• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of geneticsGeneticsCurrent IssueInformation for AuthorsEditorial BoardSubscribeSubmit a Manuscript
Genetics. Jun 2005; 170(2): 613–630.
PMCID: PMC1450392

Origins of Host-Specific Populations of the Blast Pathogen Magnaporthe oryzae in Crop Domestication With Subsequent Expansion of Pandemic Clones on Rice and Weeds of Rice


Rice, as a widely and intensively cultivated crop, should be a target for parasite host shifts and a source for shifts to co-occurring weeds. Magnaporthe oryzae, of the M. grisea species complex, is the most important fungal pathogen of rice, with a high degree of host specificity. On the basis of 10 loci from six of its seven linkage groups, 37 multilocus haplotypes among 497 isolates of M. oryzae from rice and other grasses were identified. Phylogenetic relationships among isolates from rice (Oryza sativa), millet (Setaria spp.), cutgrass (Leersia hexandra), and torpedo grass (Panicum repens) were predominantly tree like, consistent with a lack of recombination, but from other hosts were reticulate, consistent with recombination. The single origin of rice-infecting M. oryzae followed a host shift from a Setaria millet and was closely followed by additional shifts to weeds of rice, cutgrass, and torpedo grass. Two independent estimators of divergence time indicate that these host shifts predate the Green Revolution and could be associated with rice domestication. The rice-infecting lineage is characterized by high copy number of the transposable element MGR586 (Pot3) and, except in two haplotypes, by a loss of AVR-Co39. Both mating types have been retained in ancestral, well-distributed rice-infecting haplotypes 10 (mainly temperate) and 14 (mainly tropical), but only one mating type was recovered from several derived, geographically restricted haplotypes. There is evidence of a common origin of both ACE1 virulence genotypes in haplotype 14. Host-haplotype association is evidenced by low pathogenicity on hosts associated with other haplotypes.

WIDELY distributed, high-density populations, such as those of cultivated crops or humans, are large targets for host shifts by parasites. After crossing host species boundaries, successful shifts culminate with specialization and genetic divergence on the new host (Antonovics et al. 2002). If the parasite crosses a species barrier only once, virulence determinants would derive from this common origin. In contrast, if the target host species is invaded in multiple, independent host shifts, virulence could have multiple, fundamentally different determinants among host populations (MacLeod et al. 2001).

After successful establishment on a new host, especially one that is abundant and relatively genetically uniform, expansion of pandemic clones of the pathogen commonly follows (Tibayrenc et al. 1991; Maynard Smith et al. 1993; Brasier and Buck 2001). While we would expect that virulence determinants in a pandemic clone would remain associated and modified only through mutation, any genetic exchange with recombination would shuffle these determinants into novel combinations across the parasite population. Within large pathogen populations, recombination would increase the rate at which multiple virulence factors could combine to generate “super parasite” genotypes (Crow and Kimura 1970).

The rice blast pathosystem is a model system not only for plant-pathogen interactions (Valent 1990; Talbot 2003), but also for host shifts on a relatively contemporary timescale among grass species. We are mainly interested in the host shift to rice. While the genetic basis of the host-pathogen interaction fits the criteria for coevolution, the time frame does not appear to be consistent with cospeciation. The earliest report of a blast-like disease of rice (Oryza sativa) in the literature is 1637 in China and, in the following years, in Japan (Ou 1987). The degree of phylogenetic divergence among grass hosts is large compared to the phylogenetic divergence among pathogen genotypes from different hosts (Couch and Kohn 2002). Within the grasses are two major clades estimated to have originated at least 55 million years ago, comprising ~10,000 species (Crepet and Feldman 1991; Grass Phylogeny Working Group 2001; Bremer 2002). Genotypes of the pathogen species are highly host specific, yet in sum infect grasses from both major clades, suggesting a rapid, recent radiation on grasses. Given this time scale, we see domestication as the best opportunity for initial host shifts to rice, as large populations of cultivated rice are likely targets for pathogenic migrants from other hosts.

There is now good archeological evidence and more consensus among experts in support of an early domestication of O. sativa from its wild relative O. rufipogon in the middle Yangtze valley in China ~7000 years before present (BP) (Crawford and Shen 1998; Higman and Lu 1998; Vaughan and Morishima 2003). Chang (1976) presented evidence for early rice cultivation in the Indian Himalayas 2450–3950 BP and in China 5230 BP, proposing multiple, independent, concurrent domestication events across the region stretching from the foothills of the Indian Himalayas to southwestern and southern China. Separate domestication of the two varieties indica and japonica (Second 1982) and, alternatively, domestication of indicas from wild rice and derivation of japonicas from indicas (Oka 1988) have been proposed, but the evidence is equivocal (Lu et al. 2002; Cheng et al. 2003; Yamanaka et al. 2003).

The expansion of rice cultivation to northern China, Korea, and Japan brought two domesticated hosts of the blast pathogen, rice, and Italian millet (Setaria italica) into close contact in crop complexes between 7000 BP in China and 3000 BP in Korea and Japan (Crawford 1992; Higman and Lu 1998). Domesticated rice spread throughout Asia, the Mediterranean (~2300 BP), Madagascar (~1500 BP), Southern Europe (500–600 BP), and, more recently, as a result of European colonization, to Africa and North and South America (~300–500 BP) (Oka 1988). Introduction of O. sativa to Africa also brought it in contact with the more recently domesticated (~3500 BP), closely related, West African domesticated species, O. glaberrima (Chang 1976). Intercontinental dispersal of rice genotypes continues by commercial trade, official distribution of new rice cultivars from centers of plant breeding, and unofficial (in the pocket) exchange. Dispersal of rice can also disperse rice blast (Long et al. 2001). Even if rare, seed transmission has probably introduced, and reintroduced, rice blast to several countries.

The Green Revolution of the 1960s popularized the first high-yielding, semidwarf rice cultivars, abruptly changing the diversity and frequency of rice genotypes throughout Asia (Khush 2001). The responsiveness of these varieties to nitrogen fertilizer led to increased inputs that increased the susceptibility of rice to blast and set the stage for the invasion of novel blast genotypes into modern rice (Ou 1987). Existing blast pathogen populations may have undergone a bottleneck with reduction in the diversity of rice genotypes. Monoculture conditions would have facilitated the expansion of pandemic blast clones. While the Green Revolution may have caused a shift in blast populations, the “revolution” was incremental. It is likely that some blast genotypes predating the Green Revolution, or their descendants, still exist.

At either end of the timescale for invasion by the blast pathogen—7000 or 35 years ago—models would predict that the high frequency and density of rice crops, compared to other hosts, would favor specialization for rice among invading parasite genotypes (Fry 1996; Whitlock 1996). Host shifts are the most plausible explanation for the epidemics of blast in the mid-1980s on wheat (Triticum aestivum) in Brazil and in the early 1990s on perennial ryegrass (Lolium perenne) in Pennsylvania (Landschoot and Hoyland 1992; Oh et al. 2002).

The biological characteristics of the blast pathogen should enable invasions leading to host shifts. Rice blast, and gray leaf spot of grasses, is caused by Magnaporthe oryzae, a taxon recently segregated from M. grisea, a species complex of haploid ascomycetous fungi (Couch and Kohn 2002). Hosts of the M. grisea-M. oryzae species complex include a number of weeds of rice, such as torpedo grass (Panicum repens) and cutgrass (Leersia hexandra), and cultivated grasses such as S. italica (Ou 1987). Most reproduction is asexual by mitotically produced spores (conidia) from lesions on above-ground plant parts. A single lesion can produce 2000–6000 conidia/day for up to 14 days, with multiple cycles of infection and reproduction during one growing season. Some conidia may disperse beyond the usual 1-m range (Ou 1987). The initial stages of infection are relatively nonspecific and could permit colonization of novel host species (Hamer et al. 1988; Peng and Shishiyama 1988; Heath et al. 1990, 1992). Although sexual reproduction occurs in vitro (isolates from Eleusine tend to show high fertility), the morphological structures, perithecia, are not observed in nature (Hebert 1971; Silué and Nottéghem 1990; Zeigler 1998). There is indirect evidence for genetic exchange and recombination in some rice-infecting populations (Valent et al. 1986; Kumar et al. 1999; Kato et al. 2000). Predominantly asexual reproduction would facilitate the emergence of pandemic clones specialized for pathogenicity against rice cultivars.

Overcoming both general and specific plant defenses potentially limits host-species shifts. Rice presents both partial and complete resistance toward M. oryzae (Marchetti 1983; Marchetti et al. 1987; Wang et al. 1989, 1994). Partial resistance reduces the overall disease severity and is usually controlled by multiple loci (Nottéghem et al. 1994). Complete resistance toward some genotypes of M. oryzae is conferred by major genes in rice (R-genes) that operate in a gene-for-gene manner (Silué et al. 1992; Valent and Chumley 1994). Here, a plant resistance gene functions by recognizing pathogen genotypes carrying a corresponding avirulence (AVR) gene. Mutation or deletion in AVR genes permits the pathogen to evade detection by the plant, potentially allowing infection. A number of AVR genes recognized by rice R-genes have been identified in M. oryzae (Nottéghem et al. 1994; Farman and Leong 1998; Orbach et al. 2000; hnert et al. 2004). R-genes are generally regarded to be determinants of host cultivar specificity, i.e., resistance in some but not all rice varieties or cultivars (Ronald 1997; Jia et al. 2000). AVR/R-gene interactions, however, can also be determinants of host-species specificity, e.g., virulence against Eragrostis curvula (weeping lovegrass) mediated by PWL2 (Sweigard et al. 1995). AVR genes associated with pathogen genotypes from hosts other than rice may be recognized by rice R-genes (Valent et al. 1991; Farman et al. 2002), preventing establishment on rice.

Consistent with the impediments to overcoming host defenses, the available molecular evidence suggests a single acquisition of pathogenicity on rice (Hamer et al. 1989; Borromeo et al. 1993; Shull and Hamer 1994). High copy numbers of the repetitive elements MGR586 (Pot3) and MGR583 characterize rice-infecting isolates (Hamer et al. 1989). Exceptions to this general pattern may represent additional host shifts with limited frequency and distribution (Leong et al. 1994). The clonal structure observed in rice-infecting populations (Levy et al. 1991; Xia et al. 1993; Zeigler 1998) supports the theory that contemporary clonal lineages derive from a single ancestral, sexual, phylogenetic lineage. Short of having in hand archeological samples of infected plants, which do not appear to exist, a chronology of host shifts and specialization in M. oryzae can best be inferred phylogenetically on the basis of contemporary samples.

