Genetic Variation in Fusarium Section Liseola from No-Till Maize in Argentina

Abstract

Strains of Fusarium species belonging to section Liseola cause stalk and ear rot of maize and produce important mycotoxins, such as fumonisins. We isolated two species, Fusarium verticillioides (Gibberella fujikuroi mating population A) and Fusarium proliferatum (G. fujikuroi mating population D) from maize cultivated under no-till conditions at five locations in the Córdoba province of Argentina. We determined the effective population number for mating population A (N_e) and found that the N_e for mating type was 89% of the count (total population) and that the N_e for male or hermaphrodite status was 36%. Thus, the number of strains that can function as the female parent limits N_e, and sexual reproduction needs to occur only once every 54 to 220 asexual generations to maintain this level of sexual fertility. Our results indicate that the fungal populations isolated from no-till maize are similar to those recovered from maize managed with conventional tillage. We placed 36 strains from mating population A into 28 vegetative compatibility groups (VCGs). Of the 13 strains belonging to five multimember VCGs, only 2 isolates belonging to one VCG were clones based on amplified fragment length polymorphism (AFLP) fingerprints. Members of the other four multimember VCGs had an average similarity index of 0.89, and members of one VCG were no more closely related to other members of the same VCG than they were to other members of the population as a whole. This finding suggests that the common assumption that strains in the same VCG are either clonal or very closely related needs to be examined in more detail. The variability observed with AFLPs and VCGs suggests that sexual reproduction may occur more frequently than estimated by N_e.

Argentina produces approximately 15,000,000 tons of maize per year, primarily in three provinces, Buenos Aires, Santa Fe, and Córdoba; nearly one-half of this production, 48%, is exported (38). Until 1990, conventional tillage practices dominated Argentinean maize production, but no-till cropping practices were used for 78% of the 1995-1996 crop in the southern portion of Córdoba province (14). No-till has advantages in terms of fuel consumption and erosion control but also can significantly alter the relative frequency of the pathogens present, as determined for some soilborne pathogens (1). Fusarium verticillioides (Sacc.) Nirenberg (= Fusarium moniliforme Sheldon = Gibberella fujikuroi mating population A; teleomorph, Gibberella moniliformis Wineland) is the dominant Fusarium species, accounting for up to 100% of the Fusarium in freshly harvested maize planted under conventional tillage conditions in Argentina, although Fusarium proliferatum (Matsushima) Nirenberg (= G. fujikuroi mating population D; teleomorph, G. fujikuroi (Sawada) Wollenw. var. intermedia Kuhlman) (2.3 to 36%) and Fusarium subglutinans (Wollenweber et Reinking) Nelson, Toussoun, et Marasas (= G. fujikuroi mating population E; teleomorph, Gibberella subglutinans (Edwards) Nelson, Toussoun, et Marasas) (0.3 to 41%) also may appear at significant frequencies (5, 11, 37, 39). All of these species are heterothallic and have dimictic mating systems (17). These species can produce mycotoxins, such as fumonisins, fusaric acid, beauvericin, fusaproliferin, and moniliformin (2). The fumonisins are the most prominent of these toxins and are produced primarily by F. verticillioides and F. proliferatum (35). These toxins are known to occur naturally in Argentinean maize and to be produced in laboratory cultures by Argentinean isolates of F. verticillioides and F. proliferatum (5, 6, 37, 39).

Comparing fungal field populations can be difficult. The effective population number (N_e) provides an estimate of a population's size relative to the size of a randomly mating population (4, 24), and the equations used depend upon the reproductive constraints operating within the population. This statistic is usually used to evaluate populations when mating is not random and when individual members of the population do not contribute equally to the gene pool of the succeeding generation. For heterothallic ascomycetes, the relative frequencies of the mating type alleles (maximum N_e occurs when frequencies are equal) and the proportion of self-sterile hermaphrodites (maximum N_e occurs when all strains are of this type) are the primary constraints on N_e. We hypothesized that no-till maize, with its increased field surface residue, might provide a greater opportunity for sexual recombination than that available in a conventionally tilled field, where the stubble is either incorporated or removed.

