Molecular variability in the maize grey leaf spot pathogens in Brazil

Isolates of Cercospora species from leaves displaying symptoms of grey leaf spot were collected in maize-producing areas of south-central Brazil in 2001 and 2002. Restriction digests of the internal transcribed spacer region of rDNA detected the presence of the same two Cercospora species described on maize in the United States, namely C. zeae-maydis and the recently described species, C. zeina. Genetic variability among isolates was assessed by analysing 104 amplified fragment length polymorphism loci. Cluster analysis confirmed the genetic separation of isolates into two species with a mean similarity of 35%. Similarity levels within species were high, averaging 93% and 92% among isolates of C. zeae-maydis and C. zeina, respectively. The mean genetic similarity between C. zeae-maydis and C. zeina and two isolates of C. sorghi f. sp. maydis was 45% and 35%, respectively. Results of this study showed that populations of the grey leaf spot pathogens in Brazil are similar to those in the United States regarding species composition and that C. zeina is also present in Brazil.

Maize grey leaf spot disease is caused by Cercospora zeae-maydis, but Cercospora sorghi f. sp.maydis is occasionally isolated from infected leaves (Chupp, 1953).Wang et al. (1998) demonstrated that populations of the grey leaf spot pathogen in the U.S. are comprised of two sibling species, which they designated C. zeae-maydis group I and group II.Despite their morphological similarity, the two groups are substantially divergent based on analysis of amplified fragment length polymorphism (AFLP) data.In addition to the distinct AFLP profiles, the two groups can be distinguished by the restriction patterns of the internal transcribed spacer (ITS)-5.8SrDNA region.In comparative analyses of ITS-5.8SrDNA nucleotide sequences of Cercospora species and anamorphs of Mycosphaerella, Goodwin et al. (2001) concluded that groups I and II represent different species.Recently, Crous et al. (2006) applied DNA phylogeny of nucleotide sequences of five nuclear-encoded genes and analysed morphological characteristics of isolates of the grey leaf spot pathogen from South Africa and compared them with isolates of the pathogen from the U.S.A.They confirmed the distinct separation of two species, provided complete descriptions, and named the African pathogen Cercospora zeina, replacing the previous designation of C. zeae-maydis group II.
In the U.S., populations of C. zeae-maydis (group I) and C. zeina (group II) are sympatric in portions of the maize-producing regions, although C. zeina is restricted to the eastern third of the country (Wang et al., 1998).Significantly, the African population of the grey leaf spot pathogen is comprised solely of C. zeina (Dunkle & Levy, 2000;Okori et al., 2003), although the possibility of a third maize pathogen more closely aligned with C. apii and C. beticola was suggested by Crous et al. (2006).The considerable genetic divergence between the species, their differential geographical distribution, and slightly greater haplotype diversity of African isolates, together with historical records, led Dunkle and Levy (2000) to propose that C. zeina originated in Africa, although they did not rule out the possibility of independent origins in the U.S. and Africa.
The occurrence of grey leaf spot in Brazil was first reported in 1934 (Viégas, 1945), but it remained a disease of secondary importance until 2001, when severe outbreaks were reported in the central states (Fantin et al., 2001).The reasons for this shift in the severity and distribution of the disease are unclear but may be due to the introduction or emergence of more aggressive genotypes of the pathogen or to changes in maize hybrids or cropping systems or a combination of those factors.Thus, the objective of this study was to assess the genetic diversity of Brazilian isolates of the grey leaf spot pathogens collected in 2001/2002.Our results suggest that well-established populations of both species are present in Brazil and that within-species genetic diversity is very low.This is the first report of the occurrence of C. zeina in Brazil and assessment of the genetic variability among isolates of the grey leaf spot pathogen.
