Genetic Similarity of Flag Shoot and Ascospore Subpopulations of Erysiphe necator in Italy

Abstract

The overwintering mode of the grape powdery mildew fungus, Erysiphe necator (syn. Uncinula necator), as mycelium in dormant buds (resulting in symptoms known as flag shoots) or as ascospores in cleistothecia, affects the temporal dynamics of epidemics early in the growing season. We tested whether distinct genetic groups (I and III) identified previously in E. necator correlate to overwintering modes in two vineyards in Tuscany, Italy, to determine whether diagnostic genetic markers could be used to predict overwintering. Samples from one vineyard were collected from flag shoots; the other vineyard, 60 km away, had no flag shoots, and mildew colonies were assumed to be derived from ascospores. Genetic markers putatively diagnostic for groups I and III showed that both types were common in the flag shoot subpopulation. Both genetic types were found in the ascospore population, although group III was dominant. We did not find strong genetic differentiation between the two subpopulations based on inter-simple sequence repeat markers. Although there was significant (P < 0.001) genetic differentiation between these subpopulations in 1997 and when 1997 and 1998 subpopulations were pooled (θ = 0.214 and 0.150, respectively), no differentiation was evident between vineyards in 1998 (θ = 0.138, P = 0.872). Moreover, we did not observe distinct lineages corresponding to overwintering modes, as observed in previous studies. We could not determine if differentiation resulted from biological differences or restricted gene flow between the two vineyards. Our samples were taken from both subpopulations early in the epidemic, while previous studies confounded overwintering mode and sampling time. These results do not support a strong correlation between overwintering and genetic groups, highlighting the need to base population biology studies on sound biological and epidemiological knowledge.

Studies of the population biology of grape powdery mildew, caused by Erysiphe necator (syn. Uncinula necator), have described two genetically distinct groups, or biotypes, that correspond to the type of primary inoculum or to the time of sampling during an epidemic (7, 9, 15, 20). Group I (sensu Délye) isolates were recovered early in the season from flag shoots, i.e., symptomatic shoots that result from mycelium overwintering in buds that were infected during the previous season (17, 18). Group I, characterized by low genotypic diversity and only one mating type, was interpreted as being clonal (7). Group III isolates were recovered later in the growing season from leaves and/or fruit clusters and were inferred to have originated from ascospores, which overwinter in cleistothecia (16). Group III isolates were more diverse, and both mating types were present (7). The frequencies of the two groups change during powdery mildew epidemics, with group I declining in frequency and group III dominating later in the epidemic (7, 10). The two groups differ at multiple genetic loci and were reported to lack interfertility (7). Délye and colleagues (7, 9) speculated that the two groups were introduced into Europe at different times and have remained isolated ecologically and genetically.

Other studies of the population genetics of E. necator, however, have shown conflicting results to those of Délye and colleagues (7, 9). For instance, Miazzi et al. (15) and Cortesi et al. (4) found that isolates collected from flag shoots in Italy had 1:1 mating-type ratios. Furthermore, a flag shoot subpopulation in one vineyard had high genotypic diversity and a multilocus genetic structure consistent with the hypothesis that at least some sexual reproduction (4) occurs in this population. Thus, the two groups may not correspond strictly to different reproductive and overwintering modes. Although the two groups of E. necator appear to be genetically differentiated, there are reports that these groups are interfertile under laboratory conditions and can produce viable ascospores (5, 6, 15, 20). Mating between groups in the field would eventually homogenize their genetic differences. Additionally, the correlation between genetic groups and overwintering mode is not perfect: a few isolates sampled from flag shoots in two different studies were genetically indistinguishable from isolates in group III, which comprises isolates thought to originate from ascospores (7, 9).

The onset of powdery mildew epidemics is markedly affected by the type of primary inoculum. Infections from ascospores depend strongly on favorable climatic factors, such as rainfall to disperse the inoculum, whereas mycelium overwintering in buds can colonize emerging shoots under less restrictive environmental conditions (13, 18). Because of the epidemiological significance of inoculum type, reliable prediction of this parameter needs to be evaluated before genetic markers are adopted as a management tool. Cortesi et al. (4) previously analyzed the genetic structure of a subpopulation comprising only isolates collected from flag shoots early in the epidemic. In the current study, we extended our sampling to a subpopulation from a different location where no flag shoots were found and the primary inoculum derived from overwintering ascospores. Our overall goal was to test the hypothesis that genetic groups of E. necator correspond to overwintering modes. Specifically, we determined (i) how frequently each genetic group was found in each epidemiologically defined subpopulation and (ii) whether the flag shoot and ascospore subpopulations were genetically differentiated.

