Genetic dissection of MutL complexes in Arabidopsis meiosis

Abstract

During meiosis, homologous chromosomes exchange genetic material through crossing over. The main crossover pathway relies on ZMM proteins, including ZIP4 and HEI10, and is typically resolved by the MLH1/MLH3 heterodimer, MutLγ. Our analysis shows that while MUS81 may partially compensate for MutLγ loss, its role remains uncertain. However, our multiple mutant analysis shows that MUS81 is unlikely to be the sole resolvase of ZMM-protected recombination intermediates when MutLγ is absent. Comparing genome-wide crossover maps of mlh1 mutants with ZMM-deficient mutants and lines with varying HEI10 levels reveals that crossover interference persists in mlh1 but is weakened. The significant crossover reduction in mlh1 also increases aneuploidy in offspring. The loss of MutLγ can be suppressed by eliminating the FANCM helicase. Combined with the lower-than-expected chiasma frequency, this suggests that in MutLγ absence, some ZMM-protected intermediates are ultimately resolved by DNA helicases and/or their complexes with Top3α. Elevated MLH1 or MLH3 expression moderately increases crossover frequency, while their misregulation drastically reduces crossover numbers and plant fertility, highlighting the importance for tight control of MLH1/MLH3 levels. By contrast, PMS1, a component of the MutLα endonuclease, appears uninvolved in crossing over. Together, these findings demonstrate the unique role of MutLγ in ZMM-dependent crossover regulation.

Graphical Abstract

Figure 1.

mutLγ mutants exhibit milder fertility and meiotic phenotypes than zmm mutants. (A) Representative pictures of siliques for wild-type accessions (Ws and Col), zmm mutants (zip4-2 and hei10-2), and mutLγ mutants (mlh1-1,mlh1-4,mlh3-1, and mlh3-2). Scale bar, 1 cm. Fertility assays for mutLγ mutants compared to the wild types and zmm mutants as assessed by seed set (B), silique length (C), and pollen viability (D). The P values were estimated using one way analysis of variance (ANOVA) and Tukey HSD tests (Supplementary Tables S10–S12). n = 5–10 for panels (B) and (C), and n = 3 for panel (D). Cytological characterization of metaphase I meiocytes showing the frequency of cells with 0–5 bivalents (E), along with representative chromosome spreads from metaphase I meiocytes for Col (n = 27), zip4-2 (n = 53),hei10-2 (n = 50), and mlh1-4 (n = 57) (F). Scale bar, 10 μm.

Figure 2.

Local meiotic crossover recombination is decreased in mutLγ mutants. (A) Schematic representation of crossover frequency scoring via fluorescent seed-based system. Arrowheads mark the fluorescent markers delimitating a specific genomic interval. These markers are introduced into the relevant lines via crossing, and recombination frequency is determined by the segregation of the fluorescent markers. (B) Chromosome map displaying the fluorescent intervals used in panels (C) and (D). Crossover frequency (cM) for three mlh1 alleles, two mlh3 alleles, and one pms1 allele in the chromosome 3 subtelomeric region 420 and pericentromeric region 3.9 (C), and for mlh1-4 and mlh3-1 in the subtelomeric regions of chromosome 1 (1.18), and chromosome 5 (5.1) (D). Each dot represents measurements from one individual. The center line of a boxplot marks the median; the upper and lower bounds indicate the 75th and 25th percentiles, respectively. The P values were estimated using a Welch’s t-test.

Figure 3.

The progeny of Col/Ler mlh1 mutants exhibit a significant incidence of trisomy. Fertility assays for hei10-2/+ and mlh1 mutants compared to the wild-type Col/Ler hybrids as assessed by seed set (A) and silique length (B). The P values were estimated using a Welch’s t-test with n = 6–7. (C) Ploidy analysis for the mlh1 Col/Ler F₂ population as assessed by the ploidy level for each sequenced chromosome in F₂ individuals. (D) Example plot illustrating trisomy of chromosome 4, detected based on mean sequencing coverage calculated in 100 kb windows.

Figure 4.

mlh1 mutants show a reduced number of crossover events genome-wide. (A) Differences in total crossover numbers for each chromosome in mlh1 and hei10/+ versus wild type. Mean crossover numbers are shown, while whiskers define SEs. (B) Comparative representation of crossover frequency per F₂ individual within 300 kb windows, averaged along proportionally scaled chromosome arm, orientated from telomere (TEL) to centromere (CEN). SNP density per 300 kb is shaded in gray. The mean values are shown by horizontal dashed lines. (C) Similar to panel (B) but showing crossover frequency plotted along all five Arabidopsis chromosomes, averaged in 300 kb windows. Vertical dashed lines indicate centromere positions. (D) Relative reduction in crossover frequency within chromosome arms and pericentromeres in mlh1 and hei10/+ mutants, versus wild type. Each dot represents a single 300 kb window from panel (C). The center line of a boxplot indicates the median; the upper and lower bounds indicate the 75th and 25th percentiles, respectively. The pericentromeres defined as regions with higher than average DNA methylation, which surround the centromeres. The statistical significance was assessed by one-way ANOVA followed by Tukey HSD test. (E) Genome-wide correlation coefficient matrices of crossover distributions, calculated in adjacent 300 kb windows. Data for wild type, msh2, HEI10-OE, msh2 fancm zip4, msh2 recq4, recq4 HEI10-OE, recq4, and fancm zip4 from [17, 55, 59, 70, 71].

