Age-Dependent Alterations in Meiotic Recombination Cause Chromosome Segregation Errors in Spermatocytes

Abstract

Faithful chromosome segregation in meiosis requires crossover (CO) recombination, which is regulated to ensure at least one CO per homolog pair. We investigate the failure to ensure COs in juvenile male mice. By monitoring recombination genome-wide using cytological assays and at hotspots using molecular assays, we show that juvenile mouse spermatocytes have fewer COs relative to adults. Analysis of recombination in the absence of MLH3 provides evidence for greater utilization in juveniles of pathways involving structure-selective nucleases and alternative complexes, which can act upon precursors to generate noncrossovers (NCOs) at the expense of COs. We propose that some designated CO sites fail to mature efficiently in juveniles owing to inappropriate activity of these alternative repair pathways, leading to chromosome mis-segregation. We also find lower MutLγ focus density in juvenile human spermatocytes, suggesting that weaker CO maturation efficiency may explain why younger men have a higher risk of fathering children with Down syndrome.

Keywords: DNA repair; aneuploidy; chromosome segregation; crossing over; gene conversion; germ cells; homologous recombination; meiosis; noncrossovers; spermatocytes.

Figure 1. 1^st round spermatocytes have altered processing of recombination intermediates and fewer crossovers

(A) DMC1 foci at zygonema (N=2). Mean ± SD: Adult 230 ± 73, 1^st round 216 ± 64. (B) RPA2 foci at pachynema and quantification of RPA2 foci at zygonema and pachynema; EZ, early zygonema, LZ, late zygonema, P, pachynema, (N=2). Mean ± SD: Adult EZ 290 ± 78, Adult LZ 264 ± 63, Adult P 163 ± 54, 1^st round EZ 164 ± 53, 1^st round LZ 206 ± 55, 1^st round P 98 ± 55. (C) MLH1 foci at mid-pachynema; white arrowheads, unsynapsed X and Y chromosomes, (adults N=5, 1^st round N=3). Mean ± SD: Adult 22.9 ± 2.4, 1^st round 21.8 ± 1.9. (D) Metaphase I cells and quantification of the fraction of nuclei containing univalents (any), sex chromosome univalents only (XY), autosome univalent only (Auto.), and sex and autosome univalents (XY+auto); black arrowheads, unpaired X and Y chromosomes (univalents), arrows, unpaired small autosomes (adults N=7, 1^st round N=3). P value is Fisher’s exact test, two tailed. (E) Left, fraction of nuclei in which all bivalents have MLH1 foci (0) or one, two, three, and four or more bivalents lack an MLH1 focus in adult and 1^st round spermatocytes. Right, fraction of bivalents lacking an MLH1 focus on the longest (Chr1 and Chr2) and shortest (Chr16, 17, 18, 19) chromosomes, (N=3). P value is Fisher’s exact test, two tailed. Note: sample size for the longest autosomes is insufficient to determine whether there is more bivalents lacking an MLH1 focus in 1^st round compared to adult spermatocytes. Scale bar 10 μm. Except where noted, P values, Mann-Whitney, two tailed. For multiple comparisons, Kruskal-Wallis test followed by Dunn’s correction. Black bars are mean ± SD and numerical values are reported above.

Figure 2. Molecular and cytological analysis of purified 4C spermatocytes from multiple rounds of spermatogenesis

(A) Representative FACS profiles of testicular cells from Mlh3^+/−and Mlh3^−/− animals; note the lack of Mlh3^−/− 2C and 1C populations. (B) Molecular assays for detecting recombination outcomes. Testicular cells from BxD F1 hybrid mice were isolated and subjected to FACS to isolate 4C cells. Extracted DNA was used in CO and NCO assays; filled circles, polymorphisms on the B (blue) or D (red) chromosomes; open circles, either B or D polymorphisms; blue/red arrowheads, primers specific to either the B or D allele, respectively; black arrowheads, universal primers (U). (C) Comparison of CO and NCO frequency and distribution at A3 (N=6) and 59.5 (N=2) in adult Mlh3^+/− 4C spermatocytes. Top, CO breakpoint frequency. Bottom, NCO gene conversion frequency at specific polymorphisms. Top ticks, genotyped polymorphisms. Frequency is mean ± SD. cM/Mb, centimorgans per megabase. (D) Diagram depicting key events in testicular development in the 1^st three rounds of spermatogenesis. Color change from rose to blue symbolizes a temperature shift from 37°C to 33°C. (E) RPA2 foci at pachynema during the indicated rounds of spermatogenesis (N=2). Mean ± SD: Adult 163 ± 54, 1^st round 98 ± 55, 2^nd round 94 ± 46, 3^rd round 145 ± 58. (F) MLH1 foci during the indicated rounds of spermatogenesis (adult N=5, 1^st round N=3, 2^nd round N=7, 3^rd round N=2). Mean ± SD: Adult 22.9 ± 2.4, 1^st round 21.8 ± 1.9, 2^nd round 20.7 ± 2.5, 3^rd round 21.2 ± 2.3. P values, Kruskal-Wallis test followed by Dunn’s correction. (G) Fraction of nuclei in which all bivalents have MLH1 foci (0) or one, two, three, and four or more bivalents lack an MLH1 focus in the indicated rounds of spermatogenesis. Black bars are mean ± SD. Adult and 1^st round data are reproduced from Figure 1. See also Figure S1.