MATERIALS AND METHODS

Study sites and sampling.

Two commercial vineyards were sampled for powdery mildew in Tuscany: Santa Cristina and Fornace. A flag shoot subpopulation was sampled in a Santa Cristina vineyard located in Montefiridolfi (Florence province) annually from 1994 to 2001; this sample was previously analyzed genetically and epidemiologically (4-6). In contrast to Santa Cristina, no flag shoots were observed in Fornace (Montalcino, Siena province), a 0.5-ha vineyard surrounded mostly by olive plantations approximately 60 km from Santa Cristina. This vineyard and nearby vineyards were monitored in 1995 and 1996 at 10-day intervals, from just after budbreak until bloom, and no flag shoots were found, although cleistothecia were abundant (2). In this study, we continued to monitor Fornace and nearby vineyards for flag shoots. In 1997 and 1998, we visited Fornace biweekly, beginning just after budbreak in mid-May until bloom in mid-June. Based on past observations in other parts of Tuscany, flag shoots appear for several weeks after budbreak and all are visible by bloom in early/mid June (5). In both vineyards, we estimated the number of cleistothecia on bark and leaves as described previously (2). We also acquired daily temperature and rainfall data from the Agenzia Regionale Sviluppo Innovazione Agricolo-forestale—Firenze from the end of March to the end of June for each site to predict ascospore release and infection events (2).

Isolation of E. necator from flag shoots in Santa Cristina from 1997 to 2001 was described previously (4). In Fornace, we monitored vines for mildew colonies at 2-week intervals early in the epidemic, from the end of April to the end of June. We particularly examined the abaxial surfaces of the leaves because they are more exposed to E. necator ascospore discharge in cordon-trained, spur-pruned vines than are the adaxial surfaces (16). Infected leaves were collected in early June, about 10 days after the appearance of the first mildew colonies. From each leaf with either single or isolated and well-defined colonies, we obtained single-conidial isolates as described previously (4). In total, we obtained 81 flag shoot isolates over 5 years from Santa Cristina (4) and 69 isolates from foliage in Fornace in 2 years. Small samples are a consequence of the logistical constraints of culturing and maintaining mildew isolates on grape seedlings and the difficulties of producing enough fungal tissues for DNA extractions.

Mating type assay.

Mating types of isolates from Fornace were determined by crossing them with tester isolates of both mating types on grape seedlings as described previously (4). Mating type data for Santa Cristina were reported previously (4). Mating type testers were originally collected in Fornace, which is assumed to be an ascospore subpopulation. We used tester isolates M22 and M26 for mating type “+” and M20 and M23 for mating type “-”. All four tester isolates were paired with each field isolate to determine whether cleistothecia were produced. Viability of ascospores in mature cleistothecia was determined by staining with fluorescein diacetate as described previously (3).

Identification of genetic groups I and III.

Conidia and mycelium grown for 10 to 15 days on grape leaves were harvested and DNA extracted as previously described (4). All isolates were genotyped with two molecular markers that had been developed to identify the two genetic groups (10, 14).

