While it remains a matter of some debate, rapid sequence evolution of the coding sequences of duplicate genes is characteristic for early phases past duplication, but long established duplicates generally evolve under constraint, much like the rest of the coding genome. As for coding sequences, it may be possible to infer evolutionary rate, selection, and constraint via contrasts between duplicate gene divergence in the 5 prime regions and in the corresponding synonymous site divergence in the coding regions. Finding elevated rates for the 5 prime regions of duplicated genes, in addition to the coding regions, would enable statements regarding the early processes of duplicate gene evolution. Here, 1 kb of each of the 5 prime regulatory regions of Drosophila melanogaster duplicate gene pairs were mapped onto one another to isolate shared sequence blocks. Genetic distances within shared sequence blocks (d5’) were found to increase as a function of synonymous (dS), and to a lesser extend, amino-acid (dA) site divergence between duplicates. The rate d5’/dS was found to rapidly decay from values > 1 in young duplicate pairs (dS < 0.3) to 0.28 or less in older duplicates (dS > 0.8). Such rapid rates of 5 prime evolution exceeding 1 (~neutral) predominantly were found to occur in duplicate pairs with low amino-acid site divergence and that tended to be co-regulated when assayed on microarrays. Conceivably, functional redundancy and relaxation of selective constraint facilitates subsequent positive selection on the 5 prime regions of young duplicate genes. This might promote the evolution of new functions (neofunctionalization) or division of labor among duplicate genes (subfunctionalization). In contrast, similar to the vast portion of the non-coding genome, the 5 prime regions of long-established gene duplicates appear to evolve under selective constraint, indicating that these long-established gene duplicates have assumed critical functions.
gene duplication; gene expression; selection; promoter evolution
