Centromeric repeats of the Western European house mouse I: high sequence diversity among monomers at local and global spatial scales

Previous work found that the centromeric repeats of the Western European house mouse (Mus musculus domesticus) are composed predominantly of a 120 bp monomer that is shared by the X and autosomes. Polymorphism in length and sequence was also reported. Here I quantified the length and sequence polymorphism of the centromeric repeats found on the X and autosomes. The levels of local and global sequence variation were also compared. I found three length variants: a 64mer, 112mer and 120mer with relative frequencies of 2.4%, 8.6%, and 89%, respectively. There was substantial sequence variation within all three length variants with a rank-order of: 64mer < 120mer < 112mer. The 64mer was never found alone on long Sanger traces, and was arranged predominantly as a 176 bp higher-order repeat composed of a 64/112mer dimer. Reanalysis of archived ChIP-seq reads found that all three length variants were enriched with the foundational centromere protein CENP-A, but the enrichment was far higher for the 120mer. This pattern indicates that only the 120mer contributes substantially to the functional centromeres, i.e., to the kinetochore-binding, centric cores of the centromeric repeat arrays. Despite only moderate sequence divergence among random pairs of 120mers (averaging 5.9%), other measures of sequence diversity were exceptionally high: i) variant richness (numerical diversity) -on average, one new sequence variant was observed every 4th additional monomer randomly sampled (in N = 7.2 x 103 monomers), and ii) variant evenness -all of the nearly 2 x 103 observed sequence variants were at low frequency, with the most common variant having a frequency of only 5.7%. I next used long Sanger trace data from the Mouse Genome Project to assess the pattern of monomer diversity among neighboring 120mers. Unexpectedly, side-by-side monomers were rarely identical in sequence, and sequence divergence between these neighbors was nearly as high as that between random pairs taken from the genome-wide pool of all 120mers. I also used long Sanger traces to determine sequence variation among neighborhoods of 5 contiguous 120 bp monomers. Sequence diversity within these small regions typically spanned most of the entire range of that found genome-wide. Despite high sequence variation within these neighborhoods, the density of monomers with functional binding motifs for CENP-B (i.e., b-boxes with sequence NTTCGNNNNANNCGGGN) was strongly conserved at about 50%. The overarching pattern of monomer structure at the centromeric repeats of this subspecies is: i) high homogeneity in the density CENP-B binding sites, and ii) high heterogeneity in monomer sequence at both local and global levels.