Anne-Lise Boixel’s research while affiliated with University of Paris-Saclay and other places

Fungal plant pathogens pose significant threats to food security and are dominant components of global agroecosystems. Despite wide environmental distributions, many fungal populations are locally adapted to cultivated hosts, fungicide uses, and climatic conditions. Understanding the mechanisms through which pathogens overcome control measures and acclimate to new environments is essential for predicting the future impact of crop diseases. Our study focused on Zymoseptoria tritici, which causes the Septoria tritici blotch (STB) disease, representing one of the main constraints on wheat production worldwide. The control of STB is increasingly challenging due to Z. tritici's rapid evolution and adaptation to environmental conditions. We addressed this challenge by leveraging Illumina sequencing data obtained from 240 Z. tritici isolates sampled from eight Euro-Mediterranean countries, representing the diverse environmental heterogeneity of wheat-growing regions. Through Genotype-Environment Association (GEA) and Redundancy Analyses (RDA), we identified candidate genes linked to Z. tritici's adaptation to diverse climatic conditions. Furthermore, our study revealed discernible signatures of selection at both the genome-wide and candidate gene levels, further enriching our understanding of the pathogen's evolutionary trajectory. This extensive population genomic analyses underscores key genetic determinants involved in the adaptive potential of this important wheat pathogen.

Plant resistances impose strong selective pressure on plant pathogen populations through the deployment of resistance genes, which leads to the emergence of new virulences. The pathogen adaptation also involves other parasitic fitness traits, especially aggressiveness components. A previous study on Puccinia triticina, the causal agent of wheat leaf rust, revealed that the distribution frequency of virulences in the French pathogen population cannot be fully explained by the major resistance genes deployed in the landscape. From 2012 to 2015, two dominant pathotypes (distinguished by their combination of virulences) were equally frequent despite the theoretical advantage conferred to one pathotype (166 317 0) by its virulence to Lr3, frequent in the cultivated landscape, whereas the other (106 314 0) is avirulent to this gene. To explain this apparent contradiction, we assessed three components of aggressiveness — infection efficiency, latency period and sporulation capacity — for 23 isolates representative of the most frequent genotype within each pathotype (106 314 0-G2 and 166 317 0-G1, identified by their combination of microsatellite markers). We tested these isolates on seedlings of Michigan Amber, a ‘naive’ wheat cultivar that has never been grown in the landscape, Apache, a ’neutral‘ cultivar with no selection effect on the landscape-pathotype pattern, and several cultivars that were frequently grown. We found that 106 314 0-G2 was more aggressive than 166 317 0-G1, with a consistency for the three components of aggressiveness. Our results show that aggressiveness plays a significant role in driving evolution in pathogen populations by acting as a selective advantage, even offsetting the disadvantage of lacking virulence towards a major Lr gene. Higher aggressiveness represents a competitive advantage that is likely even more pronounced when exhibited at the landscape scale as the expression of its multiple components is amplified by the polycyclic nature of epidemics.

Cultivar resistance is a major asset for the management of crop diseases and can play an important role in agroecological transition. However, the wide deployment of a reduced number of resistance genes can lead to a rapid adaptation of pathogen populations and to a loss of resistance efficiency. The objective of this study was to characterize and discuss different trajectories of adult plant ratings for resistance to yellow rust in French wheat cultivars between 1963 and 2018. Among 719 cultivars assessed for at least 2 years, 590 cultivars showed no variation in their resistance scores, despite a mean of 4.3 years and up to 33 years of assessment. A set of descriptive variables was computed in order to compare the evolution of resistance score of 129 cultivars that experienced resistance variation. We applied a principal component analysis and a hierarchical clustering on principal components to this subdataset to constitute clusters corresponding to different cultivar profiles. Clusters C1 and C2 had small resistance variations (1–2 points on a 1–9 scale); Cluster C3 had long assessment durations and several small drops in resistance score and could be associated with quantitative resistance erosion; Cluster C4 included major drops in resistance score (4–5 points), often associated with known breakdowns of major resistance genes. Cases of limited drops in resistance score as a known resistance gene was broken down suggest the presence of efficient adult plant resistance. We discuss the use of information extracted from this dataset and methods to further explore sources of resistance to yellow rust present in French cultivars.