For the first marker, groups I and III were distinguished in a nested PCR that used allele-specific primers for a point mutation in the gene for eburicol 14 α-demethylase (CYP51) (10). In the first round of PCR, we combined the primers U14DM and M1I (10). Each 25-μl reaction contained 2.5 μl of 10× reaction buffer (Promega, Milan, Italy), 2.5 mM MgCl₂, 200 μM (each) deoxynucleoside triphosphates (Promega), 0.8 μM concentrations of each primer, 1 μl of 50 μl of fungal extracted DNA (approximately 5 to 10 ng), and 1 U of Taq DNA polymerase (Promega). PCRs were performed with an i-Cycler thermal cycler (Bio-Rad Laboratories, Milan, Italy) using the following conditions: initial denaturation for 1 min at 94°C and 37 cycles of PCR amplification, each consisting of 30 s of denaturation at 94°C, 1 min of annealing at 56°C, and 1 min of extension at 72°C, followed by a final extension for 5 min at 72°C. Subsequently, 1-μl aliquots of the first-round PCR mixture were subjected to a second round of PCR amplification using the allele-specific primers MUT2(I-II) and MUT2(III) (10) in association with U14DM. The reaction mixture was the same as that described above except for the final concentration of each primer, 0.6 μM, and 0.5 U of Taq DNA polymerase. DNA was amplified with the same protocol described above except that the annealing temperatures were 52°C and 53°C for MUT2(I-II) with U14DM and MUT2(III) with U14DM, respectively.

For the second marker, the two genetic groups were distinguished following PCR with a primer pair, UnE and UnF, that amplifies a sequence-characterized amplified region (SCAR) (14). PCRs were carried out in a volume of 25 μl containing 2.5 μl of 10× reaction buffer, 2.5 mM MgCl₂, 75 μM (each) deoxynucleoside triphosphates, 0.5 μM concentrations of each primer, 1 μl of 50 μl of fungal extracted DNA (approximately 5 to 10 ng), and 1 U Taq DNA polymerase. Amplifications were conducted with 35 cycles for 30 s at 94°C, 1 min at 58°C, and 1 min at 72°C, with an initial denaturation for 4 min at 95°C and a final extension of 7 min at 72°C (14).

PCR products obtained from CYP51 or from the SCAR marker were analyzed by electrophoresis in 1% and 2% agarose gels stained with ethidium bromide in Tris-borate-EDTA buffer, respectively. DNA amplifications were performed twice from a sample of independent DNA preparations to ensure the repeatability of the markers used, and DNA preparations were made at different times to ensure independence. DNA from two isolates from France, BR 1 for group I and BR 10 for group III, was used as positive controls in each PCR amplification.

Population structure analyses.

Each isolate was genotyped by PCR amplification of inter-simple sequence repeat (ISSR) markers (19) and the intron-splice junction marker with primer R1 (11); for convenience, we collectively refer to these as ISSR markers. E. necator is haploid, and each PCR product is assumed to represent a single locus. These markers are known to be reproducible, and the Santa Cristina isolates were genotyped previously (4).

Analyses of population structure were made with MultiLocus 1.2 (1). To test for genetic differentiation between vineyards, we estimated Weir's θ (21), the genetic diversity attributable to differentiation among subpopulations. To compare our results with previous studies, we used ISSR haplotypes to construct unrooted phenograms using the neighbor-joining method, with and without 1,000 bootstrap samples of the original data set. A consensus tree was obtained from the 1,000 bootstrap trees using a consensus analysis, with the majority rule method. Neighbor-joining analyses were carried out in PHYLIP (12).

RESULTS

Ascospore subpopulation in Fornace.

No flag shoots were found in Fornace or in neighboring vineyards during the course of this study. Cleistothecia were abundant on both bark and dry leaves in 1997 and 1998 at levels similar to those previously reported (2, 5). There were six ascospore infection events prior to sampling in 1997 and 11 prior to sampling in 1998. Based on these results, we concluded that isolates obtained from foliage in Fornace were derived almost exclusively from ascospores. We refer to this sample as an ascospore subpopulation.

Mating type polymorphism.

More than 90% of the Fornace isolates produced cleistothecia when crossed with one or the other of the two mating-type tester isolates. Mating type ratios (MAT+:MAT−:sterile) were 21:20:0 in 1997 and 9:16:2 in 1998 and were not significantly different from 1:1 (χ² = 0.01, P = 0.912 and 0.51, P = 0.475, respectively).

Identification of genetic groups I and III.

