Joint control of meiotic crossover patterning by the synaptonemal complex and HEI10 dosage

Abstract

Meiotic crossovers are limited in number and are prevented from occurring close to each other by crossover interference. In many species, crossover number is subject to sexual dimorphism, and a lower crossover number is associated with shorter chromosome axes lengths. How this patterning is imposed remains poorly understood. Here, we show that overexpression of the Arabidopsis pro-crossover protein HEI10 increases crossovers but maintains some interference and sexual dimorphism. Disrupting the synaptonemal complex by mutating ZYP1 also leads to an increase in crossovers but, in contrast, abolishes interference and disrupts the link between chromosome axis length and crossovers. Crucially, combining HEI10 overexpression and zyp1 mutation leads to a massive and unprecedented increase in crossovers. These observations support and can be predicted by, a recently proposed model in which HEI10 diffusion along the synaptonemal complex drives a coarsening process leading to well-spaced crossover-promoting foci, providing a mechanism for crossover patterning.

Fig. 1. Massive increase in crossovers through combination of zyp1 mutation and HEI10 overexpression.

A MLH1 foci in Col wild type and zyp1 HEI10^{oe homo} meiocytes. Following immunolocalization, REC8 (Purple) and HEI10 (Supplementary Fig. 1) were imaged with STED while MLH1 (green) was imaged with confocal microscopy. The maximum intensity projection is shown. Scale bar = 1 µm. B Corresponding MLH1-HEI10 foci quantification, in female and male, inbred Col and hybrid Col/Ler. The HEI10 transgene originates from the C2 line and is either homozygous (HEI10^{oe het}) or heterozygous (HEI10^{oe homo}). Each dot is an individual cell,; circles and triangles are females and males, respectively. the mean is indicated by a bar and a number on the top; The number of analyzed cells are indicted into brackets. P values are from one-way ANOVA followed by Fisher’s LSD test. C Experimental design for construction of female and male hybrid populations for sequencing. Created with BioRender.com. D The number of COs per chromatid set transmitted by female and male gametes of wild type, HEI10^oe, zyp1, and zyp1 HEI10^oe. Each point is a BC1/gamete, circles and triangles are females and males, respectively. The means are indicated by horizontal dashed lines and numbers on the top. P values are from one-way ANOVA followed by Fisher’s LSD test. The population size is shown in parentheses. E Correlation analysis between mean number of COs per transmitted chromatid and chromosome size (Mb). Error bars are the 90% confidence intervals of the mean. Pearson’s correlation coefficients are shown in parentheses. The sample sizes, n, are identical to panel D. F Genotypes are shown for representative transmitted chromatid sets in wild type and mutants, and for extreme cases in zyp1 HEI10^oe. Centromere positions are indicated by white points. Source data are provided as a Source Data file.

Fig. 2. CO distribution and interference analysis in female and male wild type, HEI10^oe, zyp1, and zyp1 HEI10^oe.

The distribution of COs on chromosome 1 in A female and B male of wild type, HEI10^oe, zyp1, and zyp1 HEI10^oe. The other chromosomes and genomic features are shown in Supplementary Fig. 4. C–F Distribution of distances between two COs for chromosomes with exactly two COs (Supplementary Fig. 7). The gray bar represents the expected distribution of COs without interference, calculated by permutation analysis of COs (see methods). The number of analyzed CO pairs and the p value from the two-sided Mann–Whitney test between the expected and observed are indicated. G–J CoC curves in female and male meiosis of wild type, HEI10^oe, zyp1, and zyp1 HEI10^oe, respectively. Chromosomes were divided into 13 intervals, for calculating the mean coefficient of coincidence of each pair of intervals. Source data are provided as a Source Data file.

Fig. 3. Analysis of meiotic and fertility defects.

