SUMMARYGenetic resistance to pathogens is important for sustainable maintenance of crop yields. Recent biotechnologies offer alternative approaches to generate resistant plants by compensating for the lack of natural resistance. Tomato (Solanum lycopersicum) and related species offer a model in which natural and TILLINGinduced potyvirus resistance alleles may be compared. For resistance based on translation initiation factor eIF4E1, we confirm that the natural allele Sh-eIF4E1 PI24 -pot1, isolated from the wild tomato species Solanum habrochaites, is associated with a wide spectrum of resistance to both potato virus Y and tobacco etch virus isolates. In contrast, a null allele of the same gene, isolated through a TILLING strategy in cultivated tomato S. lycopersicum, is associated with a much narrower resistance spectrum. Introgressing the null allele into S. habrochaites did not extend its resistance spectrum, indicating that the genetic background is not responsible for the broad resistance. Instead, the different types of eIF4E1 mutations affect the levels of eIF4E2 differently, suggesting that eIF4E2 is also involved in potyvirus resistance. Indeed, combining two null mutations affecting eIF4E1 and eIF4E2 re-establishes a wide resistance spectrum in cultivated tomato, but to the detriment of plant development. These results highlight redundancy effects within the eIF4E gene family, where regulation of expression alters susceptibility or resistance to potyviruses. For crop improvement, using loss-of-function alleles to generate resistance may be counter-productive if they narrow the resistance spectrum and limit growth. It may be more effective to use alleles encoding functional variants similar to those found in natural diversity.

Allele mining on susceptibility factors offers opportunities to find new sources of resistance among crop wild relatives for breeding purposes. As a proof of concept, we used available RNAseq data to investigate polymorphisms among the four tomato genes encoding translation initiation factors [eIF4E1 and eIF4E2, eIFiso4E and the related gene new cap-binding protein(nCBP)] to look for new potential resistance alleles to potyviruses. By analysing polymorphism among RNAseq data obtained for 20 tomato accessions, 10 belonging to the cultivated type Solanum lycopersicum and 10 belonging to the closest related wild species Solanum pimpinellifolium, we isolated one new eIF4E1 allele, in the S. pimpinellifolium LA0411 accession, which encodes a potential new resistance allele, mainly due to a polymorphism associated with an amino acid change within eIF4E1 region II. We confirmed that this new allele, pot12, is indeed associated with resistance to potato virus Y, although with a restricted resistance spectrum and a very low durability potential. This suggests that mutations occurring in eIF4E region II only may not be sufficient to provide efficient and durable resistance in plants. However, our study emphasizes the opportunity brought by RNAseq data to mine for new resistance alleles. Moreover, this approach could be extended to seek for putative new resistance alleles by screening for variant forms of susceptibility genes encoding plant host proteins known to interact with viral proteins.