Local and sex-specific biases in crossover vs. noncrossover outcomes at meiotic recombination hot spots in mice

Abstract

Meiotic recombination initiated by programmed double-strand breaks (DSBs) yields two types of interhomolog recombination products, crossovers and noncrossovers, but what determines whether a DSB will yield a crossover or noncrossover is not understood. In this study, we analyzed the influence of sex and chromosomal location on mammalian recombination outcomes by constructing fine-scale recombination maps in both males and females at two mouse hot spots located in different regions of the same chromosome. These include the most comprehensive maps of recombination hot spots in oocytes to date. One hot spot, located centrally on chromosome 1, behaved similarly in male and female meiosis: Crossovers and noncrossovers formed at comparable levels and ratios in both sexes. In contrast, at a distal hot spot, crossovers were recovered only in males even though noncrossovers were obtained at similar frequencies in both sexes. These findings reveal an example of extreme sex-specific bias in recombination outcome. We further found that estimates of relative DSB levels are surprisingly poor predictors of relative crossover frequencies between hot spots in males. Our results demonstrate that the outcome of mammalian meiotic recombination can be biased, that this bias can vary depending on location and cellular context, and that DSB frequency is not the only determinant of crossover frequency.

Keywords: Spo11; crossing over; gene conversion; hot spot; meiosis; recombination.

Figure 1.

Recombination at the central hot spot in males and females. (A) Overview of recombination hot spots in this study. (Top) Schematic of mouse chromosome 1, with hot spot positions indicated as green stars. The central hot spot is located between base pairs 78,589,305 and 78,592,399. The distal hot spot is located between base pairs 185,265,469 and 185,269,517 (build 38 for both). (Bottom) Both hot spots show H3 Lys4 trimethylation (H3K4me3) ChIP-seq (chromatin immunoprecipitation [ChIP] combined with deep sequencing) signals (Baker et al. 2014) in testes of mice expressing the B6 version of PRDM9 (orange trace) but not the PRDM9 version found in the CAST/EiJ strain (purple trace, which mostly overlaps the horizontal axis). (Bottom left) The 11-bp motif predicted to bind B6 (and A/J)-encoded PRDM9 (J Lange and S. Tischfield, pers. comm.) is found in both hot spots (pink dots) and is identical in the B6 and A/J strains for each hot spot. The depicted regions in these and all subsequent hot spot graphs are 78,589,447–78,592,447 for the central hot spot and 185,265,656–185,268,656 for the distal hot spot. (B) Assays to amplify and identify crossovers and noncrossovers. Filled circles represent sequence polymorphisms (red or blue for the two parental genotypes and gray for amplified but not-yet-defined internal polymorphisms), and arrowheads represent PCR primers (red or blue for allele-specific primers and gray for universal). (Panel i) In the crossover (CO) assay, two sequential rounds of allele-specific PCR selectively amplify recombinant DNA molecules from small pools of sperm or oocyte DNA from an F1 hybrid animal. Recombination frequencies are estimated from the observed fraction of pools that yield amplification products. Next, internal polymorphisms in each amplified recombinant DNA molecule are genotyped by hybridization with allele-specific oligonucleotides (ASOs) to map the location of the crossover breakpoint. (Panel ii) In the noncrossover/crossover (NCO/CO) assay, smaller pools of sperm or oocyte DNA are amplified using nested primers that are specific for one haplotype in combination with nested universal primers. In contrast to the crossover assay, both nonrecombinant and recombinant DNA molecules are amplified, with the majority (>95%) being nonrecombinants. Subsequent hybridization of amplification products to ASOs that are specific for alleles from the nonselected haplotype identifies pools containing crossovers and noncrossovers. (C) Crossovers in males. (Top graph) Crossover breakpoint maps are shown for crossover molecules amplified with allele-specific primers in the B6-to-A/J orientation (dark blue) or the A/J-to-B6 orientation (light blue). Cumulative crossover distributions are shown below. Tested polymorphisms are indicated as ticks at the top. Multiple polymorphisms contained within a single ASO are indicated as a single green tick. Numbers of observed crossovers and Poisson-corrected crossover frequencies (±SD) are indicated. (D) Similar distributions of crossover breakpoints in males and females. Data from both orientations of the allele-specific PCR were pooled separately for males (blue) and females (red). (E) Total noncrossovers in males. Total relative noncrossover frequencies from all four orientations of the PCRs (normalized for co-conversions) at the tested polymorphisms are shown as blue bars. The crossover breakpoint map in males is shown for comparison (light gray). Number of total observed noncrossovers and Poisson-corrected total noncrossover frequency (±SD) are indicated. (F,G) Noncrossovers on the B6 (F) and A/J (G) chromosomes in males. (Top) Relative noncrossover frequencies (normalized for coconversions) on the B6 (F) and A/J (G) chromosomes at the tested polymorphisms are shown as blue bars. The crossover breakpoint map in males is shown for comparison (light gray). Number of observed noncrossovers and Poisson-corrected noncrossover frequency (±SD) are indicated. (Bottom) Noncrossover gene conversion tracts on the B6 (F) and A/J (G) chromosomes. (H) Crossovers in females. Crossover breakpoint maps (top graph) and cumulative crossover distributions (bottom graph) are shown for the B6-to-A/J orientation (light red) and the A/J-to-B6 orientation (dark red) of allele-specific PCR. (I) Noncrossovers in females. Relative noncrossover frequencies and noncrossover gene conversion tracts from PCR in the universal-to-B6 orientation are presented as for males.