Figure 3. Higher SSN activity in 1^st round spermatocytes

(A) Model of the proposed DSB repair pathways during meiotic recombination. SDSA, synthesis-dependent strand annealing; dHJ, double-Holliday junction; SSNs, structure-selective nucleases. (B) DMC1 foci at zygonema (N=2, except Mlh3^−/− 1^st round N=3) and RPA2 foci at pachynema (WT N=2, Mlh3^−/− N=3) in WT (gray) and Mlh3^−/− (blue) adult and 1^st round spermatocytes. DMC1 foci mean ± SD: Adult Mlh3^+/+ 230 ± 73, Adult Mlh3^−/− 224 ± 60, 1^st round Mlh3^+/+ 216 ± 64, 1^st round Mlh3^−/− 244 ± 51. RPA2 foci mean ± SD: Adult Mlh3^+/+ 163 ± 54, Adult Mlh3^−/− 152 ± 51, 1^st round Mlh3^+/+ 98 ± 55, 1^st round Mlh3^−/− 117 ± 52. WT data is reproduced here from Figure 1. P values, Kruskal-Wallis test followed by Dunn’s correction. Black bars are mean ± SD. (C) Top, metaphase spreads from the indicated stages and genotypes. Black arrowheads, bivalents. Bottom left, bivalents in metaphase I (adult N=7, Mlh3^−/− adult N=4, Mlh3^−/− 1^st round N=4). Mean ± SD: Adult 20 ± 0.3, Adult Mlh3^−/− 1.7 ± 1.2, 1^st round Mlh3^−/− 1.9 ± 1.6. Bottom right, distribution of bivalents per nucleus in Mlh3^−/− adult and 1^st round spermatocytes. P value, Chi-squared, two-tailed. Black bars are mean ± SD. See also Figure S2 and S3.

Figure 4. Altered frequency and distribution of noncrossovers at 59.5 with age and genotype

(A) Top, histogram of total NCO frequency at specific polymorphisms (n_tot) (adult N=2, 1^st round N=6, 3^rd round N=4); Bottom, plots of representative NCOs (n_plot) from the indicated ages in WT and Mlh3^+/− spermatocytes. All NCOs observed, regardless of age, are short conversions of a single polymorphism (singletons). Here and elsewhere unless noted, the number of n_plot shown and corresponding length of the plotted maps is proportional to the NCO frequencies. (B) Frequency and distribution of singleton NCOs were unaltered between spermatocytes with and without MLH3 (compare A with B). (C) In Mlh3^−/− males, a second class of long NCOs were also observed. 1^st and 3^rd round spermatocytes had more long NCOs. Mean average tract lengths ± SD are indicated. Asterisks, cloned NCOs. See also Figure S5, S6, and S7.

Figure 5. Juvenile human spermatocytes have a lower density of MLH1 foci

(A) Representative images at the indicated stages and quantification of MLH1 foci at pachynema for individual A (A), B (B), and the average (A & B). Mean ± SD: A 56.9 ± 6.5, B 54.3 ± 8.4. Black bars are mean ± SD. (B) Comparison of spermatocytes from adults, oocytes, and spermatocytes from juveniles for MLH1 foci and total SC per nucleus, mean ± SD, and MLH1 foci per μm SC (MLH1 focus density), mean ± confidence interval. P values for MLH1 foci and total SC are derived by ANOVA with Tukey HSD Post-hoc test. P values for MLH1 focus density are derived from the confidence interval. Data from oocytes and adult spermatocytes are from (Gruhn et al., 2013; Wang et al., 2017). (C) Model of juvenile crossover maturation inefficiency. In adults, designated CO precursors are efficiently matured into COs by MLH1/3-dependent resolution. In juveniles, a fraction of designated CO precursors may be inappropriately targeted for 1: resolution by SSNs to generate both COs and NCOs or 2: acted upon by alternative NCO-specific pathways. Shorter chromosomes usually have only a single designated CO between homologs, whereas longer chromosomes often have more than one. Thus, CO maturation inefficiency will disproportionately cause mis-segregation of shorter chromosomes.