Evolutionary patterns and structural divergence of CENH3 in legumes: Implications for haploid induction breeding

Abstract Background The centromeric histone variant CENH3 is crucial for chromosome segregation and haploid induction in plants, yet its evolutionary patterns in legumes remain poorly characterized. Methods We investigated CENH3 phylogeny and molecular evolution across legumes, focusing on Vicia using phylogenetic reconstruction, sequence alignment, and evolutionary selection analyses. Results Our phylogenetic reconstruction delineated legume species into two major clades (A and B) and revealed a profound contrast between the hypervariable N‐terminal tail and the highly conserved histone fold domain (HFD). Within the HFD, the CENP‐A targeting domain (CATD) exhibited absolute functional constraint, while the N‐terminus demonstrated remarkable evolutionary plasticity. In the genus Vicia , although the exon‐intron structure was entirely conserved, substantial sequence polymorphism was identified. Comparative analysis between Vicia sativa L. and Vicia villosa Roth. highlighted species‐specific epitope divergence alongside conserved centromere localization. Evolutionary analyses revealed that CENH3 is predominantly under purifying selection, with localized positive selection in specific lineages, whereas canonical H3 exhibited a binary selection pattern dependent on phylogenetic distance. Conclusions Our findings elucidate the evolutionary dynamics of CENH3 in legumes and identify the highly conserved yet functionally distinct CATD as a promising, specific target for developing efficient haploid induction systems through genome editing.