A single nucleotide polymorphism in the β-tubulin gene distinguishing two genotypes of Erysiphe necator expressing different symptoms on grapevine



The biotrophic fungus Erysiphe necator (formerly Uncinula necator), for which two genetic groups have been described in European vineyards, is the causal agent of grapevine powdery mildew. By analysing the pathogen population with respect to polymorphism in the sequence of the β-tubulin gene, which distinguishes two groups of isolates, a new tool was developed for epidemiological and population studies and tested in the vineyard. As in many ascomycetes, the β-tubulin gene of E. necator (Entub) includes six introns and seven exons and encodes a 447-amino-acid protein. A single nucleotide polymorphism (SNP) in the intron-3 region of the Entub gene distinguished two genetic groups (A and B). This method was used to examine differences in the ratio of the two groups from a total of 289 grape powdery mildew samples collected at the beginning of the growing season from either flag shoots or leaves with sparse-spot symptoms in four different vineyards. The SNP in the intron-3 region of the β-tubulin gene, similar to SNPs in the CYP51 gene, was associated with genotypes A and B of E. necator and confirmed the existence of two sympatric populations of the pathogen in the French vineyards. Differences in the relative proportions of each group varied with the presence or absence of flag-shoot symptoms and with the region in which isolates had been collected.


Powdery mildew, caused by the biotrophic fungus Erysiphe necator (formerly Uncinula necator) (Braun et al., 2002), is a major disease of grapevine (Vitis vinifera). The pathogen, a haploid heterothallic ascomycete, overwinters either as resting mycelium within dormant buds that can reinitiate growth after budbreak and colonize young flag shoots (without sexual reproduction) (Sall & Wrysinski, 1982; Rumbolz & Gubler, 2005), or on the bark of grapevines as cleistothecia from which primary infections are initiated on susceptible tissues by released ascospores during the spring (Gadoury & Pearson, 1988). Although both flag shoots and cleistothecia are observed in Europe, cleistothecia are commonly considered to be the main source of inoculum for primary infection. However, depending on the cultivar and/or climatic conditions, infections originating from overwintering mycelium in dormant buds may generally be considered the main source of primary infections in French Mediterranean vineyards.

Studies in France and Australia based on RAPD, PCR-RFLP or specific PCR approaches have distinguished two genetic groups or biotypes within E. necator, and a third in India (Délye et al., 1997a, 1999; Evans et al., 1997). In this study, groups A and B are considered, corresponding to groups I and III, respectively, as defined by Délye et al. (1997a). In European vineyards, the two groups correspond to early symptoms on flag shoots (group A) or to foliar symptoms collected later on in the growing season (group B) (Délye & Corio-Costet, 1998; Miazzi et al., 2003). A comparison of the sequence of the gene encoding for eburicol-C14-demethylase in sterol biosynthesis (CYP51) between the two groups of E. necator isolates led to the identification of three point mutations or SNPs at codon 37 (Gly-Val substitution), codon 156 (Ile-Thr substitution) and codon 493 (silent mutation) that distinguish the two groups (Délye et al., 1997b,c). The occurrence of three SNPs in CYP51 was surprising, since this gene is highly conserved and is essential for survival of the fungus (Aoyama et al., 1996). Allele-specific PCR tools have previously been used in vineyards where samples that were identified as group A were associated with flag-shoot symptoms (Délye et al., 1999).

Combining specific markers of functional genes which could be under selection pressure with near-neutral markers such as microsatellites is important in the study of the evolution of pathogen populations (Brumfield et al., 2003). In the case of E. necator the sterol C-14-demethylase gene (CYP51), which is involved in fungicide resistance, provides such a functional gene under selection pressure (Délye et al., 1999). Hence, to increase the number of specific PCR markers for the two genetic groups A and B, the linkage between groups was examined using the level of polymorphism of the β-tubulin gene in a way similar to that previously used for the CYP51 markers. SNPs represent the most widespread type of sequence variation in genomes and are excellent tools for studying speciation and demographic history (Brumfield et al., 2003).

The β-tubulin gene is generally expressed at high levels during cell growth and its nucleotide sequence has frequently been used in phylogenetic analyses to distinguish between different species and/or different subgroups within a species (Edelman & Staben, 1994; Keeling et al., 2000; O'Donnell et al., 2000; Banke et al., 2004). Polymorphism in intron 3 has often been exploited to assess similarity and relatedness even among closely linked sets of isolates (O'Donnell et al., 1998; Widmer et al., 1998; Tooley et al., 2001).

