Project description:We report 2,649 crossover breakpoints identified by single-cell DNA sequencing of 217 sperm in a B6xCAST F1 mouse, which is heterozygous at Prdm9, with one wild-type CAST allele and one allele found in human populations (bred from a genetically engineered mother, Davies et al Nature 2012). Separately, we have reported testis DMC1 enrichment in the same mouse (GSE124991). We also infer H3K4me3 intensities at DMC1 hotspots, based on the raw H3K4me3 ChIP-seq data published under GSE119727 (Li et al). We identify several factors, including PRDM9 binding on the repair-template homologue, telomere proximity and local GC-content, that affect the probability that a DSB is repaired as a crossover. We further show that these factors also influence the time it takes for the site of a DSB to find and engage its homologue, with rapidly-engaging sites being more likely to be repaired as crossovers.
Project description:We report testis DMC1 enrichment at DSB (Double-Strand Break) hotspots in a B6xCAST F1 mouse, which is heterozygous at Prdm9, with one wild-type CAST allele and one allele found in human populations (bred from genetically engineered mother, Davies et al Nature 2012). We also report H3K4me3 intensities at these hotspots, based on the raw data published under GSE119727 (Li et al). Separately, we have reported 2,649 crossover locations identified by single-cell DNA sequencing of 217 sperm from the same hybrid mouse. We identify several factors, including PRDM9 binding on the repair-template homologue, telomere proximity and local GC-content, that affect the probability that a DSB is repaired as a crossover. We further show that these factors also influence the time it takes for the site of a DSB to find and engage its homologue, with rapidly-engaging sites being more likely to be repaired as crossovers.
Project description:Meiotic recombination is initiated by programmed DNA double-strand breaks (DSBs) and proceeds via binding of RPA, RAD51 and DMC1 to single-stranded DNA (ssDNA) substrates created after the formation of DSBs. Here, we report high-resolution in vivo maps of RPA and RAD51 binding in meiosis, mapping their binding locations and lifespans in a B6 and a genetically modified B6xCAST F1 mouse. We ascribe signals separately to the individual homologous chromosomes in the hybrid mouse, thereby separating the signal of binding to the chromosome where DSBs occurred and the chromosome that was used as template for repair. Together with super-resolution microscopy and DMC1 binding maps, we show that DMC1 and RAD51 have distinct spatial localization on ssDNA: whereas DMC1 binds near the break-site, RAD51 binds away from it. We characterize the D-loop, a critical intermediate bound by RPA, in vivo. These data show that DMC1, not RAD51, performs strand exchange in mammalian meiosis. We find that the localisation of D-loop intermediates is similar in crossover and non-crossover pathways, with a longer lifespan for crossover-destined intermediates. These findings answer long-standing questions about the molecular intermediates of recombination.
Project description:Mammalian meiotic recombination proceeds via repair of hundreds of programmed DNA double-strand breaks (DSBs). This process requires choreographed binding of RPA, DMC1 and RAD51 to single-stranded DNA (ssDNA) substrates and in vivo binding maps of these proteins provide insights into the underlying molecular mechanisms. When assayed in F1-hybrid mice, these maps can distinguish the broken chromosome from the homologous chromosome used as template for repair, which reveals further mechanistic detail and enables the structure of the recombination intermediates to be inferred. By applying CRISPR/Cas9 mutagenesis directly on F1-hybrid embryos, we have extended this powerful analysis technique to explore the molecular detail of recombination when a key component is knocked-out. As a proof-of-concept, we have generated biallelic knockouts of Dmc1 and built maps of meiotic binding of RAD51 and RPA in these knockout hybrid mice. Dmc1 mutants undergo meiotic arrest and comparison of these maps with those from wild-type mice is informative about the structure and timing of recombination intermediates in both genotypes. We confirm a complete abrogation of strand exchange in Dmc1 mutants, and observe a redistribution of RAD51 binding across both the distal and proximal ends of the resected DNA. We observe unexpected RPA and DMC1 binding in the wild-type, which suggests multiple rounds of strand invasion with template-switching in mouse. The methodology used involves direct phenotyping of hybrid “founder” mice following CRISPR mutagenesis and provides a high-throughput approach for the analysis of gene function during meiotic recombination, at low animal cost.
Project description:We report testis H3K4me3 enrichment in an F1 male from a C57BL/6J (B6) x CAST/Eij (CAST) cross (B6 mother, CAST father). This mouse is heterozygous at PRDM9 for a humanized allele (Davies et al. Nature 2016) and the CAST allele. After filtering of promoter H3K4me3 regions, these data serve as a measure of PRDM9 binding enrichment on each homologue. We found that both crossovers and non-crossovers (observed by sequencing F2/F4/F5 genomic DNA) are depleted at "asymmetric" Double-Strand Break hotspots where PRDM9 primarily binds only one of the two homologues. This proves that PRDM9 plays an important role in promoting inter-homologue interactions and can explain why increasing PRDM9 binding asymmetry predicts hybrid infertility. See Li, Bitoun, Altemose et al. 2018 (pending) for a complete summary.