linked to RYMV genome (VPg) (Hébrard et al., 2008). During this work, biological and molecular aspects of the interactions between rice and RYMV were studied for resistance alleles rymv1-2 and rymv1-3. Results obtained show that resistance to RYMV can be sustainable if resistant varieties are deployed by taking into account the virus diversity in the deployment zone. Material and methods Rice varieties, inoculum sources and plant inoculations The two rice varieties having the resistance allele rymv1-2 (Gigante and Bekarosaka) and the variety Tog5681 having the rymv1-3 allele were used. The experiments were conducted in insect-free screen houses at 28°C and 60% relative humidity. Representative virus isolates of the RYMV molecular diversity from different countries of the entire continent were used. For each virus source, the virus inoculum was prepared by crushing infected * Corresponding author (email: [email protected]). Theme 5: Integrated management of pests, diseases and weeds in rice-based systems Traoré et al.: RYMV diversity and genetic control of RYMV 5.3.2 Second Africa Rice Congress, Bamako, Mali, 22–26 March 2010: Innovation and Partnerships to Realize Africa’s Rice Potential leaves in phosphate buffer 0.1M pH 7.2 (with 10 ml of buffer per 1 g of leaves). Carborundum 600-grid was added to the extract, which was then used to rub the leaves of young rice plants (aged 14 days). Appearance of symptoms was monitored every week for 6 months. Determining VPg sequences and directed mutagenesis The viral protein genome-linked was amplified by reverse transcription polymerase chain reaction (RT-PCR) as described by Pinel-Galzi et al. (2007). Nucleotide sequences were then determined by direct sequencing of amplification products. Sequence collection and analysis were carried out using bioinformatics software package Lasergene DNAStar. Mutations associated with resistance avoidance were identified. To validate the role of each of these mutations in the virus pathogenicity, corresponding changes were introduced in an infectious clone through direct mutagenesis (Hébrard et al., 2008; Brugidou et al., 1995). The infectious clone thus modified was then inoculated to rice plants and symptoms monitored as previously. Results and discussion Mutations associated with rymv1-2 allele avoidance In total, 10 isolates out of 114 were able to avoid the resistance in the highly resistant rice variety Gigante. The VPg sequences (41 sequences) were determined in the plants that presented symptoms and were compared with those of the original isolates. Five nonsynonym mutations were located in a short region of 15 amino acids (aa) out of the 79 aa that make the VPg (Table 1). The implication of these mutations in avoiding the resistant allele rymv1-2 was demonstrated by directed mutagenesis (Pinel-Galzi et al., 2007). Candidate mutations were separately introduced in an infectious clone of RYMV which could not avoid resistance (Brugidou et al., 1995). All candidate mutations introduced in the infectious clone allowed the avoidance of allele rymv1-2 in Gigante or Bekarosaka. The most frequent mutation was that of position 48. At this position, the arginine residue (R) was replaced by six other possible amino acids: tryptophan (W), glycine (G), glutamic acid (E), valine (V) or threonine (T). Monitoring the appearance of these mutations over time has shown a major scenario bringing into play R, G and E residues. In a first phase, the R residue of the avirulent isolate was replaced by the G residue (making the isolate virulent). In a second phase, the G residue was replaced by an E residue that, when fixed, was never replaced. Thus, the E residue appeared as more adapted to rymv1-2 allele avoidance. Table 1. Mutations of the viral protein genome-linked (VPg) involved in rymv1-2 allele avoidance RYMV isolate Strain Sequences after avoidance Amino acid positions in the VPg†