Deploying disease-resistant cultivars is one of the most effective control strategies to manage crop diseases such as wheat leaf rust, caused by Puccinia triticina. After harvest, this biotrophic fungal pathogen can survive on wheat volunteers present at landscape scale and constitute a local source of primary inoculum for the next cropping season. In this study, we characterised the diversity of P. triticina populations surveyed on wheat volunteer seedlings for six consecutive years (2007–2012) at the landscape scale. A total of 642 leaf rust samples classified in 52 virulence profiles (pathotypes) were collected within a fixed 5-km radius. The pathotype composition (identity and abundance) of field-collected populations was analyzed according to the distance between the surveyed wheat plots and to the cultivars of origin of isolates. Our study emphasised the high diversity of P. triticina populations on wheat volunteers at the landscape scale. We observed an impact of cultivar of origin on pathogen population composition. Levels of population diversity differed between cultivars and their deployment in the study area. Our results suggest that wheat volunteers could provide a significant though highly variable contribution to the composition of primary inoculum and subsequent initiation of leaf rust epidemics.

Plant resistances impose strong selective pressure on plant pathogen populations through the deployment of major resistance genes, which leads to the emergence of new virulences. The pathogen adaptation also involves other life-history or parasitic fitness traits, especially aggressiveness components. A previous study on Puccinia triticina, the causal agent of wheat leaf rust, revealed that the distribution frequency of virulences in the French pathogen population cannot be fully explained by the major resistance genes deployed in the landscape. From 2012 to 2015, two major pathotypes (groups of isolates with the same combinations of virulences)-166 317 0 and 106 314 0-were equally represented in the landscape, despite the theoretical advantage conferred to 166 317 0 by its virulence to Lr3 frequent in the cultivated landscape, whereas 106 314 0 is avirulent to this gene. To explain this apparent contradiction, we assessed three components of aggressiveness-infection efficiency, latency period and sporulation capacity-for 23 isolates representative of the most frequent pathotype-genotype combination (named 'pathogenotype') within each pathotype. We tested these isolates on seedlings of Michigan Amber, a 'naive' wheat cultivar that has never been grown in the landscape, Apache, a 'neutral' cultivar with no selection effect on the landscape-pathotype pattern, and several cultivars that were frequently grown. We found that pathogenotype 106 314 0-G2 was more aggressive than 166 317 0-G1, with a consistency for the three components of aggressiveness. Our results show that aggressiveness plays a significant role in driving evolution in pathogen populations by acting as a selective advantage, even offsetting the disadvantage of 2 lacking virulence towards a major Lr gene. Higher aggressiveness represents a competitive advantage that is likely even more pronounced when exhibited at the landscape scale as the expression of its multiple components is amplified by the polycyclic nature of epidemics.

Plant pathogens are constantly evolving and adapting to their environment, including their host. Vir-ulence alleles emerge, and then increase, and sometimes decrease in frequency within pathogen populations in response to the fluctuating selection pressures imposed by the deployment of resistance genes. In some cases, these strong selection pressures cannot fully explain the evolution observed in pathogen populations. A previous study on the French population of Puccinia triticina, the causal agent of wheat leaf rust, showed that two major pathotypes-groups of isolates with a particular combination of virulences-predominated but then declined over the 2005-2016 period. The relative dynamics and the domination of these two pathotypes-166 317 0 and 106 314 0-, relative to the other pathotypes present in the population at a low frequency although compatible, i.e. virulent on several varieties deployed, could not be explained solely by the frequency of Lr genes in the landscape. Within these two pathotypes, we identified two main genotypes that emerged in succession. We assessed three components of aggressiveness-infection efficiency, latency period and sporulation capacity-for 44 isolates representative of the four P. triticina pathotype-genotype combinations. We showed, for both pathotypes, that the more recent genotypes were more aggressive than the older ones. Our findings were highly consistent for the various components of aggressiveness for pathotype 166 317 0 grown on Michigan Amber-a 'naive' cultivar never grown in the landscape-or on Apache-a 'neutral' cultivar, which does not affect the pathotype frequency in the landscape and therefore was postulated to have no or minor selection effect on the population composition. For pathotype 106 314 0, the most recent genotype had a shorter latency period on several of the cultivars most frequently grown in the landscape, but not on 'neutral' and 'naive' cultivars. We conclude that the quantitative components of aggressiveness can be significant drivers of evolution in pathogen populations. A gain in aggressiveness stopped the decline in frequency of a pathotype, and subsequently allowed an increase in frequency of this pathotype in the pathogen population, providing evidence that adaptation to a changing varietal landscape not only affects virulence but can also lead to changes in aggressiveness.

