Untying Gordian knots: The evolution and biogeography of the large European apomictic polyploid Ranunculus auricomus plant complex
by Kevin Karbstein née Glass
Date of Examination:2021-10-08
Date of issue:2021-10-19
Advisor:Prof. Dr. Elvira Hörandl
Referee:Prof. Dr. Elvira Hörandl
Referee:PD Dr. Sven Bradler
Files in this item
Name:Dissertation_Kevin_Karbstein_2021_final_fina...pdf
Size:64.8Mb
Format:PDF
Abstract
English
Polyploidy, the presence of two or more full genomic complements, repeatedly occurs across the tree of life. In plants, not only the economic but particularly the evolutionary importance is overwhelming. Polyploidization events, probably connected to key innovations (e.g., vessel elements or the carpel), occurred frequently in the evolutionary history of flowering plants, which are the most species-rich group in the plant kingdom (ca. 370,000 species) and contain 30–70% neopolyploids. Polyploidy and hybridization (i.e., allopolyploidy) are particularly considered to create biotypes with novel genomic compositions and to be key factors for subsequent speciation and macroevolution. In plants, both processes are frequently connected to apomixis, i.e., the reproduction via asexually-formed seeds. However, the enigmatic phenomenon of plant speciation accompanied by polyploidy and apomixis is still poorly understood despite tremendous progress in the field of genomics. The question of “What is a species?” is of highest priority for evolutionary biologists: Species are the fundamental units for biodiversity, and further evolutionary and ecological research relies on well-defined entities. Evolutionarily young plant species complexes offer a unique opportunity to study plant speciation and accompanying processes. They usually comprise a few sexual progenitor species, and numerous polyploid, partly apomictic, hybrid derivatives. In apomictic lineages, the lack of recombination and cross-fertilization can result in numerous clonal lineages with fixed morphological and ecological traits (agamospecies). Nevertheless, even recognizing and delimiting the sexual progenitors of species complexes is methodically challenging due to low genetic divergence, possible hybrid origins, ongoing gene flow, and/or incomplete lineage sorting (ILS). Integrative approaches using both genomic and morphometric data for disentangling the young progenitors are still lacking so far. The biogeography and evolution of those plant complexes is even more challenging. Apomicts frequently occupy larger areas or more northern regions compared to their sexual relatives, a phenomenon called geographical parthenogenesis (GP). GP patterns usually have a Pleistocene context because climatic range shifts in temperate to boreal zones offered frequent opportunities for interspecific hybridization, probably giving rise to apomixis in the Northern Hemisphere. Factors shaping GP patterns are still controversially discussed. GP has been widely attributed to advantages of apomicts caused by polyploidy and uniparental reproduction, i.e., fixed levels of high heterozygosity leading to increased stress tolerance, and self-fertility leading to better colonizing capabilities. On the one hand, complex interactions of genome-wide heterozygosity, ploidy, reproduction mode (sexual versus asexual), and climatic environmental factors shaping GP have not been studied enough. On the other hand, potential disadvantages of sexual progenitors due to their breeding system on fitness and genetic diversity have received even less attention. Finally, alongside biogeography, the reticulate relationships and genome composition and evolution of young, large polyploid plant species complexes have not yet been deciphered comprehensively. Besides challenges attributed to numerous numbers of polyploidization and hybridization events, bioinformatic analyses are also often hampered by missing information on progenitors, ploidy levels, and reproduction modes. The European apomictic polyploid Ranunculus auricomus (goldilock buttercup) plant complex is well-suited to study all the aforementioned issues. The majority of goldilock buttercups probably arose from hybridization of a few sexual progenitors, leading to more than 800 described, morphologically highly diverse agamospecies. Sexuals are estimated to have speciated less than 1.0 million years ago, and agamospecies are probably much younger. In this thesis, using R. auricomus as a model system, I examined the recalcitrant and hitherto poorly understood