Rapid chromosomal evolution and polycentromeric drive in sedges and rushes

The chromosomes of most eukaryotes have a single centromere, a specialised region involved in chromosome partitioning to daughter cells. Inter-chromosomal rearrangement risks generating chromosomes with two centromeres or none, disrupting segregation. Holocentric chromosomes, with centromere function distributed along the chromosome, are hypothesised to better tolerate inter-chromosomal rearrangement. However, evidence linking centromere organisation to rearrangement rate has been lacking. Sedges and rushes (Cyperaceae and Juncaceae) are specifically polycentric: they have several satellite-based centromeres per chromosome, whereas centromere location in "asatellitic" holocentrics is solely defined epigenetically. Using 36 chromosome-level genomes, we quantify extraordinary rearrangement rates and annotate candidate polycentromeres. Under our polycentromeric drive model, we expected satellite sequence, polycentromere architecture, and karyotype to evolve to exploit biased segregation during asymmetric meiosis. We find satellite turnover but also deep sequence conservation. Polycentromeric organisation seems constrained by chromosome size, and rearrangements are more stable when chromosomes have fewer polycentromeres, despite breakpoint regions being enriched for polycentromeres. Notably, we find putative monocentromeres in Carex myosuroides, which would be the first evidence for reversion to monocentricity in eukaryotes, and no identifiable polycentromeres in Cyperus rotundus, possibly a transition to asatellitic holocentricity. Overall, we demonstrate that this clade is powerful for linking centromere organisation to genome evolution.