Exploring patterns of cytonuclear incompatibility in Pelargonium section Ciconium

In this thesis I studied cause and effect of Cyto Nuclear Incompatibility (CNI) induced speciation in plants. By combining comparative genomics, phylogenetics and bioinformatics with large scale crossing experiments I was able to infer part of the genetic machinery underlying CNI, determine plastid inheritance, study plastome evolution in detail and establish a possible mechanism for explaining chlorosis.After introducing (Chapter 1) the phenomenon of CNI, its consequences for plant evolution and speciation, and Pelargonium section Ciconium as a model system, I investigate patterns of CNI in a comprehensive series of P. sect. Ciconium (Geraniaceae) interspecific crosses with a single parental population of P. x hortorum in Chapter two. I deduce that one or more nuclear genomic alleles are involved in the expression and management of chloroplasts, and that these act throughout the different life stages. It appears to be the case that the number of alleles needed to explain the observed segregation patters increases with phylogenetic distance between the parent species. Finally, I established that biparental inheritance of chloroplasts is widespread and common across P. sect. Ciconium. All species have the potential to transmit their plastids either paternally or maternally.In Chapter three I use the variation on abundancies of the repetitive part of the genome (the non-coding DNA collectively known as the ‘repeatome’) as a proxy for the nuclear genotype in our crossing experiments (Chapters two and five). In addition, I use genomic repeat abundancy and repeat sequence similarity to reconstruct phylogenetic trees. Hereby the repeat clusters are characters, and their abundancies states. The interpretation of abundancy variation must be done with care, because autopolyploids can cause ‘false’ sister species relationships to occur in the reconstructed tree, based on their genomic sizes, not on shared history. Nevertheless, when I included sequence read variation in the analysis as well, it became clear that the ‘Core-Ciconium’ species share >20 large repeat clusters when compared to P. elongatum and that the species P. multibracteatum, P. quinquelobatum and P. yemenense sp. nov., occurring outside the Cape Floristic Region (CFR), again share several large synapomorphic repeats.In Chapter four, I use a near complete selection of species from P. sect. Ciconium to study plastome evolution in detail. I create a detailed picture of the variation that occurs in P. sect. Ciconium plastomes and present evidence that rrn genes (especially rrn23), encoding ribosomal RNA, and ribosomal proteins have undergone concerted evolution. Rrn23 displays more variation within section P. sect. Ciconium than it does between angiosperm orders. I speculate that one of the consequences of plastome evolution in P. sect. Ciconium (and presumably Pelargonium as well) is the presence of a, structurally, altered plastidial ribosome. In this Chapter I further describe that two of the three available plastid encoded polymerase (PEP) subunits (rpoB and rpoC1) are under positive selection in Ciconium suggesting strong involvement in plastome evolution.Chapter five draws, in part, on the results reported in the previous Chapters. I continued to determine inheritance of plastids, but report newly developed markers for the mitochondria as well. I confirmed the findings from Chapter two that plastids inherit via both parental lines, but biparental inheritance of mitochondria turned out to be rarer. In this Chapter I present three more interspecific crossing series and describe more chlorosis patterns and their associated plastid types. I investigate the occurrence of chlorosis from the chloroplast perspective and find that changes in physico-chemical properties of resulting rpoB peptides occurring in P. sect. Ciconium, by and large, correlate with the observed patterns of chlorosis in F1 interspecific offspring. There is a strong indication that the Ciconium-unique sequence insertions in rpo genes encode parts of the polymerase that are in direct contact with template DNA during transcription. If so, then rpo variation, or PEP structure variation, must be considered, alongside nuclear genomic changes reported in other studies, when understanding plastome evolution in Geraniaceae.In chapter six I discuss the results from the previous chapters in the context of the Bateson-Dobzhansky-Müller model of speciation. I propose a refinement of the model, in which different kinds of selective pressures act at different moments in time, for Pelargonium. I discuss that P. section Ciconium is, likely in an evolutionary phase where CNI is the driver for speciation