Integrating Molecular Phylogenies and Developmental Genetics: a Gesneriaceae Case Study

We have analysed and characterised the phylogenetic potential of a nuclear developmental gene, cycloidea (originally isolated from Antirrhinum), involved in the development of floral zygomorphy. We have compared the evolution of part of this putative single copy gene in Old World Gesneriaceae with two contrasting DNA sequence regions, using two sets of data (a ‘genus’ data set and a ‘species’ data set); the chloroplast trnL(UAA) intron and the spacer between the trnL (UAA) 3′ exon and trnF (GAA) were relatively conserved and suitable for phylogenetic reconstruction at genus level. The multicopy internal transcribed spacers (ITS1 and ITS2) of nuclear ribosomal DNA in contrast appear to be evolving about five times faster and are suitable for resolution at the species level. The putative homologue of cycloidea (Gcyc) has an intermediate substitution rate about three times faster than the chloroplast intron/spacer region. However, the level of pairwise sequence divergence of Gcyc is higher than that of ITS at very low levels of divergence. This difference in apparent rate of molecular evolution between ITS and Gcyc at different levels of the taxonomic hierarchy we attribute to the process of molecular drive in the multicopy ITS. At lower levels of divergence (e.g. between closely related species) fixation of genetic changes in the multicopy ribosomal DNA acts as a restraint on evolutionary rate, whereas third codon position changes in coding single copy nuclear (scnDNA) genes are unconstrained. However, at high levels of divergence (e.g. between general, scnDNA evolution is more functionally constrained than that of ITS and Gcyc therefore varies less. The small restraining effect of concerted evolution is not noticeable at these levels of sequence divergence. All three regions appear to evolve in a clock-like manner and are found to be suitable for phylogenetic reconstruction by parsimony, resulting in the same or similar topologies. We have examined the Gcyc sequences of three species that have reverted to actinomorphy from a zygomorphic condition. The gene appears to be intact and therefore, by implication, functional in these species. Furthermore, in one of these clades there has been a reversion back to zygomorphy which also implies that the gene is intact. We therefore suggest that in naturally occurring actinomorphic Gesneriaceae Gcyc continues to have a functional role, but zygomorphy is reduced by modifying genes. There is no convincing evidence that Gcyc evolves faster in actinomorphic lineages.