Most recent assessments of fungal pathogens have used multilocus markers to characterize populations (for reviews see references 30 and 41). Vegetative compatibility groups (VCGs) have been used to estimate genetic variability within Fusarium field populations (8, 10, 21). Members of the same VCG are identical at a set of at least 10 different vic loci that are dispersed throughout the genome (21). For ease of analysis, members of the same VCG are usually assumed to be clones, but this assumption has not been rigorously tested (23).

Multilocus molecular markers, including restriction fragment length polymorphisms (10, 18) and PCR-based random amplified polymorphic DNAs (9, 15, 29), also have proven to be useful for characterizing some populations of fungal pathogens. Recently, another PCR-based method, amplified fragment length polymorphism (AFLP) (40), has been used to characterize fungal populations (12, 27, 42). AFLPs can provide complex marker profiles with no prior cloning or sequence data, and it is possible to generate a very large number of markers (as many as 40 to 50 markers per primer pair depending on the genome size and the primers selected) much more quickly than is possible with typical random amplified polymorphic DNA or restriction fragment length polymorphism analyses. The usually large number of useful AFLP markers permits a more robust statistical analysis—correlations between subsets of AFLPs appear to be relatively high (12, 42)—and, presumably, represents greater coverage and dispersal of markers across the genome. The utility, repeatability, and efficiency of the AFLP technique are leading to broader application of this technique to analysis of fungal populations (12, 27, 42).

Our objectives in this study were (i) to identify common species belonging to Fusarium section Liseola in maize grown under no-till conditions in Córdoba province, Argentina, (ii) to estimate the N_e for G. fujikuroi mating population A in the fields examined, and (iii) to determine the relative number of clones within the population, as assessed by the number of VCGs and the AFLPs observed.

MATERIALS AND METHODS

Strain recovery and identification.

Isolates of Fusarium species were recovered from seeds from five maize fields managed under no-till conditions in Córdoba province during the 1995-1996 growing season (Fig. 1). Samples (42 ears per sample) from each field were collected, hand shelled, pooled, and used for Fusarium species isolation. The fields were within a 50-km-diameter region, and we assumed that the strains belonged to a single population for analytical purposes. One hundred maize seeds from each sample were plated on a medium containing pentachloronitrobenzene (33). Isolates were subcultured as single spores by dilution plating and then identified by using the morphological criteria of Nelson et al. (34) and differences in isozyme band patterns (13). Ninety-six strains were randomly selected from a total of 712 isolates for further analysis.

FIG. 1.

Strain recovery and analysis scheme used in this study.

Mating type and female fertility.

Crosses to confirm mating populations and to identify mating types were made in triplicate on carrot agar by using the protocol of Klittich and Leslie (19) with standard tester strains A-00149 (MATA-1), A-00999 (MATA-2), D-04853 (MATD-2), and D-04854 (MATD-1) as female parents and the uncharacterized field isolates as male parents. A cross was scored positive only if we observed perithecia oozing a cirrhus of ascospores. Female fertility of the field strains belonging to mating population A was determined in crosses in which the field isolates were the female parents and the standard testers were the male parents. N_e was calculated based on mating type ratios [N_e(mt)] and the proportion of female fertile strains [N_e(f)] by using the equations of Leslie and Klein (24), as was the average number of asexual generations per sexual generation.

Vegetative compatibility tests.

Thirty-six of the strains belonging to G. fujikuroi mating population A were tested for vegetative compatibility (7). nit mutants were generated on minimal medium plus 3% KClO₃ and were assigned to one of three phenotypic classes (nit1, nit3, or NitM) based on their growth on several media containing differential nitrogen sources. Pairings to test for vegetative compatibility were made in 24-well hybridoma plates as previously described by Klittich and Leslie (20).

DNA isolation and AFLPs.