Figure 5.

Crossover interference is weakened but maintained in mlh1. (A) cis-DCO distances calculated from parental–heterozygous–parental genotype transitions for WT, mlh1,hei10/+,HEI10-OE, and msh2 fancm zip4. The center line of a boxplot indicates the median, while the upper and lower bounds indicate the 75th and 25th percentiles, respectively. Each dot represents one cis-inter-crossover distance. The P values were estimated using Wilcoxon Signed-Rank test. (B–F) Histograms illustrating the distribution of inter-crossover distances for observed data (colored bars) versus randomly generated distances (gray bars) across five genotypes. The curves represent gamma-fitted distributions for the observed data (colored curves) and the expected data (gray curves). To assess the statistical significance of the differences between the observed and expected distributions, bootstrap resampling was applied to estimate the shape v parameter, and the Mann–Whitney U test was used to calculate significance. (B) Wild type (n = 1139), (C) mlh1 (n = 63), (D) hei10/+ (n = 88), (E) HEI10-OE (n = 404), and (F) msh2 fancm zip4 (n = 194). Median inter-crossover distances for each genotype are marked with dashed lines in matching colors. Data for panels (B), (E), and (F) from [17, 55, 59]. (G) Coefficient of coincidence (CoC) analysis. The CoC is shown on the Y-axis, with the inter-interval distance in Mb on the X-axis. A CoC value of 1 indicates independent crossover occurrence, while a value near 0 suggests crossover interference.

Figure 6.

MUS81 endonuclease partially compensates for the loss of MutLγ.Fertility as assessed by seed set (A) and pollen viability (B). The P values were estimated using one-way ANOVA and the Tukey HSD tests (Supplementary Tables S13 and S14). Sample sizes: n = 5–11 in panel (A), and n = 3 in panel (B). (C–E) Cytological characterization of metaphase I meiocytes. (C) Chromosome representations of 5S and 45S rDNA positions in Col and Ws genetic backgrounds. (D) Representative pictures of chromosome spreads of metaphase I meiocytes stained with DAPI and labeled with FISH against 5S (red) and 45S (green) rDNA. Scale bar, 10 μm. (E) Frequency of ring, rod, and univalent chromosome configuration in submetacentric chromosomes 1, 3, and 5, acrocentric chromosomes 2 and 4, and the average across all chromosomes. The number of characterized meiocytes for each genotype is as follows: Col (n = 69), Ws (n = 50), mlh1-1 (n = 53), mlh3-1 (n = 108), mlh1-1 mlh3-1 (n = 34), mus81-2 (n = 102), mlh1-1 mlh3-1 mus81-2 (n = 44), zip4-2 (n = 255), zip4-2 mus81-1 (n = 57), fancm-9 (n = 52), hei10-2 (n = 29), and mlh1-1 mlh3-1 fancm-9 (n = 55). (F) Mean chiasma counts with 95% confidence intervals for selected genotypes. Significance was assessed using a one-way ANOVA with Tukey’s HSD test (Supplementary Table S15). Non-significant values are not displayed.

Figure 7.

Increased MutLγ expression boosts crossovers, but its missregulation drastically reduces crossover numbers and plant fertility. (A) Crossover frequency in inbred Col/Col and hybrid Col/Ler F₁ plants with additional MLH1 or MLH3 copies under their respective native promoters, measured in 420 subtelomeric region of chromosome 3. Each data point represents a measurement from one plant. Independent transformant lines for each genotype are indicated by a different data point shape. (B) As for panel (A) but measured in 3.9 pericentromeric region. (C) Crossover frequency of MLH1,MLH3, and PMS1 overexpressors under the control of DMC1 promoter at the T₁ generation, measured in 420. Each data point represents a measurement from one plant. (D) Fertility assays for MLH1 and MLH3 overexpressors, under DMC1 promoter, assessed by seed set. Each data point represents the average of five siliques from a single plant. The P values were estimated using one-way ANOVA and the Tukey HSD tests (Supplementary Table S18). (E) As in panel (D) but showing pollen viability. Each data point represents a measurement of 500 pollen grains from one plant. The P values were estimated using one-way ANOVA and the Tukey HSD tests (Supplementary Table S19). (F) Crossover frequency in inbred Col/Col and hybrid Col/Ler F₁ plants overexpressing additional MLH1 or MLH3 copies under DMC1 promoter, measured in the 420 interval. (G) Expression levels in F₁ Col × pDMC1::MLH3 plants shown in panel (F) as determined via RT–qPCR for both MLH1 and MLH3 in each plant. Each dot represents one biological replicate (one cross of a T₂ plant). Bar plots show the average fold change relative to wild type. (H) Crossover frequency in inbred Col/Col F₁ plants with additional MLH1 and MLH3 copies (double overexpressors) under their native or DMC1 promoter, measured in the 420 interval. In this figure each dot represents a measurement from one individual. For crossover frequency (A–C, F–H), the center line of a boxplot indicates the median; the upper and lower bounds indicate the 75th and 25th percentiles, respectively, and the P values were estimated using the Welch’s t-test.