The aims of this study were as follows: (i) to obtain β-tubulin gene sequences of E. necator isolates belonging to groups A and B for subsequent studies on gene expression; (ii) to compare sequences between the two genotypes; (iii) to produce a new marker for the analysis of population structure dynamics using samples from vineyards, with a sensitivity similar to that of a PCR assay, reproducible, and complementing studies based on the CYP51 gene; (iv) to assess the presence of genotypes A and B on different leaf symptoms (flag shoots or lesions on leaves) at the beginning of a grapevine powdery mildew epidemic; and (v) to confirm the sympatric nature of the two populations in the same vineyard.

In order to meet these objectives, two pairs of primers were developed that amplify part of the β-tubulin gene of E. necator. These were then used to analyse a range of samples taken from different epidemics in different vineyards. This procedure aimed to provide rapid identification and distinction between the two groups (A or B) in field samples without the lengthy process of isolating and culturing the fungus on detached grapevine leaves.

Materials and methods

Erysiphe necator isolates and samples

Fourteen E. necator isolates (Table 1) from mildewed grape leaves from French vineyards were used to assess the variation in the β-tubulin gene sequence. An additional 48 isolates obtained from the authors’ own collection and identified previously as belonging to either genetic group A (25) or B (23), were used to assess the performance of the PCR tools (Table 1). A collection of 289 samples was taken from several plots in two vineyards, either with a Mediterranean climate (Carcassonne, Aude and Nîmes, Languedoc-Roussillon), or with an oceanic climatic influence (Begadan and St-Ciers, Gironde) (Table 2). The samples were used to test for the presence of the two genetic groups and to examine their ratios. All samples were collected during the first few days of the powdery mildew epidemics during 1999 or 2000, from flag shoots or leaves with sparse spot symptoms. For each date of sampling, all visible symptoms on leaves were harvested.

Table 1. Erysiphe necator isolates belonging to genetic groups A and B, used for cloning and sequencing the β-tubulin gene (in bold) and isolates used to validate the PCR assays. Isolates of group A (25) all originate from flag-shoot symptoms and isolates of group B (23) were from sparse spots on leaves (23)
IsolatesGeographic originDate of isolation (month/year)Genetic group
2B3, 2B17FranceAude05/2000A
B101, 171,FranceAude05/2001A
201, 202, 210,FranceAude05/2001A
211, 302, 702,FranceAude05/2001A
704, 1208FranceAude05/2001A
BR30, 33FranceAude05/2002A
CC50, CC4,FranceGironde07/2000A
C31, CC25FranceGironde07/2000A
CPH3, CPH8FranceCorbière07/2000A
SC15, SC2FranceSaint Ciers07/1999B
2G12, 2G13FranceAude07/2001B
B113, 114,FranceAude08/2001B
117, 119, 146,FranceAude08/2001B
212, 301, 508,FranceAude08/2001B
808, 901, 1110FranceAude08/2001B
CC43, CC34FranceGironde07/2000B
Table 2.  Origins of the 289 samples of Erysiphe necator collected on leaves at the beginning of epidemics from flag-shoot symptoms or sparse spots on leaves of different cultivars in vineyards with a Mediterranean climate (Carcassonne and Nîmes) or under oceanic climatic influence (Bégadan and St-Ciers)
Flag shootSparse leaf spot
Carcassonne, Aude199925 MayCarignan1921
Carcassonne, Aude200028 JuneChardonnay 041
Nîmes, Languedoc-Roussillon200025 MayCarignan44 8
Nîmes, Languedoc-Roussillon2000 7 JuneGrenache 124
Begadan, Gironde199910 JuneCabernet 073
St-Ciers, Gironde1999 7 JuneMerlot 058

Inoculation procedure for β-tubulin gene cloning

Powdery mildew isolates were inoculated under sterile conditions on decontaminated grape leaves (cv. Cabernet Sauvignon) in Petri dishes as previously described (Délye & Corio-Costet, 1998). Leaves were placed at the bottom of a Plexiglas settling tower and conidia were blown in at the top from sporulating leaves (600–800 conidia per cm2 of leaf). Inoculated leaves were incubated for 14 days at 22°C under 16 h illumination (25 µE m2 s−1) per day.