A collection of 69 monoconidial isolates of the grey leaf spot pathogen was obtained by direct isolation from diseased maize leaves randomly collected from production areas in the south, southeast, and west-central regions of Brazil (Table 1).Conidia were removed from the lesions, transferred onto potato dextrose agar (PDA) plates, and incubated for 3 to 5 days at 27 °C until typical colonies developed.For DNA extraction, mycelium was obtained from stationary liquid cultures grown in 200 mL of PD (potato-dextrose) medium for 14 days under continuous light at 27 °C.The DNA from four American and two African isolates of the grey leaf spot pathogen, previously classified into genetic groups I or II (Wang et al., 1998), and two isolates of C. sorghi f. sp.maydis were included in the analysis for genetic comparative purposes (Table 1).
To evaluate the production of cercosporin, cultures were grown on PDA and incubated at 27 °C for 14 days in constant light.Colonies secreting a reddish-purple pigment into the medium were scored positive for cercosporin production.
DNA was extracted from approximately 100 mg of freeze-dried mycelium according to Reader and Broda (1985).The isolates were analysed by restriction fragment length polymorphism (RFLP) of the ITS-5.8SrDNA region after digestion with TaqI (New England Biolabs) as described by Wang et al. (1998).In addition, individual digestions with MseI, BfaI, DdeI (New England -Biolab), HhaI, and HaeII (Fermentas) were also tested to investigate the existence of unique restriction patterns that may indicate additional species or genetic variability.The ITS-5.8S rDNA fragment was amplified by PCR with primers ITS4 (5' TCC TCC GCT TAT TGA TAT GC 3') and ITS5 (5' GGA AGT AAA AGT CGT AAC AAG G 3') (Dunkle and Carson, 1998).The reaction was carried out in a final volume of 50 μL according to Miller et al. (1999).Digestions were performed in a final volume of 20 μL with 5 μL of the PCR product and 8 units of each restriction endonuclease.Restriction fragments were separated by electrophoresis in a 3% agarose gel.AFLP amplifications were performed according to Vos et al. (1995) with the same primer combinations used by Wang et al. (1998).Genomic DNA was digested with EcoRI and MseI and ligated with doublestranded adaptors.Preamplification was performed with the EcoRI primer adapter plus the selective base A and with the MseI primer adapter plus the selective base A or C. Preamplified DNA fragments were diluted 1:5 (v/v) with ultra pure water and used as templates for the selective amplifications with primer EcoRI + A and either MseI + AT or MseI + CA.PCR products were separated by electrophoresis at 80W and a maximum temperature of 50 °C for 4 h in a 6% polyacrylamide gel in 1X TBE buffer.The gel was stained with silver nitrate and developed in sodium carbonate (Creste et al., 2001).Enzyme digestion and AFLP amplifications were carried out twice with a subset of eleven isolates in order to evaluate the reproducibility of the AFLP profiles.
For each isolate, AFLP loci were visually scored as present or absent and the AFLP profile was converted into binary data.Genetic similarity between isolates was calculated using the Dice coefficient based on paired comparisons of AFLP profiles, as described by Levy et al. (1993).A phenetic cluster (dendrogram) was obtained by the unweighted pair-group method with arithmetic averaging (UPGMA).Analyses were performed with the NTSYSpc version 1.8 statistical software (Exeter Software, Setauket, NY), and the dendrogram was constructed with the MEGA II -version 2.1 software program (Kumar et al., 2001).
