Inheritance of resistance to carboxylic acid amide (CAA) fungicides in Plasmopara viticola




Mandipropamid is a new mandelic acid amide fungicide expressing high activity against foliar infecting oomycetes, including the grapevine downy mildew, Plasmopara viticola. Because cross-resistance with the valinamide fungicides iprovalicarb and benthiavalicarb and the cinnamic acid amide fungicides dimethomorph and flumorph was postulated, all five compounds are classified as carboxylic acid amide (CAA) fungicides. To support this classification, cross-resistance among these compounds with field isolates and the segregation of resistance in F1 and F2 progeny of P. viticola were evaluated. A bimodal distribution of sensitivity in field isolates and cross-resistance among all CAAs for the vast majority of isolates were detected. Crosses between sensitive (s) and CAA-resistant (r) isolates of opposite mating types, P1 and P2, yielded abundant oospores. All F1-progeny isolates were sensitive to CAAs (s:r segregation 1:0), whereas in F2 progeny segregation of about 9:1 (s:r) was observed suggesting that resistance to CAA fungicides is controlled by two recessive nuclear genes. Mating type segregated in a ratio P1:P2 of c. 2:1 in F1 and 1:1 in F2 progeny. In the same crosses, resistance to the phenylamide fungicide mefenoxam segregated in a ratio of c. 1:3:2 (sensitive:intermediate:resistant), reflecting the monogenic, semidominant nature of resistance. The risk of resistance in P. viticola was classified as high for phenylamide and moderate for CAA fungicides. This is the first report on the inheritance of phenotypic traits in P. viticola.


Downy mildew of grape caused by Plasmopara viticola is considered one of the most devastating grapevine diseases occurring in over 90 countries (Emmett et al., 1992). Disease management usually consists of different cultural practices, warning systems, resistant cultivars and chemical control. Fungicides such as the phenylamides (e.g. mefenoxam), QoIs (strobilurins, e.g. azoxystrobin and famoxadone), cymoxanil, fosetyl-Al and multi-site inhibitors (e.g. mancozeb, folpet and chlorothalonil) remain the most widely used management tools against downy mildew of grapes (Emmett et al., 1992; Gisi, 2002), although resistance has evolved to several classes of fungicides. Plasmopara viticola is a high-risk pathogen in terms of likelihood of resistance evolution (FRAC classification); therefore, it is important to continuously discover and develop fungicides with a new mode of action. Recently, a new valine amide carbamate fungicide, iprovalicarb, was introduced to the market for downy mildew control (Stenzel et al., 1998), and in terms of biological and cytological mode of action showed many similarities to dimethomorph, an older compound of the cinnamic acid amide class of fungicides (Albert et al., 1991; Cohen et al., 1995; Albert & Heinen, 1996). A close analogue of iprovalicarb, benthiavalicarb, was described and will be available for P. viticola control soon (Miyake et al., 2003). In 2005, a new mandelic acid amide derivative, mandipropamid, was presented for the effective control of oomycete pathogens (Huggenberger et al., 2005).

For the introduction of a new fungicide to the market, data have to be generated describing the baseline sensitivity of field populations as well as cross-resistance information to existing fungicides. When resistant isolates occur, either in field populations or as artificial mutants, basic information on inheritance of resistance and risk assessment have to be given for the registration document according to EPPO guideline 1/213 and the principles elucidated in more detail by Gisi & Staehle-Csech (1988) and Brent & Hollomon (1998). Dimethomorph-resistant isolates of P. viticola were claimed to already exist in French vineyards in 1994 (Chabane et al., 1996) and again in 2000 (H. Steva, personal communication). Since then, sensitivity monitoring performed by agrochemical companies has repeatedly revealed the presence of isolates resistant to dimethomorph and iprovalicarb in certain grape-growing regions of France and Germany (FRAC, 2005).

