Sexual intraspecific recombination but not de novo origin governs the genesis of new apomictic genotypes in Potentilla puberula (Rosaceae).

Apomixis - asexual reproduction via seeds - might arise de novo following polyploidisation events, or via reproductive transfer of apomixis. Both processes can be obtained within species or via hybridisation. We aimed to determine the origin of apomictic genotypes in Potentilla puberula, a rosaceous species showing reproductive differentiation with ploidy: sexual tetraploids and apomictic penta- to octoploids, which regularly co-occur in sympatry. The study is based on 726 individuals, comprising all cytotypes, collected from 138 populations in the Eastern European Alps. We established relationships of cytotypes based on AFLP fingerprinting and cpDNA sequencing to test (1) whether the apomicts are of recurrent allopolyploid origin or originated from within the species via autopolyploidy, and (2) whether there are indications for reproductive transfer versus de novo origin of apomixis. Three principal pathways were identified which explain the origin of new apomictic genotypes, all involving at least one apomictic parent and thus compatible with the idea of reproductive transfer of the apomictic trait to the progeny: (1) self-fertilisation of unreduced egg cells in apomicts; (2) cross-fertilisation among apomicts; and (3) occasionally, heteroploid crosses among sexuals and apomicts. Autopolyploids derived from tetraploid sexuals were repeatedly observed, but did not express apomixis. Finally, our results suggest no role of other species in the origin of extant apomictic genotypes of P. puberula, although local hybrids with P. crantzii were identified. In conclusion, our results show that the formation of new apomictic genotypes required a genetic contribution from at least one apomictic parent. This finding is in accordance with the idea that apomixis is inheritable in P. puberula. On the contrary, lack of apomixis in penta- and hexaploids derived from sexual backgrounds did not support the hypothesis of a de novo origin of apomixis. Relatively high frequency of remnant sexuality in the apomicts involving different cytological pathways of seed formation can explain their high cytological and genotypic diversity. Finally, lack of global introgression from a third taxon is in support of P. puberula as a concise, although highly diverse, species.