The species composition of the grey leaf spot pathogen population was determined by restriction analysis of the PCR-amplified ITS-rDNA region (Wang et al., 1998).Digestion of the 520 bp long fragment amplified from all isolates with TaqI generated three different restriction patterns, which exactly match the patterns (not shown) of C. zeae-maydis, C. zeina, and C. sorghi described by Wang et al. (1998).In all, 41 isolates were classified as C. zeae-maydis (group I type), 28 as C. zeina (group II type), and two as C. sorghi f. sp.maydis.As expected, the control African isolates 5-34 and 5-24, and the U.S. isolate 19 exhibited the expected C. zeina restriction pattern, whereas the three other U.S. isolates (1-5, 14B, and 11B) exhibited the C. zeae-maydis pattern.Of the other five restriction enzymes tested, only BfaI failed to digest the ITS fragment; HhaI produced an identical RFLP pattern for all isolates; MseI distinguished isolates of C. zeae-maydis from isolates of C. zeina but did not distinguish C. zeina from C. sorghi f. sp.maydis, whereas restriction patterns from HaeII and DdeI digestions distinguished C. zeae-maydis and C. zeina, but did not distinguish C. zeae-maydis from C. sorghi f. sp.maydis.Isolates of C. zeae-maydis and C. zeina were detected in samples from all states, except Goiás, where only C. zeina isolates were detected.
The genetic variability among isolates comprising each species was determined by AFLP analysis.Both primer combinations generated over 150 AFLP loci, but for comparative analyses only those fragments ranging from 50 to 350 bp, a total of 104 AFLP loci, were included in the analysis.These scored loci were also present in the replicate AFLP amplifications performed with a sub-set of eleven isolates.Isolates grouped into two clusters, with a mean genetic similarity of 35% between them (Figure 1).
As reported by Wang et al. (1998), the clusters matched the grouping of isolates according to their ITS digestion patterns.The mean similarity among isolates within the clusters was 92% and 93.5% for C. zeina and C. zeae-maydis,  The AFLP analyses and the two distinct ITS-5.8SrDNA restriction patterns indicated the presence of two divergent species that correspond to those reported to occur in the U.S. Isolates of both species were recovered from samples taken from locations separated by as far as 1,500 km.The widespread occurrence of such highly divergent groups suggests that the grey leaf spot epidemics in Brazil were caused by factors other than a sudden change in the genetic composition in the population of the grey leaf spot pathogen, or by a recent introduction of more aggressive genotypes.Among the possible factors are the relatively recent modifications in maize cultivation in Brazil, which include the widespread use of susceptible hybrids, the extension of the total cropping season resulting from an increase in acreage of late-season crops notably in the west-central regions, and the adoption of the reducedtillage cropping system known to favour the survival of the pathogen (Payne et al., 1987;de Nazareno et al., 1993).Together, these practices may have contributed to a build-up of inoculum that led to epidemic.Thus, it appears that the history of occurrence of grey leaf spot in Brazil is similar to that in the U.S., where the pathogen was not regarded as economically important for several years after it was first reported, but became a major pathogen concomitant with incorporation of changes in cultivation (Roane et al., 1974;Latterell and Rossi, 1983;Ward et al., 1999).
The within-species levels of genetic similaritiy among Brazilian isolates of both species were nearly identical to those reported for U.S. isolates: 93 and 92% for Brazilian isolates compared to 93 and 94% for U.S. isolates of C. zeae-maydis and C. zeina, respectively (Wang et al., 1998;Dunkle and Levy, 2000).However, the variability between C. zeae-maydis and C. zeina was slightly higher in the U.S., where the species were separated by a mean genetic distance of 80%, compared to Brazil, where the mean genetic distance was 65%.Although both studies used the same AFLP primer combinations and analysed a similar number of loci (111 in the U.S. study and 104 in this study), the discrepancy may reflect a sampling effect of the specific AFLP loci scored and of the number of isolates analysed.Among the 91 isolates included in the U.S. study, only 12 were C. zeina isolates compared to 28 of the 69 Brazilian isolates.Nevertheless, the extent of genetic variability observed within and between Brazilian species supports the conclusions of Wang et al. (1998) that the two groups they defined in the U.S. represent distinct gene pools and are likely to be different species, as suggested by Goodwin et al. (2001) and recently confirmed by Crous et al. (2006).