A prerequisite for investigating the inheritance of fungicide resistance in P. viticola is the availability of single-sporangiophore isolates which are well defined in terms of phenotype (e.g. mating type and sensitivity to fungicides) and genotype (amplified fragment length polymorphism (AFLP) genotype). Because the heterothallic nature of P. viticola was discovered only recently (Wong et al., 2001), sexual crosses and studies on the segregation of any phenotypic trait have not been carried out for this pathogen yet. Therefore, in the present study, defined P1 and P2 mating-type isolates from field populations had first to be selected (Scherer & Gisi, 2006) and then methods for oospore germination and production of progeny isolates developed. During this selection isolates were randomly crossed to determine the mating type and tested on leaf discs for sensitivity to different fungicides. Crosses were made between defined P1 and P2 isolates by co-inoculation of grape leaves to produce F1- and F2-progeny isolates. With this procedure the segregation pattern of fungicide resistance was determined, which was the basis for assessing the risk and extent of resistance in P. viticola to different classes of fungicides.

In this paper, two major aspects of resistance risk evaluation in P. viticola are described: first, cross-resistance in field isolates among mandipropamid, iprovalicarb, benthiavalicarb and dimethomorph, all members of the newly defined fungicide class, the carboxylic acid amide (CAA) fungicides, as compared to the phenylamide fungicide mefenoxam; and secondly, the inheritance of resistance to CAA fungicides in F1- and F2-progeny isolates as compared to that of mefenoxam resistance.

Materials and methods

Collection of isolates

Field isolates of P. viticola were collected in 2004 either from leaves from 139 different sites across different European countries (France, 58; Germany, 17; Switzerland, 22; Austria, 3; Italy, 4; Spain, 27; and Portugal, 8) or from air samples (airborne sporangia) in France (total of 40 samples from regions Rhône, Languedoc and Gascogne/Armagnac). To ensure genetic uniformity, single-sporangiophore isolates were produced by collecting sporangia from a single sporangiophore using a fine needle. The isolates were propagated and maintained on detached leaves of glasshouse-grown Vitis vinifera cv. Gutedel plants.

Characteristics of parental isolates (F0)

To obtain parental (F0) isolates with defined mating type (Fig. 1), nine randomly selected single-sporangiophore isolates were assessed for fitness parameters such as lesion size, sporulation capacity and oospore production. For oospore production, the isolates were crossed in the following manner: leaf discs (14 mm diameter) were placed, lower surface upwards, on moistened filter paper in Petri dishes (5 cm diameter). Six replicate discs per Petri dish were inoculated with three 7·5-µL droplets containing a mixture of 2·5 × 104 sporangia mL−1 of each isolate to be paired. The inoculated leaf discs were incubated in a growth chamber at 18°C (12 h light per day). After 21 and 45 days, leaf discs were examined microscopically for the presence of oospores. If no oospores were detected, the leaf discs were clarified with 1% KOH and re-examined (Scherer & Gisi, 2006). Four isolates with strong fertility (rapid oospore production in crosses) and high aggressiveness (intensive sporulation after 8 days), two of each mating type, were selected as reference isolates.

Figure 1.

Steps for investigating segregation of resistance to mandipropamid in single-sporangiophore isolates of Plasmopara viticola.

Sensitivity assay

The sensitivity assay was conducted in 24-well plates with 1 mL of 0·5% water agar in each well. Leaf discs (15 mm diameter) were punched from the third and fourth leaves of grape cv. Gutedel plants and placed upside down onto the agar. Fungicide dilutions (from a 1000-mg a.i. L−1 stock solution disolved in dimethyl sulfoxide, DMSO) were made to give concentrations of 300, 100, 30, etc. down to 0·01 mg a.i. L−1). Four CAA fungicides: mandipropamid, dimethomorph, iprovalicarb and benthiavalicarb (Fig. 2) were included in the tests as well as the phenylamide fungicide mefenoxam. The fungicides were applied with Tecan spray equipment (Genesis). One day after fungicide application the discs were spray-inoculated with a sporangial suspension of 50 000 sporangia mL−1. Plates were closed with a lid and incubated in a climate chamber (as described above) for 6 days. Disease assessment was by visual estimation of the percentage of the leaf disc area infected. The data were analysed using the agstat program for determining EC50 values.