We could not identify the genotype of 18 of 138 (13%) isolates with the allele-specific primers MUT2(I-II) and MUT2(III) of CYP51 because we failed to amplify DNA in primary PCR. However, all isolates could be genotyped with the UnE-UnF SCAR primers (Table 1). Genetic groups determined by the two markers were concordant with the expected results for the tester isolates BR 1 and BR 10 but discordant for 24 of 120 (20%) field isolates for which we obtained data for both markers (Table 2). All mating-type testers were in group III by both sets of markers. Based on the concordant data, both genetic groups occurred in both subpopulations. Almost all isolates from Fornace (51 of 52, for which both markers were concordant) were in group III, which is consistent with epidemiological evidence for the absence of flag shoots and the presence of cleistothecia and viable ascospores. In Santa Cristina, 14 isolates were genotyped as group III by both markers even though this sample comprised individuals obtained exclusively from flag shoots (4) (Table 1). In addition, both mating types were found among isolates of both genetic groups. Among the 31 group I concordant isolates, the mating-type ratio was 16:11:4 (MAT+:MAT−:sterile), and among the 70 group III isolates, this ratio was 34:32:4.

TABLE 1.

Distribution of genetic groups I and III (sensu Délye) of E. necator based on CYP51 allele-specific (10) and SCAR (14) markers

Subpopulation and yr	n	No. of isolates of indicated genetic group (sensu Délye)
		CYP51 allele specific			SCAR marker
		Group I	Group III	NDa	Group I	Group III
Santa Cristina (flag shoots)b
1997	26	10	7	9	23	3
1998	34	11	22	1	24	10
1999	16	5	7	4	15	1
2000	3	2	0	1	3	0
2001	2	2	0	0	2	0
Total	81	30	36	15	67	14
Fornace (ascospores)b
1997	29	0	27	2	0	29
1998	28	2	25	1	2	26
Total	57	2	52	3	2	55

^aND, not determined. We failed to amplify DNA in primary PCRs in repeated attempts.

^bSource of overwintering inoculum (see the text).

TABLE 2.

Concordance of genetic groups I and III (sensu Délye) of E. necator in two subpopulations^b

Subpopulation	CYP51 marker	No. of isolates with SCAR marker:
		Group I	Group III
Santa Cristina	Group I	30a	0
Fornace		1a	1
Santa Cristina	Group III	22	14a
Fornace		1	51a
Santa Cristina	Not determined	15	0
Fornace		0	3

^aIsolates with concordant data for both markers.

^bConcordance was determined by two different genetic markers, CYP51 (10) and a SCAR marker (14). Unmarked cells indicate either discordancy or that no data were obtained for CYP51.

Analysis of population structure.

ISSR markers from amplifications of DNA extracted from 53 isolates from Fornace were generally similar to those for isolates from Santa Cristina with respect the number and size of amplicons (4). We scored 72 fragment sizes, each considered a separate presence/absence locus (see Table S1 in the supplemental material). The same multilocus haplotype was found in Fornace in only three isolates (F98-46, F98-55, and F98-66); however, one of them differed in mating type. In Santa Cristina (4), the same multilocus haplotype was found in two isolates in 1997 (SC97-26 and SC97-30); both isolates had the same mating type. Two haplotypes were recovered in different years (SC98-14 and SC99-14, SC97-40 and SC98-3), and each isolate of the pair had the same mating type. All other isolates had unique haplotypes (4).

Based on ISSR allele frequencies, Santa Cristina and Fornace were significantly differentiated in 1997 (θ = 0.214, P = 0.001) and when 1997 and 1998 data were pooled (θ = 0.150, P < 0.001), indicating some restriction in gene flow. However, the two vineyards were not significantly differentiated in 1998 (θ = 0.138, P = 0.872) and neither were the 1997 and 1998 samples from Fornace (θ = 0.029, P = 0.999). Allele frequencies were significantly different in both years for CYP51 (1997, χ² = 17.34, P < 0.001; 1998, χ² = 5.88, P = 0.015) and the SCAR marker (1997, χ² = 40.53, P < 0.001; 1998, χ² = 22.84, P < 0.001). When haplotype data were used to construct a neighbor-joining phenogram, no pattern of separation was evident between the subpopulations (data not shown). Furthermore, there was little support for any internal branches, indicating essentially no lineage structuring to this population.

