Limited phylogenetic distribution of a long tandem-repeat cluster in the mitochondrial control region in Bubo (Aves, Strigidae) and cluster variation in Blakiston's fish owl (Bubo blakistoni).

To investigate the phylogenetic position of Blakiston's fish owl (Bubo blakistoni), we sequenced the mitochondrial (mt) DNA control region and cytochrome b (cyt b) for nine Bubo species. Maximum-likelihood analyses of combined control region and cyt b sequences, and cyt b sequences alone, showed that species formerly placed in genus Ketupa comprise a monophyletic group. Unexpectedly, we discovered a long cluster of 20-25 tandem repeat units 77 or 78bp long in the third control region domain in four of the nine Bubo species for which the control region was sequenced (B. blakistoni, B. flavipes, and B. ketupu in the Ketupa clade; B. lacteus), leading to overall control region lengths of 3.0-3.8kpb estimated from agarose gel electrophoresis. The control region in B. lacteus is the longest (3.8kbp) reported to date in vertebrates. Sequencing of eight repeat units at each end of the cluster in 20 B. blakistoni individuals detected several types of repeat units 77 or 78bp long, and six patterns in the order of unit types. The occurrence of a repeat cluster in all three species examined in the Ketupa clade suggests their common ancestor also had a cluster, whereas a maximum parsimony tree showed repeat-unit types grouping by species, rather than by paralog groups, suggesting independent origins of the clusters. We reconcile these results with a turnover model, in which the range in cluster-length variation and unit types at the 5' end are hypothetically functionally constrained by the protein-binding function of the control region, but otherwise there is a continual turnover of units in evolutionary time, with new unit types arising through mutations, proliferating by duplication of single and double repeat blocks, and being lost through deletion. Estimated free energies for reconstructed secondary structures of single and especially pairs of repeat units were higher than for homologous single-unit blocks in species lacking a repeat cluster, supporting slipped-strand mispairing as the mechanism of cluster turnover.