DNA and RNA extractions

Total genomic DNA extraction was performed as previously described (Délye et al., 1995). Mycelium frozen in liquid nitrogen was ground into a powder and RNA was extracted with a commercial kit (RNAble; Eurobio).

Cloning and sequencing of β-tubulin gene

From consensus sequences, degenerated primers Tub3 (5′-GGCXAARGGXCAYTAYACXGA-3′, 512-fold degenerate) and Rtub4 (5′-TGYTGXGTXARYTCXGGXAC-3′, 2048-fold degenerate) were designed corresponding to the highly conserved amino acid sequences 101-WAKGHYT-107 and 287-PELTQQ-292 of fungal β-tubulin. For PCR reactions, 20 µL of reaction mixture contained 130 mm Tris-HCl, 32 mm (NH4)2SO4, 4 mm MgCl2, 200 µm each dNTP, approximately 10 ng template DNA, 0·5 U Silverstar DNA polymerase (Eurogentec S.A.) and 0·25 µm of each of the primers Tub3 and Rtub4. Amplification was performed in 37 cycles consisting of 30 s of denaturation at 95°C, 1 min of annealing at 58°C and 1·5 min of extension at 72°C on a DNA thermal cycler (Crocodile III; Appligene Oncor). This PCR resulted in a single major product of 574 bp, which was cloned in Escherichia coli (XL1 blue) using the pGEM-T vector system (Promega). For each E. necator isolate used in cloning and sequencing experiments, 10 DNA inserts were sequenced on both strands by the MWG Biotech sequencing service (http://www.mwg-biotech.com/).

The complete cDNA sequence was generated after RNA extraction. About 1 µg of total RNA was obtained for each mg of mycelium and reverse transcription was performed using a commercial kit (5′–3′ Race Kit; Roche). Primers Btun1 and Rbtun2 (Table 3) were used for reverse transcription reactions to amplify the β-tubulin gene. The products were cloned and sequenced as described above. The genomic DNA sequence was obtained by sequencing the Btun10 and Rbtun15 PCR products (Table 3).

Table 3.  Sequence of the primers used for β-tubulin gene cloning, allele-specific PCR and NAS-PCR
PrimersPrimer sequenceCoordinatesa (bp)Annealing temp. (˚C)
  • a

    Relative to translation start site (ATG).

Rbtun25′-CGGCGGACAACATCTAAGAC-3′ 662–64359
TbioA5′-TTTCTGTACGTAGAT-3′ 195–20937·6
TbioB5′-TTTCTGTACGTAGAC-3′ 195–20937·9

Nested allele-specific PCR (NAS-PCR)

Allele-specific PCR of the β-tubulin gene

Mycelium and conidia from 48 isolates of E. necator from the reference collection (Table 1) and 289 samples from vineyards were used. Primer Btun10 was designed from a noncoding sequence including a part of the ATG codon of Entub, while Rltub was designed from a coding sequence including part of intron 5. These primers amplify a 457 bp DNA fragment of the β-tubulin gene in first-round PCR (Table 3). PCR mixtures as described above were submitted, on a DNA thermal cycler (Appligene), to one cycle of 3·5 min of denaturation at 95°C, followed by 40 cycles of amplification, each consisting of 30 s of denaturation at 95°C, 1 min of annealing at 60°C and 1 min of extension at 72°C. Finally, one cycle of 8·5 min of extension at 72°C was performed, then 1 µL of PCR product was used in the second PCR run. Allele-specific PCR primers (TbioA and TbioB) were designed on the assumption that a 3′ mismatch does not prime in PCR at a specific annealing temperature (Zirnstein et al., 1999). Primers TbioA and TbioB were designed specifically to prime Entub sequences containing a T or a C at nucleotide 209 (Table 3). Primer Rltub was based on an Entub sequence that was identical in all Entub sequences studied. In the second PCR round, samples contained 0·25 U of Silverstar DNA polymerase (Eurogentec S.A.) and the allele-specific PCR primers at a final concentration of 0·25 µm. Amplification for NAS-PCR was performed in a reaction mixture containing 130 mm Tris-HCl, 32 mm (NH4)2SO4, 10 mm MgCl2 and 200 µm of each dNTP. Approximately 10 ng of template DNA, plus primers Rltub, TbioA and TbioB at 0·1 µm each, were submitted to 30 cycles of PCR amplification, each consisting of 30 s of denaturation at 95°C, 1 min of annealing at 52°C or 58°C and 1 min of extension at 72°C.

