The first steps of plant mating system evolution

Abstract

Plants are strikingly complex and diverse organisms. Evolutionary transitions in plant mating systems are frequent, particularly towards selfing. Mating systems are important because they correspond to the mating behavior of a population or species, and they may affect ecological interactions with antagonists and mutualists or define the ability to invade new environments. Therefore, plants are ideal organisms to study ecology and evolution.Empirical studies and theoretical models have tried to explain the distribution, maintenance and evolution of plant mating systems. Traditionally, inbreeding depression is regarded as the main barrier to the evolution of selfing, and its magnitude may be influenced by environ- mental conditions. Moreover, mating systems are also widely associated with morphological characteristics, for instance with a selfing syndrome in flowers. Also, ecological partners that establish interactions with plants may influence plant performance and contribute to selection on traits related to reproduction and mating.However, several gaps are still present in our current knowledge on plant mating systems. For example, the traditional inbreeding depression threshold of 50 % that is considered to act as a barrier to the evolution of selfing has not been reasonably tested. Similarly, the effect of environmental stress on inbreeding depression is still in debate. Few studies report lifetime estimates of inbreeding depression for perennial self-incompatible species. Studies tested investigated the selfing syndrome by usually accounting for few floral traits. Moreover, few studies produced empirical data on pollinator-mediated importance on floral trait selection and mating system variation, especially in model systems which have recently evolved into selfing. Furthermore, few studies have simultaneously addressed the effect of antagonistic and mutualistic interactions on plant performance.To contribute to the knowledge on inbreeding depression and plant mating system evolution, I performed a greenhouse experiment in which I estimated inbreeding depression on lifecycle traits of six outcrossing and five selfing Arabidopsis lyrata North American populations. I found that cumulative lifetime inbreeding depression was ∼ 30 % and it did not differ between mating systems. I found no evidence for purging. To test whether inbreeding depression could be magnified by environmental stress, I assessed inbreeding depression in plants exposed to drought stress and with induced defenses. Treatments reduced overall performance, but not inbreeding depression. My results suggest that inbreeding depression in A. lyrata is overall low, independent from mating system variation. My results therefore suggest that the lack of a strong barrier ( i.e. low inbreeding depression) may have facilitated the evolution of selfing in North American A. lyrata . Finally, the fact that stress did not magnify inbreeding depression in my experiment suggests that low inbreeding depression was not an artifact that can be attributed to beneficial greenhouse conditions.To test whether the evolution of selfing leads to a flower selfing syndrome, I performed a greenhouse experiment with six selfing and six outcrossing populations of the North American A. lyrata. I measured over ten floral traits with conventional and geometric morphometric methods to assess which floral traits are possibly under first selection after the evolution of selfing. Selfers had 9.2 % smaller corollas, 8.4 % longer pistils and 21.5 % lower P/O ratios than outcrossers, but there were no differences in shape, floral integration and herkogamy between mating systems. Most variation in floral traits was actually explained by population genetic background rather than by mating system. My results suggest that the evolution of selfing has lead to reductions in corolla size and lower P/O ratios in North American A. lyrata , but has not lead to changes in shape.To investigate the importance of pollination on mating system and floral trait evolution, I conducted a common-garden experiment using five selfing and six outcrossing populations of the North American A. lyrata . I first estimated floral traits and then performed pollinator observations on potted plants to test whether mating system or floral trait variation explained pollinator visitation. Mating system did not explain visitation in A. lyrata , but population genetic clustering and daytime did. My results suggest that after the recent evolution of selfing, pollinators may still visit selfing and outcrossing populations in similar frequencies. My results do not support that pollinators select flower size or shape in A. lyrata .To assess whether ecological interactions are important selective drivers that shape plant reproductive traits, I performed a field-experiment using the Gentiana - Phengaris - Myrmica system. To assess herbivory, I analyzed Phengaris alcon oviposition on Gentiana asclepiadea in relation to vegetation characteristics, and plant- and ant-host traits. To assess pollination, I made pollinator observations and tested the effect of flower size on pollinator visitation. Finally, to determine whether antagonistic selection by herbivores and pollinators affected reproductive traits, I estimated fruit size and seed set of plants in “open pollination” and “pollen supplementation” treatments. Phengaris alcon oviposition was explained by the surrounding vegetation around plant-hosts, but also by Myrmica and Gentiana vegetative traits. However, my results do not support that P. alcon imposes a strong selection on floral traits. The putative pollinator groups bumblebees and syprhid flies, on the other hand, imposed directional selection towards wider and shorter corollas. Overall there was a signal of stabilizing selection for intermediate flower sizes, which may be the result of a balance between the selection imposed by pollinators I found in this study, and opposing selection by generalist herbivores (not assessed here).In summary, I conclude that inbreeding depression does not always provide a strong barrier to the evolution of selfing in plants. Moreover, the early evolution of selfing may lead to subtle changes in flower morphology, which suggests that selection for a selfing syndrome at least partly evolves from standing genetic variation. Pollinator-mediated selection on floral traits may drive the evolution of floral morphology. The result of direct and indirect ecological interactions between plants with their herbivores and pollinators should influence plant performance and ultimately mating system evolution.