Adaptation of a plant pathogen to partial host resistance: selection for greater aggressiveness in grapevine downy mildew

Abstract

An understanding of the evolution of pathogen quantitative traits in response to host selective pressures is essential for the development of durable management strategies for resistant crops. However, we still lack experimental data on the effects of partial host resistance on multiple phenotypic traits (aggressiveness) and evolutionary strategies in pathogens. We performed a cross-inoculation experiment with four grapevine hosts and 103 isolates of grapevine downy mildew (Plasmopara viticola) sampled from susceptible and partially resistant grapevine varieties. We analysed the neutral and adaptive genetic differentiation of five quantitative traits relating to pathogen transmission. Isolates from resistant hosts were more aggressive than isolates from susceptible hosts, as they had a shorter latency period and higher levels of spore production. This pattern of adaptation contrasted with the lack of neutral genetic differentiation, providing evidence for directional selection. No specificity for a particular host variety was detected. Adapted isolates had traits that were advantageous on all resistant varieties. There was no fitness cost associated with this genetic adaptation, but several trade-offs between pathogen traits were observed. These results should improve the accuracy of prediction of fitness trajectories for this biotrophic pathogen, an essential element for the modelling of durable deployment strategies for resistant varieties.

Keywords: Vitis vinifera; erosion; evolvability; fitness cost; host specificity; obligate plant pathogen; quantitative resistance; virulence.

Figure 1

Principal component analysis (PCA) on multilocus genotypes based on 32 microsatellite markers, for the 103 Plasmopara viticola isolates. Pathogen origins are indicated in colour on the PCA. ‘SUS’: 49 isolates collected from the susceptible host (pale green); ‘RES’: 54 isolates collected from partially resistant hosts (dark green).

Figure 2

Sporulation dynamics of Plasmopara viticola isolates from partially resistant hosts (‘RES’; dark green; n = 54) and susceptible host (‘SUS’; light green; n = 49). Isolates were used to inoculate (A) Vitis vinifera cv. Cabernet sauvignon, (B) Regent, (C) Prior and (D) Bronner. On the left‐hand side of the figure, the mean percentage (±SE) of the leaf disc area displaying sporulation (experimental data obtained from image analysis) is presented from 0 to 5 days post‐inoculation (dpi). On the right‐hand side of the figure, sporulation curves resulting from the adjustment of a logistic model (mean ± 95% confidence intervals) for the experimental data are presented for each pathogen origin (‘SUS’ and ‘RES’). The relative sporulating leaf disc area presented on the y‐axis is the proportion of the leaf disc displaying sporulation on day x divided by the sporulating area on day 5. Horizontal lines correspond to 50% of final sporulation (5 dpi). From these curves, we estimated T ₅₀ as the time to reach the 50% and sporulation rate, and the slope at T ₅₀ (Table 2). See the text for details of the logistic model and Table 3 for statistical differences between inoculated hosts and pathogen origins.

Figure 3

Quantitative pathogenicity traits of Plasmopara viticola isolates. The pathogen origin ‘RES’ includes isolates collected from resistant varieties (dark green; solid line; n = 54) and the pathogen origin ‘SUS’ includes isolates collected from susceptible Vitis vinifera cultivars (light green; dashed line; n = 49). Isolates were used to inoculate Bronner, Prior, Regent and V. vinifera cv. Cabernet sauvignon (Cab. sauv.). For each quantitative trait (A: spore production; B: spore size and C: latency period), mean ± SE values of isolates by pathogen origin and frequency density plots are presented for each inoculated host on the left‐ and right‐hand side panels, respectively. See Table 3 for full model statistics. Post hoc least squares mean differences between pathogen origins (‘SUS’ and ‘RES’) on each inoculated host were determined and the significance of these differences is indicated as follows: ***P < 0.0001; **P < 0.001; *P < 0.05.

Figure 4

Quantitative trait correlations for 103 Plasmopara viticola isolates collected from susceptible grapevine cultivars (‘SUS’, triangles; n = 49) and partially resistant grapevine varieties (‘RES’, circles, n = 54). Spore production decreased with (A) increasing spore size, (B) increasing latency period duration and (C) increasing sporulation rate. Isolates were used to inoculate Bronner (green), Prior (purple), Regent (red) and Vitis vinifera cv. Cabernet sauvignon (blue). The overall correlation is indicated with a black solid line. Full statistics for all correlations tested are presented in Table 4.

Figure 5

Specificity of Plasmopara viticola isolates for resistant grapevine varieties. None of the differences between local and foreign isolates was significant (main pathogenicity traits), on any of the inoculated hosts (Table 5). Mean (±SD) values for quantitative pathogenicity traits (A: spore production; B: spore size; C: latency period) are presented for each inoculated resistant host, for local and foreign isolates. Local isolates were collected from a focal resistant variety and used to inoculate the same variety (Bronner: n = 9; Prior: n = 10; Regent: n = 7), whereas foreign isolates were collected from a focal resistant variety and used to inoculate a different variety (Bronner: n = 37; Prior: n = 44; Regent: n = 46).