A–F DAPI-stained meiotic chromosome spreads from Col/Ler male meiocytes in wild type (A, B) and zyp1 HEI10^oe (C–F). A, C, E Metaphase I. B, D, F Metaphase II. C, D Normal chromosome configurations in zyp1 HEI10^oe. E, F Rare abnormal chromosome configurations in zyp1 HEI10^oe. Scale bar = 10 µm. G, H Representative cleared fruits of wild type Col and zyp1 HEI10^oe mutants. I Corresponding quantification of fertility. Each dot represents the fertility of an individual plant, measured as the number of seeds per fruits averaged on ten fruits. The red bar shows the mean. All plants were siblings grown together in a growth chamber. The number n of analyzed plants is indicated and P values are one-way ANOVA followed by Fisher’s LSD test. J The percentage of aneuploid samples detected in each population (Supplementary Figs. 9 and 10). The proportion of aneuploid samples in each population is shown on top of the bars. Source data are provided as a Source Data file.

Fig. 4. Analysis of SC/axis lengths in female and male meiocytes.

A–D REC8 immunolocalization in female and male meiocytes of wild type and zyp1 HEI10^{oe homo} (Col). Imaging was done with 3D-STED and the projection is shown. Scale bar = 1 µm. E–H REC8 signal was traced in 3D. Each bivalent pair is color-coded. I–L Individual trace of the longest chromosome (presumably chromosome 1), with start-to-end color code. M Measurement of the total SC length. Each dot is the SC length of an individual cell. Circles and triangles are females and males, respectively. The bars indicate the mean. One-way ANOVA followed by Sidak correction showed that SCs were systematically longer in males than in females (p < 0.000001). The same test did not detect any differences between any of the pairs of males of different genotypes (p > 0.7). For females, none of the pairwise comparisons were significantly different (p > 0.13) except in Col/Ler HEI10^oe that was lower than Col/Ler zyp1 (p = 0.006) and Col zyp1 (p = 0.008). Note that variations in slide preparation and exact meiotic stage may affect this result. The number n of analyzed cells in indicated. N Correlation analysis between the mean number of COs per chromosome and SC length (µm) in Col/Ler background. SCs were attributed to specific chromosomes based on their length (e.g., the longest was presumably chromosome 1). Pearson’s correlation coefficients are shown in parentheses. The number of samples in shown in Fig. 4M for SC length and in Fig. 1D for crossovers per chromatid. The relationship between the mean number of MLH1 foci per cell and total SC length per cell in O Col background and P Col/Ler background. The number of samples in shown in Fig. 4M for SC length and in Fig. 1B for MLH1 foci per cell. Q The relationship between the mean number of COs and SC length in Col/Ler background. The number of samples in shown in Fig. 4M for SC length and in Fig. 1D for crossovers per chromatid set. The 90% confidence intervals are indicated as error bars. Source data are provided as a Source Data file.

Fig. 5. Model of crossover patterning via HEI10 coarsening.

HEI10 (red) is captured at the middle of the SC and coarsens into large pro-CO foci. The number of large pro-CO foci is determined by SC length (heterochiasmy), and HEI10 expression levels. HEI10 overexpression increases CO number, and weakens interference but maintains heterochiasmy. In absence of an SC (zyp1), HEI10 is exchanged directly between the foci and the nucleoplasm abolishing both interference and heterochiasmy, and the number of foci depends on HEI10 expression level. Created with BioRender.com.

Fig. 6. A coarsening model for crossover designation explains the measured data.

A Number of MLH1 foci predicted by the model compared to the experimental measurements shown in Fig. 1B. Error bars denote 90% confidence of the mean. B Number of COs per chromatid predicted by the model compared to the experimental measurements shown in Fig. 1D. The respective chromatids are labeled and error bars denote 90% confidence of the mean. C–H Predicted distributions of distances between two COs for chromatids with exactly two COs; compare to Fig. 2C–F. Means are indicated by vertical dashed lines. I–L Predicted coefficient of coincidence curves; compare to Fig. 2G–J. A–L Numerical details are given in the Methods. Numerical predictions were determined from $n=10000$ (except $n=1000$ in panel A) independent repetitions.