Our goal is to evaluate the adaptive abilities of V. inaequalis to different resistance factors (qualitative vs. quantitative, single vs. in combination) and to understand the role and the determinisms of agressiveness (qualitative component of pathogenicity) in adaptation.
1.1 Genetic determinism of V. inaequalis host specificity on firethorn and apple (FunAdpat 2016-2018, funded by RFI)
Molecular determinants of co-evolution between fungal pathogens and their host plants have been extensively studied. However, adaptation of pathogens to new host plants is poorly understood and no clear molecular mechanism has been reported. We aimed at identifying and characterizing genes that determine fungal adaptation to different host plants using the biotrophic V. inaequalis fungus as a model. Within V. inaequalis, we described two formae speciales: V. inaequalis f. sp. pomi (POMI) that can infect apple (Malus sp.) but not firethorn (Pyracantha sp) and V. inaequalis f. sp. pyracantha (PYR), that can infect firethorn but not apple (Le Cam et al., 2002). Genome analysis allowed to easily distinguish these two formae specialesand it is very likely that V. inaequalis f.sp. pyracantha arose from a host jump. We combined comparative genomics, transcriptomics, genetics and pathogenicity tests to investigate host specificity of V. inaequalis on firethorn and apple. This approach allowed to distinguish the two formae speciales either in terms of presence-absence polymorphism or highly divergent coding sequences. Surprisingly, out of 57 non-redundant candidate genes, 55 were co-located on the same chromosome. The progeny from a cross between the two formae speciales exhibited a gradient of virulence on firethorn and unexpectedly, none of them could parasitize apple. This finding contrasts with commonly accepted views, and shows that hybridization can be detrimental to fungal pathogens.
1.2 Impact of apple domestication and its globalization on genetic architecture of the life history traits and pathogenicity of V. inaequalis (GANDALF funded by ANR; ESCAPE, 2016-2019, funded by RFI and AVANCE, 2018-2020, funded by SPE-INRAE)
Domesticated apple (Malus domestica) and V. inaequalis, are both descendants of their respective wild ancestors in Central Asia. The fungus was present in Central Asia before domestication of apple (i.e 7,000 years ago) on M. sieversii, the wild ancestor of apple trees. Although the role of domestication in shaping genetic diversity of plants is well described, its evolutionary impact on their associated pathogens remains unclear. Using population genomics, we showed that populations of V. inaequalis sampled in wild forests of Kazakhstan (wild strains) diverged ca. 4,500 years ago without gene flow from those sampled in anthropomorphic areas on both wild and domestic apple trees (agricultural strains), before entering into secondary contact about a century ago, probably when domestic apple trees were introduced in Central Asia. Agricultural strains have higher sporulation rate and bigger asexual spores than wild strains; they are able to cause disease on both M. domestica and M. sieversii, whereas wild strains are able to cause disease only on M. domestica; agricultural strains are also more aggressive than wild strains on M. sieversii (SA-E07-5). All these phenotypic differences can be considered as the result of a pestification process that reflects the mirroring process of pathogen adaptation in response to host domestication (Saleh et al.,2014) Indeed, plant and animal pathogens were not actually domesticated by humans, but they are supposed to have experienced adatative changes in response to the domestication of their host. For instance, pathogens may acquire more virulence genes as their host introgress resistance genes This pestification may explain that agricultural strains are invading wild forests in the TianShan mountains and represents a serious threat for M. sieversii as they can outcompete wild forms of the same pathogen. We identified a Small Secreted Protein (SSP) (V_081690_319 locus) as a potential avirulence gene that could partially explained that only agricultural strains are able to infect M. domestica (Figure 1).
The secondary contact between agricultural and wild strains in Kazakhstan results in the production of hybrid strains. Genomic analyses of these hybrids and comparison with hybrids produced in vitro from a cross between a wild and an agricultural strain showed that the natural hybrids are only F1, meaning a complete reproductive isolation due to hybrid sterility. We also detected single and pairwise segregation distortions in the progenies obtained in vitro, potentially linked to barriers against gene flow between agricultural and wild strains. One 15 Kb genomic region is impermeable to gene flow and included a candidate gene (velvet Vos A) that could be involved in a Dobzhansky-Muller incompatibility.
During a six-month visit in R. Nielsen’s lab in University of Berkeley (California, USA), C. Lemaire analyzed all the genomic footprints of pestification in the agricultural genetic lineage, which led to the first complete description of a pathogen pestification at the genomic level. The agricultural genetic lineage is expected to have experienced many successive adaptations during host domestication and agriculture development and let several genomics signatures of selective sweeps. Using haplotype-based statistics we found 5 genomic regions of selective sweeps in the agricultural lineage and none in the wild genetic lineage (Figure 2). Interestingly, when considering local recombination rates and polymorphism in these swepped regions, we found that all of them were probably introgressed from other wild populations (infecting M. orientalis in Caucasia and M. sylvestris in Europe) rather than acquired by spontaneous advantageous mutation. At last, comparing local recombination rates at SSP vs other genes, we showed a significantly higher recombination rate in SSP vs. to other genes in agricultural lineage, whereas there was no significant difference between recombination rates in wild lineage. All these results allowed us to claim that introgressions from other populations were essential in the pestification process in V. inaequalis.
Figure 1: The proportion of Malus domestica ancestry in the apple tree (y axis) plotted against the proportion of agricultural-type ancestry in Venturia inaequalis strains (x axis). For each of the 189 host-pathogen pairs for which the genotype at the V_081690_319 locus was available, the V. inaequalis strains carrying the A allele were colored in grey and the ones carrying the G allele in red. Red allele (G) is fixed in strains infecting M. domestica (upper right corner). This locus partially explained that only agricultural strains are able to infect M. domestica.
