Haplotypic divergence coupled with lack of diversity at the Arabidopsis thaliana alcohol dehydrogenase locus: roles for both balancing and directional selection?

Abstract

Abstract We designate a region of the alcohol dehydrogenase locus (Adh) of the weedy crucifer, Arabidopsis thaliana, as "hypervariable" on the basis of a comparison of sequences from ecotypes Columbia and Landsberg. We found eight synonymous and two replacement mutations in the first 262 nucleotides of exon 4, and an additional two mutations in the contiguous region of intron 3. The rest of the sequence (2611 bp) has just three mutations, all of them confined to noncoding regions. Our survey of the hypervariable region among 37 ecotypes of A. thaliana revealed two predominant haplotypes, corresponding to the Columbia and Landsberg sequences. We identified five additional haplotypes and 4 additional segregating sites. The lack of haplotype diversity is presumably in part a function of low rates of recombination between haplotypes conferred by A. thaliana's tendency to self-fertilize. However, an analysis in 32 ecotypes of 12 genome-wide polymorphic markers distinguishing Columbia and Landsberg ecotypes indicated levels of outcrossing sufficient at least to erode linkage disequilibrium between dispersed markers. We discuss possible evolutionary explanations for the coupled observation of marked divergence within the hypervariable region and a lack of haplotype diversity among ecotypes. The sequence of the region for closely related species argues against the possibility that one allele is the product of introgression. We note (1) that several loss of function mutations (both naturally and chemically induced) map to the hypervariable region, and (2) the presence of two amino acid replacement polymorphisms, one of which causes the mobility difference between the two major classes of A. thaliana Adh electrophoretic alleles. We argue that protein polymorphism in such a functionally significant part of the molecule may be subject to balancing selection. The observed pattern of extensive divergence between the alleles is consistent with this explanation because balancing selection on a particular site maintains linked neutral polymorphisms at intermediate frequencies.