Impact of genetic drift, selection and accumulation level on virus adaptation to its host plants: Impact of genetic drift on PVY adaptation

Resistance to pathogens, defined as the capacity of a host to decrease its pathogen load (Raberg et al., 2007; Restif and Koella, 2004), is widespread in plants. However, resistance efficiency, specificity and genetic determinism are highly variable across genotypes of a given plant species. To date, plant breeders have mostly created resistant cultivars using resistance mechanisms showing monogenic inheritance and a high efficiency level, often called ‘qualitative resistance’. Unfortunately, the protection conferred by such resistance genes is often poorly durable, and resistance ‘breakdowns’ caused by pathogen evolution can be frequent and rapid (Garcia‐Arenal and McDonald, 2003; McDonald and Linde, 2002). In the case of viruses, the breakdown of a major resistance gene recently introgressed into commercial plant cultivars and used by growers can be schematically divided into three major steps (Gomez et al., 2009; Moury et al., 2011). If we assume that no resistance‐breaking variant is initially present, the first step is the appearance of such variants from a wild‐type (WT) virus population. The second step is within‐plant colonization and accumulation of the resistance‐breaking variants in competition with the rest of the virus population (i.e. WT variants). The last step is the transmission of the resistance‐breaking variants to other plants, allowing epidemics to develop in plant cultivars carrying the resistance gene. Different evolutionary forces rule these three steps. The appearance of resistance‐breaking variants usually involves a small number of nucleotide substitutions in the so‐called avirulence factor encoded by the viral genome (Harrison, 2002; Moury et al., 2011). More rarely, recombination may be required for this step (Diaz et al., 2004; Miras et al., 2014). Then, the accumulation of resistance‐breaking variants within plants depends on selection and genetic drift. Selection favours the variants with highest fitness, i.e. growth rates when considering the within‐plant scale, increasing their frequency over time. This deterministic force is usually evaluated with the selection coefficient, defined as the difference in fitness between two variants. By contrast, genetic drift acts in the same way on all variants of the population, introducing random fluctuations in the dynamics of variant frequencies (Charlesworth, 2009). This stochastic force is commonly evaluated using the effective population size N e, defined as the size of an idealized population (i.e. a panmictic population of constant size with discrete generations), which would show the same degree of randomness in the evolution of variant frequencies as the observed population (Kimura and Crow, 1963; Wright, 1931). In plant viruses, the intensity of genetic drift is modulated by bottlenecks occurring at multiple steps of virus infection (French and Stenger, 2003; Gutierrez et al., 2012, 2010 ; Sacristan et al., 2003; Zwart and Elena, 2015). Modelling approaches have estimated that the evolutionary forces acting at the within‐plant scale, especially the mutational pathway involved in resistance breakdown and the fitness cost associated with the resistance‐breaking mutation(s), account for about 50% of the risk of resistance breakdown in the field (Fabre et al., 2009, 2012b, 2015). Experimental data have shown that these two factors are indeed good predictors of the risk of resistance breakdown (Fabre et al., 2012a; Harrison, 2002; Janzac et al., 2009). The remaining 50% depend on factors related to virus epidemiology and thus mostly to the third step of resistance breakdown. One way to avoid or delay the breakdown of monogenic qualitative resistance is to combine the resistance gene with a suitable genetic background. Combined with a partially resistant genetic background, a major resistance gene can show a significant increase in durability, as demonstrated experimentally for resistances targeting an RNA virus (Palloix et al., 2009), a fungus (Brun et al., 2010) and a nematode (Fournet et al., 2013). Indeed, the host genetic background can affect the level of resistance to pathogens and the intensity of different evolutionary forces undergone by pathogen populations (Lannou, 2012). In the case of the Potato virus Y (PVY, genus Potyvirus; family Potyviridae)–pepper (Capsicum annuum; family Solanaceae) pathosystem, Quenouille et al. (2013, 2015 ) showed a significant correlation between the breakdown frequency of a major resistance gene [the pvr23 gene, encoding a eukaryotic translation initiation factor 4E (eIF4E)] and the capacity of the virus to accumulate in the plant, i.e. the additional resistance level conferred by the plant genetic background. Assuming an identical virus mutation rate between plant genotypes, they hypothesized that within‐plant virus accumulation was linked to the total number of virus replications during plant infection and, consequently, to the probability of appearance of the resistance‐breaking mutations. Using a progeny of pepper genotypes carrying the same major resistance gene, but contrasting genetic backgrounds, Quenouille et al. (2014) mapped the quantitative trait loci (QTLs) controlling either within‐plant virus accumulation or the frequency of breakdown of the major resistance gene in the pepper genome. The two QTLs controlling virus accumulation co‐localized with some of the QTLs controlling the frequency of breakdown of the major resistance gene, which provided a genetic explanation for the observed correlation between the two traits. Further, by comparing two pepper genotypes carrying the same major resistance gene, associated with either a partially resistant or a susceptible genetic background, Quenouille et al. (2013) showed that the selection of the most adapted resistance‐breaking PVY mutants was slower and/or rarer in plants with a partially resistant genetic background than in those with a susceptible genetic background. This slower and/or rarer selection may result from: (i) the smaller selection coefficient of the adapted mutants; and/or (ii) a higher genetic drift in the plants with a partially resistant genetic background (Charlesworth, 2009; Feder et al., 2016; Quenouille et al., 2013; Rouzine et al., 2001). The aim of this study was to disentangle the role and relative importance of the three factors, virus accumulation, selection coefficient between virus variants and virus effective population size, on the breakdown of the major resistance gene in order to foster the breeding of plant cultivars with durable virus resistance.