DNAs were isolated from cultures grown on complete medium (7) in shake flasks by using the cetyltrimethylammonium bromide method of Murray and Thompson (32), as modified by Kérenyi et al. (17). AFLPs were generated by the method described by Vos et al. (40), as modified by Zeller et al. (42), by using genomic DNAs digested with EcoRI and MseI. We amplified particular AFLP fragments by using 2-base extensions from both the EcoRI (GG or TT) and MseI (AC or CA) ends of the fragments. AFLP bands were scored manually for each of the three primer-pair combinations; the presence or absence of each band was determined for each individual. Bands of the same molecular size in different individuals were assumed to represent the same allele, and each band was treated as a single independent locus with two alleles.

We estimated the genetic distance between isolates with data from three AFLP primer-pair combinations by using unweighted pair grouping by mathematical average algorithims (SAS, version 6.11 for PC; SAS Institute, Cary, N.C.).

RESULTS AND DISCUSSION

Biological species and mating type.

The 96 strains examined were assigned to either the A (70 strains) or D (26 strains) mating populations of G. fujikuroi. Both of these mating populations have previously been recovered from maize in Argentina and elsewhere (2, 5, 6, 8, 16, 22). Thus, the A and D mating populations of G. fujikuroi are dominant on no-till maize grain at harvest time, just as they are on maize cultivated with traditional tillage methods. Toxigenically, representatives of these mating populations have been shown to produce fumonisins, fusaproliferin, moniliformin, and beauvericin (25, 26, 31), although so far strains from Argentina have been chemically confirmed to produce only fumonisins B₁, B₂, and B₃.

The larger sample of 712 isolates contained 2 to 3% G. fujikuroi mating population E. In other locations in Argentina, strains belonging to this mating population may constitute as much as 36% of the population (A. Torres, M. M. Reynoso, F. Rojo, M. L. Ramirez, and S. Chulze, submitted for publication). We attribute the relative scarcity of strains of mating population E in our study to differences in climate rather than to differences in tillage practices. The Córdoba region has a temperate climate with a drier winter (annual precipitation, 600 to 700 mm) and a lower elevation than does semitropical Jujuy province (from which previous samples were obtained [5, 6]). Jujuy province has a semitropical highland climate (elevation, approximately 3,000 m) with annual precipitation of 220 to 340 mm and less extreme but generally cooler temperatures (mean annual temperature, 9.4 to 10.5°C) than those found in Córdoba province.

The mating types segregated 23:47 MATA-1–MATA-2 for isolates from the A mating population and 13:13 MATD-1–MATD-2 for isolates from the D mating population. The mating type ratio in the A mating population was the inverse of the ratio reported by Leslie and Klein (24) but similar to the ratio reported by Park et al. (36) for isolates recovered from Korean maize. We found that 8 of 70 strains from mating population A were fertile as the female parent.

N_e.

N_e was calculated by using mating type and female fertility data for the strains from G. fujikuroi mating population A. The sample size for the D mating population was too small for a reliable estimate of either N_e parameter to be made. In the sample for G. fujikuroi mating population A, N_e(mt) was 88% of the count and N_e(f) was 37%. Thus, the limited proportion (11%) of hermaphrodites in the population had significantly more impact on N_e than the skewed mating type ratio (47:23) had (the probability that the MATA-1/MATA-2 ratio was 1:1 was <1%). The N_e(mt) value was different from the value for strains of the A mating population from Tanzania (28) but similar to the value for strains from the United States (24). N_e(f) was much lower in this Argentinean population than it was in comparable populations of mating population A organisms from both the United States and Tanzania (24, 28). Indeed, this Argentinean population more closely resembled the relatively infertile populations of mating population F organisms from the United States (24) or mating population H organisms from South Africa (3). Depending on assumptions with respect to selection against female-fertile strains during vegetative growth and the mutation rate at which female fertility is lost, the average number of asexual generations per sexual generation was between 54 and 220, and the frequency of hermaphrodites in such an equilibrium population could range from 1 to 33% (observed value, 11%). These results suggest that sexual reproduction does not occur frequently in the population analyzed.

