Genetic diversity of the blast fungus , Magnaporthe grisea ( Hebert ) Barr , in Burkina Faso

Trapping nurseries trialed at two screening sites in Burkina Faso appeared to be an effective tool to characterize the virulence spectrum of blast populations using limited equipment and reduced labor. It made it possible to identify the best site to be used for screening for durable resistance. The effectiveness of some resistance genes indicated that they could be pyramided to provide durable resistance to blast fungus in Burkina Faso. The study also revealed the possible existence of new pathotypes in Burkina Faso. Fifty-five isolates of the blast fungus, Magnaporthe grisea, collected from the nurseries and rice fields were analyzed using random amplified polymorphic DNA (RAPD) PCR. Five major groups (Mg-1, Mg-2, Mg-3 Mg-4 and Mg-5) were defined. Mg-1, Mg-2 and Mg-3 were the largest groups representing, 30.9, 25.5 and 30.9% of the 55 isolates analyzed. Only 9.1 and 3.6% belong to Mg-4 and Mg-5, respectively. Our results confirmed that RAPD PCR offers an inexpensive and speedy means of generating markers for analyzing the population structure of the blast fungus.


INTRODUCTION
Blast disease is caused by Magnaporthe grisea (Hebert) Barr., anamorphe Pyricularia grisea Sacc.(Rossman et al., 1990) is an important fungal disease of rice known to occur in most rice producing areas of the world (Ou, 1985).The disease results in yield loss as high as 70 -80% (Ou, 1985) when predisposition factors (high mean temperature values, degree of relative humidity higher than 85 -89%, presence of dew, drought stress and excessive nitrogen fertilization) favor epidemic development (Piotti et al., 2005).Losses of nearly 80% have been reported in certain years in West Africa (Delassus, 1973).Particularly dangerous in upland rice, it also causes serious damage in rainfed lowland and irrigated systems in Burkina Faso, mainly when farmers seek to intensify production by the use of improved varieties and fertilizers (Séré, 1999).Therefore, blast constitutes one of the main constraints to intensification for increasing rice *Corresponding author.E-mail: y.sere@cgiar.org.Fax: +229 21 35 05 56; Tel.: +229 21 35 01 88, production.
The use of resistant varieties is the most economical and effective way of controlling rice blast mainly in resource-poor farmers' fields.Unfortunately, the causal fungus is able to overcome this resistance within two to three years after these plants are cultivated widely (Babujee and Gnanamanickham, 2000).The breakdown in resistance has been attributed to the high variability of the pathogen and there are numerous reports that this diversity may be due to continuous generation of novel pathogenic variation.Knowledge of the genetic variation within and among populations is an important component of understanding the population biology of pathogenic fungi and infers the impact of driving force influencing the evolution of pathogen populations (McDonald et al., 1997).Therefore, information on population diversity may be used for developing strategies to increase the durability of resistance (Xia et al., 2000).
There has been a tremendous accumulation of knowledge on pathogen population diversity in the past decade especially with the development of molecular techniques.In particular, DNA fingerprinting was used 12 Pi N°4 Pita2+Pish (2) 13 Fukunishiki Piz + Pish (2) 14 Bl 1 Pib + Pish (1)

K1 Pita
(1)  Inukai et al. (1994) extensively to reveal the genetic diversity of plant pathogens needed for effective deployment of resistance and to identify shifts in races or population structures that might occur (Javan-Nikkhah et al., 2004).However, the studies of genetic diversity of M. grisea in Burkina Faso have not previously been critically conducted, as very little research information is available.Therefore, the aim of this current study is to assess the pathological diversity of the blast pathogen at two screening sites and conduct genetic fingerprinting of M. grisea isolates for the first time in Burkina Faso using random amplified polymorphic DNA (RAPD) markers.It is anticipated that this study will lead to a better understanding of the diversity and distribution of blast pathogens both from rice and non-rice hosts, and to its potential application in rice breeding programs aiming at development of durable blast-resistant rice cultivars.Séré et al. 2569

Assessment of the pathological diversity
Trapping nurseries were conducted at two screening sites in Burkina Faso (Farako-Ba research station in upland condition, and Banfora Lowland rice experimental site) using 19 cultivars bearing known resistance genes, 4 released varieties and 2 varieties as susceptible checks (Table 1).Symptoms were recorded on each variety by considering plants without symptoms or with only hypersensitive lesions as resistant to the blast population on both sites (Thinlay et al., 2000).