We explored the origin of rice-infecting populations of M. oryzae with a multilocus phylogenetic analysis of 497 isolates of M. oryzae from culture collections and population samples. The origin was inferred for pandemic populations of M. oryzae established on rice, as well as the pattern of host shifts among closely related haplotypes from rice and other grasses. To test the hypotheses of one invasion of rice, parsimony, maximum-likelihood (ML), and Bayesian methods were implemented independently. From this, the timing of the host shift to rice was estimated. Several independent methods detected and localized conflict indicative of recombination. Detecting conflict was necessary (i) to implement statistical parsimony, (ii) to determine the combinability of loci, and (iii) to test the hypothesis of pandemic clonality associated with host shifts to rice. The phylogenetic framework was also used as a basis for assessing the role of AVR genes in determining host-species specificity and for determining the distribution of mating types. Pathogenicity tests were performed to assess host specificity in isolates representing phylogenetic lineages most closely related to rice-infecting populations.


Fungal isolates:

Isolates used in this study included both population samples and isolates from culture collections. Isolates collected by B. Couch were derived from single spores from single, sporulating lesions. Leaves containing lesions were incubated overnight in a moist chamber at room temperature to induce sporulation. Spores from a single lesion were streaked on water agar (Difco, Sparks, MD; 15 g agar/liter of distilled H2O) and incubated overnight to allow spores to germinate. Single, germinating spores were subcultured on potato dextrose agar (PDA; Difco). Isolates were maintained on PDA. Storage of isolates was on filter paper inoculated and colonized by the fungus and then dried and stored in a desiccator at −20°.

DNA extraction:

Isolates were grown in either 10 or 20 ml of liquid Fries medium (30 g sucrose, 5 g ammonium tartrate, 1.0 g NH4NO3, 0.5 g MgSO4•7 H2O, 1.0 g KH2PO4, 0.1 g NaCl, 0.13 g CaCl2•2H2O, 0.02 g FeSO4•7H2O, 1.0 g yeast extract/liter of H2O) on a rotary shaker for 3–5 days. Mycelium was harvested by filtration and dried under a vacuum. DNA was extracted from ~15 mg of dried mycelium following the procedure of Zolan and Pukkila (1986).

Marker development:

Genomic regions containing DNA sequence polymorphisms were identified by direct sequencing of PCR products amplified from 19 reference isolates from five hosts: Eleusine coracana (goosegrass; 2 isolates), Eragrostis curvula (weeping lovegrass; 2), perennial ryegrass (2), rice (11), and Setaria millets (2). A total of 26 regions were screened. Eleven genomic regions containing portions of genes were amplified using previously published PCR primers: histone 3, histone 4, β-tubulin, aldehyde dehydrogenase, phosphate permease (Glass and Donaldson 1995), translation elongation factor 1 α, calmodulin, chitin synthase 1, actin, a ras protein, and the intergenic spacer region of the nuclear ribosomal DNA repeat (Carbone and Kohn 1999). Five genomic regions containing portions of genes were amplified using PCR primers designed from M. grisea sequences accessioned to GenBank: a xylanase gene (xyn22), a hydrophobin (MPG1), a nitrogen regulatory protein (NUT1), an ergosterol biosysnthesis gene (ERG2), and a G-protein α-subunit (MAGA). Eleven additional genomic regions were amplified using PCR primers designed from M. grisea genomic clones accessioned to GenBank that lacked significant homology to known genes. Primers designed in this study that amplify regions containing polymorphisms within the set of 21 reference isolates are in Table 1. Primers were designed with the aid of GENEWORKS (IntelliGenetics, Mountainview, CA). Polymorphic regions were localized to chromosomal location by performing BLASTN searches against the M. grisea genome sequence, assembly 2.1, produced by the Broad Institute, Massachusetts Institute of Technology (http://www.broad.mit.edu/annotation/fungi/magnaporthe).

Polymerase chain reaction primers designed fromM. grisea sensu lato genes or genomic clones

Polymerase chain reaction and sequencing:

PCR reactions were performed using either a Perkin-Elmer 9600 or a Perkin-Elmer 9700 thermocycler (Perkin-Elmer, Foster City, CA). PCR reactions contained 10 mm Tris-HCl (pH 8.3), 50 mm KCl, 200 μm of each dNTP, 0.5 μm of each oligonucleotide primer, 0.5 units AmpliTaq DNA polymerase (Roche Molecular Systems, Branchburg, NJ), 2.0 mm MgCl2, and 10 μl of a 1:200 dilution of genomic DNA in a total volume of 20 μl. Thermocycling conditions were as follows: 95° for 8 min, 35 cycles of 95° for 15 sec, 55° for 20 sec, 72° for 60 sec, followed by 72° for 5 min and a 4° hold. PCR products were purified using QIAquick spin columns (QIAGEN, Mississauga, ON) following the manufacturer's instructions. Sequencing reactions were performed using an ABI Prism Big Dye terminator cycle sequencing kit (Applied Biosystems, Foster City, CA), following the manufacturer's instructions with the exception that only 25% of the recommended volume of terminator mix was used. Sequencing reactions were separated on an ABI 310 Genetic Analyzer. Electropherograms were interpreted with SEQUENCE ANALYSIS SOFTWARE v. 3.3 (Applied Biosystems). Forward and reverse sequences were imported into SEQUENCHER v. 4.1 (Gene Codes, Ann Arbor, MI), assembled into contigs, and visually checked for errors.

Single-strand conformation polymorphisms:

Single-strand conformation polymorphisms (SSCPs) (Orita et al. 1989) were used to screen isolates for DNA sequence polymorphisms in 10 of the genomic regions identified as polymorphic: actin (ACT), BAC6, β-tubulin (βt-1), calmodulin (CAL), CH7-BAC7, CH7-BAC9, chitin synthase 1 (CHS), translation elongation factor 1α (EF-1α), MPG1, and NUT1. For SSCP screening, the βt-1 region was divided into two overlapping fragments. The 5′ region was amplified using the βt-1a primer (Glass and Donaldson 1995) and βt1-256R (5′-GGTCTGGATGTTGTTGGGGATCC-3′). The 3′ region was amplified using the βt-1b primer (Glass and Donaldson 1995) and βt1-271F (5′-TGAAGGAGGTCGAGGACCAG-3′). PCR reactions for SSCP analysis were performed in a 12-μl reaction volume using primer pairs end labeled with 33P using T4 kinase (GIBCO BRL, Gaithersburg, MD), following manufacturer's instructions. Following thermocycling, 4 μl of stop dye [95% formamide, 10 mm EDTA (pH 8.0), 0.1% bromophenol blue, and 0.1% xylene cyanol] was added to each reaction and samples were heated to 95° for 8 min and chilled to 4°. One microliter of each sample was loaded on a 30 cm × 39 cm × 0.4 mm vertical polyacrylamide gel (6% polyacrylamide, 0.12% bisacrylamide, and 0.5 × tris-borate-EDTA). Gels were prepared with or without the addition of glycerol to a final concentration of 7%. Samples were separated at 150 V for 19–30 hr. Gels were blotted on filter paper and dried. Autoradiography was performed with Kodak BioMax film (Eastman Kodak, Rochester, NY) at room temperature. Three representatives of each allelic class were then sequenced to confirm their identity. For alleles with a frequency of less than three, all representatives were sequenced.

Phylogenetic analysis:

Sequences for each haplotype were aligned using CLUSTAL X (Thompson et al. 1994) and edited manually using SEQUENCE ALIGNMENT EDITOR (Se-Al) v.1.0 α1 by Andrew Rambaut (http://evolve.zoo.ox.ac.uk/software.html). For coding indels, the complex indel-coding method of Simmons and Ochoterena (2000) was employed.

The basic units for phylogenetic analysis were multilocus haplotypes. Individuals sharing a haplotype had the same SSCP alleles for all 10 genomic regions screened in this study. Individuals were clustered into multilocus haplotypes using the neighbor-joining clustering method implemented in PAUP* v.4.0b10 for Macintosh (Swofford 2002). Since each SSCP allele was characterized by sequencing, each haplotype was represented by a unique multilocus DNA sequence.

Phylogenetic relationships among haplotypes were inferred using maximum-parsimony (MP) and ML methods implemented in PAUP* (Swofford 2002). Data from each genomic region were first analyzed alone, and then data for all regions were combined in a single analysis. For the MP analysis, heuristic searches were performed using starting trees obtained by stepwise addition, a simple addition sequence, and the branch-swapping algorithm of tree-bisection-reconnection. Visual inspection of branch-length tables, lists of character changes, and homoplasy matrices generated in PAUP* were used to locate parts of trees where there was phylogenetic conflict and to identify the haplotypes responsible for this conflict. Partitioned Bremer support values were calculated using the program TREEROT v. 2 (Sorenson 1999). For the PAUP tree searches specified by the TREEROT command file, 40 random addition replicates were performed for each search and the MaxTrees limit was set to 5000. Network reconstruction using statistical parsimony (Templeton et al. 1992) was performed using the program TCS v. 1.13 (Clement et al. 2000). For ML analysis, heuristic searches were performed using the Hasegawa, Kishino, and Yano (HKY) model of nucleotide substitution, identified using MODELTEST v. 3.06 (Posada and Crandall 1998) with empirically determined nucleotide frequencies and a molecular clock not enforced.

Bayesian analysis was performed using MRBAYES v. 3.0B3 (Huelsenbeck and Ronquist 2001). The settings used for the Metropolis Coupled Markov Chain Monte Carlo sampling were as follows: trees were sampled every 1000 generations from four chains run for 10,000,000 generations with a temperature setting for the heated chains of 0.2. The standard 4 × 4 model of DNA substitution was used. This gave an output of 10,000 trees, 999 of which were discarded as the “burn-in” for the Markov chain, leaving 9001 trees for estimation of the posterior probabilities for tree topologies.

Phylogenetic compatibility among polymorphic sites was visualized by generating a compatibility matrix using RETICULATE (Jakobsen and Easteal 1996). Split decomposition (Bandelt and Dress 1992) was used to visualize the tree-like and reticulate relationships among haplotypes. The program SPLITSTREE v. 2.4 (Dress et al. 1996) was used to construct a splits graph for the haplotypes. Iterations were performed until all haplotypes were resolved. The statistical support for each split was evaluated using bootstrapping (100 replicates).

Dating divergence:

The assumption of a molecular clock was tested using the 1-d.f. method of Tajima (1993). Two independent methods were used to estimate the divergence time for the host shift from Setaria millet to rice (Thorne et al. 1998; Sanderson 2002). First, in a Bayesian analysis the a priori estimate of the mean time between the tip and root of the tree was set to 10,000 years before present; according to the program manual, the standard deviation for this estimate is set equal to the mean (http://statgen.ncsu.edu/thorne/multidivtime.html). The burn-in value for the Markov chain was set at 100,000 cycles and divergence time estimates were based on 10,000 samples taken at intervals of 100 cycles from a single Markov chain. The burn-in was estimated by running the program with the burn-in set to zero, graphing node age estimates, and estimating the point at which the Markov chains converged on a “stable” value. In the second analysis, with the aid of r8s (http://ginger.ucdavis.edu/r8s/) the divergence time was estimated using the Langley-Fitch method with the truncated Newton optimization algorithm. Two separate runs were performed. In both runs the minimum age for the host shift from Setaria millet was set at 30 years BP (the age of our oldest sample from haplotype 10). For the first run, the root age was fixed at 10,000 years BP. In the second run, the root age was constrained to be between 10,000 and 40 years BP.

