Selection on length mutations after frameshift can explain the origin and retention of the AP3/DEF-like paralogues in Impatiens

Evolution of class B genes through gene duplication has been proposed as an evolutionary mechanism that contributed to the enormous floral diversity. Frameshift mutations are a likely mechanism to explain the divergent C-terminal sequences of MIKC gene subfamilies. So far, the inferences for frameshifts and selective pressures on the C-terminal domain are made for old duplications for which the exact selective pressures are obscured by evolutionary time. This motivated us to study an example of a recent duplication, which allows us to consider in more detail the selective pressures that are involved after duplication. We find that after duplication and frameshift of Impatiens class B genes, the individual codons show no evidence for adaptive selection. It is rather the length of the C-terminal domain that either is strictly conserved or varies strongly. This suggests a role for the length of the C-terminal domain in the retention of duplicated genes.