DISCUSSION

The primary goal of this study was to determine if epidemiologically defined subpopulations of E. necator also were genetically distinct subpopulations. To address this question, we sampled from subpopulations where we knew the sources of primary inoculum (4) and we sampled these subpopulations at the same time, early in the epidemic. In previous studies (7, 9, 15, 20), isolates recovered from flag shoots (and some leaves [see Miazzi et al. {15}]) early in the epidemic were genetically distinct from isolates collected later from foliage and fruit clusters. We did not find that the two subpopulations, which represent distinct overwintering modes and sources of primary inoculum, were genetically distinct.

A major difference between our study and previous studies of E. necator is that we sampled flag shoot and ascospore subpopulations at the same time, early in the epidemic. Previous studies (7, 9, 10, 15) sampled flag shoots early in the epidemic and then sampled leaves and fruit later in the season to obtain isolates putatively derived from ascospores. The sequential sampling strategy confounds time during the epidemic with possible overwintering modes and sources of primary inoculum. Miazzi et al. (15) sampled from leaves and clusters early in the epidemic and found only group I isolates, although they did not attribute these isolates to either ascospore infections or secondary colonies derived from flag shoots. By sampling only at the beginning of the epidemic and by controlling for inoculum source by sampling flag shoots in one subpopulation and confirming their absence in the other, we eliminated the time factor. The correlation of distinct genetic groups to overwintering claimed previously (7, 9) could be the result of limited sampling. Even with small sample sizes, however, these earlier studies identified exceptions to the strict correlation between genetic types and sampling time. Determining the temporal dynamics in genetic composition of E. necator populations will require more systemic sampling early and late in epidemics in the same vineyards.

We used two distinct genetic markers to identify genetic groups in E. necator. These two markers were perfectly correlated to genetic groups in previous studies (10, 14), which is why they were developed as diagnostic markers. However, these same markers gave conflicting results for 20% of the isolates in this study (Table 2). This discrepancy could result if recombination occurs between the diagnostic loci. Recombination is consistent with previous observations that the two genetic groups are interfertile in the laboratory (4-6, 15, 20). In our hands, moreover, we failed to obtain primary PCR products of CYP51 for 13% of our isolates, despite repeated attempts (Table 1). We do not know if the lack of amplification represents a genetic difference, e.g., a null allele. Regardless, the value of these markers is questionable for populations in Tuscany.

The flag shoot and ascospore subpopulations do not form distinct haplotype lineages, but there is evidence for genetic differentiation between Santa Cristina and Fornace based on ISSR allele frequencies (θ = 0.214 and 0.150 in 1997 and both years pooled, respectively). However, no differentiation was evident between vineyards in 1998 (θ = 0.138, P = 0.872). The markers diagnostic for previously defined genetic groups had different allele frequencies in the two vineyards in both years.

In Fornace, the alleles associated with genetic group III (sensu Délye) (Table 1) dominated, which is consistent with predictions that group III is more representative of ascospore subpopulations. Differentiation of genetic markers between subpopulations might be expected to be weak because Santa Cristina and Fornace are separated by only 60 km. Previous studies of E. necator in Europe found that rapid amplification of DNA ends haplotypes were more similar within than between vineyards (7-9, 15), although no such correlations were found in Australia with restriction fragment length polymorphism and ISSR markers (20). In the present study, the overwintering mode is confounded with location, which makes it impossible to determine if genetic differentiation is associated with biological differences in overwintering or with geographic location. Additional subpopulations of both overwintering types are needed to address this question, although it may be difficult to find replicated pairs of vineyards, in close proximity, with and without flag shoots.

Our results show that the genetic types defined by Délye et al. are not useful for predicting overwintering mode in populations of E. necator in Tuscany. This study clearly leaves some important questions unanswered. We know that group III isolates can form flag shoots in the field, but we do not have direct evidence that group I isolates can form ascospores and survive in cleistothecia in the field. One way to answer this question is to sample ascospores directly from cleistothecia at the beginning of epidemics and genotype the resulting colonies. We also do not know what factors affect the formation of flag shoots. For example, the lack of flag shoots in Fornace may be due to environmental factors or disease management practices rather than to genetic characteristics of the pathogen population. It also is not clear that the relative frequencies of genetic groups are constant throughout an epidemic. Evaluation of the temporal stability of a population requires that isolates be collected both early and late in the season and from flag shoots and foliage. Thus, considerable work remains to be done to understand the population biology of this pathogen.