PCR test for presence or absence of AVR genes:

The presence or absence of three AVR genes, AVR-IRAT7 (ACE1), AVR-Pita, and AVR-Co39, was assessed using a plus/minus PCR test. A total of 59 isolates were screened. This group consisted of representatives from 28 haplotypes. The strains that were screened and their haplotype designations, indicated in parentheses, were: (1) G8, (2) G17, (3) VII-765, (3) 1152, (4) GF2-1, (4) G48, (5) Arcadia, (6) SJ5-1-2, (6) TT15-1, (7) SA1-1-1, (7) SA7-2-2, (8) SA2-1-1, (9) TT10-2, (10) BK19, (10) Guy11, (10) 75A36, (10) V-10, (11) 92A20, (11) 93M5, (11) 93T7, (11) A119, (12) BK6, (12) BK8, (12) ML39, (12) V33, (13) ML2, (13) ML7, (14) 75L23, (14) ML56, (14) ML91, (14) V-2, (15) Os/Or1-1-1, (15) Os/Or7-1-1, (15) Os/Or-30-1-1, (16) C9227-6, (16) DH1-2, (16) DH2-2, (16) DR1-2-1, (17) ML8, (18) DR42-3-2, (18) SA5-1-1, (19) GrF1, (19) McConnell, (20) 330, (20) 365, (21) Br81, (22) SAG00T12, (22) SAG00T6, (23) SAG00T3, (23) SAG00T4, (24) G199, (25) V-749, (25) VII-758, (26) V58, (26) V60, (27) VII-739, (27) VII953-1, (28) 1122, and (28) 94-115. For each isolate, PCR amplifications were performed using primers specific for each of the AVR genes. To ensure that the PCR primers amplified the expected product, representative amplicons were directly sequenced. For both AVR-Pita and AVR1-Co39, two independent amplifications were performed. As a positive control for DNA quality, amplifications were also performed using the primers for the actin region. Lack of amplification was interpreted as evidence for loss of function of the AVR gene due to deletion or rearrangement of the gene.

PCR primers were designed from GenBank sequences for AVR-Pita (AF207841) and AVR1-Co39 (AF463528). Amplifications using the AVR-Pita primers Pita-F (5′-CGCCTTTTATTGGTTTAATTCG-3′) and Pita-R (5′-CCTCCATTCCAACACTAACG-3′) were performed using an annealing temperature of 60°. Amplifications using the AVR-Co39 primers Co39F-327 (5′-TGCGATATAATGGCCAAACA-3′) and Co39R-800 (5′-GACCGATCTGTCGGGAAGTA-3′) were performed using an annealing temperature of 55°.

The distribution of ACE1 (hnert et al. 2004) was assessed using allele-specific PCR (Fudal 2004). The two major virulent ACE1 genotypes (ACE1-vir1, ACE1-vir2) and the widely distributed ACE1 avirulent allele (ACE1-avr) could be detected using two different PCR reactions, one for the virulent alleles and one for the avirulent allele, which were performed in replicate using the following primers: the ACE1-avr allele was specifically amplified using the primer pairs Guy11-X438+ (5′-GTTTCCGACATACTTTGCGCCC-3′) and Guy11-i1O- (5′-GAGCCGACGTAGAGTTTTGGG-3′). The ACE1-vir1 allele was specifically amplified using the primer pairs CM28-X438+ (5′-TTTCCCGACCTACTTTGCACCG-3′) and CM28-i1O- (5′-GACCCCACGTACAGTTTGGCA-3′). The ACE1-vir2 genotype was characterized by two PCR reactions performed with CM28 and Guy11 primers, respectively; the isolate has the ACE1 vir2 genotype if there is an amplicon for both reactions. Representative isolates with known ACE1 genotypes (CM28, ACE1-vir1; PH14, ACE1-vir2; Guy11, ACE1-avr) were included as positive controls and water as a negative control in all replicates.

Mating-type frequencies among haplotypes:

Clones in the plasmid vectors SK43 and SK44 of the mating-type idiomorphs (Kang et al. 1994) were purified using an alkaline lysis procedure (Sambrook et al. 1989) and were then digested with EcoRI and BglI following the manufacturer's instructions (Life Technologies, GIBCO BRL, Burlington, ON) to liberate the mating-type-specific sequences to be used as probes. These probe sequences were separated from plasmid DNA by gel electrophoresis and purified using a QIAquick gel extraction kit (QIAGEN) following the manufacturer's instructions. A 2.4-kb fragment from SK43 contained the MAT1-1-specific sequence, and a 2.3-kb fragment from SK44 contained the MAT1-2-specific sequence. Probes were labeled with 32P using a nick translation kit (Invitrogen BV, Groningen, The Netherlands). Hybridization was performed at 60°. Autoradiography was performed with a Storm PhosphorImager (Amersham Bioscience, Piscataway, NJ).

For each isolate of M. oryzae, 10 μl of mini-prep DNA was digested with EcoRI (MBI Fermentas, Flamborough, ON) and separated on a 24-cm, 0.8% agarose gel for 45 min at 4 V/cm. Following electrophoresis, DNAs were blotted on GeneScreen Plus hybridization transfer membranes (Perkin-Elmer Life Sciences, Boston) following the manufacturer's instructions.

Pathogenicity testing:

Two sets of pathogenicity tests were performed under standard inoculation conditions for this pathogen (Valent et al. 1991) or (Sallaud et al. 2003) by B. Valent (Kansas State University) and by D. Tharreau (Centre de Coopération Internationale en Recherche Agronomique pour le Développement), respectively. In the first test, isolates assayed were representatives of haplotypes 6 (DR 22-1-1, isolated from P. repens), 10 (Guy 11 from rice), 18 (DR 52-2-1 from L. hexandra), and 19 (McConnell from S. viridis). A highly pathogenic field strain from China, O-137, was included as a positive control. Inoculations were performed on rice, variety YT16 (a highly susceptible variety), and on P. repens (accession no. 338659 from Morocco obtained from the USDA Plant Genetic Resources Conservation Unity, Griffin, GA). In the second test, isolates assayed were representatives of haplotypes 19 (Gr F1, Gr F2, Gr F3, Gr F4, and McConnell), haplotype 5 (Arcadia from S. viridis), haplotype 4 (Gr F2-1 from S. faberii), and haplotypes 18 (Dr 51-2-3 from L. hexandra). The rice-infecting isolate PH14 (PO6-6) was used as positive control. The rice cultivars, listed from most susceptible to most resistant, were Sariceltik, Maratelli, IR1529, and IR 64. Disease severity was scored on a scale from 0 to 5 on the basis of lesion types defined by Valent et al. (1991). Strains producing lesions of types 0 and 1, which do not sporulate, are considered nonpathogenic or avirulent. Strains producing lesions of type 2 are very weakly pathogenic although potentially capable of sporulating. Strains producing lesion types 3–5, which are capable of sporulating under conducive conditions, are considered pathogenic (Valent et al. 1991). Strains that produce type 5 lesions are considered highly pathogenic on the particular host genotype on which the reaction is observed. Some strains produce a variety of lesion types on a single plant. This reaction is termed “mesothetic” and scoring reflects the predominant lesion type.


Marker development:

Of the PCR primers tested, 10 amplified regions were polymorphic within the set of reference isolates: βt-1, EF-1α, CAL, CHS, ACT, a hydrophobin (MPG1), a nitrogen regulatory protein (NUT1), BAC6, CH7-BAC7, and CH7-BAC9. These regions are distributed throughout the genome and were localized to six of the seven M. grisea chromosomes as indicated in parentheses: CAL (II), CHS (III), EF-1α (III), ACT (IV), BAC6 (IV), βt-1 (VI), CH7-BAC7 (VII), CH7-BAC9 (VII), MPG1 (VII), and NUT1 (VII).

Haplotype identification and DNA sequence polymorphisms:

A total of 37 unique multilocus haplotypes were identified by SSCPs and verified by sequencing in 497 isolates of M. oryzae (Table 2). A total of 83 base substitutions and 13 indels were identified in 3266 bp of aligned sequence. Two polymorphic sites, a C-G transversion in CH7-BAC7 and a 1-bp indel in EF1, could not be resolved using SSCPs and were excluded from all analyses. Of the 96 polymorphic sites, 63 were phylogenetically informative, i.e., shared by two or more haplotypes. The other 33 sites were unique to one haplotype.

Summary of the frequency, host, and geographic association of the 37 multi-locus haplotypes identified inM. oryzae

Parsimony analysis of haplotypes:

Maximum-parsimony phylogenies inferred by analyzing data for each of the 10 loci separately contained very little homoplasy. Homoplasy occurs when a character state present in two taxa is not derived from a single common ancestor. In phylogenetic analyses of DNA sequences, homoplasious nucleotide sites are those in which identical mutations are inferred to have occurred multiple times. Single, most parsimonious, minimum-length trees were inferred using base substitutions from 7 of the 10 loci: ACT (2 steps), BAC6 (3 steps), BT1 (6 steps), CAL (5 steps), CH7-BAC7 (22 steps), CHS (2 steps), EF1 (4 steps), and NUT1 (3 steps). A single, most parsimonious tree of 24 steps, 1 step longer than the minimum-length tree, was inferred for the MPG1 region. The CH7-BAC9 region contained the most homoplasy. When parsimony analysis was conducted using only base substitutions, three most parsimonious trees of 14 steps were inferred, each 2 steps longer than the minimum-length tree. Two base positions were found to be homoplasious. These base positions are part of a group of three adjacent polymorphic nucleotide sites, positions 2112–2114 in the multiple sequence alignment, that may be a mutational hotspot subject to multiple hits.

