Centromere evolution in annual and perennial soybeans and its implication for hybridization in cultivated species

Centromeres are comprised of repeats and centromeric specific histone H3 (CENH3), and they play essential roles in maintaining genome stability and defining chromosome karyotypes. Previous studies in monocotyledons show inter-species diversity of centromeric repeats and relatively stable transmission of CENH3 regions in polyploids and hybrids. Meanwhile, different CENH3 loading patterns exist in tetraploid cotton and soybean hybrid offsprings from their diploid donors. However, the dynamics of centromeric repeats and CENH3 regions during evolution and hybridization in soybean remain largely unknown.

By comparative analysis of centromere structure between perennial and annual soybeans, we find that retrotransposons and tandem repeats are the dominant centromeric specific repeats in perennial and annual soybeans, respectively. Both types of repeats exhibit low sequence similarity and minimal exchange during evolution, suggesting distinct mechanisms for centromere chromatin assembly. Furthermore, in perennial tetraploid soybeans, we detect the highest frequency of centromere repositioning among all reported polyploid plants when compared to their diploid progenitors. Interestingly, we identify diverse CENH3 loading patterns in F1 hybrids derived from different intra-species crosses, particularly those newly formed and parentally biased CENH3 loading. Different parental lines exhibit various capacities to induce distinct CENH3 loading patterns in F1 hybrids. Our results reveal recurrent CENH3 loading dynamics during soybean polyploidization and intra-species hybridization, suggesting active centromere reorganization during genetic transmission.

Our findings reveal substantial turnover of centromeric repeats between perennial and annual soybeans, but also demonstrate various centromere repositioning in soybean polyploids and intra-species hybrids. This work expands our understanding in centromere biology.