Allele-specific PCR of CYP51

New primers 14 DM (5′-ATGTACATTGCTGACATTTTGTCGG-3′) and M1I (5′-CGCTATCTCTCGATCAGG-3′) amplified a 1038-bp DNA fragment of a CYP51 gene in a first round. PCR mixtures of 20 µL containing 70 mm Tris-HCl, 17 mm (NH4)2SO4, 2 mm MgCl2, 0·2 mg mL−1 bovine serum albumin, 0·05% w/v polyethylene-ether W1, 200 µm of each dNTP, 0·4 U Silverstar polymerase, 0·2 µm of each primer and 10 ng template DNA were submitted to 37 cycles of PCR amplification, each consisting of 30 s of denaturation at 95°C, 30 s of annealing at 54°C and 1·5 min of extension at 72°C. Subsequently, 1 µL of PCR product was used in a second PCR run. Allele-specific PCR primers U14DM, MUT2 (group A) and MUT2 (group B) were used as described by Délye et al. (1999). In the second PCR round, samples contained 0·05 U of Silverstar DNA polymerase and the allele-specific PCR primers at a final concentration of 0·1 µm. Amplification for NAS-PCR was carried out in a reaction mixture as described above. Approximately 10 ng of template DNA, plus primers U14DM, MUT2 (A) and MUT2 (B) at 0·1 µm each were submitted to 30 cycles of PCR amplification, each consisting of 30 s of denaturation at 95°C, 1 min of annealing at 52 or 56°C and 1 min of extension at 72°C.

All amplified DNA fragments were visualized under UV light after electrophoresis at 100 V on ethidium bromide-stained (0·4 µg mL−1) 1% agarose gels in 0·5X Tris-borate-EDTA buffer. In all experiments appropriate negative controls containing no template DNA were subjected to the same procedure to exclude or detect any possible contamination or carryover. DNA extracted from six fungi [Botryotinia fuckeliana (anamorph Botrytis cinerea), Alternaria alternata, Fusarium sp., Geotrichum sp., Phaeomoniella chlamydospora and Eutypa lata] and the oomycete Plasmopara viticola that may also be found on grape leaves and berries were tested for amplification using NAS-PCR. No PCR products were detected for any of these organisms (results not shown).

Sequence analysis

For sequence alignment, β-tubulin DNA sequences were translated into amino acids and aligned using ClustalW (Thompson et al., 1994). Based on this amino acid alignment, DNA sequences were aligned with protal2dna (http://bioweb.pasteur.fr/seqanal/interfaces/protal2dna.html).

Nucleotide sequence accession numbers

All sequences of variant β-tubulin genes of E. necator (Entub) and cDNA reported in this paper were deposited under a single accession number (AY074934) in the GenBank nucleotide sequence database. The EMBL accession numbers of cDNA tubulin nucleotide sequences used in this study are: Blumeria graminis (formerly Erysiphe graminis) X51326, B. fuckeliana Z69263, Erysiphe pisi X81961, and Rhynchosporium secalis X81046. The GenBank accession number of CYP51 of E. necator is AF042067.


Nucleotide sequence and organization of the Entub gene

Amplified PCR products of DNAs extracted from 14 isolates of E. necator appeared as a single band of the expected size (1945 bp, data not shown). After the cloning of PCR fragments, the β-tubulin gene was sequenced. Sequences of the gene from group A (2B3, 2B7, FM023, AOG11, CC12, CC50, CPH3) and group B isolates (FMN22, 2B5, CC20, SC15, FNB12, FNI11, CC43) were identical, except at one nucleotide position in the intron-3-region sequence. Reverse transcription PCR (RT-PCR) was used to confirm that the putative introns were excised from transcripts. A fragment of the expected size (approximately 1·7 kb) was obtained and cloned. Sequence analysis showed that six putative introns were excised. The inferred 447-amino-acid protein encoded by the 1344-bp coding sequence of the 2000-bp sequence has a predicted molecular mass of 49·87 kDa.