Data from each locus analyzed individually provided very little phylogenetic resolution among haplotypes because many haplotypes shared identical sequences for some loci. To increase the resolution, data from all loci were combined in a single maximum-parsimony analysis. Using only base substitutions, 160 most parsimonious trees (MPTs) were inferred, each with a length of 118, 36 steps longer than the minimum-length tree (82 steps). Each tree had a consistency index (CI) of 0.695. One of the 160 MPTs is shown (Figure 1). Maximum-likelihood analysis was also performed: maximum likelihood resolved identical branching to those parts of the parsimony tree that were free of phylogenetic conflict. When indels were included in the analysis, 80 MPTs were inferred. Each tree had a length of 142 steps, 44 steps longer than the minimum-length tree (98 steps). Each tree had a CI of 0.690. Additional trees were not recovered with heuristic searches performed with random addition of taxa in 10 replicates. Partitioned Bremer support indices were also calculated for each branch in the tree (Figure 1). The loci BT1, MPG1, and BAC6 have negative Bremer support values for a number of branches in this tree: BT1 at four branches, MPG1 at four branches, and BAC6 at two branches. Positive Bremer support values indicate that a branch is supported by data from a given locus and larger values are indicative of greater support. A negative value indicates that a branch is not supported by data from that locus. Larger negative values indicate greater conflict. Conflict among loci, as evidenced by branches with both positive and negative Bremer support values, was localized in most interior branches, as well as the bipartition containing haplotypes 2, 24–28, 30, 36, and 37. In contrast, in the bipartition-containing haplotypes 3–19, 29, and 31–34, there are only small negative Bremer support values, indicating that this part of the tree was relatively free of conflict. Conflict among specific loci was also apparent. BAC6 and CH7-BAC7 were in conflict at one node. Both BT1 and MPG1 were in conflict with four other loci: CH7-BAC7, CH7-BAC9, EF1, and NUT1.

Figure 1.
One of 160 MPTs for the combined data from all 10 loci. Numerical designations prefixed by an “H” refer to a total of 37 unique multilocus haplotypes from 497 isolates of M. oryzae. The host genus or genera from which each haplotype was ...

To identify haplotypes responsible for phylogenetic conflict within the multilocus analysis, a homoplasy matrix was generated for the MP tree shown in Figure 1 (data not shown). There was very little conflict among haplotypes 3–19, 29, and 31–34, as evidenced by low or zero homoplasy values. In contrast, the other haplotypes had higher homoplasy values in the range of 6–22, indicative of more conflict among these haplotypes, given this tree topology.

Phylogenetic conflict in this phylogeny was localized in the portions of the tree including haplotypes from E. curvula, Eleusine spp., L. perenne, Brachiaria mutica, and Stenotaphrum secundatum. The absence of conflict among loci permits the relationships among haplotypes 3–19, 29, and 31–34 from O. sativa, Setaria millets, P. repens, and L. hexandra to be represented by a tree. In contrast, relationships among the remaining haplotypes may reflect a history of recombination better represented as a reticulate network than as a tree.

Compatibility matrix analysis:

Phylogenetic conflict is due either to recombination, in this case among phylogenetically informative sites, or to convergent evolution. A compatibility matrix was constructed to visualize patterns of phylogenetic compatibility among informative sites both within and among loci. Patterns of incompatibility consistent with recombination or convergent evolution were identified by visual inspection of the matrix. The compatibility matrix (Figure 2) is a visual representation of Hudson's four-gamete test performed on all pairwise combinations of sites (Hudson and Kaplan 1985). Incompatibility among sites indicates that they support conflicting bipartitions of taxa, with large blocks of incompatibility generally indicating recombination. As resolved in the phylogenetic analysis of each locus, the matrix showed that sites within a locus were generally compatible with the exception of the CH7-BAC9 and MPG1 regions, where there were large blocks of incompatible sites. Large blocks of incompatibility were detected among loci on the same chromosome as well as among loci on different chromosomes. The distances, in base pairs, between loci on the same chromosome were calculated using the complete genome sequence (Figure 2). Some distances could not be determined precisely due to gaps in the genome sequences: in these cases, minimum distances are given. In a second analysis (matrix not shown) performed on haplotypes 3–19, 29, and 31–34 from 21 polymorphic sites representing nine loci, only 2 sites were incompatible: site 1874 from the CH7-BAC7 region and site 2199 from the CHS region. Although seven of the nine loci contained multiple polymorphic sites, no blocks of incompatible sites were observed. This pattern of incompatibility is due either to a single recombination event between these loci, indicating low levels of recombination, or to the occurrence of multiple substitutions at 1 of the 2 sites.

Figure 2.
A pairwise compatibility matrix for all phylogenetically informative sites from 9 of the 10 polymorphic loci generated using RETICULATE (Jakobsen and Easteal 1996). Incompatible sites that support conflicting phylogenies are indicated by solid boxes. ...

Network analysis:

The method of split decomposition (Bandelt and Dress 1992) was used to visualize the tree-like and non-tree-like relationships among haplotypes. A splits graph was constructed for the 37 M. oryzae haplotypes (Figure 3) using only base substitutions. As seen with the pattern of compatibility and homoplasy resolved in the parsimony analysis, a predominantly tree-like relationship among haplotypes 3–19, 29, and 31–34 was resolved. The reticulate relationships among the remaining haplotypes (1, 2, 20, 21–23, 24–28, 35, and 36) are depicted by multiple closed loops in the splits graph consistent with recombination.

Figure 3.
A splits graph for M. oryzae haplotypes based on combined data from all 10 loci. Numerical designations refer to each of a total of 37 multilocus haplotypes from 497 isolates of M. oryzae. Indels were excluded from the analysis. Nodes containing sampled ...

Since a tree-like relationship could be inferred among haplotypes 3–19, 29, and 31–34, this phylogeny was converted to a network by means of statistical parsimony (Figure 4). Only base substitutions were included in the analysis. Haplotypes 10 and 14 from rice were the most frequent and widely distributed haplotypes in the sample. These two common, rice-infecting haplotypes belong to interior nodes in the network and have many mutational derivatives. Regardless of the root position for the network, the common, widespread, rice-infecting haplotypes appear to have a single origin. Midpoint rooting was used to polarize this network and determine the direction of host shifts. The root was placed at the midpoint of the longest branch in the phylogeny (the midpoint root). Since we are representing infraspecific relationships in our network, we assume that haplotypes that occur at tips of the network, such as H8, are more recently derived than internal haplotypes, such as H10. Midpoint rooting was the only reasonable way of rooting this network and places the root at an internal, ancestral node within the network. Using another M. oryzae haplotype for rooting was impossible, given the reticulate relationship among isolates from different hosts and the arbitrariness of choosing one haplotype as an outgroup. The most closely related species, M. grisea, was too divergent for use as an outgroup. Coalescent rooting was not reliable, given the differences in sampling among hosts. Our conclusion of a single host shift to rice is not affected by the placement of the root at any other internal or tip node. Placement of the root at different internal nodes, however, would alter the inferred pattern of host shifts. This method placed the root between haplotypes 3 and 19. Given this root position, the pattern of host shifts in M. oryzae is from Setaria millets to rice and from rice to L. hexandra and P. repens. The phylogenetic conflict between sites from CH7-BAC9 and CHS that was resolved in the compatibility analysis is represented in this network as a single closed loop.

Figure 4.
The statistical parsimony network constructed for haplotypes 3–19, 29, and 31–34. Each haplotype or group of haplotypes is represented by a solid circle. The relative frequency of each haplotype is linearly related to the area of each ...

Bayesian analysis:

Bayesian analyses corroborated the branching order resolved in the network analysis. A Bayesian phylogeny was constructed from the combined data for 10 loci for all haplotypes (Figure 5). The main features of this phylogeny are strong support (P = 1.000) for the clade-containing haplotypes 3–19, 29, and 31–34. This clade contains isolates from Setaria millets, L. hexandra, P. repens, and most of the rice isolates. If we assume that the root position falls between haplotypes 3 and 19, then haplotypes from Setaria millets are basal to the clade-containing haplotypes from rice, L. hexandra, and P. repens (P = 0.710). Haplotypes from both L. hexandra and P. repens are derived from rice-infecting haplotypes (P = 1.000). A single Setaria millet-infecting haplotype (H4) also appears to be derived from a rice-infecting haplotype. This analysis suggests the following pattern of host shifts: Setaria millet to rice, followed by shifts from rice to L. hexandra and P. repens.

Figure 5.
The Bayesian consensus tree for M. oryzae haplotypes inferred from the combined data from all 10 loci. Numerical designations prefixed by an “H” refer to a total of 37 unique multilocus haplotypes from 497 isolates of M. oryzae. The values ...

Dating divergence of the rice-infecting lineage:

Two events in the history of rice cultivation that may have facilitated a host shift to rice were domestication [~7000 years ago; (Crawford and Shen 1998)] and the Green Revolution (35 years ago). We attempted to determine which event was most likely associated with the origin of rice-infecting populations by dating the divergence of rice-infecting haplotypes. If molecular evolution in the lineage containing rice-infecting isolates is clock like, the number of base substitutions separating two haplotypes should be proportional to their divergence times. A molecular evolutionary clock hypothesis could not be rejected for isolates from rice, Panicum, and Leersia (P ≥ 0.75); therefore, estimates of the timing of divergence of rice-infecting isolates were generated using rate smoothing. Calibration of the molecular clock to estimate divergence times in years from the number of substitutions separating two haplotypes requires an estimate of the age for at least one node in a phylogeny. For the Bayesian analysis we used an a priori estimate of 10,000 years BP for the age of the node separating haplotypes from Setaria millet from haplotypes from rice, L. hexandra, and P. repens. In this analysis the standard deviation of the estimate was set to 10,000 years BP. This time estimate was chosen as a maximum age for this node because it predates rice domestication, but still falls within the time frame of use of rice by humans (12,000 BP). Using the method of Sanderson (2002) implemented in the program r8s, two estimates for the age of the node separating haplotypes from Setaria millet from haplotypes from rice, L. hexandra, and P. repens were used; the age was fixed at 10,000 BP, and the age was constrained to be between 10,000 and 40 BP. Both methods for estimating divergence time gave comparable results. The Bayesian method of Thorne et al. (1998) resolved a divergence time of 7291 BP (95% C.I. = 26,042–453 BP). Using the method of Sanderson (2002), a divergence time of 5373 BP was estimated when the root age was fixed at 10,000 BP. When the root age was allowed to vary between 10,000 and 40 BP, the divergence time was estimated at 2574 BP. Both dating methods were in agreement in supporting an earlier divergence time, rather than a later time associated with the Green Revolution.

Distribution of AVR genes:

The distribution of AVR-Pita and AVR1-Co39 was evaluated using a simple plus/minus PCR test (Figure 5). Failure to amplify a gene was taken as evidence for the loss of gene function as a result of complete or partial deletion or rearrangement. To address the issue of false negatives due to low-quality DNA, PCR amplifications of the actin region were performed on all DNA dilutions. The actin region was amplified in all isolates, indicating that DNA quality was sufficient for PCR amplification. It is possible, however, that mutation in or around the primer-binding site could also have contributed to false negatives. The distribution of ACE1 was evaluated using allele-specific PCR. A single loss of function for AVR-Co39 is associated with the major Oryza-infecting lineage, excluding haplotypes 9 and 26. In contrast, the loss of AVR1-Pita is not associated with a single clade on any host. The two ACE1 virulent genotypes (vir1 and vir2) were detected only in haplotype H14 and its derived haplotype H16 (Figure 4), suggesting a common origin for both virulent genotypes. H16 has only vir1, consistent with an origin of this virulence genotype in a subpopulation of H14.

