An Analysis of the Population Genetic Structure and Species History of Drosophila Melanogaster and Drosophila Simulans Using Restriction Fragment Length Polymorphisms of Mitochondrial DNA
Animal mitochondrial DNA (mtDNA) has several features that give it great utility in the study of geographic structure of natural populations. Its small size and covalently closed circular conformation make it easy to purify. Strict maternal inheritance and homoplasmy makes the effective copy number of mtDNA as little as 1/4 that of nuclear loci; this renders populational complements of mtDNA less susceptible to the homogenizing effects of gene flow due to migration, and more susceptible to founder effects due to fluctuations in effective population size. The absence of recombination allows for comparatively simple reconstruction of genealogies of mtDNA variants. A comparatively high rate of base substitution assures that most species will have enough mtDNA variants to make population genetic inferences meaningful.
Finally, sequence variants of mtDNA harboured in natural populations are presumed to be selectively neutral (i.e. has no effect on organismal fitness). This means that the distribution of mtDNA variants in populations will be due entirely to the stochastic and deterministic effects of species history, and not to natural selection. As a result, mtDNA can be used to infer species history directly and, after allowing for differences in modes of transmission, to infer the action of natural selection on other genetically determined factors like allozymes. The research of this dissertation was to survey world-wide natural populations of Drosophila melanogaster and its sibling species D. simulans for restriction fragment length polymorphisms of mtDNA. These cosmopolitan species have been widely studied for variation of allozymes and many other genetically determined factors (e.g. chromosomal inversions, morphometric characters).
A total of 144 isofemale lines of D. melanogaster were analyzed from 18 geographic populations. Considerable size variation was observed in this sample. Most size variation occurs in the major non-coding region (A+T-rich region), yielding a total size range of 18.2 kbp to 19.9 kbp. The occurrence of several size variants among all haplotypes indicates that the rate of size mutation is quite high. Further, the frequency distribution of size variants suggests that there is selection against larger sized mtDNA molecules; a replication advantage to smaller sized molecules is a likely explanation. Finally, there is evidence for 'small-scale' size variation (i.e. a total range of 20 bp) in the coding region of mtDNA. However, it could not be determined if the observed mobility variation of the fragments in question was actually due to addition/deletion of DNA or due to conformational effects of particular base substitutions.
Using 10 restriction enzymes, 23 restriction haplotypes were observed in D. melanogaster mtDNA. The phylogeographic distribution of these haplotypes allowed populations to be roughly divided into three longitudinal regions. Euro-African populations showed high intra- and inter-populational diversity and were inferred to be the oldest. Far East populations showed low intra-populational diversity but high inter-populational diversity; these populations similarly have a lengthy, but more complex, history. Western Hemisphere populations have low intra- and inter-populational diversity, and were inferred to be the youngest. A colonization history of this species is proposed. The species history suggested by mtDNA alone is quite similar to one proposed from the collective results of studies on several genetically-influenced traits in D. melanogaster (David and Capy 1988). mtDNA analysis is therefore shown to be a very efficient means by which to study species history. The very different histories of these regions reinforces the notion that the parallel latitudinal clines of allozyme alleles observed in this species (Singh et al 1982) are due to natural selection and are not historical.
A total of 79 isofemale lines of D. simulans from 14 geographic populations (13 continental populations, and the Seychelles Islands) were surveyed. The 10 restriction enzymes revealed a very discontinuous distribution of variation. The continental populations harboured only four haplotypes, while the Seychelles population harboured two haplotypes that were very different from the continental ones. The discontinuous distribution suggests that D. simulans evolved as a series of allopatric ancestral populations, as has been previously suggested. The relative lack of diversity among the continental mtDNAs, coupled with the observation that none of the four continental haplotypes are unique to a single population, strongly suggests that continental populations of D. simulans result from a recent world-wide expansion from one of the ancestral populations, possibly concurrent with the expansion into the western Hemisphere by D. melanogaster. A recent expansion would explain the lower degree of allozyme population structure seen in D. simulans than in D. melanogaster, and is more parsimonious than the alternative explanations of a narrower niche-width or the adoption of a general purpose genotype.