Analysis of deduced amino acid sequences

The β-tubulin gene of E. necator encoded a polypeptide having extensive homology with that of other filamentous fungi, including E. pisi (98·9% amino acid identity), B. fuckeliana (98·4%) and B. graminis (97%). Comparison of amino acid sequences between E. pisi, B. graminis and E. necator showed that only five amino acids differed at positions within the protein, e.g. M-47 is methionine at position 47 (M-47, A-126, E-436, Q-441 and L-443) for E. pisi and 14 for B. graminis (M-47, D-209, F-294, Q-335, V-338, S-339, S-420, E-427, E-436, P-438, V-441, S-442, N-443 and E-445). Most of the nonconserved amino acids were clustered at the C-terminal end of the protein. The six introns in the Entub gene interrupted nucleotide sequences encoding highly conserved domains of the β-tubulin gene. Five introns were located in a region between the fourth and the 53rd amino acids of the β-tubulin gene, and the sixth was at amino acid 317.

Codon usage

Four codons were lacking in Entub codon usage (TCG, AGC, CTG and CGG). Three of these had G in the third position. Two of them (TCG and AGC) coded for serine, whereas CTG and CGG coded for leucine and arginine, respectively. The pattern of codon usage of this gene indicates a strong bias. The codons for which T was in third position were used three times more often than codons with C in the third position (i.e. cysteine, phenylalanine, histidine, asparagine, aspartic acid). This was also the case for CYP51, the only other gene so far cloned in E. necator (Délye et al., 1997c).

Intron polymorphism

Although β-tubulin is a highly conserved gene, a distinct genetic polymorphism was observed within the intron region, where nucleotide sequences were different in the two groups of E. necator. All isolates from group A exhibited a single T-to-C change at nucleotide 209 of the intron-3 region. Sequence data from the β-tubulin intron-3 region made it possible to design PCR primers to distinguish between the two genotypes, A and B, of E. necator.

Nested PCR

An allele-specific PCR test based on a single nucleotide polymorphism (SNP) in the intron-3 region was performed on E. necator reference and field isolates. Primers Btun10 and Rbtun2 (Table 3) were used for the first PCR amplification of part of the tubulin gene from DNAs extracted from 48 isolates of E. necator and from various fungi to assess the specificity of amplification. Figure 1 shows the second PCR amplification and the responses of 10 isolates of group A (lanes 1–10), 10 isolates of group B (lanes 11–20) and seven samples containing a mixture of isolates belonging to groups A and B (lanes 21–27). The largest monomorphic PCR fragment was always a 457-bp fragment amplified with primers Btun10 and Rltub. Allele-specific amplifications with primers TbioA and Rltub (Fig. 1, top) yielded an amplified DNA fragment of the expected size (275 bp) for group A isolates (lanes 1–10), but not group B isolates (lanes 11–20). Amplification of group B DNA was obtained when primers TbioB and Rltub were used (Fig. 1, bottom, lanes 11–27). Forty-eight isolates from the laboratory collection were tested using this method and gave the same results as the CYP51 PCR assay detection of groups A or B (Table 4).

Figure 1.

Nested allele-specific PCR products obtained from 27 DNA samples by using primers R1tub and TbioA (top) and primers R1tub and TbioB (bottom) in second-round PCR amplification. DNA was extracted from sample fields. Arrows indicate the position of the 275-bp DNA fragment amplified from isolates belonging to genetic group A (top) (lanes 1–10 and lanes 21–27) or group B (bottom) (lanes 11–27). Lane T contained a water control (no DNA). Lane M contained a molecular weight marker (100-bp ladder).

Table 4.  Comparison of NAS-PCR tests on Erysiphe necator isolates with CYP51 primers and TUB2 primers
Allele-specific group AAllele-specific group BAllele-specific group AAllele-specific group B
AO1, AOG11, 2B3, 2B17, B101, 171, 201, 202, 210, 211, 302, 702, 704, 1208, BR30, BR33, CC12, CC50, CC4, CC31, CC25, CPH3, CPH8, FMO23, 2G4++
B113, B114, B117, B119, B146, B212, B301, B508, B808, B901, B1110, 2B5, CC43, CC34, D20, FMN22, FNB12, FNI11, FNI22, 2G12, 2G13, SC2, SC15++

