Evolution of plant RNA polymerase IV/V genes: evidence of subneofunctionalization of duplicated NRPD2/NRPE2-like paralogs in Viola (Violaceae)

<jats:title>Abstract</jats:title> <jats:sec> <jats:title>Background</jats:title> <jats:p>DNA-dependent RNA polymerase IV and V (Pol IV and V) are multi-subunit enzymes occurring in plants. The origin of Pol V, specific to angiosperms, from Pol IV, which is present in all land plants, is linked to the duplication of the gene encoding the largest subunit and the subsequent subneofunctionalization of the two paralogs (<jats:italic>NRPD1</jats:italic> and <jats:italic>NRPE1</jats:italic>). Additional duplication of the second-largest subunit, <jats:italic>NRPD2/NRPE2</jats:italic>, has happened independently in at least some eudicot lineages, but its paralogs are often subject to concerted evolution and gene death and little is known about their evolution nor their affinity with Pol IV and Pol V.</jats:p> </jats:sec> <jats:sec> <jats:title>Results</jats:title> <jats:p>We sequenced a ~1500 bp <jats:italic>NRPD2/E2</jats:italic>-like fragment from 18 <jats:italic>Viola</jats:italic> species, mostly paleopolyploids, and 6 non-<jats:italic>Viola</jats:italic> Violaceae species. Incongruence between the <jats:italic>NRPD2/E2</jats:italic>-like gene phylogeny and species phylogeny indicates a first duplication of <jats:italic>NRPD2</jats:italic> relatively basally in Violaceae, with subsequent sorting of paralogs in the descendants, followed by a second duplication in the common ancestor of <jats:italic>Viola</jats:italic> and <jats:italic>Allexis</jats:italic>. In <jats:italic>Viola</jats:italic>, the mutation pattern suggested (sub-) neofunctionalization of the two <jats:italic>NRPD2/E2</jats:italic>-like paralogs, <jats:italic>NRPD2/E2-a</jats:italic> and <jats:italic>NRPD2/E2-b</jats:italic>. The <jats:italic>d</jats:italic> <jats:sub> <jats:italic>N</jats:italic> </jats:sub>/<jats:italic>d</jats:italic> <jats:sub> <jats:italic>S</jats:italic> </jats:sub>ratios indicated that a 54 bp region exerted strong positive selection for both paralogs immediately following duplication. This 54 bp region encodes a domain that is involved in the binding of the Nrpd2 subunit with other Pol IV/V subunits, and may be important for correct recognition of subunits specific to Pol IV and Pol V. Across all <jats:italic>Viola</jats:italic> taxa 73 <jats:italic>NRPD2/E2</jats:italic>-like sequences were obtained, of which 23 (32%) were putative pseudogenes - all occurring in polyploids. The <jats:italic>NRPD2</jats:italic> duplication was conserved in all lineages except the diploid MELVIO clade, in which <jats:italic>NRPD2/E2-b</jats:italic> was lost, and its allopolyploid derivates from hybridization with the CHAM clade, section <jats:italic>Viola</jats:italic> and section <jats:italic>Melanium</jats:italic>, in which <jats:italic>NRPD2/E2-a</jats:italic> occurred in multiple copies while <jats:italic>NRPD2/E2-b</jats:italic> paralogs were either absent or pseudogenized.</jats:p> </jats:sec> <jats:sec> <jats:title>Conclusions</jats:title> <jats:p>Following the relatively recent split of Pol IV and Pol V, our data indicate that these two multi-subunit enzymes are still in the process of specialization and each acquiring fully subfunctionalized copies of their subunit genes. Even after specialization, the <jats:italic>NRPD2/E2</jats:italic>-like paralogs are prone to pseudogenization and gene conversion and <jats:italic>NRPD2</jats:italic> and <jats:italic>NRPE2</jats:italic> copy number is a highly dynamic process modulated by allopolyploidy and gene death.</jats:p> </jats:sec>