This study also demonstrated a high genetic similarity among Brazilian, U.S., and African isolates.The mean similarity between Brazilian and U.S. isolates of C. zeaemaydis was 90%, a value comparable to the variability found within this species in both countries, as discussed above.Brazilian isolates of C. zeina were on average 90.2% and 89% genetically similar to the U.S. and African isolates of the same species, respectively.Similarly, the East African population of C. zeina was found to lack population structure and was indistinguishable from the U.S. C. zeina (group II) population based on AFLP and RFLP analyses (Okori et al., 2003).Thus, there appears to be limited genetic variability within populations of C. zeina worldwide, which is consistent with the likelihood that this species reproduces asexually and indicates that geographically separated populations do not differ substantially from the founding population, which is proposed to be in Africa (Dunkle and Levy, 2000;Crous et al. 2006).
In the U.S., C. zeina isolates are restricted to the eastern part of the country, whereas C. zeae-maydis is widespread throughout the maize-growing regions.In Brazil, C. zeae-maydis and C. zeina are sympatric in the maizeproducing regions of the country that were sampled.On the other hand, isolates of C. zeae-maydis have not been recovered from Africa (Dunkle and Levy, 2000;Okori et al., 2003;Crous et al., 2006), suggesting that the grey leaf spot pathogen was not introduced into that continent from the U.S. as proposed by Ward et al. (1999).Further considerations, such as historical perspective of the occurrence of the disease on both continents and the relatively greater haplotype diversity in the African population, provide convincing evidence that C. zeina was introduced into the U.S. from Africa or from an unidentified source (Dunkle and Levy, 2000).The chronological records of grey leaf spot in Brazil provide no definitive clues as to which of the two species was the predecessor there, or whether both species have been present since the first observation in 1934.However, the presence of C. zeina in Brazil offers a second alternative to account for its occurrence in the U.S. as well as in Africa.Crous et al. (2006) offered the tenable hypothesis that C. zeina is native to Africa as a pathogen on another indigenous host, such as sorghum, and recently evolved pathogenicity to maize.In any event, the limited genetic variability among the C. zeina populations makes untenable a hypothesis proposing independent origins to explain its occurrence on three continents.
Nevertheless, the most obvious and consistent features distinguishing the two species are the faster growth and production of the reddish-pigmented phytotoxin cercosporin only by isolates of C. zeae-maydis.
Grey leaf spot can be managed by the use of genetically resistant hybrids.Thus, inferences on the durability of resistance genes are of great interest to plant breeders.Despite significant interactions between host genotypes and environments regarding resistance to Cercospora zeaemaydis (Thompson et al., 1987;Bubeck et al., 1993;Lehmensiek et al., 2001), the low genetic variability within and between species of the grey leaf spot pathogen assessed from populations from three very distinct geographical areas (U.S.A., Africa, and Brazil) suggests that variability in disease severity is related to environmental factors rather than to genetic variation in the pathogen.Given the relative genetic uniformity of the respective asexual pathogen populations, the sources of quantitative resistance in maize germplasm will most likely remain effective in Brazil, U.S.A., and Africa.

Figure 1 -
Figure 1 -Genetic relationships among Brazilian isolates of Cercospora zeae-maydis, C. zeina and C. sorghi f. sp.maydis based on the unweighted pair-group method with arithmetic averages (UPGMA) analysis of AFLP data.

Table 1 -
Identification and origin of Cercospora spp.isolates used in this study.
respectively.The mean similarity among C. zeae-maydis isolates from Brazil and the United States was 90%, whereas among C. zeina isolates the mean similarity was 90.2%.The mean similarity between African and Brazilian isolates of C. zeina was 89%.A third group, consisting of the two C. sorghi f. sp.maydis isolates, displayed mean similarities of 45% and 35% with isolates of C. zeaemaydis and C. zeina, respectively.All 69 isolates grew well on PDA, producing greenish grey mycelium.The growth rate of C. zeae-maydis isolates was consistently and noticeably faster than that of C. zeina.Of the 41 isolates of C. zeae-maydis analysed, 26 (63%) produced cercosporin, but none of the 28 isolates of C. zeina produced the toxin.