Fine monitoring of the effects of grapevine resistance loci on the development of Plasmopara viticola

Plasmopara viticola, the agent of downy mildew native to North America, was introduced into Europe in the second half of the 19 century. It rapidly devastated all the European vineyards (Gessler et al., 2011). For decades, massive applications of fungicides were used to control downy mildew, because the European grapevine varieties display no or little resistance. In this context, the development of resistant varieties is a promising strategy for limiting the use of fungicides. Over the last twenty years, more than thirty genetic factors, derived mainly from Vitis species, have been identified and some of them are considered as major genes. The strategy of pyramiding resistance factors in a variety is used to limit the risk of resistance breakdown by pathogen isolates in the vineyard. In this strategy, the choice of different resistance genes to be combined together in the same variety is crucial to ensure resistance durability. In this study, we analyzed the effect of four resistance factors on the P. viticola development: Rpv1, located on linkage group (LG) 12 and inherited from V. rotundifolia (Merdinoglu et al., 2003); Rpv3.1, located on LG 18, inherited from V. rupestris (Welter et al,. 2007; Bellin et al., 2009; Di Gaspero et al., 2012; Foria et al., 2020); Rpv3.3 located on LG 18, inherited from V. labrusca or V. riparia (Di Gaspero et al., 2012; Vezzulli et al. 2019) and Rpv10 located on LG 9 and inherited from the Asian species V. amurensis (Schwander et al., 2012). For this, we used a grapevine population in which the four resistance factors (loci) segregated. Among the fifteen possible combinations of these loci, the study focused only on the individual loci to analyse the effect of each of them compared to the susceptible genotype (with no resistance factor). For this purpose, three approaches were implemented: macroscopic symptoms assessment, cytological observations of developmental stages and monitoring of the pathogen biomass by quantification of P. viticola-specific lipids.


Introduction
Plasmopara viticola, the agent of downy mildew native to North America, was introduced into Europe in the second half of the 19 th century. It rapidly devastated all the European vineyards (Gessler et al., 2011). For decades, massive applications of fungicides were used to control downy mildew, because the European grapevine varieties display no or little resistance. In this context, the development of resistant varieties is a promising strategy for limiting the use of fungicides. Over the last twenty years, more than thirty genetic factors, derived mainly from Vitis species, have been identified and some of them are considered as major genes. The strategy of pyramiding resistance factors in a variety is used to limit the risk of resistance breakdown by pathogen isolates in the vineyard. In this strategy, the choice of different resistance genes to be combined together in the same variety is crucial to ensure resistance durability. In this study, we analyzed the effect of four resistance factors on the P. viticola development: Rpv1, located on linkage group (LG) 12 and inherited from V. rotundifolia (Merdinoglu et al., 2003); Rpv3.1, located on LG 18, inherited from V. rupestris (Welter et  Rpv10 located on LG 9 and inherited from the Asian species V. amurensis (Schwander et al., 2012). For this, we used a grapevine population in which the four resistance factors (loci) segregated. Among the fifteen possible combinations of these loci, the study focused only on the individual loci to analyse the effect of each of them compared to the susceptible genotype (with no resistance factor). For this purpose, three approaches were implemented: macroscopic symptoms assessment, cytological observations of developmental stages and monitoring of the pathogen biomass by quantification of P. viticola-specific lipids.

Materials and methods
Plant material and experimental design. For each Rpv locus, three genotypes were selected among the population and for each genotype, three plants were used. Plants, derived from green cuttings, were grown on potting soil in a greenhouse at 22-19°C (day-night) and a photoperiod of 16 h light until reaching a 12 leaf stage. For each plant, the leaves, number 4 and 5, (starting from the apex) were excised and transferred to the laboratory. Leaf discs (2 cm diameter) were produced. All the leaf discs were placed in 12 microwell plates. They were inoculated by spraying a sporangia solution (5x10 4 sp/ml -1 ) or with water. For each leaf, 5 discs were sprayed with the P. viticola "Colmar" isolate and 5 discs with the "Lednice" isolate. This isolate was able to overcome the Rpv3.1 locus (Peressotti et al, 2010). The remaining leaf discs were sprayed with water, as a control. Then, the plates were incubated in a growth chamber (21°C, 80% relative humidity, 50 µmol/m 2 /s light intensity). Leaf discs were collected at 24, 48, 72, 96 hours and 6 days post infection. Each leaf disc was cut in 2 halves, one for the cytological study and one for the pathogen biomass assessment.  Monitoring of the pathogen biomass by quantification of P.viticola-specific lipids. The assessment of the pathogen biomass by relative quantification of P. viticola lipids (Negrel et al, 2018), was monitored at 24, 48, 72 and 96 hours post infection. These lipids included ceramides and derivatives of arachidonic and eicosapentaenoic acid which were not detected in healthy grapevine plants. Prior to extraction, each leaf disc was freeze-dried, weighed and powdered. Metabolites were extracted with MeOH/CHCl3 (1/1, v/v) using 15 μL per mg dry weight. The suspension of leaf powder in MeOH/CHCl3 was then sonicated for 15min in an ultrasound bath. Two milliliters of ultrapure water were added to allow phase separation. After centrifugation at 15,000 g for 10 min, 100 μL of the chloroform phase was recovered, diluted with 100 μL of MeOH and analyzed by UHPLC-MS according to Negrel et al. (2018). The results were presented as the total lipids detected for each Rpv locus and for each P. viticola isolate, corresponding to the sum of peak areas resulting from individually quantified lipids. Symptom assessments by using the OIV 452-1 scale.