An intron-3 allele-specific PCR assay was used to analyse 289 mildewed samples collected at the beginning of epidemics from flag shoots or leaves with sparse spots. All mildewed tissues were collected at the same stage of epidemics (Table 5). In both the Carcassonne and Nîmes vineyards, numerous samples from flag shoots of cv. Carignan were all identified as belonging to genotype A, except one sample from Carcassonne. On cv. Grenache grown at Nîmes, only one sample from a flag shoot was of genotype A. In the St-Ciers and Begadan vineyards, no flag-shoot symptoms were present and identification of the genotype showed that all samples belonged to genotype B. Similarly, for samples collected from early symptoms in Carcassonne (cv. Chardonnay), no mildewed flag shoots were present and all samples were identified as genotype B.

Table 5.  Nested PCR detection of group A and group B of Erysiphe necator from samples collected on flag-shoot or sparse-spot symptoms on leaves of different cultivars in vineyards with Mediterranean climate (Carcassonne and Nîmes) or under oceanic climatic influence (Bégadan and St-Ciers)
NaSymptombLocalityYearCultivarGenetic groupcTbioAd (number of samples)Tbio B (number of samples)
  • a

    Number of samples.

  • b

    F, flag-shoot symptom; SP, sparse-spot symptom.

  • c

    Identification of genotype with CYP51 nested PCR.

  • d

    Detection with TbioA and TbioB nested PCR.

  • e

    18 samples identified as group A and one as group B.

  • f

    Different vineyards.

19FCarcassonne, Aude (1)f1999CarignanA/Be18 1
21SP   B 021
41SPCarcassonne, Aude (2)2000ChardonnayB 041
44FNîmes, Languedoc-Roussillon (1)2000CarignanA44 0
8SP   B 0 8
1FNîmes, Languedoc-Roussillon (2)2000GrenacheA 1 0
24SP   B 024
73SPBégadan, Gironde1999CabernetB 073
58SPSt-Ciers, Gironde1999MerlotB 058


This study indicates the potential of the β-tubulin gene of E. necator as a useful DNA marker for studying the polymorphism of powdery mildew in local populations or for phylogeographic analysis. In powdery mildews, which include E. necator, E. pisi and B. graminis, the number of introns is similar at identical positions (Ayliffe et al., 2001). The intron regions are mainly located at the beginning of the gene corresponding to the N-terminal domain of the protein, where the amino acid residues are involved in nucleotide binding and lateral contact domains (Nogales et al., 1998; Keskin et al., 2002). Erysiphe necator used 56 of the possible 61 codons and those ending in T were more frequent than those ending in C, as is the case in Neurospora crassa (Edelman & Staben, 1994). The polymorphism of the β-tubulin gene was used to help explain the relationships among different Claviceps spp. or Cryptosporidium (Sulaiman et al., 1999; Tooley et al., 2001), as well as supplying variation for phylogenetic analysis of populations of organisms such as Mycosphaerella graminicola and Fusarium graminearum (Chen & McDonald, 1996; O'Donnell et al., 2000; Banke et al., 2004).

As with the comparison of the DNA sequence of the CYP51 gene of E. necator, an SNP was found in the intron-3 region at nucleotide (nt) 209 which distinguished the two groups A and B. Allele-specific PCR amplification of the β-tubulin sequence of DNA obtained from 48 isolates of E. necator and 289 field samples demonstrated that the polymorphism between the two groups (A and B) of E. necator corresponded to CYP51 gene polymorphisms (Délye et al., 1997c, 1999). Data obtained at the beginning of the grapevine powdery mildew epidemic from field samples collected from different symptoms (flag shoots or leaf colonies) in different locations and years revealed that the majority of the isolates from flag shoots belonged to group A (64), with one exception. This exception may have been the result of either ascospore release synchronized with the development of the shoot at the beginning of the growing season, or of genotype B isolates overwintering in dormant buds.

