Interploid reproductive isolation in the Campanula rotundifolia polyploid complex

Polyploidy, or whole-genome duplication, has long been considered an important factor in plant speciation due to the tendency of newly formed polyploids to be reproductively isolated from their diploid progenitors. Recent evidence has called into question both the incidence of polyploidy-induced reproductive isolation, particularly among higher ploidy levels, and the role of whole-genome duplication in the diversification of angiosperms. To better understand the role of polyploidy in plant divergence and speciation, I used the Campanula rotundifolia autopolyploid complex to compare reproductive isolation between diploids (2X) and tetraploids (4X), and between tetraploids and hexaploids (6X). I evaluated postzygotic isolation, prezygotic isolation due to pollinator preference and pollen choice, and interploid reproduction occurs in nature. Because the evolution and mating systems of the complex are poorly understood, I constructed chloroplast and nuclear phylogenies and performed a survey of self-compatibility. These demonstrated that both tetraploids and hexaploids had multiple origins, and that self-incompatibility was gradually lost with repeated genome duplication and as the complex expanded from Europe into North America. Reciprocal 2X-4X and 4X-6X crosses demonstrated that while diploids and tetraploids have high postzygotic isolation, tetraploids and hexaploids are capable of interbreeding. I then confirmed these results, and the fertility of newly formed tetraploids and pentaploids, in backcrosses. Artificial mixed-ploidy arrays exposed to natural pollinators demonstrated that pollinator-mediated reproductive isolation was modest in C. rotundifolia. These arrays further supported two different patterns of reproduction based on parental ploidy, and demonstrated that interploid reproduction is most likely for rare cytotypes. Lastly, surveys of natural mixed-ploidy contact zones found interploid reproduction in 4X-6X contact zones, as demonstrated by extensive presence of pentaploids and pentaploid-derived aneuploids, and by lack of genetic differentiation between tetraploid and hexaploid parents. By contrast, 2X-4X contact zones contained no triploids and no evidence of gene flow aside from one spontaneous neo-tetraploid. Taken together, these studies clearly demonstrate that reproductive isolation in polyploid complexes follows two patterns: strong (but not complete) isolation between diploids and polyploids, and weak or no isolation between polyploids. The strong reproductive isolation found between diploids and tetraploids supports the “instant speciation” hypothesis, but ongoing gene flow between tetraploids and hexaploids demonstrates that this concept is not universal. These findings suggest that interploid gene flow, particularly between different polyploid cytotypes, may constrain divergence, and may provide a causal mechanism for lower observed diversification of polyploids relative to diploids.