Figure 2.

Chemical structures of carboxylic acid amide (CAA) fungicides (common structural parts in bold).

Crossing experiments and sexual recombination

Two isolates with known mating types of P1 and P2 and sensitivity to mandipropamid and mefenoxam were chosen for crossing experiments (Fig. 1). Isolate D5, collected in 2001 in Germany, was a P1 mating-type isolate and sensitive to mandipropamid, whereas CH05·2, collected in Switzerland, was P2 and resistant to mandipropamid. Both isolates showed intermediate resistance to mefenoxam. The production of oospores was as described above for mating-type determination. Leaf discs containing oospores were stored for 8 weeks at room temperature (19°C) on dry filter paper. Afterwards, 10–15 dry leaf discs were homogenized in 5 mL cold distilled water; the oospore concentration was about 500–1800 mL−1. The homogenate was diluted with 7·5 mL water and mixed with 2 g ground perlite in 9-cm-diameter Petri dishes as described by Rubin & Cohen (2006) for Phytophthora infestans. For germination of oospores and infection with P. viticola, very young grape leaves were floated on the perlite/oospore mixture and plates incubated for 7–10 days at 19°C in a growth chamber (14 h light per day). Microscopic observations revealed that under such conditions oospores germinated by formation of a macrosporangium releasing zoospores into the substrate which consequently produced infections in the leaves. When first sporulating spots were observed, e.g. sporangiophores alongside the main veins on the upper leaf surface, the leaves were transferred to new Petri dishes with moist filter paper and incubated for another week. Sporangia from one sporangiophore were then isolated and transferred into one water droplet (5 µL) on new leaves for multiplication of the isolates. Such isolates were tested for sensitivity to fungicides and for mating type as described above. In this way, F1-progeny isolates were produced of which pairs of isolates with opposite mating types (siblings) were crossed to produce F2-progeny isolates (Fig. 1). Unfortunately, backcrosses could not be produced, because the zoospores released by the macrosporangia repeatedly failed to infect the leaves floated on the oospore/perlite mixture, possibly because of inbreeding depression and loss of pathogenicity.

Molecular characterization

Two microsatellite (SSR, simple sequence repeat) loci, ISA and CES, each with three alleles, were used for evaluating the ploidy of the two parental isolates (F0), CH05·2 and D5, and the success of sexual recombination to produce pure F1-progeny isolates (especially the five isolates II.8, II.10, II.25, II.26 and II.29 used for production of F2 progeny). The methods used for SSR analysis are described in detail elsewhere (Gobbin et al., 2003; Scherer & Gisi, 2006). DNA fingerprinting of isolates was performed with AFLP as described by Vos et al. (1995). EcoR1 primers with two selective bases were labelled with the fluorescent dye FAM (Microsynth). The fluorescent-labelled fragments were visualized on a 3130 Genetic Analyser (Applied Biosystems). The peak pattern was analysed using genemapper v3·7 software (Applied Biosystems). A total of 16 primer combinations were used, yielding 439 polymorphic AFLP markers. The selection of markers was made on the basis of differences in polymorphisms between the two parental isolates CH05·2 and D5. Genetic similarity and upgma cluster analysis for AFLP genotypes were performed by using the computer package phylip (Felsenstein, 1995), with the Jaccard distance coefficient and neighbour-joining cluster analysis. Radial trees were visualized using treeview software (


In 2004, P. viticola samples were collected either from infected leaves or airborne sporangia across Europe and tested for sensitivity to mandipropamid. All samples originated from trial sites, except those from France, of which 25 leaf samples came from trial sites, 33 from commercial fields and 40 from randomly collected airborne samples. The EC50 values of the sensitive isolates (n = 98) ranged from 0·03 to 3·6 mg L−1 (variation of about 100-fold) with a median EC50 of 0·95 mg L−1 (Fig. 3). Clearly separated from the sensitive isolates were 41 resistant isolates with an EC50 > 300 mg L−1 (Fig. 3) of which 37 came from trial sites in France, Germany and Switzerland, where CAA fungicides have been used for many years. The resistance factor calculated on the basis of EC50 values of mandipropamid for single isolates ranged between 100 and 10 000.