The frequencies of the two groups in the vineyards were dependent on host cultivar and field location. In the Bordeaux vineyards, observation of mildewed flag shoots and detection of group A were rare. Conversely, in the Mediterranean vineyards, the two groups were present in different ratios depending on cultivar and/or geographic location. The results indicate the presence of sympatric populations of E. necator within the same vineyard, as well as within the same plant. Similar results were found in other plant pathogens, e.g. Phialophora gregata (Meng et al., 2005). The comparison between the two regions is interesting, because, as with P. gregata, the two groups A and B of grapevine powdery mildew are associated with different symptoms and show cultivar preference. However, artificial infections suggest that isolates from group A are capable of infecting various grapevine cultivars, such as Chardonnay, Merlot or Cabernet Sauvignon (unpublished data). Some studies suggest that, as powdery mildew epidemics progress in a season, flag shoots decline in frequency, which is consistent with putative ecological differences between the two groups, which have distinct niches (Délye et al., 1999; Miazzi et al., 2003). Recent work on spatial analysis of flag-shoot populations has shown that vines with flag shoots are aggregated within and between years, suggesting dispersion over relatively short distances (Cortesi et al., 2004). This suggests that genotypes A and B of E. necator exploit different ecological niches in a way similar to that described for Botrytis cinerea, for which two sympatric species were identified (Giraud et al., 1997; Martinez et al., 2005). Indeed, changes in the relative frequencies of the subpopulations of B. cinerea (vacuma and transposa) in vineyards were shown to be correlated with differences in aggressiveness on grape berries, and corresponded to the presence (transposa) or absence (vacuma) of a specific transposon. Change in subpopulations dynamics was most likely caused by differences in phytopathogenic fitness. The results of the present study suggest similar behaviour for the two powdery mildew genetic groups, and differences in aggressiveness could explain why genotype A was not present everywhere, depending on cultivar and environmental conditions. Flag-shoot symptoms were prevalent at the beginning of epidemics and group A of E. necator was found on cvs Cinsaut, Carignan and Merlot in French Mediterranean vineyards. In contrast, flag shoots and group A isolates were rarely present in northern European vineyards and in the vineyards of Bordeaux. This was corroborated by the observation that, under controlled subculturing of isolates on Cabernet Sauvignon leaves, the isolates belonging to genotype A quickly lost their capacity to sporulate. One hypothesis is that the two groups have different dispersal and overwintering strategies. Genotype A, isolated from flag shoots, might overwinter asexually in future dormant buds as soon as the berries develop on grape clusters. Genotype A would then appear on young shoots after budbreak the following year. In contrast, group B isolates might overwinter using two modes of conservations, but dominantly involving sexual reproduction and the production of cleistothecia. The following year, genotype B could spread by ascopores, resulting in numerous sparsely plots over a larger geographical area. It has been observed in other species that the balance between sexual and asexual reproduction within one species (ecotype) may depend on geographic distribution and environmental conditions (Articus et al., 2002). The hypothesis that the flag-shoot subpopulation of E. necator may reproduce clonally needs further testing. However, specific molecular markers delimiting two genetically distinct groups lead to the conclusion that, in French vineyards, recombination events could be rare or nonexistent between the two groups. These results suggest the beginning of a subspeciation process in the E. necator species similar to the one described in cereal powdery mildew, in which different formae speciales exist showing variations in rDNA-ITS and the β-tubulin gene (Wyand & Brown, 2003).

In conclusion: (i) for the first time, the gene encoding β-tubulin was cloned and sequenced from E. necator, and it is the second known gene of this pathogen; (ii) a point mutation was identified in the intron-3 region of this gene that appears to discriminate two types of distinct symptoms of grapevine powdery mildew and genotypes of E. necator; (iii) the results reported here show that allele-specific PCR based on the β-tubulin gene intron 3 SNP of E. necator can be used for future spatiotemporal studies in combination with other molecular markers (e.g. the CYP51 gene, ITS1, microsatellites); and (iv) this gene could be useful for testing phylogenetic hypotheses and for tracing historical patterns of gene flow in association with other DNA sequence loci.


This work was supported by research grant B4116 from the Aquitaine Region, by the Institute of Vine Sciences and Wine of Bordeaux (ISVV), and by the French National Institute for Agricultural Research (France). We are grateful to G. Brarda (Aude, Chamber of Agriculture, Carcassonne, France), B. Molot (Nîmes, ITV, France), M. Raynal (Bordeaux, ITV, France), C. Renaud for providing samples in 1999 and 2000, and S. Richart-Cervera, E. Cazier, L. Douence and J. Jolivet for technical assistance. We would also like to thank S. Savary, D. Bailey and C. Délye for correcting the manuscript.