Little is known about the impact of host immunity on sexual reproduction in fungal pathogens. In particular, it is unclear whether crossing requires both sexual partners to infect living plant tissues. We addressed this issue in a three-year experiment investigating different scenarios of Zymoseptoria tritici crosses according to the virulence (‘vir’) or avirulence (‘avr’) of the parents against a qualitative resistance gene. Co-inoculations (‘vir × vir’, ‘avr × vir’, ‘avr × avr’) and single inoculations were performed on a wheat cultivar carrying the Stb16q resistance gene (Cellule) and a susceptible cultivar (Apache), in the greenhouse. We assessed the intensity of asexual reproduction by scoring disease severity, and the intensity of sexual reproduction by counting the ascospores discharged from wheat residues. As expected, disease severity was more intense on Cellule for ‘vir × vir’ co-inoculations than for ‘avr × vir’ co-inoculations, with no disease for ‘avr × avr’. However, all types of co-inoculation yielded sexual offspring, whether or not the parental strains caused plant symptoms. Parenthood was confirmed by genotyping (SSR markers), and the occurrence of crosses between (co-)inoculated and exogenous strains (other strains from the experiment, or from far away) was determined. We showed that symptomatic asexual infection was not required for a strain to participate in sexual reproduction, and deduced from this result that avirulent strains could be maintained asymptomatically “on” or “in” leaf tissues of plants carrying the corresponding resistant gene for long enough to reproduce sexually. In two of the three years, the intensity of sexual reproduction did not differ between the three types of co-inoculation in Cellule, suggesting that crosses involving avirulent strains are not anecdotal. We discuss the possible mechanisms explaining the maintenance of avirulence in Z. tritici populations and the potential impact of particular resistance deployments such as cultivar mixtures for limiting resistance breakdown.

Plant pathogens are constantly evolving and adapting to their environment, including their host. Virulence alleles emerge, and then increase, and sometimes decrease in frequency within pathogen populations in response to the fluctuating selection pressures imposed by the deployment of resistance genes. In some cases, these strong selection pressures cannot fully explain the evolution observed in pathogen populations. A previous study on the French population of Puccinia triticina , the causal agent of wheat leaf rust, showed that two major pathotypes — groups of isolates with the same combinations of virulences — predominated but then declined over the 2005-2016 period. The relative dynamics of these two pathotypes — 166 317 0 and 106 314 0 — relative to the others present in the population could not be explained solely by the frequency of Lr genes in the landscape. Within these two pathotypes, we identified two main genotypes that emerged in succession. We assessed three components of aggressiveness — infection efficiency, latency period and sporulation capacity — for 44 isolates representative of the four P. triticina pathotype-genotype combinations. We showed, for both pathotypes, that the most recent genotypes were more aggressive than the older ones. Our findings were highly consistent for the various components of aggressiveness for pathotype 166 317 0 grown on Michigan Amber — a 'naive' cultivar never grown in the landscape — or on Apache — a 'neutral' cultivar, with no selection effect on the landscape-pathotype pattern. For pathotype 106 314 0, the most recent genotype was more aggressive on several of the cultivars most frequently grown in the landscape, but not on 'neutral' and 'naive' cultivars, and only in terms of its latency period. We conclude that the quantitative components of aggressiveness can be significant drivers of evolution in pathogen populations. A gain in aggressiveness allowed the maintenance of a declining pathotype, and even further expansion of that pathotype, in the pathogen population, providing evidence that virulence alone is not sufficient, aggressiveness also being required for the adaptation of a pathogen to a changing varietal landscape.