Magnaporthe grisea Genes for Pathogenicity and Virulence Through a Series of Backcrosses

We have identified genes for pathogenicity toward rice (Oryza sativa) and genes for virulence toward specific rice cultivars in the plant pathogenic fungus Magnaporthe grisea. A genetic cross was conducted between the weeping lovegrass (Eragrostis curvula) pathogen 4091-5-8, a highly fertile, hermaphroditic laboratory strain, and the rice pathogen 0-135, a poorly fertile, female-sterile field isolate that infects weeping lovegrass as well as rice. A six-generation backcrossing scheme was then undertaken with the rice pathogen as the recurrent parent. One goal of these crosses was to generate rice pathogenic progeny with the high fertility characteristic of strain 4091-5-8, which would permit rigorous genetic analysis of rice pathogens. Therefore, progeny strains to be used as parents for backcross generations were chosen only on the basis of fertility. The ratios of pathogenic to nonpathogenic (and virulent to avirulent) progeny through the backcross generations suggested that the starting parent strains differ in two types of genes that control the ability to infect rice. First, they differ by polygenic factors that determine the extent of lesion development achieved by those progeny that infect rice. These genes do not appear to play a role in infection of weeping lovegrass because both parents and all progeny infect weeping lovegrass. Second, the parents differ by simple Mendelian determinants, “avirulence genes,” that govern virulence toward specific rice cultivars in all-or-none fashion. Several crosses confirm the segregation of three unlinked avirulence genes, AvrI-C039, Avr l M20I and Avrl-YAMO, alleles of which determine avirulence on rice cultivars CO39, M201, and Yashiro-mochi, respectively. Interestingly, avirulence alleles of AvrI-C039, Avrl-M2OI and Avr l YAM0 were inherited from the parent strain 4091-5-8, which is a nonpathogen of rice. Middle repetitive DNA sequences (“MGR sequences”), present in approximately 40-50 copies in the genome of the rice pathogen parent, and in very low copy number in the genome of the nonpathogen of rice, were used as physical markers to monitor restoration of the rice pathogen genetic background during introgression of fertility. The introgression of highest levels of fertility into the most successful rice pathogen progeny was incomplete by the sixth generation, perhaps a consequence of genetic linkage between genes for fertility and genes for rice pathogenicity. One chromosomal DNA segment with MGR sequence homology appeared to be linked to the gene Avr l -C039 . Finally, many of the crosses described in this paper exhibited a characteristic common to many crosses involving M. grisea rice pathogen field isolates. That is, pigment-defective mutants frequently appeared among the progeny. T HE haploid heterothallic ascomycete Magnaporthe grisea (Hebert) Barr [anamorph, Pyricularia grisea Sacc., formerly Pyricularia oryzae Cavara (ROSSMAN, HOWARD and VALENT 1990)] includes pathogens of many grasses, although individual field isolates are limited to one or a few host species (KATO 1978; Ou 1985; MACKILL and BONMAN 1986). Strains of M . grisea cause the serious disease of rice called blast (OU 1985). Among isolates pathogenic to rice, hundreds of “races” have been identified as defined by their virulence or avirulence toward particular rice cultivars. Cultivars of rice that differ from one another by the presence or absence of dominant blast resistance genes have been developed (YAMADA et al. 1976), but the resistance of any one of these cultivars is effective only against certain races of the pathogen. The fungus shows a high degree of variability in the Genetics 127: 87-101 (January, 1991) field; new races frequently appear with the ability to infect previously resistant rice cultivars. Individual M. grisea strains produce distinctive and reproducible symptoms on particular host plants under a given environmental regime (LATTERELL 1975; Ou 1985). Pathogenicity is a complex phenotype, involving such distinct components as infection efficiency (determines lesion number), rate of lesion development, extent of colonization (determines lesion size), and efficiency of sporulation. Genetic analysis of host-pathogen interactions, accomplished mainly in systems with obligate fungal pathogens such as the rusts and the mildews, confirms the complexity of the interactions (DAY 1974; CRUTE 1986). Quantitative pathogenicity factors determine at least some aspects of the interaction between the pathogen and its host (CATEN et al. 1984; CRUTE 1986). “Major genes,” 88 B. Valent, L. Farrall and F. G. Chumley single genes with large effects on a host-pathogen interaction, are also common. Except for preliminary reports of genetic analysis of host specificity (YAEGASH1 and ASAGA 1981; LEUNG et al. 1988; ELLINGBOE, WU and ROBERTSON 1990), little is known of the genetic basis of M . grisea pathogenicity and host specificity. The ability of M. grisea strains to undergo genetic crosses is also a complex phenotype. As with other heterothallic ascomycetes, compatibility for mating is governed by alleles of the mating type locus, M a t l . At least one parent must be female fertile, with the capacity to produce perithecia, and the two parents together must be able to