SELFADAPT · Genomic of adaptation in selfing polyploid species: the case of Capsella bursa-pastoris

Genomic of adaptation in selfing polyploid species: the case of Capsella bursa-pastoris

Many flowering plants possess both male and female functions on a same individual. Among them, self-fertilization, the ability to fertilize an ovule with its own pollen, is rather widespread and it evolved many times from outcrossing ancestors. Self-fertilization can be selected for because it secures reproduction when mates are limited and enhance gene transmission as two copies of an individual are passed on to the offspring through seeds (both male and female copies come from the same parent) under selfing versus only one copy under outcrossing (the maternal copy). However, on the long run, by reducing genetic mixing, selfing reduces genetic diversity, adaptive potential and allows the accumulation of slightly deleterious mutations. Thus, selfing is supposed to be an evolutionary dead-end, all the more so that reversion towards outcrossing is very unlikely. However, paradoxically, several current selfers seem to be successful in the short term and are associated with range expansion, some being considered as weeds and invaders. This could be due to their demographic advantage thanks to reproductive assurance but demographic expansion also requires adaptation to new local environmental conditions (e.g. phenology must match new climatic conditions). In addition, evolution of selfing is often associated with the evolution of polyploidy, which could boost adaptive potential of selfing species and contribute to explain their transitory success. How the short-term advantage and the long-term disadvantage of selfing interact and shape the fate of selfing is still largely unknown. How selfing and polyploidy affects adaptation to new environmental conditions is thus a central issue to tackle this paradox. The central objective of the project was to understand how polyploid selfing species adapt to their environment and at which scale, using Capsella bursa-pastoris as a model species. C. bursa-pastoris is a tetraploid annual selfing weed with a worldwide distribution (except in the tropics) that combines all the typical characteristics of a recently successful tetraploid selfing species with paradoxically low genetic diversity and evidence of accumulation of weakly deleterious mutations. More specifically, we wanted to address the two main questions: - What is the genetic basis of adaptation in a polyploid selfing species? Does adaptation proceed from new mutations or standing variation, from change in protein or in gene expression, and what is the role of gene duplicates? - Do adaptation patterns vary with geographic scales? Especially, is adaptation more efficient locally than globally and do the genetic basis depend on geographic scale We used a combination of approaches to study patterns of adaptation at the genomic, gene expression and phenotypic levels. First, used a polymorphism dataset to establish the population genetic structure and the demographic history of the species. Second, we used full genome-sequencing data to detect selection at the genomic levels across the species range. Third, we used transcriptome-sequencing data to assess how patterns of expression vary across the species range and whether it can be involved in local adaptation. Four, we analysed the data of common garden experiment to test for local adaptation at the phenotypic level. Finally, we designed experiments to assess the competitive ability of the species across the species range and in comparison with the three other Capsella species. We found that in Eurasia C. bursa-pastoris clustered into three distinct genetic groups: a Middle East cluster, one cluster grouping samples from Western Europe and Southeastern Siberia, and a third one centered on Eastern Asia. Population genetics analysis also revealed that C. bursa-pastoris underwent a typical colonization history likely starting from the Middle East towards Europe and followed by successive human-mediated expansions into Eastern Asia, along with several genetic bottlenecks reducing genetic diversity across the colonization front from West to East. Based on expression levels, we found the same clustering patterns and that most differences observed could be explained by genetic drift due to the demographic history, suggesting low levels of local adaptation at the gene expression level. Similarly, analysis at the phenotypic level shows little evidence of local adaptation. Interestingly, the eastern most populations performed the worst in any environments (especially Chinese accessions performed the worst even when grown in China). Finally, genomic scan approaches revealed less signature of positive selection in the most eastern regions. Overall these results suggest that, contrary to our initial hypothesis, range expansion of C. bursa-pastoris has not come along with efficient local adaptation but instead genomic and phenotypic deterioration, which is in agreement with recent models predicted accumulation of deleterious mutations during range expansion (called the expansion load). Based on these results and on previous theoretical predicting that deleterious mutations could only matter under competitive conditions, we hypothesized that the ecological success of C. bursa-pastoris could be partly due to the fact that it mainly lives in non-competitive conditions. We also hypothesized that, compared to diploid selfers, polyploidy could also contribute to buffering the negative effects of selfing. The results of the competition experiments we carried out supported our predictions. The competitive ability of C. bursa-pastoris is intermediate between those of the diploid outcrosser (C. grandiflora) and the diploid selfers (C. rubella and C. orientalis) but decreases from West to East, following the colonization front. This suggests that C. bursa-pastoris benefits from the reproductive assurance of selfing and from the buffering effect of polyploidy, which has allowed its exceptional expansion. The other side of the coin is that colonization lead to the loss of diversity, accumulation of deleterious mutations and decline in its competitive ability. On the long term this could explain how a short-term advantage can reverse into a long-term threat. The results of the project shed a new light on the evolution of selfing species and suggest that the interplay between ecological and genomic conditions could explain the fate of selfing species. This is especially relevant for the management of weed and crop species, among which selfing and polyploidy are frequent, and in the context of global changes induced by human activities that affect the ecological conditions that species encountered. Our project suggests new directions to tackle these challenges by better integrating ecological and genomic approaches.

Data: CORDIS, © European Union