Patterns of Mitochondrial DNA Strand Asymmetry Correlate With Phylogeny.

Abstract

In the current era of megabase DNA sequencing, there is a need for analytical methods that extend beyond the level of single genes or gene clusters and allow us to characterize genome features more globally. Following up on the observations of Lobry (1, 2) concerning “GC skew” [relative enhancement of G in one strand over a given DNA segment: (G C)/(G + C)], we have begun to examine “purine excess” in long DNA sequences. This function-defined as the sum of all purines minus the sum of all pyrimidines encountered in a sequence up to the given position-displays the overall strand asymmetry. In bacterial genomes, the minima and maxima of the purine excess curves correlate closely with, respectively, the origins and termini of replication (3). We have now calculated the purine-excess curves for 52 complete mitochondrial DNA sequences available from GenBank. They cover the entire range of eukaryotes, from yeast to human, and display intriguing patterns that correlate with phylogenetic relationships. The curves for closely related organisms (e.g., sets of primates, pinnipeds, marsupials, fish, insects, or sea urchins) virtually superpose, whereas curves from more distant heterogeneous groups (e.g., crustaceans plus insects) have recognizable similarities, but also clear differences of detail. Characteristically, the larger “ancestral” mtDNAs from single-celled organisms have “choppier” curves (see Fig. 1) with many sharp peaks and troughs, whereas the curves become smoother as one advances to the vertebrates with their “derived” mtDNAs of about 15 kb. Over the range from Latimeria to Homo-including 29 species-the purine ex-