Pathogenicity testing:

In all pathogenicity tests the rice-infecting controls were generally highly pathogenic toward rice (Table 3). Nonrice isolates were generally avirulent or weakly pathogenic toward rice. Some isolates did produce lesions on rice cultivars but these lesions were of a less severe type or fewer in number than those produced by the rice-infecting controls. The isolate Dr 51-2-3 from cutgrass produced type 3 lesions on the variety IR1529-380-3 and a few type 3 lesions on Sariceltik. The isolates Gr F2-1, Arcadia, and Gr F4 from S. viridis produced some lesions on the more susceptible cultivars Maratelli and Sariceltik. Isolates from S. viridis and P. repens were not pathogenic toward rice and produced some lesions of the type capable of sporulating on P. repens. Even though strain O-137 was highly pathogenic toward rice, it was only weakly pathogenic on P. repens.

Pathogenicity testing ofM. oryzae isolates fromO. sativa,S. italica,L. hexandra, andP. repens

Mating-type testing:

Both MAT1-1 and MAT1-2 were identified within rice-infecting populations of M. oryzae with MAT1-2 being the most frequent (Table 4). The common, widely distributed haplotypes, 10 and 14, were composed of isolates with either MAT1-1 or MAT1-2. Haplotypes derived from 10 and 14 (Figure 4) included isolates with only one mating type. Only single mating types were identified within populations from torpedo grass (MAT1-2) and cutgrass (MAT1-1).

Distribution and frequency of mating-type idiomorphs amongM. oryzae haplotypes fromO. sativa,P. repens, andL. hexandra


Contrary to the expectation that rice should be a large target for multiple, continuing, parasite host shifts, we resolved a single, relatively old origin from Setaria millet for a pandemic lineage of M. oryzae on rice. In turn, rice crops have been a source for invasions of the common rice field weeds, cutgrass and torpedo grass. Nearly all rice-infecting isolates fell within a single phylogenetic lineage, as would be expected if the contemporary, globally distributed, rice-infecting populations of M. oryzae originated as a result of a single host shift to rice. The branching phylogenetic relationship of rice-infecting haplotypes is evidence of clonal amplification of this lineage following invasion of rice. Reticulate relationships indicating recombination were not observed among these haplotypes. Isolates from torpedo grass and cutgrass compose two distinct lineages derived from a rice-infecting ancestor. The host specificity associated with these lineages appears to be maintained by differences in pathogenicity among isolates from different hosts. Isolates from hosts other than rice either were not pathogenic toward rice or were much less pathogenic when compared with typical rice-infecting isolates (Table 3). Although serial culturing could reduce pathogenicity, control isolates, also serially cultured, demonstrated expected pathogenicity. Among several possible mechanisms for host specialization, the pathogen avirulence determinant AVR-Co39 was widespread in haplotypes from hosts other than rice, potentially preventing infection of rice cultivars equipped with the corresponding resistance gene Pi-Co39.

A single origin of a pandemic lineage of M. oryzae on rice could represent (i) one invasion of rice, (ii) multiple invasions with the descendants of one invasion surviving genetic drift, or (iii) multiple invasions mostly undetected in our sampling due to their limited geographical distribution. During the course of this study the sample was increased three times in increments of ~100 isolates without resolving additional host jumps, making alternative (iii) unlikely. Sampling from diverse locations should have resolved additional host shifts where new genotypes have become established on rice even if they have not become globally distributed. A potential bias in this sample is that very weakly pathogenic isolates would be underrepresented since only plants with visible lesions, capable of sporulating, were sampled. It is possible that changes in deployment of rice cultivars and agronomic practices post-Green Revolution resulted in a bottleneck consistent with alternative (ii), reducing genetic diversity in the pathogen and obliterating traces of pathogen lineages other than the one rice-infecting lineage detected in this study. This explanation would not undermine a time estimate of lineage divergence predating the Green Revolution.

The observed direction of host shifts in M. oryzae is especially interesting, considering that both millet and rice were domesticated in China and appear to have partially co-occurred early in the history of agriculture in Asia. Italian millet is the domesticated species, with a primary gene pool comprising two weedy Setaria millets, S. viridis and S. faberi, which are recognized by some as distinct species (de Wet et al. 1979; Dekker 2003). Although the widespread contact between Italian millet and rice in China occurred only recently and within written historical time, there is some archeological evidence for earlier co-occurrence of these two crops (Chang 1986; Crawford 1992; Higman and Lu 1998).

The sequence of host shifts was robust even with alternative root positions, and the relationship of haplotypes in the lineage including millet, rice, cutgrass, and torpedo grass was not obscured by phylogenetic conflict. Midpoint rooting resolved a single root position between haplotypes 3 and 19 (Figure 4). The relationships among these haplotypes are well supported by parsimony, network, and Bayesian methods (Figures 1, ,3,3, ,4,4, and and5).5). The only conflict in this part of the phylogeny is represented by the closed loop containing haplotypes 4, 5, 10, and 19, due to one base substitution (Figure 4). A single origin for rice-infecting haplotypes is still resolved regardless of the cause of this conflict, either recombination or multiple hits.

In addition to the single dominant lineage of haplotypes associated with rice, a lineage was resolved comprising several haplotypes from goosegrass (Eleusine), but also including one haplotype (26) associated with rice (Figures 1, ,3,3, and and5).5). Haplotype 26 was sampled at low frequency, as 2 of 112 isolates from India, and was not recovered in subsequent sampling from Southeast Asia. Haplotype 26 may represent a decaying chain of infection in which the parasite has crossed a species boundary but cannot sustain a positive reproductive rate (May et al. 2001).

The pandemic spread of a single HIV subtype suggests that the accidents of genetic drift have dominated the population structure on a global scale, resulting in a single widely distributed lineage (Sharp et al. 1994; Chen et al. 1997; Hutchinson 2001). If the pattern seen in HIV is a model for M. oryzae, additional host shifts to rice may have evaded detection because haplotypes recently established on rice could have limited dispersal. If remnants of additional host shifts exist, finding them will require regional sampling without predictive data for selecting localities. The association between early domestication of rice and Setaria millet suggests that the host shift to rice occurred in China, where evidence for additional shifts may yet be found.

In the statistical parsimony analysis we included only base substitutions. Indels were excluded from the analysis because they are not accommodated by the substitution model (Templeton et al. 1992). When indels were added to the network inferred from statistical parsimony, more haplotypes associated with rice were distinguished and a single 4-bp indel appeared to distinguish rice-infecting haplotypes (excluding haplotypes 9 and 26) from all other haplotypes. Inclusion of this character necessitates inferring an unsampled common ancestor (with no host association) for the haplotypes from rice, cutgrass, and torpedo grass. In contrast, in all analyses in which only base substitutions were used, haplotypes 10, 13, and 29 were clearly resolved as ancestral to haplotypes from cutgrass and torpedo grass. These haplotypes show nearly complete association with rice; haplotype 13 consists of 2 isolates from rice, haplotype 29 consists of 1 isolate from rice, and haplotype 10 consists of 126 isolates, 123 from rice. In haplotype 10, 2 of the 3 isolates from hosts other than rice were from barley, but were pathogenic on rice, suggesting that they represented invasions of barley from rice (D. Tharreau, unpublished results). We propose the most parsimonious explanation that both this indel and the high copy number of Pot3 evolved in rice-infecting populations following the host shifts from rice to torpedo grass and cutgrass. We are assuming that an unsampled ancestor of H10 had a low copy number of Pot3 and lacked the indel. This haplotype gave rise to the cutgrass and torpedo grass populations. Following host shifts to cutgrass and torpedo grass, a small deletion occurred in the CAL locus and the copy number of Pot3 increased.

Amplification of Pot3 also appears to have occurred in the rice-infecting lineage following host shifts from rice to cutgrass and torpedo grass. In the present study, isolates from rice have a high copy number of this element while isolates from cutgrass and torpedo grass have a much lower copy number (data not shown). Isolates from Setaria millets have been reported to show variable hybridization to Pot3, with some isolates showing weak or no hybridization (Farman 2002) and others showing a medium copy number in Setaria millet isolates but lower than that in rice isolates (Eto et al. 2001; Kang et al. 2001; Tosa et al. 2004). An independent amplification of Pot3 in some Pennisetum isolates with low genetic similarity to isolates from Setaria millets and rice has also been reported (Kang et al. 2001). These observations suggest that amplification of transposable elements may occur frequently and, on a contemporary timescale, in populations on a variety of hosts. The high copy number of this element in all populations from rice supports a single origin for rice-infecting populations and also a history of asexual reproduction. Prolonged asexual reproduction in ascomycetous fungi favors the amplification of repetitive elements since transposon inactivation requires a sexual cycle (Cambareri et al. 1989), although sexual reproduction may stimulate transposition of some elements such as MAGGY (Eto et al. 2001).

Multilocus analysis of M. oryzae by means of parsimony, network, and Bayesian methods all resolved a pattern of clonality in haplotypes from rice and their close relatives from Setaria millets, cutgrass, and torpedo grass and a pattern of recombination among haplotypes from other hosts. This was done by sequential analyses, first on data from each locus and then on combined data from all loci, which were used to identify the cause of homoplasy, to identify haplotypes responsible for homoplasy, and to localize homoplasy to specific regions within the phylogenies inferred for all haplotypes.

Patterns of incompatibility among polymorphic sites from all loci, represented in a compatibility matrix, indicate that recombination is likely to be responsible for the majority of the homoplasy observed in these data (Carbone and Kohn 2001). Convergent evolution of SSCP alleles is not an important source of homoplasy in this study; no cases of convergent evolution were identified by direct sequencing of SSCP alleles and comparable levels of conflict among loci were resolved when the analysis was restricted to isolates for which all loci were sequenced (data not shown). Sites are considered incompatible if two nucleotides at different sites support conflicting phylogenetic topologies. Incompatibility among sites is due either to recombination among sites or to convergent evolution resulting from multiple substitutions. Several arguments support recombination in this data set. First, to develop the large blocks of incompatible sites identified among the loci CH7-BAC7, CH7-BAC9, and MPG1 in the compatibility analysis (Figure 2), it is unlikely that the alternative to recombination, multiple hits, would occur repeatedly in adjoining nucleotide positions, given an infinite-sites or nearly infinite-sites model of DNA sequence evolution (Kimura 1969; Hudson and Kaplan 1985). On the other hand, in comparisons of other pairs of loci with fewer polymorphic sites, it is difficult to distinguish recombination from multiple hits as the cause of incompatibility. But given that multiple hits are expected to be rare, the presence of clusters of incompatible sites, as in this data set, is more likely due to recombination. Further, the majority of these loci with fewer polymorphisms occur on different chromosomes among which segregation should be random. The small numbers of substitutions in these loci do not suggest that they fall within mutational hotspots, in which multiple hits would be expected to occur with higher frequency. Finally, no polymorphic sites that had more than two bases segregating within the sample were identified, consistent with an infinite-sites model of DNA sequence evolution (Kimura 1969).