Figure 3.

Sensitivity to mandipropamid (EC50) of Plasmopara viticola bulk isolates from leaves collected at different trial sites in Europe in 2004 (isolates sorted by ascending EC50 values).

For evaluating intrinsic activity and cross-resistance, a group of 45 sensitive and resistant isolates covering the entire sensitivity range was tested against the four CAAs mandipropamid, dimethomorph, iprovalicarb and benthiavalicarb. Mandipropamid and benthiavalicarb were similarly active (median EC50: 1·05 and 1·10 mg L−1, respectively), whereas iprovalicarb (EC50: 6·2 mg L−1) and dimethomorph (EC50: 10·3 mg L−1) were less active. Cross-resistance among all four CAAs was observed for the vast majority of the isolates (regression coefficient r2: 0·71–0·95, P < 0·05) (Fig. 4). In contrast, no cross-resistance was found between the CAA fungicide dimethomorph and the phenylamide fungicide mefenoxam for a total of 40 isolates collected from airborne sporangia in France (Fig. 5). For this group of isolates, the ratio of CAA-sensitive to CAA-resistant isolates was about 1:1, whereas for mefenoxam, the ratio between sensitive, intermediate and resistant isolates was 1·5:2:1 (Fig. 5). Three of the 40 isolates carried combined resistance to both dimethomorph and mefenoxam (multiple resistance), without obvious evidence for fitness costs.

Figure 4.

Cross-resistance (correlation between EC50 values) among mandipropamid, dimethomorph, iprovalicarb and benthiavalicarb for Plasmopara viticola isolates collected in Europe in 2004 (n= 45).

Figure 5.

Sensitivity (EC50 values) to dimethomorph (DME) and mefenoxam (MFX) of Plasmopara viticola isolates randomly collected from air samples in 2004 in France (Rhône, Languedoc, Gascogne/Armagnac). Definition of sensitivity classes in this test: for mefenoxam: sensitive: EC50 < 1 mg L−1; intermediate: EC50 1–10 mg L−1; resistant: EC50 > 10 mg L−1. For dimethomorph: isolates considered resistant if EC50 > 20 mg L−1. Discriminatory doses are indicated by horizontal lines (for mefenoxam) and vertical line (for dimethomorph), respectively. Arrows indicate EC50 values  > 300 mg L−1.

The ultimate test of whether resistance to different fungicides was genetically linked was the determination of the segregation pattern in F1 and F2 progeny. When a CAA-sensitive and a CAA-resistant single-sporangiophore isolate (with different mating types, P1 vs. P2) were co-inoculated in droplets onto the same leaf, allowing sexual recombination and oospore formation (cross II), all 33 F1-progeny isolates were sensitive to CAA fungicides with a variation of about 50-fold between the most and the least sensitive offspring (Fig. 6). No resistant isolates were found (s:r segregation 1:0). The same behaviour was also found in two other P1 × P2 combinations (data not shown because fewer progeny isolates were obtained). When two F1 isolates (siblings) were crossed (three different crosses K, G and J), the majority of F2-progeny isolates were again sensitive to CAA fungicides, with only 1–5 isolates per cross being resistant (Fig. 7). The segregation of resistance (s:r) in the F2 progeny was 24:1, 22:1 and 16:5 in crosses K, G and J, respectively, yielding an overall segregation of 62:7 (approx. 9:1).

Figure 6.

Sensitivity (EC50 values) to mandipropamid and mating type (P1, P2) of F1 progeny isolates (n= 33) produced from cross II between a sensitive (D5) and a resistant (CH05·2) parent (black columns) of Plasmopara viticola (leaf-disc assay). Isolates with encircled number were selected for the F2 generation. Bars indicate minimum and maximum values of three replicates. EC50 of resistant parent (CH05·2) > 300 mg L−1. Mating type of isolates with question mark was not determined.

Figure 7.