Figure 2: Detection of selective sweeps in the kazakh agricultural and wild lineages of V. inaequalis using the haplotype-based normalized XPEHH statistics. The blue line indicates null values of XPEHH. Dashed lines defines thresholds above which (for positive values) and below which (for negative values) the XPEHH statistics is considered as significantly different from zero, that is where a selective sweep signature likely occurs. Positive significant values indicate the occurence of a selective sweep in the agricultural lineage of V. inaequalis in Kazakhstan, as negative values indicates the occurence of a selective sweep in the wild lineage. This figure shows that selectve sweeps are only detected in the agricultural lineage, likely accounting for a pestification syndrom in this population.
A video is available at : https://www.youtube.com/watch?v=2wocn2TB6dg&ab_channel=INRAE
1.3 Impact of apple breeding for scab resistance genes and use of plant resistance inducers on V. inaequalis populations
1.3.1 Apple major resistance genes
Apple production is one of the cropping systems that requires the most pesticides. Up to 40 treatments are needed each year just to combat the devastating scab. For decades, apple breeders around the world have been exploiting genetic resources of numerous wild Malus spp to identify novel resistance (R) genes. To date, the Rvi6 gene introgressed from the crabapple M. floribunba, is the sole R gene that has been deployed with commercial success in apple orchards. However, emergence of virulent populations responsible for resistance breakdowns occurred in orchards since 2000’s. The rapid emergence of new virulent isolates is often hypothesized to arise from de novo mutations within avirulent populations already present in the agrosystem, as a response to the introduction of new resistant hosts. However, conclusions of our previous studies on the breakdown of the apple Rvi6 resistance gene, which is the most used resistance gene in worldwide apple breeding programs, do not match this scenario (Gladieux et al., 2011; Leroy et al., 2014). During this mandate, analysis of the genetic variation in the population virulent on Rvi6 cultivars at the European scale revealed that this virulent population emerged from standing genetic variation existing on the progenitor M. floribunda outside of orchards ca 24,000 years ago. Interestingly, we) showed that subsequent secondary contact between divergent non-agricultural and agricultural pathogen populations led to important evolutionary and epidemiological changes in pathogens (SA07-4). Indeed, hybrids exhibited a higher aggressiveness variance and we observed an invasion of the virulent trait in orchards. By means of genetics, genomics and functional approaches, we then identified the corresponding fungal AvrRvi6 gene, the first avirulence gene to be cloned to date in V. inaequalis (Figure 3). Taking advantage of a global collection of V. inaequalis, we revealed the existence of virulent alleles on Malus floribunda, M. sylvestris, M. orientalis and M. sieversii that are endemic species in Europe, Caucasia and Central Asia respectively. We also confirmed the lack of de novo resistance breakdown in European orchards suggesting that AvrRiv6 could be essential for the fungus fitness. Interestingly, analysis of virulent strains overcoming Rvi6 cultivars and wild Malus sp. in the USA provided by Papp et al (2020) revealed that their virulent alleles are the same as those identified in Europe (not published). This suggests that the Rvi6 gene could be much more durable than previously expected in countries and continents without wild Malus species. This novel scenario of virulence emergence in agriculture reveals that wild apple species used in breeding resistance programs are reservoirs of virulence. This is problematic because it is expected to happen faster than de novo mutations. This finding will challenge plant resistance breeding methods and management strategies.
Figure 3: Demonstration that the candidate gene g11287 is the AvrRvi6 gene using the Agrobacterium tumefaciens transient transformation assay (ATTA) on Nicotiana benthamiana. A hypersensitive response (HR) was induced in the tobacco when the Rvi6 resistance gene was co-expressed with the candidate gene g11287. The co-expression of the Cladosporium fulvum avirulence gene CfAvr4 and the corresponding resistance gene Cf4 was used as a positive control. Picture was taken at five days post inoculation.
A video is available at : https://www.youtube.com/watch?v=rXXBqgm-lLw&ab_channel=INRAE
1.3.2 Apple Resistance QTLs (ARAMIS 2013-2015, funded by Metaprogramm SMACH INRAE and FUNMAGAZINE 2015-2017, funded by RFI)
Quantitative resistances based on QTLs (Quantitative Trait Locus) are generally supposed to be more sustainable than R genes. Thanks to experiments in both controlled and field conditions over five to eight years, we showed, in collaboration with RESPOM, a decrease in the efficiency of quantitative resistance over time in orchards in relation to the selection of more aggressive strains of V. inaequalis. We then performed a cross between two strains of V. inaequalis, with low and high aggressiveness towards these QTLs, respectively to identify genetic determinants of QTL overcomings. Genotyping 137 of these progenies with a 70K SNP microarray and phenotyping disease on apple trees in controlled conditions allowed us to identify one candidate gene involved in the aggressiveness towards one of the QTLs. This gene encodes for a small secreted protein. Unexpectedly, this is a small secreted Protein that is also an AvrRvi5 gene candidate identified in the group of J. Bowen (Pers. Comm J. Bowen, PFR NZ). Collaboration is underway to conclude on the role of this SSP.
1.3.3 Durability of plant resistance inducers (PRI) (TAVINNOV 2018-2021, funded by Metaprogramm SMACH INRAE)
Considering the risk of adaptation of V. inaequalis to plant resistance triggered by PRI (Plant resistance inducers) in orchards we started an interdisciplinary project with agronomists at INRAE Avignon (PSH team) and with RESPOM team to evaluate durability of PRI efficacy. We showed that natural variation sensitivity of strains to the PRI tested (Bion) pre-exists in orchards indicating a risk to select the most adapted strains over time leading to a loss of PRI efficacy.