The complementary analyses using parsimony, network, and Bayesian methods resolved patterns of recombination consistent with the compatibility matrix. Predominantly tree-like relationships, expected with clonality (Carbone et al. 1999), were inferred among haplotypes from wheat, rice, Setaria millets, cutgrass, and torpedo grass. Reticulate relationships, expected with recombination, were required to infer relationships among the remaining haplotypes from other hosts. The resolution of tree-like relationships among rice-infecting haplotypes is consistent with previous reports based on other types of data that support that M. oryzae has a predominantly clonal population structure and is known to reproduce only asexually (Ou 1987; Levy et al. 1993; Park et al. 2003).

We detected negligible, ambiguous signals of recombination in the rice-Setaria millet-cutgrass-torpedo grass lineage. Two sites were found to be incompatible, due to either recombination or convergent evolution. If recombination is the cause of this incompatibility, it appears to have occurred prior to the divergence of haplotypes from rice, cutgrass, and torpedo grass (Figure 4). The detection of both mating types in each of two haplotypes, MAT1-1 and MAT1-2 in haplotypes 10 and 14, is further support of the hypothesis of historical but not contemporary recombination. Although both mating types are present in some populations, this in itself is not indicative of contemporary sexual reproduction. In the absence of selection or drift, both mating types could be maintained. Since no incompatibility was detected among phylogenetically informative sites in these lineages, the presence of both mating types is not strong evidence for active sexual reproduction.

The main difficulty in detecting recombination is the low power of the four-gamete test (Hudson and Kaplan 1985). The probability of detecting recombination increases with the number of segregating sites on either side of the recombination breakpoint and the recombination rate between regions (Hudson and Kaplan 1985). We chose multiple markers located on different chromosomes or distributed along the same chromosome to increase the number of potentially segregating sites screened, the number of intervals in which recombination can occur, and the rate of recombination among loci, given that the recombination rate should be maximal among unlinked loci. The incompatibility among closely linked loci on the same chromosome suggests that the recombination detected is due to true sexual reproduction rather than parasexuality, which results most frequently in recombination among chromosomes (Pontecorvo 1956). Recombination within some rice-infecting populations has been reported, as evidenced by allelic associations in pairwise comparisons, and multilocus variance analysis of DNA fingerprints with the multilocus probe MGR586 (Pot3) and single-locus probes (Zeigler et al. 1997; Zeigler 1998). However, the best evidence for recombination within rice-infecting populations would be recovery of perithecia from naturally infected hosts in the field, with ascospores showing segregation with respect to parental types. To date, reports of female sexual fertility and high genetic diversity in pathogen populations in southern China and Thailand seem to be the best evidence available of at least local sexual reproduction against a background of predominant clonality (Hayashi et al. 1997; Mekwatanakarn et al. 1999).

A history of recombination has probably obscured the origins of host-specific populations of M. oryzae on perennial ryegrass and wheat. A possible strategy is to identify closely related populations on other hosts, possibly tall fescue (Festuca arundinacea) for perennial ryegrass (Urashima et al. 1999; Viji et al. 2001; Farman 2002; Tosa et al. 2004), among which tree-like patterns of relationship may exist.

In addition to revealing patterns of recombination, the multilocus approach detected divergence among isolates from different hosts not previously detected with a single locus. For example, wheat and goosegrass (Eleusine) isolates cannot be distinguished solely on the basis of the intergenic transcribed spacer of the nuclear rDNA repeat, but are distinguished on the basis of some RFLP markers, including the FOSBURY retrotransposon specific to Eleusine (Viji et al. 2001; Tosa et al. 2004). In the present study, isolates from the two hosts are distinguished as a number of host-specific, multilocus haplotypes whose phylogenetic relationships are obscured by recombination.

Host shifts to rice from other grasses may be limited by either AVR-R interactions preventing initiation of a host shift or the subsequent inability to become established on rice. The presence of AVR1-Co39 in nonrice haplotypes and its absence in haplotypes from rice corroborate the earlier findings of Farman (2002). The presence of AVR1-Co39 in haplotypes 6, 7, 8, and 9 suggests that this gene was lost in the rice-infecting lineage following the host shift to topedo grass but does not necessarily imply that the loss of this gene is essential for rice parasitism. Although the rice R-gene that recognizes AVR1-Co39 has been defeated by the pathogen, it may still constrain additional host shifts to rice. Following the initial stages of the host shift to rice, lack of fitness of the invader on the new host could impede establishment. Pathogenicity tests support this argument; in our greenhouse test, nonrice isolates were generally not pathogenic toward rice. The avirulent or weakly pathogenic reactions toward rice observed for haplotypes from Setaria millet, torpedo grass, and cutgrass suggest that they are poorly adapted to rice, and although they may be able to establish some level of infection in vitro, they may not be able to become established on rice.

Changes in the virulence spectra of rice-infecting isolates also occur within rice-infecting populations. Breakdown of complete resistance, conferred by major resistance genes, often occurs rapidly in the field in response to the release of new, blast-resistant cultivars (Nottéghem et al. 1994). Our results suggest that gain of virulence toward rice varieties with the resistance gene corresponding to ACE1 has occurred in a single rice-infecting haplotype. The two ACE1 virulent genotypes (ACE1-vir1 and ACE1-vir2) were detected only in isolates from haplotype 14, suggesting a common origin for both virulent genotypes. Haplotype 16, descended from haplotype 14, has only ACE1-vir1, suggesting that it was derived from a subpopulation of haplotype 14, carrying the ACE1-vir1 allele. The amplification of transposable elements such as Pot3 in the genomes of rice-infecting isolates and the resultant genetic instability could result in deletion of AVR genes and in the rapid evolution of virulence toward newly deployed R-genes.

Given that M. oryzae on rice seems to have a single origin and has been moved around the world in association with rice cultivation, what genetic determinants are unique to the rice-infecting lineage? What mechanisms in rice could potentially prevent the establishment of new M. oryzae genotypes? Our understanding of the genetic determinants conveying virulence and pathogenicity in the rice-infecting lineage of M. oryzae will deepen by considering the origins and dynamics of these determinants implicit in the evolutionary relationships among haplotypes from other hosts. Presumably the lineages of Setaria millets, torpedo grass, and cutgrass are genetically very similar to the lineage from rice since they are closely related. They are, however, expected to lack some or all of the genetic determinants responsible for high fitness on rice. Representatives of these haplotypes would be prime candidates as parents in crosses to identify the genetic determinants of fitness on rice common to all rice-infecting haplotypes.


We thank C. Vera-Cruz and E. Borromeo for their help with sampling in the Philippines as well as M. Marchetti and all those who contributed isolates to this study. For technical support we thank Caroline Sirjusingh and Andrea Chan at the University of Toronto, Henri Adreit and Joelle Milazzo at Centre de Coopération Internationale en Recherche Agronomique pour le Développement (CIRAD), and Melinda Dalby at Kansas State University. This work was supported by a Discovery Research Grant from the Natural Sciences and Engineering Research Council of Canada to L. M. Kohn and by scholarships to B. Couch from the Ontario Graduate Scholarship program. Part of this work was supported by the European Community RESIDIV ICA4-CT2000-30021 project grants to the CIRAD and Centre National de la Recherche Scientifique teams. A United States Department of Agriculture-Cooperative State Research and Extension Service grant to M. Milgroom (L. M. Kohn coapplicant) supported the development of markers and fieldwork of B. Couch.


Sequence data from this article have been deposited with the EMBL/GenBank Data Libraries under accession nos. AY944075, AY944076, AY944077, AY944078, AY944079, AY944080, AY944081, AY944082, AY944083, AY944084, AY944085, AY944086, AY944087, AY944088, AY944089, AY944090, AY944091, AY944092, AY944093, AY944094, AY944095, AY944096, AY944097, AY944098, AY944099, AY944100, AY944101, AY944102, AY944103, AY944104, AY944105, AY944106, AY944107, AY944108, AY944109, AY944110, AY944111, AY944112, AY944113, AY944114, AY944115, AY944116, AY944117, AY944118, AY944119, AY944120, AY944121, AY944122, AY944123, AY944124, AY944125, AY944126, AY944127, AY944128, AY944129, AY944130, AY944131 and AY944359, AY944360, AY944361, AY944362, AY944363, AY944364, AY944365, AY944366, AY944367, AY944368, AY944369, AY944370, AY944371, AY944372, AY944373, AY944374, AY944375, AY944376, AY944377, AY944378, AY944379, AY944380, AY944381, AY944382, AY944383, AY944384, AY944385, AY944386, AY944387, AY944388, AY944389, AY944390, AY944391, AY944392, AY944393, AY944394, AY944395, AY944396, AY944397, AY944398, AY944399, AY944400, AY944401, AY944402, AY944403, AY944404, AY944405.


