Phylogeny and comparative chloroplast genomics of the Campanulaceae

Abstract

The Campanulaceae is a large, nearly cosmopolitan family, chiefly of temperate regions, with a huge range of variation in morphology, breeding systems, ecology, pollen and seed morphology. The monophyly of the family is well-established, but intrafamiliar and generic circumscription remains controversial. I reconstructed a molecular phylogeny of Campanulaceae using the chloroplast genes atpB, matK and rbcL, sampling a wide diversity of the family, and used the phylogeny to clarify major family relationships and address a number of other problems related to taxonomy and biogeography of the family. Phylogenetic analyses showed strong evidence that there are two major clades in the family, the platycodonoids, and a larger clade comprised of the wahlenbergioids and campanuloids. I also found that a number of historical genera are not monophyletic. In addition, phylogenetic evidence suggests that some oceanic island taxa arose from single introductions, and others from multiple introductions. The phylogenetic evidence also suggests that the North American Campanulaceae are non-monophyletic. The Campanulaceae is also known for having highly rearranged chloroplast genomes, based on earlier gene mapping and restriction site studies. Chloroplast genome organization, gene order and content are highly conserved among most land plants. I sequenced the complete chloroplast genome sequences of two members of the Campanulaceae, Trachelium caeruleum L. and Platycodon grandiflorus (Jacq.) A.DC., and Pseudonemacladus oppositifolius (BL Rob.) McVaugh, a member of the closely related Nemacladaceae. These genomes are far more rearranged than originally suggested, with multiple inversions, gene duplications and losses, and possible transpositions. Recombination between repeats or tRNAs has been suggested as two means of chloroplast genome rearrangements. In a comparison with other angiosperms, the Trachelium, Platycodon, and Pseudonemacladus chloroplast genomes have both the highest number and largest repeated intergenic sequences, and these are concentrated near rearrangement endpoints. Genes for tRNAs occur at many but not all inversion endpoints, so it is possible that some combination of repeats and tRNAs may have mediated recombination leading to these rearrangements. This does not preclude the possibility that some other mechanisms of structural evolution may also have participated in the extraordinary chloroplast genome evolution in these taxa.