Common models for the evolution of duplicated genes after genome duplication are subfunctionalization, neofunctionalization, and pseudogenization. Although the crucial roles of cis-regulatory mutations in subfunctionalization are well-documented, their involvement in pseudogenization and/or neofunctionalization remains unclear. We addressed this issue by investigating the evolution of duplicated homeobox genes, six6.L and six6.S, in the allotetraploid frog Xenopus laevis. Based on a comparative expression analysis, we observed similar eye-specific expression patterns for the two loci and their single ortholog in the ancestral-type diploid species Xenopus tropicalis. However, we detected lower levels of six6.S expression than six6.L expression. The six6.S enhancer sequence was more highly diverged from the orthologous enhancer of X. tropicalis than the six6.L enhancer, and showed weaker activity in a transgenic reporter assay. Based on a phylogenetic analysis of the protein sequences, we observed greater divergence between X. tropicalis Six6 and Six6.S than between X. tropicalis Six6 and Six6.L, and the observed mutations were reminiscent of a microphthalmia mutation in human SIX6. Misexpression experiments showed that six6.S has weaker eye-enlarging activity than six6.L, and targeted disruption of six6.L reduced the eye size more significantly than that of six6.S. These results suggest that enhancer attenuation stimulates the accumulation of hypomorphic coding mutations, or vice versa, in one duplicated gene copy and facilitates pseudogenization. We also underscore the value of the allotetraploid genome of X. laevis as a resource for studying latent pathogenic mutations.

The CRISPR/Cas9 method has become popular for gene disruption experiments in Xenopus laevis. However, the experimental conditions that influence the efficiency of CRISPR/Cas9 remain unclear. To that end, we developed an image analysis technique for the semi‐quantitative evaluation of the pigment phenotype resulting from the disruption of tyrosinase genes in X. laevis using a CRISPR/Cas9 approach, and then examined the effects of varying five experimental parameters (timing of the CRISPR reagent injection into developing embryos; amount of Cas9 mRNA in the injection reagent; total injection volume per embryo; number of injection sites per embryo; and the culture temperature of the injected embryos) on the gene disruption efficiency. The results of this systematic analysis suggest that the highest possible efficiency of target gene disruption can be achieved by injecting a total of 20 nL of the CRISPR reagent containing 1500 pg of Cas9 mRNA or 4 ng of Cas9 protein into two separate locations (10 nL each) of one‐cell stage embryos cultured at 22°C. This study also highlights the importance of balancing the experimental parameters for increasing gene disruption efficiency and provides valuable insights into the optimal conditions for applying the CRISPR/Cas9 system to new experimental organisms.