Mechanisms of pericentromeric crossover inhibition

Abstract

Meiosis maintains ploidy during sexual reproduction and provides genetic diversity through recombination. During meiotic divisions, a single DNA replication event precedes two sequential (reductional then equational) cell divisions. Proper homologue segregation at meiosis I requires (1) crossovers between homologous chromosomes (2) sister kinetochore mono-orientation (3) centromere-proximal (pericentromeric) cohesin protection.

Crossover formation is spatially regulated at centromeres and telomeres. The pericentromere, defined as a cohesin-rich region flanking each centromere, displays reduced crossover frequency when compared to chromosome arms. Where pericentromeric crossovers do form, chromosomes have a high frequency of mis-segregation. Prior studies in Saccharomyces cerevisiae have shown that the Ctf19 kinetochore protein is key to coordinating pericentromeric crossover inhibition, through dual roles in pericentromeric cohesin recruitment and reducing DSB formation. Despite this knowledge, exactly how cohesin regulates crossover formation at pericentromeres remains unknown.

Unpublished work found that the pericentromeric cohesin regulator, shugoshin (Sgo1) also prevents crossovers at the pericentromere. Sgo1 was identified for its role in protecting pericentromeric cohesin in metaphase I, where it recruits a trimeric protein phosphatase 2A (PP2A-Rts1) to de-phosphorylate the cohesin subunit Rec8. As a consequence, cohesin at the pericentromere is separase-resistant and remains chromosome-bound until metaphase II. In this thesis, I hypothesised that Sgo1-(PP2ARts1)- dependent de-phosphorylation of pericentromeric Rec8 might also be important in meiotic prophase I to prevent crossover formation in this region. I found that a sgo1- 3A mutant, which loses its PP2A-Rts1 interaction, had increased pericentromeric crossovers, lending some support to the hypothesis. I also generated several other Sgo1-(PP2A-Rts1) interaction mutants, then carried out phenotypic and biochemical characterisation. Two mutations, sgo1-4A and rts1-5A, abolished Sgo1-PP2A interaction as inferred by mass spectrometry, yet pericentromeric cohesin protection could be maintained past metaphase I. This suggests that this critical meiosis I event is highly robust towards reductions in Sgo1-(PP2A-Rts1) interaction. Conversely, the rts1-R446E and rts1 anchor-away mutants did disrupt meiosis II sister chromatid segregation, likely due to perturbations to the Sgo1-(PP2A-Rts1) interaction. Surprisingly, these same mutations did not significantly impact pericentromeric crossover frequency. This raises the idea that Sgo1’s role in pericentromeric crossover inhibition could be modulated through interactors other than PP2A-Rts1.

Separately, I found that a rec8-14D allele, which mimics constitutive phosphorylation on sites normally de-phosphorylated by Sgo1-(PP2A-Rts1), resulted in increased crossover formation at both a centromere and a chromosome arm. Previous studies reported that a rec8-24A mutant, which has additional phospho-sites mutated to alanine, generated a reduced number of inter-homologue crossovers compared to wild-type REC8. These data point towards a role for balanced Rec8 phosphorylation levels in maintaining crossover frequencies. To understand how Rec8 phosphorylation affects crossovers, I utilised both phosphomutants to test several hypotheses. Parallel to double-stranded DNA break (DSB) formation and repair is meiotic chromosome looping along with the installation of the synaptonemal complex throughout prophase I. Rec8 influences the localisation of Red1, an axial filament protein of the synaptonemal complex, and both proteins are critical for inter-homologue DSB repair. I found that both phosphomutants had elevated Rec8 and Red1 occupancies genomewide, suggesting that the recombination phenotypes observed cannot be explained by Rec8 or Red1 levels on chromosomes alone. Given the loop-axis structure of meiotic chromosomes, loop length is inversely correlated with axis length. Additionally, there is also evidence for a correlation between loop/axis length and crossover frequencies in the literature. I found that the phospho-mimetic rec8-14D had longer chromosomal loops while rec8-24A had shorter loops. These results suggest that axis length could also be altered in these mutants and this potentially could provide an explanation for the differences in crossover frequencies.

In sum, it remains to be clarified if Sgo1-dependent recruitment of PP2A-Rts1 and/ or the phosphatase activity is required for inhibiting pericentromeric crossovers. However, work here highlights the interesting possibility that maintaining a balanced level of Rec8 phosphorylation is critical for crossover homeostasis. I propose that this posttranslational modification affects some properties of Rec8-cohesin required for loop formation. This could have important consequences for the subsequent length of the chromosome axis, ultimately influencing crossover frequencies and the fidelity of meiotic chromosome segregation.School of Biological Sciences, University of Edinburg