Blast fungus collection and culture
Lesions were collected at the epidemic initiation and early in the morning when high spore production was observed on the lesions (Pinnschmidt et al., 1993).Additional samples were also collected from local and improved varieties in farmers' fields and from wild rice (Oryza longistaminata) and graminaceous weeds in the main rice-growing localities of the country (Table 2).With a sterile needle, a fragment of mycelium and conidiophores with spores was taken from sporulating lesions and moved gently on the surface of a water agar medium in Petri dishes (10 g of agar, 50 mg of penicillin and 50 mg of dihydrostreptomycin in an adequate quantity of distilled water for 1 liter of medium).The spores were then dispersed on the surface of the medium, and single ones were surrounded under a stereomicroscope.Six to ten hours later, the germinating single marked conidia were transferred to starch medium (10 g of soluble starch, 1 g of yeast extract, 10 g of agar, 50 mg of penicillin and 50 mg of dihydrostreptomycin in an adequate quantity of distilled water for 1 liter of medium).Such monoconidial in vitro cultures were grown at room temperature for 6 to 7 days before transfer for long term conservation according to method adapted from Valent et al. (1986).The conservation method used includes growing the fungus on filter papers overlaying the starch medium.After colonization, the paper discs are transferred aseptically under lamina flow in sterile paper filter bags, dried in glass boxes containing silica gel for 5 to 6 days, and then transferred into sealed plastic under vacuum.
They are stored at 4 o C.

Genomic DNA extraction
Mycelia and spores from 10 days old monoconidial culture are transferred into 75 ml of potato dextrose broth (pH 7.4) in a 250 ml conical flask and kept under constant shaking at 28 o C for 6 days.The resulting mycelia ball was freeze-dried, and DNA was extracted using the standard CTAB (cetyl trimethyl ammonium bromide) method (Hamer and Givan, 1990).

RAPD-PCR analysis
RAPD-PCR analysis (Guthrie et al., 1992) was carried out on all the isolates collected both from rice and non-rice host.DNA primers tested were purchased from Operon Technologies (Alameda, California, USA) and each is 10 nucleotides long.Two concentrations of each DNA (24 ng and 96 ng per reaction) were used to test reproducibility and eliminate sporadic amplification products from the analysis.Eighty primers (OPA, OPB, OPC and OPZ series), including PAP2 and PAP3, were screened with three isolates (MGR 272, MGR 069 and MGR 097) for their ability to amplify the M. grisea DNA.The primers that showed polymorphism were used in amplifying the DNA from all M. grisea isolates.Amplifications were performed in a 25 µl reaction mixture consisting of genomic DNA, 1X reaction buffer (Promega), 100 µM each of dATP, dCTP, dGTP, Gels were visualized by staining with ethidium bromide solution (0.5 µg/ml) and banding patterns were photographed over UV light using a red filter.

Phylogenetic analysis
Positions of unequivocally scorable RAPD bands were transformed into a binary character matrix ("1" for the presence and "0" for the absence of a band at a particular position).Pairwise distance matrices were compiled by the NTSYS-pc 2.0 software (Rohlf, 1993) using the Jaccard coefficient of similarity.Phylogenetic tree was created by the unweighted pair-group method arithmetic (UPGMA) average cluster analysis (Sneath and Sokal, 1973;Swofford and Olsen, 1990).

