2.2.42
STUDIES ON THE ORIGIN OF THE WHEAT BLAST FUNGUS: MOLECULAR ANALYSIS WITH A LINE-LIKE RETROTRANSPOSON

AS URASHIMA, Y HASHIMOTO, LD DON, M KUSABA, Y TOSA, H NAKAYASHIKI, and S MAYAMA

Faculty of Agriculture, Kobe University, Nada, Kobe 657, Japan

Background and objectives
Magnaporthe grisea is the causal agent of blast disease of many gramineous plants including common crops such as rice (Oryza sativa), foxtail millet (Setaria italica), common millet (Panicum miliaceum), finger millet (Eleusine coracana), etc. In 1985, blast disease occurred on wheat (Triticum aestivum) in Brazil and spread through all wheat-growing regions in this country. At first, the wheat blast fungus was speculated to be derived from the rice blast fungus. However, fingerprinting with MGR586, a DNA-type transposon, demonstrated that this was not the case [1]. Although data on mating compatibility and pathogenicity suggested that the finger millet blast fungus was the most similar to the wheat blast fungus [2], its origin is still unclear. In this study, we analysed phylogenic relationships among isolates from wheat, other crops, and weeds collected in Brazil using a LINE-like retrotransposon as a probe.

Results and conclusions
MGR586 is not an appropriate probe for fingerprinting of Triticum isolates since its copy number is very low in their genome. We surveyed a genomic library of an Eleusine isolate, and found a clone, pEBA18, containing a repetitive element. This element had a putative amino acid sequence with a high homology to the reverse transcriptase domain of LINE-like retrotransposons, and proved to be a member of MGR583 family. Based on the copy number of this element, isolates from various hosts were classified into three groups. The first group comprised isolates from Oryza, Setaria and Panicum, which carried this element in multiple copies. The second comprised those from Triticum, Eleusine, Digitaria, and Brachiaria, which carried it at moderate copy number. The third comprised mioga isolates, which carried no copies. These results indicate that isolates from Oryza, Panicum, and Setaria form a group distinct from Triticum isolates. Relationships within the second group were analysed by DNA fingerprinting with pEBA18.

Isolates from Digitaria horizontalis produced no common bands with Triticum isolates, indicating that they are remote from Triticum isolates. Isolates from Eleusine coracana produced only one common band with Triticum isolates, uggesting that they are not close to Triticum isolates. This result was contradictory to the data on mating compatibility and pathogenicity, but seemed to be reasonable since E. coracana has not been found in Brazil. However, isolates from Brachiaria plantaginea, a common weed in Brazil, produced fingerprints similar to those of Triticum isolates. A UPGMA dendrogram showed that Triticum isolates clustered into a single group at 75% similarity level, but had formed no clonal lineages within the group. From these results, we suggest that wheat isolates were derived from a single origin close to Brachiaria isolates, and then developed diversity under low selection pressure. There is a possibility that this diversification may be promoted by sexual recombination, since wheat isolates are highly fertile [2].

References
1. Valent B, Chumley FG, 1991. Annual Review of Phytopathology 29, 443-467.
2. Urashima AS, Igarashi S, Kato H, 1993. Plant Disease 77, 1211-1216.