  • Antonovics, J., M. Hood and J. Partain, 2002. The ecology and genetics of a host-shift: Microbotryum as a model system. Am. Nat. 160: S40–S53. [PubMed]
  • Baker, R. H., and R. DeSalle, 1997. Multiple sources of character information and the phylogeny of Hawaiian drosophilids. Syst. Biol. 46: 654–673. [PubMed]
  • Baker, R. H., X. Yu and R. DeSalle, 1998. Assessing the relative contribution of molecular and morphological characters in simultaneous analysis trees. Mol. Phylogenet. Evol. 9: 427–436. [PubMed]
  • Bandelt, H. J., and A. W. Dress, 1992. Split decomposition: a new and useful approach to phylogenetic analysis of distance data. Mol. Phylogenet. Evol. 1: 242–252. [PubMed]
  • Böhnert, H. U., I. Fudal, W. Dioh, D. Tharreau, J.-L. Nottéghem et al., 2004. A putative polyketide synthase/peptide synthetase from Magnaporthe grisea signals pathogen attack to resistant rice. Plant Cell 16: 2499–2513. [PMC free article] [PubMed]
  • Borromeo, E. S., R. J. Nelson, J. M. Bonman and H. Leung, 1993. Genetic differentiation among isolates of Pyricularia infecting rice weed hosts. Phytopathology 83: 393–399.
  • Brasier, C. M., and K. W. Buck, 2001. Rapid evolutionary changes in a globally invading fungal pathogen (Dutch elm disease). Biol. Invasions 3: 223–233.
  • Bremer, K., 2002. Gondwanan evolution of the grass alliance of families (Poales). Evolution 56: 1374–1387. [PubMed]
  • Cambareri, E. B., B. C. Jensen, E. Schabtach and E. U. Selker, 1989. Repeat-induced G-C to A-T mutations in Neurospora. Science 244: 1571–1575. [PubMed]
  • Carbone, I., and L. M. Kohn, 1999. A method for designing primer sets for speciation studies in filamentous ascomycetes. Mycologia 91: 553–556.
  • Carbone, I., and L. M. Kohn, 2001. A microbial population–species interface: nested cladistic and coalescent inference with multilocus data. Mol. Ecol. 10: 947–964. [PubMed]
  • Carbone, I., J. B. Anderson and L. M. Kohn, 1999. Patterns of descent in clonal lineages and their multilocus fingerprints are resolved with combined gene genealogies. Evolution 53: 11–21.
  • Chang, K. C., 1986 The Archaeology of Ancient China. Yale University Press, New Haven, CT.
  • Chang, T. T., 1976. The origin, evolution, cultivation, dissemination and diversification of Asian and African rices. Euphytica 25: 425–441.
  • Chen, Z., A. Luckay, D. L. Sodora, P. Telfer, P. Reed et al., 1997. Human immunodeficiency virus type 2 (HIV-2) seroprevalence and characterization of a distinct HIV-2 genetic subtype from the natural range of simian immunodeficiency virus-infected sooty mangabeys. J. Virol. 71: 3953–3960. [PMC free article] [PubMed]
  • Cheng, C., R. Motohashi, S. Tsuchimoto, Y. Fukuta, H. Ohtsubo et al., 2003. Polyphyletic origin of cultivated rice: based on the interspersion pattern of SINEs. Mol. Biol. Evol. 20: 67–75. [PubMed]
  • Clement, M., D. Posada and K. A. Crandall, 2000. TCS: a computer program to estimate gene genealogies. Mol. Ecol. 9: 1657–1660. [PubMed]
  • Couch, B. C., and L. M. Kohn, 2002. A multilocus gene genealogy concordant with host preference indicates segregation of a new species, Magnaporthe oryzae, from M. grisea. Mycologia 94: 683–693. [PubMed]
  • Crawford, G. W., 1992 Prehistoric plant domestication in East Asia, pp. 7–38 in The Origins of Agriculture: An International Perspective, edited by C. W. Cowan and P. J. Watson. Smithsonian Institution Press, Washington, DC.
  • Crawford, G. W., and C. Shen, 1998. The origins of rice agriculture: recent progress in East Asia. Antiquity 72: 858–866.
  • Crepet, W., and G. D. Feldman, 1991. The earliest remains of grasses in the fossil record. Am. J. Bot. 78: 1010–1014.
  • Crow, J. F., and M. Kimura, 1970 An Introduction to Population Genetics Theory. Harper & Row, New York.
  • Dekker, J., 2003. The foxtail (Setaria) species-group. Weed Sci. 51: 641–656.
  • de Wet, J. M. J., L. L. Oestry-Stidd and J. I. Cubero, 1979. Origins and evolution of foxtail millets (Setaria italica). J. Agric. Trop. Bot. Appl. 26: 53–64.
  • Dress, A., D. Huson and V. Moulton, 1996. Analyzing and visualizing sequence and distance data using SplitsTree. Discrete Appl. Math. 71: 95–109.
  • Eto, Y., K. Ikeda, I. Chuma, T. Kataoka, S. Kuroda et al., 2001. Comparative analyses of the distribution of various transposable elements in Pyricularia and their activity during and after the sexual cycle. Mol. Gen. Genet. 264: 565–577. [PubMed]
  • Farman, M. L., 2002. Pyricularia grisea isolates causing gray leaf spot on perennial ryegrass (Lolium perenne) in the United States: relationship to P. grisea isolates from other host plants. Phytopathology 92: 245–254. [PubMed]
  • Farman, M. L., and S. A. Leong, 1998. Chromosome walking to the AVR1-CO39 avirulence gene of Magnaporthe grisea: discrepancy between the physical and genetic maps. Genetics 150: 1049–1058. [PMC free article] [PubMed]
  • Farman, M. L., Y. Eto, T. Nakao, Y. Tosa, H. Nakayashiki et al., 2002. Analysis of the structure of the AVR1–CO39 avirulence locus in virulent rice-infecting isolates of Magnaporthe grisea. Mol. Plant Microbe Interact. 15: 6–16. [PubMed]
  • Fry, J. D., 1996. The evolution of host specialization: Are trade-offs overrated? Am. Nat. 148: S84–S107.
  • Fudal, I., 2004 Etude du gene d'avirulence ACE1 de Magnaporthe grisea, agent pathogène du riz: analyse de l'expression du gène ACE1 et évolution dans les populations de Magnaporthe grisea. Ph.D. Thesis, University of Orsay, Orsay, France.
  • Glass, N. L., and G. C. Donaldson, 1995. Development of primer sets designed for use with the PCR to amplify conserved genes from filamentous Ascomycetes. Appl. Environ. Microbiol. 61: 1323–1330. [PMC free article] [PubMed]
  • Grass Phylogeny Working Group, 2001. Phylogeny and subfamilial classification of the grasses (Poaceae). Ann. Mo. Bot. Gar. 88: 373–457.
  • Hamer, J. E., R. J. Howard, F. G. Chumley and B. Valent, 1988. A mechanism for surface attachment in spores of a plant pathogenic fungus. Science 239: 288–290. [PubMed]
  • Hamer, J. E., L. Farrall, M. J. Orbach, B. Valent and F. G. Chumley, 1989. Host species-specific conservation of a family of repeated DNA sequences in the genome of a fungal plant pathogen. Proc. Natl. Acad. Sci. USA 86: 9981–9985. [PMC free article] [PubMed]
  • Hayashi, N., C. Li, J. Li, M. Iwano, H. Naito et al., 1997. Distribution of fertile Magnaporthe grisea fungus pathogenic to rice in Yunnan Province, China. Ann. Phytopathol. Soc. Jpn. 63: 316–323.
  • Heath, M. C., R. J. Howard, B. Valent and F. G. Chumley, 1990. Interactions of two strains of Magnaporthe grisea with rice, goosegrass, and weeping lovegrass. Can. J. Bot. 68: 1627–1637.
  • Heath, M. C., R. J. Howard, B. Valent and F. G. Chumley, 1992. Ultrastructural interactions of one strain of Magnaporthe grisea with goosegrass and weeping lovegrass. Can. J. Bot. 70: 779–787.
  • Hebert, T. T., 1971. The perfect stage of Pyricularia grisea. Phytopathology 61: 83–87.
  • Higman, C., and T. L.-D. Lu, 1998. The origins and dispersal of rice cultivation. Antiquity 72: 867–877.
  • Hudson, R. R., and N. L. Kaplan, 1985. Statistical properties of the number of recombination events in the history of a sample of DNA sequences. Genetics 111: 147–164. [PMC free article] [PubMed]
  • Huelsenbeck, J. P., and F. Ronquist, 2001. MrBayes: Bayesian inference of phylogenetic trees. Bioinformatics 17: 754–755. [PubMed]
  • Hutchinson, J. F., 2001. The biology and evolution of HIV. Annu. Rev. Anthropol. 30: 85–108.
  • Jakobsen, I. B., and S. Easteal, 1996. A program for calculating and displaying compatibility matrices as an aid in determining reticulate evolution in molecular sequences. Comput. Appl. Biosci. 12: 291–295. [PubMed]
  • Jia, Y., S. A. McAdams, G. T. Bryan, H. P. Hershey and B. Valent, 2000. Direct interaction of resistance gene and avirulence gene products confers rice blast resistance. EMBO J. 19: 4004–4014. [PMC free article] [PubMed]
  • Kang, S., F. G. Chumley and B. Valent, 1994. Isolation of the mating-type genes of the phytopathogenic fungus Magnaporthe grisea using genomic subtraction. Genetics 138: 289–296. [PMC free article] [PubMed]
  • Kang, S., M. H. Lebrun, L. Farrall and B. Valent, 2001. Gain of virulence by insertion of a Pot3 transposon in a Magnaporthe grisea avirulence gene. Mol. Plant Microbe Interact. 14: 671–674. [PubMed]
  • Kato, H., M. Yamamoto, T. Yamaguchi-Ozaki, H. Kadouchi, Y. Iwamoto et al., 2000. Pathogenicity, mating ability and DNA restriction fragment length polymorphisms of Pyricularia populations isolated from Gramineae, Bambusideae and Zingiberaceae plants. J. Gen. Plant Pathol. 66(1): 30–47.
  • Khush, G. S., 2001. Green revolution: the way forward. Nat. Rev. Genet. 2: 815–822. [PubMed]
  • Kimura, M., 1969. The number of heterozygous nucleotide sites maintained in a finite population due to steady flux of mutations. Genetics 61: 893–903. [PMC free article] [PubMed]
  • Kumar, J., R. J. Nelson and R. S. Zeigler, 1999. Population structure and dynamics of Magnaporthe grisea in the Indian Himalayas. Genetics 152: 971–984. [PMC free article] [PubMed]
  • Landschoot, P. J., and B. F. Hoyland, 1992. Gray leaf spot of perennial ryegrass turf in Pennsylvania. Plant Dis. 76: 1280–1282.
  • Leong, S., M. Farman, J. Smith, A. Budde, Y. Tosa et al., 1994 Molecular genetic approach to the study of cultivar specificity in the rice blast fungus, pp. 87–110 in Rice Blast Disease, edited by R. Zeigler, S. Leong and P. Teng. CAB International, Wallingford, Oxfordshire, UK.
  • Levy, M., J. Romao, M. A. Marchetti and J. E. Hamer, 1991. DNA fingerprinting with a dispersed repeated sequence resolves pathotype diversity in the rice blast fungus. Plant Cell 3: 95–102. [PMC free article] [PubMed]
  • Levy, M., F. J. Correa Victoria, R. S. Zeigler, S. Xu and J. E. Hamer, 1993. Genetic diversity of the rice blast fungus in a disease nursery in Colombia. Phytopathology 83: 1427–1433.
