Coevolution of retroelements and tandem zinc finger genes.

ABSTRACT

Vertebrate genomes encode large and highly variable numbers of tandem C2H2 zinc finger (tandem ZF) transcription factor proteins. In mammals, most tandem ZF genes also encode a KRAB domain (KZNF proteins). Very little is known about what forces have driven the number and diversity of tandem ZF genes. Recent studies suggest that one role of KZNF proteins is to bind and repress transcription of exogenous retroviruses and their endogenous counterpart LTR retroelements. We report a striking correlation across vertebrate genomes between the number of LTR retroelements and the number of host tandem ZF genes. This correlation is specific to LTR retroelements and ZF genes and was not explained by covariation in other genomic features. We further show that recently active LTR retroelements are correlated with recent tandem ZF gene duplicates across vertebrates. On branches of the primate phylogeny, we find that the appearance of new families of endogenous retroviruses is strongly predictive of the appearance of new duplicate KZNF genes. We hypothesize that retroviral and LTR retroelement burden drives evolution of host tandem ZF genes. This hypothesis is consistent with previously described molecular evolutionary patterns in duplicate ZF genes throughout vertebrates. To further explore these patterns, we investigated 34 duplicate human KZNF gene pairs, all of which underwent an early burst of divergence in the major nucleotide contact residues of their ZF domains, followed by purifying selection in both duplicates. Our results support a host-pathogen model for tandem ZF gene evolution, in which new LTR retroelement challenges drive duplication and divergence of host tandem ZF genes.