Sensitivity to mandipropamid of F2− progeny isolates (n= 69) produced from three crosses (G, J, K; white and hatched columns) between sensitive F1 isolates (grey columns highlighted by triangles) of Plasmopara viticola. F0: black columns highlighted by arrows; sensitive (s) parent D5: EC50= 0·11 0·10 (0·05–0·22); resistant (r) parent CH05·2: EC50 > 100.

Different hypotheses were analysed statistically to test whether one or two genes were involved in CAA resistance. Because all F1 progeny isolates were sensitive, resistance might be controlled by a recessive nuclear gene or by a mitochondrial one. Since resistance re-appeared in the F2 progeny, the latter possibility was discounted. The observed 62:7 (c. 9:1) segregation of resistance in the F2 progeny could be based on two recessive genes with either one or both loci being heterozygous (Table 1). Assuming the three crosses J, G and K had the same genetic background and the crosses were analysed individually, only one locus of the resistant parent could be heterozygous for matching all conditions. Based on these results, resistance to CAA fungicides was assumed to be inherited by two recessive nuclear genes (Table 1).

Table 1.  Statistical analysis of hypothetical, expected and observed segregation pattern of CAA resistance in three F2 progenies of Plasmopara viticola
Genotype of progenyHypothetical segregationExpected segregationObserved segregationPNull hypothesis
Cross J n = 21
one recessive gene1:3 5·2515·755160·8997accepted
two recessive genes (one locus heterozygous)1:7 2·6318·385160·1171accepted
two recessive genes (both loci heterozygous)1:15 1·3119·695160·0009rejected
Cross G n = 23
one recessive gene1:3 5·7517·251220·0222rejected
two recessive genes (one locus heterozygous)1:7 2·8820·131220·2371accepted
two recessive genes (both loci heterozygous)1:15 1·4421·561220·7063accepted
Cross K n = 25
one recessive gene1:3 6·2518·751240·0153rejected
two recessive genes (one locus heterozygous)1:7 3·1321·881240·1988accepted
two recessive genes (both loci heterozygous)1:15 1·5623·441240·6421accepted
Sum of crosses n = 69
one recessive gene1:317·2551·757620·0044rejected
two recessive genes (one locus heterozygous)1:7 8·6360·387620·5542accepted
two recessive genes (both loci heterozygous)1:15 4·3164·697620·1814accepted

It could not be ruled out that one of the parental isolates, CH05·2 or D5, might be polyploid rather than diploid, as individual oomycetes mostly are. Polyploid isolates of P. viticola were detected recently in Greek populations by Rumbou & Gessler (2006), therefore, the parental isolates as well as the five F1-progeny isolates, II.8, II.10, II.25, II.26 and II.29, which were crossed for production of F2-progeny isolates, were characterized with the two SSR loci ISA and CES each containing three alleles. For both loci, CH05·2 was homozygous (allele 138 of ISA and 138 of CES), whereas D5 was heterozygous (alleles 112 and 134 of ISA and 165 and 167 of CES). All five F1-progeny isolates carried two alleles for both loci, both alleles received from either parent (Table 2). These results suggest that the five F1 isolates emerged from sexual recombination (no selfing) and were diploid, as were the parents. Further genotypic characterization of 39 F2-progeny isolates derived from the three crosses with AFLP markers demonstrated that all were different from each other and from F1 and F0 parental isolates in at least 37 out of the 439 polymorphic markers, thus excluding selfing (Fig. 8). Using 16 different primer combinations, 439 semidominant polymorphic markers were identified distinguishing the F0 parents (CH05·2 and D5). These polymorphic markers were used to compare the F1- and F2-progeny isolates by cluster analysis; the isolates derived from cross K formed a distinct cluster including one of the F0 parents (CH05·2) and were separate from crosses G and J, whereas the progeny were intermixed (Fig. 8).