VCGs and AFLPs.

We identified 28 VCGs among 36 strains, for a genetic diversity (number of VCGs/number of isolates) of 0.77. Twenty-three strains belonged to single-member VCGs, and the remaining 13 belonged to five multimember VCGs. VCG 3 contained five strains, while the other four multimember VCGs contained two strains each. VCG 3 contained strains from three different fields, while VCGs 4 and 5 each contained strains from two fields. VCGs 1 and 2 were limited to strains from a single field. VCGs 2 through 4 were associated with only a single mating type, while VCGs 1 and 5 contained one strain of each mating type.

With the 13 strains belonging to the five multimember VCGs, we identified 97 polymorphic AFLP bands among a total of 153 bands generated following amplification with three different primer pairs, as follows (number of polymorphic bands/total number of bands): EcoRI+GG and MseI+AC, 32/43; EcoRI+GG and MseI+CA, 25/42; and EcoRI+TT and MseI+CA, 40/68. Only the VCG 2 isolates were clones. Isolates belonging to the other multimember VCGs had average similarity indices of 0.80, 0.89. 0.90, and 0.78. The average similarity among the 13 strains was 86%, which was not significantly different from the similarity obtained for any of the individual heterogeneous VCGs. These results indicate that the common assumption that field strains belonging to mating population A in the same VCG are clones is not always true. These results may also indicate that there are relatively few segregating vic genes in the populations, but with 28 VCGs the number of polymorphic vic loci, assuming two alleles per locus, must be at least five and more likely is higher. These results also suggest that sexual reproduction is relatively common, as there were perhaps only two clones in our population of 70 strains from G. fujikuroi mating population A. The fact that we found strains belonging to the same VCG in different fields also indicates that our decision to treat all of the strains as members of the same population is reasonable. It also indicates that any differentiation within field populations of this fungus must occur either at a much larger scale (e.g., Argentina compared with the United States) or at a much smaller scale (e.g., exhaustive samples of different 0.25-m² areas from the same field) than we examined in this study.

In conclusion, our results indicate that the mating populations of G. fujikuroi present in no-till maize are the same as those found under conventional tillage conditions. Although we recovered relatively few strains of F. subglutinans (G. fujikuroi mating population E) in this study, the scarcity of this species could be attributed to the climate rather than to the tillage practice. We think that the potentials for contamination with fumonisins of grain from fields cultivated under conventional and no-till conditions are similar, as the two G. fujikuroi mating populations most commonly associated with fumonisin production were recovered from the no-till fields.

The results of the N_e and VCG-AFLP studies are somewhat contradictory. The N_e(f) values are the lowest recorded for any set of strains belonging to mating population A, suggesting that sexual reproduction is relatively infrequent. However, the lack of clonality in the multimember VCGs suggests that sexual reproduction is relatively common as clones are rare. The apparent difference between the AFLP and N_e estimates for recombination could have several causes. A trivial explanation is that female fertility assessments in the laboratory are not an accurate reflection of female fertility under field conditions and that more strains are fertile under field conditions than under our laboratory conditions, which resulted in an artificially low N_e. A second possibility is that the population sampled has just recently passed through a sexual reproduction cycle and the numerous genotypes generated as a result of this reproduction event have not had a chance to be lost yet through either selection or drift. A related explanation is that genotypes in this population are relatively stable and abundant and are rarely lost to drift. Under these conditions the AFLPs would be relatively uninformative in terms of providing an estimate of the amount of sexual reproduction. Finally, if migration into the population, for example on the seeds planted each year, is an important source of variation, then the variation could be high even though sexual reproduction is relatively rare. Distinguishing these hypotheses would require sampling over multiple years and monitoring of the populations being introduced into the fields. The lack of clonality in a VCG and the indication that members of the same VCG may be no more closely related than any two strains selected at random from the population suggest that further studies of the relatedness of field strains in the same VCG also are needed.