Figure 2.

Recombination at the distal hot spot in males and females. (A) Similar distributions of crossover breakpoints in males in both orientations. Crossover breakpoint maps and cumulative crossover distributions are shown as in Figure 1C. (B) Difference in crossover formation between males and females. Data from both orientations of the allele-specific PCR were pooled for males and females, presented as in Figure 1D. No crossover molecules were recovered from oocyte DNA samples. (C) Total noncrossovers in males and females, presented as in Figure 1E. (D–G). Relative noncrossover frequencies and noncrossover gene conversion tract distributions in males (D,E) and females (F,G) on the B6 (D,F) and A/J (E,G) chromosomes, presented as in Figure 1F.

Figure 3.

Estimates of relative DSB activity are an unreliable predictor of crossover frequency. (A) Comparison of SSDS reads with crossover distributions at the central and distal hot spots. Crossover breakpoint maps from spermatocytes (pooled from both orientations of allele-specific PCR) are shown in gray. Forward strand SSDS reads are shown in dark green, and reverse strand reads are in light green; total read counts (a measure of DSB activity) are indicated (data from Brick et al. 2012). The midpoint between accumulations of the forward and reverse strand reads marks the hot spot center. Note that A/J and B6 share the same Prdm9 allele, and the symmetry of crossover maps (Figs. 1C, 2A) indicates that recombination initiation occurs at comparable frequencies on both haplotypes. Thus, the SSDS and crossover maps are directly comparable even though they were generated from animals of different strain backgrounds (pure B6 for SSDS vs. A/J × B6 F1 hybrids for crossovers). Note that the Y-axis scales are the same for both hot spots. (B) Comparison of SSDS read counts and crossover frequencies at published mouse hot spots. SSDS data are from Brick et al. (2012). Filled blue circles are the hot spots from this study, filled black circles denote published hot spots assayed by allele-specific PCR of sperm DNA, and the open circle denotes a published hot spot assayed for crossing over by pedigree analysis (see Supplemental Table S5 for details). The dotted line is a least squares regression line fitted to the log transformed data. The orange box marks a group of hot spots that are within a factor of two of each other for SSDS signal but cover a much wider range in crossover activity. Note that different studies used different methods to correct recombination assays for amplifications efficiencies, so measured crossover frequencies may be underestimated to different degrees for specific hot spots (see the Materials and Methods for further details).