MutLγ enforces meiotic crossovers in Arabidopsis thaliana

Abstract

During meiosis, each chromosome pair experiences at least one crossover (CO), which directs their balanced segregation in addition to shuffling genetic information. COs tend to be away from each other, a phenomenon known as CO interference. The main biochemical pathway for CO formation, which is conserved in distant eukaryotes, involves the ZMM proteins together with the MLH1-MLH3 complex (MutLγ). Here, we aim to clarify the role of MutLγ in CO formation in Arabidopsis thaliana. We show that AtMutLγ is partially dispensable for ZMM-dependent CO formation. HEI10 large foci-that mark CO sites in wild-type-form at a normal level in mlh1 and mlh3 mutants, but are inefficiently maturated into COs. Mutating the MUS81 nuclease in either mlh1 or mlh3 leads to chromosome fragmentation, which is suppressed by further mutating the zmm msh5. This suggests that in the absence of MutLγ, recombination intermediates produced by ZMMs are resolved by MUS81, which does not ensure CO formation. Finally, CO interference is marginally affected in mlh1, which is compatible with a random sub-sampling of normally patterned CO sites. We conclude that AtMutLγ imposes designated recombination intermediates to be resolved exclusively as COs, supporting the view that MutLγ asymmetrically resolves double-Holliday junctions, yielding COs.

Graphical Abstract

Figure 1.

MLH1 and MLH3 are partially dispensable for class I CO formation. (A–D) Metaphase I chromosome spreads for both wild-type and mutant male meiosis, stained with DAPI. Pairs of homologous chromosomes connected by crossovers are referred to as bivalents, abbreviated as “b.” When crossovers are absent and homologous chromosomes are separated, they are referred to as univalents, abbreviated as “u.” Scale bar = 10 μm. (E) Quantification of the number of univalent pairs per male meiotic cell. If zero pairs of univalents were observed, the bar is not visible. The mean ±95% confidence interval is shown and the number n of analyzed cells is indicated. P values are Kuskal–Wallis and uncorrected Dunn’s test performed in Prism 10.2.0. (F–G) Immunolocalization of REC8 (purple), HEI10 (red), and MLH1 (green) at diakinesis of male meiocytes. Signal shift due to chromatic aberration was corrected using the Chromatic Aberration Wizard of the Huygens software. (F) In wild type, MLH1 and HEI10 colocalize in foci. (G) In mlh3, HEI10 foci are present (see also Fig. 3), but MLH1 foci are not detected. Scale bar = 3 μm.

Figure 2.

MUS81 becomes crucial for resolving crossovers in the absence of MutLγ. Representative images of anaphase I and metaphase II chromosome spreads from wild type (A,B) and mutant mlh1 mus81 (C, D) and mlh1 mus81 msh5 (E, F). Arrowheads indicate chromosome fragments. (G) Percentage of cells in which chromosome fragmentation was detected (black) or not (light gray). The numbers above the bar plots indicate the total number of cells analyzed for each genotype. Scale bar = 10 μm.

Figure 3.

MLH1 and MLH3 stabilize HEI10 foci. (A–I) Immunostaining of male meiocytes showing REC8 (purple) and HEI10 (green) and DNA (DAPI, gray) from pachytene to diakinesis in wild-type, mlh1, and mlh3 mutants. (J) Quantification of HEI10 foci in male meiocytes across genotypes at diplotene and diakinesis stages. Each dot represents a single cell, and the red bar indicates the mean. P values are one-way analysis of variance with uncorrected Fisher’s least significance difference. (K) Number of univalents per male meiotic cell at metaphase I. The mean ±95% confidence interval is shown, and the number n of analyzed cells is indicated.

Figure 4.

Genetic crossovers are differently affected in mlh1 and HEI10±. (A) Quantifying of the number of univalent pairs at metaphase I of male meiosis in Col/Ler F1 hybrids of three different genotypes (wild-type mlh1-/- and HEI10±). If no univalent pairs are observed, no bars are displayed. The number n of analyzed cells is indicated. Bars indicate mean ± SEM. (B) Representative images of fruits from wild-type and mlh1 mutant hybrid plants. (C) Quantification of fertility in mlh1 hybrids. Each dot represents the average number of seeds per fruit (out of 10 fruits) for an individual plant. The red line indicates the mean number of seeds per fruit for a given genotype. P values are Mann–Whitney tests. (D) Number of genetic crossovers per transmitted gamete in back-cross populations (Sina plot). Each dot corresponds to an individual female (circle) and male (triangle) gamete. Red lines indicate the mean CO number. The number of analyzed gametes and the average crossover genotype for each sex/genotype are shown. (E) Quantification of fertility in HEI10± hybrids. Each dot represents the average number of seeds per fruit (out of 10 fruits) for an individual plant. The red line indicates the mean number of seeds per fruit for a given genotype. Statistical significance was assessed using the Mann–Whitney test. (F) Quantification of HEI10/MLH1 co-foci in HEI10± male meiocytes at diplotene and diakinesis stage. Each dot represents a cell, and the horizontal red bar indicates the mean. P values are Mann–Whitney test.

Figure 5.

Crossover distribution and interference in mlh1 and HEI10±. (A) Distribution of crossovers along chromosomes in females (top) and males (bottom) of wild type, mlh1, and HEI10±. A window size of 2 Mb with a step size of 50 kb was used for plotting. The Y axis is CO frequency in centimorgan per Mb. The X-axis is the position in Mb (TAIR10). The vertical dashed line marks the position of the centromere. To gain power wild-type data from previous experiments were pooled with wild types from this study (Supplementary Fig. S2). The number of analyzed samples n (BC1 plants, or transmitted gamete) are indicated for each genotype. Intervals (2 Mb) that are significantly different between wild type and mutants are indicated by stars (chi-square with FDR correction for multiple testing). (B) CoC curves for female and male meiosis. The CoC is displayed on the Y-axis, while the X-axis shows the inter-interval distance in Mb. A CoC value of one suggests independent CO occurrence, whereas a value near zero indicates crossover interference. The analyzed samples are the same as in panel (A). (D, E) Interference length (L_int), which measures the shift in inter-CO distances due to interference. Higher L_int values indicate stronger crossover interference. Each dot is the L-int for an individual chromosome. Tests are two-tailed Mann–Whitney test.