A conserved architectural domain shapes centromere evolution in Drosophila

Abstract

Centromeres ensure faithful chromosome segregation despite being embedded within rapidly evolving repetitive DNA, a contradiction known as the centromere paradox. While centromere identity is defined by the histone variant CENP-A, how conserved function is maintained amid rapid DNA turnover remains unclear. Here, we generate highly contiguous genome assemblies from single Drosophila melanogaster individuals that, for the first time, resolve a chromosome through its centromere, linking the chromosome 3 arms within a continuous sequence. Comparative assemblies from wild-derived strains reveal extensive structural variation in pericentromeric satellites, including large-scale expansions, contractions, and sequence divergence. Despite this variation, the CENP-A-associated centromeric core exhibits conserved organization across strains. Integration of Hi-C interaction maps with sequence analyses shows that flanking dodeca satellite arrays form a spatially interacting domain that bridges both sides of the centromere, whereas adjacent Prodsat arrays are more variable and show weaker interactions. These results support a model in which rapidly evolving centromeric DNA is constrained by conserved higher-order architecture, providing a framework for reconciling the rapid evolution of centromere sequence with its conserved function.