Table 2.  Genotypic characterization (allele combination) of two F0 parental isolates (CAA-resistant CH05·2 and CAA-sensitive D5) and five CAA sensitive F1-progeny isolates of Plasmopara viticola based on the two SSR loci, ISA and CES, each with three alleles (figures are lengths in base pairs)
F0CH05·2  xx  
D5xx  xx
F1II.8x xx x
II.10 xxxx 
II.25 xxxx 
II.26 xxx x
II.29 xxx x
Figure 8.

upgma cluster analysis of 39 F2 progeny isolates (CAA-resistant isolates highlighted with circles) of Plasmopara viticola produced from three crosses (G, J, K) compared with the five F1 (grey frame) and the two F0 parents (white frame, parent CH05·2, CAA-resistant and D5,CAA sensitive) based on 439 AFLP markers. Genetic distance was calculated using the Jaccard coefficient.

For technical reasons, the mating type was not determined for all F1- (and F2-) progeny isolates. Within the tested isolates, mating-type segregation P1:P2 was 16:7 in the F1 progeny (Fig. 6) and 6:7 in the F2 progeny (data not shown). Based on the small number of tested progeny isolates, the proportion of mating-type segregation could not be evaluated properly; nevertheless, the observed segregation confirmed that sexual recombination did occur in the tested crosses.

The two parents used for inheritance of CAA resistance (sensitive D5 × resistant CH05·2 in cross II) were intermediate in sensitivity to mefenoxam. All their F1-progeny isolates were sensitive to CAA fungicides (s:r segregation 1:0), but showed the typical segregation pattern for phenylamide fungicides with a ratio of sensitive:intermediate:resistant of 1:2·7:1·8 (Fig. 9). These data show the absence of cross-resistance between CAA and phenylamide fungicides, and that there was no co-segregation of resistance to CAA and phenylamide fungicides. The last open question concerned co-segregation of resistance to different CAA fungicides. This was addressed by testing the sensitivity of a range of sensitive and resistant F2 isolates against mandipropamid, dimethomorph and iprovalicarb. The results showed a very clear co-segregation of resistance to the three molecules and also a clear cross-resistance amongst them for the F2 isolates (Figs 10 and 11).

Figure 9.

Segregation of sensitivity (EC50 values) of Plasmopara viticola isolates to mandipropamid (MPD) and mefenoxam (MFX) in the F1 progeny (MPD s:r = 1:0; MFX s:i:r = 1:2·7:1·8).

Figure 10.

Sensitivity (EC50 values) of Plasmopara viticola isolates to mandipropamid and dimethomorph in the F2 progeny and in F1 and F0 parents. Isolates are defined as resistant to mandipropamid and dimethomorph if EC50 > 5 and > 20 mg L−1, respectively.

Figure 11.

Sensitivity (EC50 values, cross-resistance) of Plasmopara viticola isolates to mandipropamid, dimethomorph and iprovalicarb in the F2 progeny. Isolates were defined as resistant to iprovalicarb and dimethomorph if EC50 > 20 mg L−1.


This paper is the first report on inheritance of phenotypic traits in P. viticola. No previous data on crosses in P. viticola with any trait are available and because of the biotrophic nature of the pathogen, it is extremely labour-intensive to produce sufficient progeny isolates to analyse the segregation pattern properly. Since the heterothallic nature of P. viticola became known only recently (Wong et al., 2001) and isolates with defined mating type were not available, new techniques had to be developed to collect single-sporangiophore isolates, to determine their mating types, named P1 and P2 (Scherer & Gisi, 2006), and most importantly to induce oosporic infection of detached grape leaves in the laboratory under controlled conditions. This technique involved 2-month-old oospores in dry leaves, mixed as a powder with perlite amended with water in Petri dishes, which were incubated together with grape leaves on top of the substrate as bait to allow zoospore release from macrosporangia and infection of the host tissue. This method allowed the production of F1- and F2-progeny isolates for studying the segregation of resistance to fungicides.