Virulence of the blast population at two screening sites
The term virulence is used to differentiate the ability of a blast fungus strain to overcome the resistance gene of a particular rice variety.The virulence spectrum of an isolate or a population refers to the number of varieties with different resistance genes to be attacked.The trapping nurseries developed at two screening sites indicated that 12 of the varieties bearing known resistance genes were attacked at Farako-Ba; meanwhile, only four developed susceptible lesions at Banfora (Table 3).The virulence spectrum of the blast fungus population was therefore broader at Farako-Ba than at Banfora.The higher diversity of the blast population at Farako-Ba might be due to the diversity of rice germplasm cultivated over the years at this experimental station as host genotype is known to influence the composition of the pathogen population (Chen et al., 1995;Park et al., 2003).Such diversity indicates that Farako-Ba is a better screening site for durable resistance than Banfora, as resistance breakdown had been most often attributed to a poor pre-release challenge by an adequate pathogen population (Mekwatanakarn et al., 2000).The M. grisea population at both Farako-Ba and Banfora sites could not match the single gene Pita in K1 and the multiple genes Pib+Pish, Pit+Piks, Pizt+Pish, Pikh+Pi1+Pita+Pi in BL1, K59, Toride 1 and Tetep, respectively.However, Yachiro Mochi bearing the same Séré et al. 2571 resistance gene as K1 (Pita) is susceptible at Farako-Ba.Therefore, the resistance of K1 is not only due to Pita but to an additional undetermined resistance gene.Furthermore, the resistance of Sha Tia Tsao at Farako-Ba might be due to an additional resistance gene since Caloro, bearing the same Piks gene, is attacked.On the other hand, all the strains that could attack Ishikari Shirokee should have been virulent on Sha Tiao Tsao.Therefore, the lack of susceptible lesions on Sha Tiao Tsao at Farako-Ba while Ishikari Shiroke was attacked means that Sha Tiao Tsao had an undermined efficient resistant gene against the Farako-Ba blast population.Such a situation is not surprising in identifying and characterizing resistant genes using Japanese differential cultivars.Some resistant genes detected using Japanese fungus strains, could not be detected with Philippines isolates.Similarly, Japanese and Kiyosawa differential varieties include some resistance genes which could be recognized by Japanese differential fungus strains.Therefore, new resistance genes are found using isolates from different origins.For instance, Fujisaka 5 was first selected as a differential variety with a single resistance gene (Pii).Later, an additional gene, Piks, was found in Fujisaka 5 (Kiyosawa and Ling, 2001).Piks was found in Japanese cultivars only with a Filipino fungus strain.In Tetep, Kiyosawa (1976) first estimated the presence of Pikh.Thereafter, Pi-1, Pita and an unknown gene were found within Tetep (Kiyosawa and Ling, 2001).In the same way, Pish was identified later in Shin 2, Kusabue, Fukunishiki, Pino4 and Bl1 previously known, respectively, to carry Piks, Pik, Piz, Pita2 and Bl1 (Inukai et al., 1994).Moreover, some of the differential lines could carry several major and environmentally-sensitive minor resistance genes causing intermediate and poorly repeatable reactions even in well-controlled conditions (Mekwatanakarn et al., 2000).

Molecular characterization of blast fungus isolates
Ten primers out of 82 primers tested showed polymorphism among individual isolates.The amplification reactions with the 10 primers generated 153 bands, 108 of them being polymorphic (Table 4) with sizes ranging between 150 and 1,000 base pairs (Figure 1).The amplified DNA distinct band pattern obtained allowing the identification of each individual.For instance, isolate MGW036 presents unique bands when its DNA is amplified with most of the primers tested (Figure 1).These bands could be used to characterize and identify it.
The study revealed that among the five groups, only Mg-1 was found at Farako-Ba while Mg-2, Mg-3 and Mg-4 were distributed in two or three 3 localities.Isolates originating from the same host plants and from the same localities belong to different groups.Similar results were obtained by Xia et al. (1993) who found four lineages from a single cultivar in two fields in a single region in the USA.Therefore, sampling for assessing blast pathogen genetic diversity should take such results into consideration in order to ensure that the individuals sampling do reflect the genetic diversity of the population.
The high distinction pattern of each isolate obtained in this study suggests possible and frequent occurrence of mutants in M. grisea in different host cells (Bronson et al., 1990;Chumley and Valent, 1990;Levy et al., 1991;Klister and Miao, 1992).The limited number of morphological and cultural characters of M. grisea, and the lack of standardization of cultural conditions and virulence tests among the different researchers have led to confusion and uncertainty in the characterization of this pathogen (Babujee and Gnanamanickham, 2000).Distinct phenotypes usually consist of isolates that are genetically less related and such identification of isolates using cultural and morphological techniques often lacks consistency and precision (Babujee and Gnanamanickham, 2000).In the current study, we have found that identification of genetic diversity in M. grisea depends on different host cells and occurrence of mutants.For instance, 17 isolates genotyped as Mg-1 were originated only from cultivated rice varieties from Farako-Ba experimental station, while 14 isolates from O. longistaminata and associated cultivated rice were genotyped as Mg-2.Furthermore, the Mg-3 genotype consists of 22 isolates originated from other rice-related weeds and associated cultivated rice varieties while two isolates that do not interact with a cultivated rice variety but originated only from O. longistaminata were genotyped Mg-4.
The possible population structure, frequency and distribution of M. grisea genotypes in Burkina Faso have been revealed by this study.RAPD markers indicated possible relationship between host origin, mutation and genetic variation among M. grisea isolates, and this demonstrated the fingerprinting and diagnostic potential of RAPD.Obviously, for these DNA band patterns to have practical meaning in the areas of plant pathology, population biology and molecular epidemiology, specific DNA bands must be related to host origins, mutation and virulence genes (Welsh and McClelland, 1990).This could be accomplished by a systematic comparison of DNA band patterns among fungi, contrasting for the different host origins, mutation and the virulence genes present.A similar approach has been used to differenttiate aggressive from non-aggressive isolates of the oilseed rape pathogen Phoma lingam (Schafer and Wostmeyer, 1992).
The DNA fingerprint defined for each race of M. grisea should be useful for epidemiological surveys, disease diagnoses, and in the identification of new virulent strains, isolates and their origin.This information could be useful in rice breeding programs aiming at development of a lineage exclusion method (Zeigler et al., 1994) in breeding for durable blast-resistant rice cultivars to different rice ecologies and localities in Burkina Faso.However, our study didn't reveal any correspondence between virulence of M. grisea and the fingerprinting group as each group is composed of several virulence strains.A study conducted by Rathour et al. (2004) indicated that no correlation was observed between RAPD pattern and the virulence characteristic of the pathogen.The degree of correlation between lineages and pathotypes varies among populations, strong in the USA, Colombia and Europe, but moderate in the Philippines and Vietnam, while no clear correlation was observed in Japan (Park et al., 2003).Therefore, the relationship between virulence and lineage seems to be less complex in countries with a short history of rice cultivation and with a limited number of lineages than in those countries with a long history and high number of lineages (Piotti et al., 2005).

