A Comprehensive Study of Genic Variation in Natural Populations of Drosophila melanogaster . IV . Mitochondrial DNA Variation and the Role of History ZIS . Selection in the Genetic Structure of Geographic Populations

Preliminary studies with restriction fragment length polymorphisms of mitochondrial DNA (mtDNA) in natural populations of Drosophila melanogaster revealed considerable variation in terms of nucleotide sequence and overall size. In this report we present data from more isofemale lines and more restriction enzymes, and explore the utility of the data in inferring a colonization history of this species. Size variation in the noncoding A + T-rich region is particularly plentiful, with size variants occurring in all restriction site haplotypes in all populations. We report here classes of small-scale mobility polymorphisms (apparent range of 20 bp) in specific restriction fragments in the coding region. The variation in one such fragment appears to be generated even more rapidly than in the noncoding region. On the basis of the distribution of restriction site haplotypes, the species range can be divided into three major regions along longitudinal lines: Euro-African populations are the most diverse and are taken to be oldest; Far East populations have a complex distribution of haplotypes; Western Hemisphere populations are the least diverse and are interpreted to be the youngest. The history inferred from mtDNA alone is remarkably similar to one based on several nuclear markers. The mtDNA haplotype distribution is also very different from that of allozymes in these same populations. We interpret this as further evidence that natural selection is still the most parsimonious explanation for the parallel latitudinal allozyme clines in this species. T HE pattern of genetic variation in natural populations of species is a result of a complex interplay of natural selection and the effects of species history. Geographical routes of colonization, gene flow, and variation of population size (as it pertains to random genetic drift) are examples of historical factors which can influence genetic differentiation of populations. Historical information, therefore, is crucial to a study of the action of natural selection on the genetic structure of a species. An important difference between the effects of natural selection and history on genetic variation is that natural selection will exert its influence differently on individual loci, while the influence of history should be more or less equivalent on all loci. This feature of historical effects allows the inference of species history in two ways: a recurrent pattern of polymorphism, divergence and fixation of a number of loci may indicate the influence of history, or we can follow the fate of a single neutral marker. Mitochondrial DNA (mtDNA) is an important tool in the analyses of phylogenies of closely related species (DESALLE and GIDDINGS 1986; GEORGE and RYDER 1986), of gene flow between hybridizable sibling speIsland, Charlottetown, Prince Edward Island, CIA 4P3, Canada. Genetics 129: 103-117 (September, 1991) ’ Present address: Department of Biology, University of Prince Edward cies and subspecies, particularly with respect to geographic zones of contact (FERRIS et a2 1984; CARR et al 1986; HARRISON, RAND and WHEELER 1987; SPOLSKY and UZZEL 1984), and population structures of individual species (SAUNDERS, KESSLER and AVISE 1985; LANSMAN et al 1983; LATORRE et al. 1988). mtDNA is appropriate because of its uniparental inheritance, nonrecombinogenicity, small size, and elevated rate of mutation (reviewed in AVISE et al. 1987). A much more contentious issue is whether intraspecific sequence variants of mtDNA are selectively neutral. The evidence for this is largely conjectural. The extreme importance of the mitochondrial gene products implies a necessary conservation of their genes. Empirical support for this notion has come from interspecies sequence studies (BROWN and SIMPSON 1982) and experiments on cyto-nuclear associations (CLARK 1985; CLARK and LYCKEGAARD 1988). It is likely that most “selectable” mutations are therefore deleterious. In this paper, we use restriction analysis of mtDNA polymorphism to infer species history in Drosophila melanogaster. In a preliminary report (HALE and SINGH 1987), we demonstrated that worldwide populations of D. melanogaster show a much greater degree of population structure for mtDNA restriction haplotypes than for allozyme variants. Here, we ex104 L. R. Hale and R. S. Singh