produce asci with viable ascospores. The fertility of M. grisea field isolates ranges from total sterility (inability to mate with any other strain), through female sterility (ability to mate only as a male), to full fertility (ability to mate as a male or as a female; VALENT et al. 1986). Even among hermaphroditic strains there appears to be a fertility continuum. Hermaphroditic strains of relatively low fertility cross with only a few other strains and produce few perithecia, and strains of the highest fertility cross with many other strains and produce large numbers of perithecia. Viability of ascospores for crosses between field isolates ranges from less than 1 % up to 10% in exceptional cases. The level of fertility of field isolates generally correlates with host specificity. For example, field isolates that infect weeping lovegrass, goosegrass (Eleusine indica) or finger millet (Eleusine coracana) typically are hermaphrodites that can mate to yield numerous viable ascospores. Hundreds of field isolates that infect rice have been tested and found to be female sterile (VALENT et al. 1986; B. VALENT, unpublished results), and those rice pathogens that do mate with hermaphroditic strains infecting grasses other than rice usually produce ascospores with poor viability. A recent unique exception to the rule of female sterility among rice pathogen field isolates is a strain, Guy-1 1, isolated by J. L. NOTTEGHEM in French Guyana (LEUNG et al. 1988). Rice pathogens of both mating types occur in the field (KATO and YAMAGUCHI 1982; YAEGASHI and YAMADA 1986). In addition to host specificity and level of fertility, M. grisea field isolates differ in middle repetitive DNA sequences. That is, rice pathogens from around the world contain a family of repeated DNA sequences, “MGR sequences,” that appear to be absent from or present in low copy number in field isolates that infect grasses other than rice (HAMER et al. 1989). MGR sequences are interspersed with single copy DNA on all chromosomes of rice pathogens. The presence of MGR sequences only in rice pathogens suggests that rice pathogens in nature are genetically isolated from nonpathogens of rice in the same geographic area. Breeding for improved fertility with M . grisea weeping lovegrass and goosegrass pathogens has produced strains with excellent properties for genetic analysis (VALENT et al. 1986; VALENT and CHUMLEY 1987). The series of crosses reported here originated with the dual purposes of determining the genetic differences between a pathogen and a nonpathogen of rice and of introgressing high fertility into laboratory strains that infect rice. The development of highly fertile laboratory strains that infect rice as well as other grass species should permit rigorous genetic analysis of host species specificity, host cultivar specificity and mechanisms of pathogenesis in M . grisea. The segregation of pathogenicity, virulence, fertility and MGR markers was investigated through six generations of backcrosses. Segregation patterns suggested simple Mendelian inheritance of genes that control specificity ( i .e . , virulence) toward rice cultivars as well as polygenic inheritance of factors that determine pathogenicity toward rice. Hermaphroditic fertile rice pathogens were generated, although no highly successful pathogens showed the very high fertility characteristic of laboratory strains that infect grasses other than rice. Most crosses reported here exhibited instability at one particular genetic locus, BUFl (CHUMLEY and VALENT 1990). All other markers observed, including mating type, segregated normally. Detection of genetic linkage between a particular MGR sequence and the avirulence gene Avr l -C039 suggested that MGR-associated restriction fragment length polymorphisms (RFLPs) might serve as physical markers for cloning genes of interest by chromosome walking. MATERIALS AND METHODS Strains: Most strains described in this paper (see Table 1) were derived from three field isolates. The fertile laboratory strain 4091-5-8 was a progeny strain from a cross between a Japanese field isolate that infects weeping lovegrass, K7679 (Mutl-2) , and a Japanese field isolate that infects finger millet and goosegrass, WGG-FA40 ( M a t l I ) (VALENT et al. 1986). Both field isolates were generously provided by HIROSHI YAECASHI of the Tohoku National Agricultural Experiment Station in Akita, Japan. Strain 4091-5-8 is a pathogen of weeping lovegrass and goosegrass, but it does not produce visible symptoms on any of the fifteen cultivars of rice tested. The field isolate 0-135, which infects rice and weeping lovegrass, was collected by BV in 1985 at the China National Rice Research Institute in Hangzhou, Zhejiang, China. A rice-pathogen laboratory strain, 6043 (LEUNG et ul. 1988), was generously provided by HEI LEUNC, Washington State University, Pullman. The laboratory strains 4360-19-1 and 4360-15-1, both pathogens of rice, were progeny from other crosses in our laboratory and will be described elsewhere. The strains commonly used in mating type tests were 4091-5-8 (Mutl-2, g) and 41 36-4-3 ( M a t l l , Ep) (VALENT and CHUMLEY 1987). Media and permanent storage of the pathogen: Isolates were grown and crossed on oatmeal agar plates, prepared as follows. Fifty grams of rolled oats were heated in 500 ml M. grisea Host Specificity Genes