  • Long, D. H., J. C. Correll, F. N. Lee and D. O. TeBeest, 2001. Rice blast epidemics initiated by infested rice grain on the soil surface. Plant Dis. 85: 612–616.
  • Lu, B.-R., K. L. Zheng, H. R. Qian and J. Y. Zhang, 2002. Genetic differentiation of wild relatives of rice as assessed by RFLP analysis. Theor. Appl. Genet. 106: 101–106. [PubMed]
  • MacLeod, A., A. Tait and C. M. R. Turner, 2001. The population genetics of Trypanosoma brucei and the origin of human infectivity. Philos. Trans. R. Soc. Lond. B Biol. Sci. 356: 1035–1044. [PMC free article] [PubMed]
  • Marchetti, M., 1983. Dilatory resistance to rice blast in USA rice. Phytopathology 73: 645–649.
  • Marchetti, M., L. Xinghua and C. Bollich, 1987. Inheritance of resistance to Pyricularia oryzae in rice cultivars grown in the United States. Phytopathology 77: 799–804.
  • May, R. M., S. Gupta and A. R. McLean, 2001. Infectious disease dynamics: What characterizes a successful invader? Philos. Trans. R. Soc. Lond. B Biol. Sci. 356: 901–910. [PMC free article] [PubMed]
  • Maynard Smith, J., N. H. Smith, M. O'Rourke and B. G. Spratt, 1993. How clonal are bacteria? Proc. Natl. Acad. Sci. USA 90: 4384–4388. [PMC free article] [PubMed]
  • Mekwatanakarn, P., W. Kositratana, T. Promraksa and R. Zeigler, 1999. Sexually fertile Magnaporthe grisea rice pathogens in Thailand. Plant Dis. 83: 939–943.
  • Nottéghem, J. L., D. Tharreau, D. Silué and E. Roumen, 1994 Present knowledge of rice resistance and strategies for Magnaporthe grisea pathogenicity and avirulence gene analysis, pp. 155–165 in Rice Blast Disease, edited by R. Zeigler, S. Leong and P. Teng. CAB International, Wallingford, Oxfordshire, UK.
  • Oh, H. S., Y. Tosa, N. Takabayashi, S. Nakagawa, R. Tomita et al., 2002. Characterization of an Avena isolate of Magnaporthe grisea and identification of a locus conditioning its specificity on oat. Can. J. Bot. 80: 1088–1095.
  • Oka, H. I., 1988 Origin of Cultivated Rice. Elsevier, Amsterdam.
  • Orbach, M. J., L. Farrell, J. A. Sweigard, F. G. Chumley and B. Valent, 2000. A telomeric avirulence gene determines efficacy for the rice blast resistance gene Pi-ta. Plant Cell 12: 2019–2032. [PMC free article] [PubMed]
  • Orita, M., Y. Suzuki, T. Sekiya and K. Hayashi, 1989. Rapid and sensitive detection of point mutations and DNA polymorphisms using the polymerase chain reaction. Genomics 5: 874–879. [PubMed]
  • Ou, S. H., 1987 Rice Diseases. The Commonwealth Mycological Institute, Kew, Surrey, UK.
  • Park, S. Y., M. G. Milgroom, S. S. Han, S. Kang and Y. H. Lee, 2003. Diversity of pathotypes and DNA fingerprint haplotypes in populations of Magnaporthe grisea in Korea over two decades. Phytopathology 93: 1378–1385. [PubMed]
  • Peng, Y. L., and J. Shishiyama, 1988. Temporal sequence of cytological events in rice leaves infected with Pyricularia oryzae. Can. J. Bot. 66: 730–735.
  • Pontecorvo, G., 1956. The parasexual cycle. Annu. Rev. Microbiol. 10: 393–400. [PubMed]
  • Posada, D., and K. A. Crandall, 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14: 817–818. [PubMed]
  • Ronald, P. C., 1997. The molecular basis of disease resistance in rice. Plant Mol. Biol. 35: 179–186. [PubMed]
  • Sallaud, C., M. Lorieux, E. Roumen, D. Tharreau, R. Berruyer et al., 2003. Identification of five new blast resistance genes in the highly blast-resistant rice variety IR64 using a QTL mapping strategy. Theor. Appl. Genet. 106: 794–803. [PubMed]
  • Sambrook, J., E. F. Fritsch and T. Maniatis, 1989 Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.
  • Sanderson, M. J., 2002. Estimating absolute rates of molecular evolution and divergence times: a penalized likelihood approach. Mol. Biol. Evol. 19: 101–109. [PubMed]
  • Second, G., 1982. Origin of the genic diversity of cultivated rice (Oryza spp.): study of the polymorphism scored at 40 isozyme loci. Jpn. J. Genet. 57: 25–57.
  • Sharp, P. M., D. L. Robertson, F. Gao and B. H. Hahn, 1994. Origins and diversity of human immunodeficiency viruses. AIDS 8: S27–S42.
  • Shull, V., and J. Hamer, 1994 Genome structure and variability in Pyricularia grisea, pp. 65–86 in Rice Blast Disease, edited by R. Zeigler, S. Leong and P. Teng. CAB International, Wallingford, Oxfordshire, UK.
  • Silué, D., and J. L. Nottéghem, 1990. Production of perithecia of Magnaporthe grisea on rice plants. Mycol. Res. 94: 1151–1152.
  • Silué, D., J. L. Nottéghem and D. Tharreau, 1992. Evidence of a gene-for-gene relationship in the Oryza sativa-Magnaporthe grisea pathosystem. Phytopathology 82: 577–580.
  • Simmons, M. P., and H. Ochoterena, 2000. Gaps as characters in sequence-based phylogenetic analysis. Syst. Biol. 49: 369–381. [PubMed]
  • Sorenson, M. D., 1999 TreeRot, Version 2. Boston University, Boston.
  • Sweigard, J. A., A. M. Carroll, S. Kang, L. Farrall, F. G. Chumley et al., 1995. Identification, cloning, and characterization of PWL2, a gene for host species specificity in the rice blast fungus. Plant Cell 7: 1221–1233. [PMC free article] [PubMed]
  • Swofford, D. L., 2002 PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods), Version 4.0. Sinauer Associates, Sunderland, MA.
  • Tajima, F., 1993. Simple methods for testing the molecular evolutionary clock hypothesis. Genetics 135: 599–607. [PMC free article] [PubMed]
  • Talbot, N. J., 2003. On the trail of a cereal killer: exploring the biology of Magnaporthe grisea. Annu. Rev. Microbiol. 57: 177–202. [PubMed]
  • Templeton, A. R., K. A. Crandall and C. F. Sing, 1992. A cladistic analysis of phenotypic association with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. III. Cladogram estimation. Genetics 132: 619–633. [PMC free article] [PubMed]
  • Thompson, J. D., D. G. Higgins and T. J. Gibson, 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22: 4673–4680. [PMC free article] [PubMed]
  • Thorne, J. L., H. Kishino and I. S. Painter, 1998. Estimating the rate of evolution of the rate of molecular evolution. Mol. Biol. Evol. 15: 1647–1657. [PubMed]
  • Tibayrenc, M., F. Kjellberg, J. Arnaud, B. Oury, S. F. Brenière et al., 1991. Are eukaryotic microorganisms clonal or sexual? A population genetics vantage. Proc. Natl. Acad. Sci. USA 88: 5129–5133. [PMC free article] [PubMed]
  • Tosa, Y., K. Hirata, H. Tamba, S. Nakagawa, I. Chuma et al., 2004. Genetic constitution and pathogenicity of Lolium isolates of Magnaporthe oryzae in comparison with host species-specific pathotypes of the blast fungus. Phytopathology 94: 454–462. [PubMed]
  • Urashima, A. S., Y. Hashimoto, L. D. Don, M. Kusaba, Y. Tosa et al., 1999. Molecular analysis of the wheat blast population in Brazil with a homolog of retrotransposon MGR583. Ann. Phytopathol. Soc. Jpn. 65: 429–436.
  • Valent, B., 1990. Rice blast as a model system for plant pathology. Phytopathology 80: 33–36.
  • Valent, B., and F. G. Chumley, 1994 Avirulence genes and mechanisms of genetic instability in the rice blast fungus, pp. 111–134 in Rice Blast Disease, edited by R. S. Zeigler, S. A. Leong and P. S. Teng. CAB International, Wallingford, Oxfordshire, UK.
  • Valent, B., M. S. Crawford, C. G. Weaver and F. G. Chumley, 1986. Genetic studies of fertility and pathogenicity in Magnaporthe grisea (Pyricularia oryzae). Iowa State J. Res. 60: 559–594.
  • Valent, B., L. Farrall and F. G. Chumley, 1991. Magnaporthe grisea genes for pathogenicity and virulence identified through a series of backcrosses. Genetics 127: 87–101. [PMC free article] [PubMed]
  • Vaughan, D. A., and H. Morishima, 2003 Biosystematics of the genus Oryza, pp. 27–65 in Rice: Origin, History, Technology, and Production, edited by C. W. Smith and R. H. Dilday. John Wiley & Sons, Hoboken, NJ.
  • Viji, G., B. Wu, S. Kang and W. Uddin, 2001. Pyricularia grisea causing gray leaf spot of perennial ryegrass turf: population structure and host specificity. Plant Dis. 85: 817–826.
  • Wang, G., D. Mackill, M. Bonman, S. McCouch, M. Champoux et al., 1994. RFLP mapping of genes conferring complete and partial resistance to blast in a durably resistant rice cultivar. Genetics 136: 1421–1434. [PMC free article] [PubMed]
  • Wang, Z., D. J. Mackill and J. M. Bonman, 1989. Inheritance of partial resistance to blast in indica rice cultivars. Crop. Sci. 29: 848–853.
  • Whitlock, M. C., 1996. The red queen beats the jack-of-all-trades: the limitations on the evolution of phenotypic plasticity and niche breadth. Am. Nat. 148: S65–S77.
  • Xia, J. Q., J. C. Correll, F. N. Lee, M. A. Marchetti and D. D. Rhoads, 1993. DNA fingerprinting to examine microgeographic variation in the Magnaporthe grisea (Pyricularia grisea) population in two rice fields in Arkansas. Phytopathology 83: 1029–1035.
  • Yamanaka, S., I. Nakamura, H. Nakai and Y.-I. Sato, 2003. Dual origin of the cultivated rice based on molecular markers of newly collected annual and perennial strains of wild rice species, Oryza nivara and O. rufipogon. Genet. Res. Crop Evol. 50: 529–538.
  • Zeigler, R. S., 1998. Recombination in Magnaporthe grisea. Annu. Rev. Phytopathol. 36: 249–275. [PubMed]
  • Zeigler, R. S., R. P. Scott, H. Leung, A. A. Bordeos, J. Kumar et al., 1997. Evidence of parasexual exchange of DNA in the rice blast fungus challenges its exclusive clonality. Phytopathology 87: 284–294. [PubMed]
  • Zolan, M. E., and P. J. Pukkila, 1986. Inheritance of DNA methylation in Coprinus cinereus. Mol. Cell. Biol. 6: 195–200. [PMC free article] [PubMed]

Articles from Genetics are provided here courtesy of Genetics Society of America
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Cited in Books
    Cited in Books
    PubMed Central articles cited in books
  • GSS
    Published GSS sequences
  • MedGen
    Related information in MedGen
  • Nucleotide
    Published Nucleotide sequences
  • PopSet
    Published population set
  • Protein
    Published protein sequences
  • PubMed
    PubMed citations for these articles
  • Taxonomy
    Related taxonomy entry
  • Taxonomy Tree
    Taxonomy Tree

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...