The data presented in this study show that resistance to CAA fungicides in field populations of P. viticola is distributed in a bimodal fashion, with isolates being either sensitive or resistant with a resistance factor of over 100. The lack of intermediate resistance suggested that resistance is either dominant or recessive. The data also show that resistance to mandipropamid is associated with resistance to dimethomorph, iprovalicarb and benthiavalicarb (all CAA fungicides), but not to mefenoxam (a phenylamide fungicide), suggesting for P. viticola cross resistance among all CAAs, thus sharing a similar mode of action/mechanism of resistance. Prior to this study, cross-resistance was not expected for all four CAA molecules because the chemical structures are quite different, although there is a certain similarity (Fig. 2). The ultimate proof for cross-resistance should derive from genetic studies in which a CAA-sensitive and a CAA-resistant isolate were crossed. With this approach, it was shown that all F1-progeny isolates were sensitive to CAAs (s:r segregation 1:0), whereas segregation of resistance in the F2 progeny (total of isolates from all three crosses) was c. 9:1 (sensitive:resistant). This segregation pattern suggests that (i) resistance to mandipropamid cannot be inherited maternally (cytoplasmatically, by a mitochondrial gene) because segregation in F2 was not 1:0 (s:r); (ii) resistance to mandipropamid is inherited recessively and is most probably controlled by two nuclear genes; (iii) resistance to all CAAs co-segregates and is therefore controlled by the same genes; and (iv) resistance to CAAs and resistance to mefenoxam do not co-segregate and are therefore controlled by different genes. Nevertheless, genes controlling resistance to both CAAs and mefenoxam may occur in the same isolate, causing multiple resistance. Cross-resistance between dimethomorph, iprovalicarb and benthiavalicarb was also described recently for P. viticola by Young et al. (2005).

Resistance to mefenoxam segregated in this study in a ratio close to 1:3:2 (sensitive:intermediate:resistant), which is typical for the well-known Mendelian segregation pattern (1:2:1) controlled by one semidominant gene (Gisi & Cohen, 1996; Knapova et al., 2002). In fact, the results generated for mefenoxam served as a positive control for segregation after sexual recombination. Therefore, the results of this study describing the inheritance of CAA resistance are considered robust, suggesting recessiveness of CAA resistance in P. viticola. The F2 progenies of the three single crosses G, J and K contained different proportions of resistant isolates, ranging from 24:1 (in K) to 22:1 (in G) and 3:1 (in J). In cross J resistance may be based on one recessive gene or two recessive genes with one locus being heterozygous, whereas in the other two crosses the observed segregation pattern can be explained by two recessive genes (with one or both loci being heterozygous) or one recessive gene affected by several modifier genes. The involvement of modifier (‘minor’) genes was used to explain the somewhat distorted segregation of metalaxyl resistance in some P. infestans crosses (Fabritius et al., 1997; Knapova et al., 2002). Assuming that all F1-progeny isolates of the three crosses have the same genetic background for resistance, they are believed to belong to the same ‘family’ yielding an overall segregation of 9:1, which is within the hypothetical limits of 7:1 and 15:1 for resistance controlled by two recessive genes (Table 1). Alternatively, the different proportions of resistant isolates in the three progenies may just be the result of too few isolates per cross to yield the same segregation pattern. Also, the different ratios of mating-type segregation, P1:P2, in F1 (about 2:1) and F2 (about 1:1) may result from the limited number of tested isolates rather than the phenomenon of balanced lethals because of a deficiency in essential genes for mating, as described for P. infestans (Judelson et al., 1995). Therefore, it is suggested that segregation of mating type in P. viticola is balanced, but to date no-one has described the inheritance pattern of mating type in P. viticola.

The recessive nature of CAA resistance controlled by two nuclear genes and the obligate mating of P. viticola each season will dilute resistance by keeping a certain proportion of sensitive individuals in the population. The intrinsic risk and extent of resistance to CAA fungicides in P. viticola is postulated to be moderate and considerably lower for CAAs than for phenylamides and QoIs, where resistance is controlled by one semidominant nuclear gene and a mitochondrial gene, respectively. Therefore, it is expected that CAA resistance in P. viticola can be managed under field conditions by using appropriate strategies such as restricted number of applications and the use of mixtures with non-cross-resistant fungicides.


The sensitivity tests done by EpiLogic (Dr F. Felsenstein) for the 40 samples collected randomly by EpiLogic in autumn 2004 in France (Fig. 5) are acknowledged.