Conclusion
In order to analyze the pathogenic population structure of the blast fungus in Burkina Faso, trapping nurseries of varieties with known resistance genes were installed.Our study indicated that such nurseries are an effective tool to characterize the virulence spectrum of blast populations.They also avoid the difficulties of collecting representative samples and minimizing variability inherent to the experimental system.The nurseries made it possible to characterize the best site to be used for screening for resistance to blast disease.The results also revealed the effectiveness of some resistance genes: Pita, Pib associated with Pish, Pit and the three genes (Piks, Pizt and Pish) associated in Tetep.Pyramiding some of these resistance genes can provide durable resistance to the blast fungus in Burkina Faso as the resistance of cultivars that carry several resistance genes should be longerlasting than those carrying single genes (Kiyosawa, 1982).However, the reaction of varieties with known resistance genes indicated that some of them may bear additional undermined resistance genes which mean that the blast fungus pathoype in Burkina Faso, as revealed by the reaction of the varieties carrying known resistant genes, is different from what was found so far in Asia.Therefore, it appeared necessary to reassess the behavior of the resistance genes identified in Asia against the Burkina Faso blast population because the precise delineation of pathogenic variability in the rice production area is a prerequisite for identifying rice genotypes with a broader resistance spectrum (Rathour et al., 2004).The  Molecular markers are used extensively to characterize plant pathogens and elucidate population genetic structure and the evolutionary relationship of plant pathogens (Traoré et al., 2005;Fargette et al., 2004).Most of the molecular studies carried out on the blast pathogen were done using restriction fragment length polymorphism (RFLP) techniques with the probe MGR-586 (Levy et al., 1991;Chen et al., 1995;Mekwatanakarn et al., 2000).However, such techniques are expensive and timeconsuming, making them prohibitive for the analysis of a large number of samples (George et al., 1998) as well as for laboratories with modest facilities.Our results confirm that DNA (RAPD) offers an inexpensive and speedy means to generate molecular markers for analyzing the population structure of blast fungus.However, further investigation is needed on the relationship between DNA fingerprinting group and pathotypes in order to develop methods to establish population structures and define efficient resistance genes for use in breeding for durable blast resistance.

Table 1 .
Identity of varieties used in trapping nurseries at Farako-Ba and Banfora.

Table 2 .
Identity of M. grisea isolates used.

Table 3 .
Reaction of 24 varieties in trapping nurseries at Farako-Ba research station and Banfora rice experimental site in Burkina Faso.

Table 4 .
Oligonucleotide primers that showed genetic discrimination among the M. grisea isolates using RAPD-PCR analysis.

Table 5 .
DNA fingerprint groups and corresponding isolates with their host plant and locality of origin