Bivalent individualization during chromosome territory formation in Drosophila spermatocytes by controlled condensin...

Reduction of genome ploidy from diploid to haploid necessitates stable pairing of homologous chromosomes into bivalents before the start of the first meiotic division. Importantly, this chromosome pairing must avoid interlocking of non-homologous chromosomes. In spermatocytes of Drosophila melanogaster, where homolog pairing does not involve synaptonemal complex formation and crossovers, associations between non-homologous chromosomes are broken up by chromosome territory formation in early spermatocytes. Extensive non-homologous associations arise from the coalescence of the large blocks of pericentromeric heterochromatin into a chromocenter and from centromere clustering. Nevertheless, during territory formation, bivalents are moved apart into spatially separate subnuclear regions. The condensin II subunits, Cap-D3 and Cap-H2, have been implicated, but the remarkable separation of bivalents during interphase might require more than just condensin II. For further characterization of this process, we have applied time-lapse imaging using fluorescent markers of centromeres, telomeres and DNA satellites in pericentromeric heterochromatin. We describe the dynamics of the disruption of centromere clusters and the chromocenter in normal spermatocytes. Mutations in Cap-D3 and Cap-H2 abolish chromocenter disruption, resulting in excessive chromosome missegregation during M I. Chromocenter persistence in the mutants is not mediated by the special system, which conjoins homologs in compensation for the absence of crossovers in Drosophila spermatocytes. However, overexpression of Cap-H2 precluded conjunction between autosomal homologs, resulting in random segregation of univalents. Interestingly, Cap-D3 and Cap-H2 mutant spermatocytes displayed conspicuous stretching of the chromocenter, as well as occasional chromocenter disruption, suggesting that territory formation might involve forces unrelated to condensin II. While the molecular basis of these forces remains to be clarified, they are not destroyed by inhibitors of F actin and microtubules. Our results indicate that condensin II activity promotes chromosome territory formation in co-operation with additional force generators and that careful co-ordination with alternative homolog conjunction is crucial.