These four
With the "Colmar" isolate ( Fig.2), sporulation was observed for all Rpv loci. However, the number of spores produced was variable. For the susceptible genotypes (S), the quantity of spores produced was the highest, whereas for the locus Rpv10, only very few spores were produced. In addition, a hypersensitive reaction (HR) was observed for all Rpv loci, however they displayed different HR characteristics in shape, size, color and number. The Rpv3.1 and Rpv10 loci developed very small necrosis, whereas Rpv1 and Rpv3.3 loci induced larger necrosis. For the genotypes with the Rpv3.1 locus and challenged with the "Lednice" isolate ( Fig.3), the sporulation level was as high as that of "S". More generally, for the susceptible genotypes (S) and for each Rpv locus, the sporulation levels seemed higher with this isolate. The HR was also less visible.  The analysis of the OIV452-1 scores (Fig.4) showed that with the "Colmar" isolate, all Rpv loci induced a partial resistance displaying scores between 4.5 (Rpv 3.3) and 7.7 for Rpv10. The scores of "S" and that of Rpv3.3 locus were very close (between 3.2 and 4.0). The scores of the Rpv1 and the Rpv3.1 loci were very similar, but below that of the Rpv10 locus. With the "Lednice" isolate, the OIV452-1 scores of all Rpv loci were lower compared to those produced by the other isolate, especially for the Rpv3.1 locus, whose score dropped from 6.8 to 3.7 and reached that of "S". Cytological analysis of the P. viticola development.
With the "Colmar" isolate ( Fig.5), the "mycelium" appeared as the most important structure for the "S", even if "branched hypha" and "collapsed hypha" or "collapsed mycelium" were sometimes observed. The same profile was observed for the Rpv3.3 locus but with a higher percentage of "branched hypha". This suggests that for this locus, the pathogen development was slightly delayed. For the Rpv1, Rpv3.1 and Rpv10 loci, "collapsed branched hypha" was the most prevalent stage, suggesting an early destruction of the pathogen structures, especially for the Rpv10 locus for which "collapsed long hypha" were also observed.
With the "Lednice" isolate ( Fig.6), the observed profiles were similar to that of the "Colmar" isolate, except for the Rpv3.3 locus. In this case, the profile was close to that of the "S" and no collapsed structures or delay in the pathogen development were observed. Interestingly, for the Rpv1 and Rpv3.3 loci, the proportion of mycelium was higher than for the "Colmar" isolate, suggesting that "Lednice" was more aggressive than the "Colmar" isolate. For the Rpv10 locus, no differences between isolates were observed.  Evaluation of the pathogen biomass by quantification of pathogen-specific lipids (Fig.7). With the "Colmar" isolate, the highest amount of total lipids was obtained for the "S" genotypes and for the Rpv3.3 locus. Conversely, the lowest amount was detected for the Rpv10 locus (20 times less than for the "S"). The amount detected for Rpv1 and Rpv3. 1 loci was respectively 11 and 8 times lower than for the "S". For all the Rpv loci, the total lipid amounts were higher with "Lednice" than with the "Colmar" isolate. A significant isolate effect was detected for the Rpv3.1 locus for which the amount measured was close to that of the "S". Interestingly, a significant effect was also observed for the Rpv1 locus.

Discussion
In this study, the first sporulation events were observed at 4 dpi for "S" and for all Rpv genotypes, albeit with different intensity. Sporulation levels ranged from high for "S" and Rpv3.3 to very low for Rpv10, with both pathogen isolates. Rpv1 and Rpv3.1 loci exhibited intermediate levels with the "Colmar" isolate, whereas sporulation of the "Lednice" isolate was higher with the Rpv3.1 locus, due to the breakdown of this locus by this isolate. According to Bellin et al, (2009), the OIV 452-1 scale rated the plant reaction represented by the necrosis pattern (flecks, spots, substomatal HR) and the sporulation level. The OIV 452-1 scores clearly reflected the symptoms induced by the different Rpv loci. However, this approach didn't explain how these symptoms were produced. Generally, symptoms are the result of the mesophyll colonization by the pathogen and the plant reaction effect. The colonization is considered as a sequence of transformations of pathogen structures: from vesicles to mycelium. At 4 dpi, both the cytology and the pathogen biomass approaches indicated a delay of the pathogen development between the Rpv genotypes and the susceptible genotypes "S". The pathogen biomass approach demonstrated that for Rpv1, Rpv3.1 and Rpv10 loci, the total amount of pathogen lipids was respectively 11, 8 and 20 times lower than for the "S" genotypes. This indicated that the pathogen development was strongly reduced by the presence of these Rpv loci but with different degrees. Moreover, the cytological approach revealed that pathogen structures were collapsed or destroyed by the presence of these loci at different time points of the pathogen development. For the Rpv10 locus, the collapsed "long hypha" stage represented 20% of the total structures observed. This indicated a rapid effect of this resistance locus after the beginning of the infection. For the Rpv1 locus, 60% of the pathogen structures were represented by the "collapsed branched hypha" stage, suggesting that the effect was delayed compared to that of Rpv10. The destruction of the pathogen structures could be the result of the defense mechanisms effect. Several recent transcriptomic studies have shown that Rpv loci are able to induce multiple defense responses. The accumulation of stilbenes seems to be a common defense response between  The cytological and the pathogen biomass approaches seemed to be complementary to produce information concerning the effect of the Rpv loci studied either during the early stages of the pathogen development or for the explanation of the macroscopic symptoms.

Conclusions
This study showed, by observing the macroscopic symptoms, that the Rpv1, Rpv3.1, Rpv3.3 and Rpv10 loci induced various resistance levels to P. viticola. The high resistance level conferred by the Rpv10 locus could be explained by a delay and a strong inhibition of the pathogen development associated to a rapid destruction of the pathogen.