phylogenetic, genomic, and biogeographical relationships of young polyploid apomictic plant complexes. I developed a comprehensive theoretical and bioinformatic workflow, starting with analyzing the evolution of the sexual progenitor species, continuing with unraveling reproduction modes and biogeography of apomictic polyploids, and ending up with revealing the reticulate origins and genome composition and evolution of the polyploid complex. Spanning up to 251 populations and 87 R. auricomus taxa Europe-wide, this work gathered data of 97,312 genomic loci (RADseq), 663 nuclear genes (target enrichment), and 71 plastid regions, and 1,474 leaf ploidy, 4,669 reproductive seed, 284 reproductive crossing (seed sets), as well as 1,593 geometric morphometric measurements. First of all, phylogenomics based on RADseq, nuclear gene, and geometric morphometric data supported the lumping of the twelve described sexual morphospecies into five newly circumscribed progenitor species. These species represent clearly distinguishable genetic main lineages or clusters, which are both well geographically isolated and morphologically differentiated: R. cassubicifolius s.l., R. envalirensis s.l., R. flabellifolius, R. marsicus, and R. notabilis s.l. Mainly within-clade reticulate relationships, missing geographical isolation, and a lack of distinctive morphological characters led to this taxonomic treatment. Interestingly, allopatric speciation events took place ca. 0.83–0.58 million years ago during a period of severe climatic oscillations, and were probably triggered by vicariance processes of a widespread European forest-understory ancestor. Sexual species re-circumscriptions were additionally supported by population crossing experiments. Besides inbreeding depression, outbreeding benefits, and sudden self-compatibility, crossings also revealed a lack of reproductive barriers among some of the formerly described morphospecies. Moreover, flow cytometric ploidy and reproductive, RADseq, and environmental data were combined into a genetically informed path analysis based on Generalized Linear Mixed Models (GLMMs). The analysis unveiled a complex European GP scenario, whereby diploids compared to polyploids showed significantly higher sexuality (percent of sexual seeds), more petals (petaloid nectary leaves), and up to three times less genome-wide heterozygosity. Surprisingly, sexuality was positively associated with solar radiation and isothermality, and heterozygosity was positively related to temperature seasonality. Results fit the southern distribution of diploid sexuals and suggest a higher resistance of polyploid apomicts to more extreme climatic conditions. Finally, a self-developed, multidisciplinary workflow incorporating all previously gathered data demonstrated, for the first time, the predominantly allopolyploid origin, genome composition, and post-origin genome evolution of the R. auricomus complex. Taxa were organized in only three to five supported, north-south distributed clades or cluster, each usually containing diploid sexual progenitor species. Allopolyploidizations involved two to three different diploid sexual subgenomes per event. Only one autotetraploid event was detected. Allotetraploids were characterized by subgenome dominance and enormous post-origin evolution, i.e., Mendelian segregation of hybrid generations, back-crossing to parents, and/or gene flow due to facultative sexuality of apomicts. Four diploid sexual progenitors and a previously unknown, nowadays extinct progenitor, probably gave rise to the more than 800 taxa of the European R. auricomus complex. Analyses also showed that the majority of analyzed polyploid agamospecies are non-monophyletic and similar morphotypes probably originated multiple times. The lack of monophyly suggests a comprehensive taxonomic revision of the entire complex. In the General Discussion, I combine my thesis results with existing plant studies on diploid sexual and polyploid apomictic phylogenetics, biogeography, and composition and genome evolution of young species complexes. I explain the taxonomic conclusions and how species complexes link micro- and macroevolutionary processes. Finally, I give conclusions of my thesis and an outlook of the project and the field of polyploid phylogenetics.
Keywords: Phylogenomics; Rad-Seq; Hyb-Seq; Ranunculus auricomus; Polyploidy; Apomixis; Speciation; Plants; Geometric Morphometrics; Geographical Parthenogenesis; Genomics; Species Concepts; Nuclear genes; Plastid regions