FIGL1 attenuates meiotic interhomolog repair and is counteracted by the RAD51 paralog XRCC2 and the chromosome axis protein ASY1 during meiosis

Abstract

Two recombinases, RAD51 and DMC1, catalyze meiotic break repair to ensure crossovers (COs) between homologous chromosomes (interhomolog) rather than between sisters (intersister). FIDGETIN-LIKE-1 (FIGL1) downregulates both recombinases. However, the understanding of how FIGL1 functions in meiotic repair remains limited. Here, we discover new genetic interactions of Arabidopsis thaliana FIGL1 that are important in vivo determinants of meiotic repair outcome. In figl1 mutants, compromising RAD51-dependent repair, either through the loss of RAD51 paralogs (RAD51B or XRCC2) or RAD54 or by inhibiting RAD51 catalytic activity, results in either unrepaired breaks or meiotic CO defects. Further, XRCC2 physically interacts with FIGL1 and partially counteracts FIGL1 activity for RAD51 focus formation. Our data indicate that RAD51-mediated repair mechanisms compensate FIGL1 dysfunction. FIGL1 is not necessary for intersister repair in dmc1 but is essential for the completion of meiotic repair in mutants such as asy1 that have impaired DMC1 functions and interhomolog bias. We show that FIGL1 attenuates interhomolog repair, and ASY1 counteracts FIGL1 to promote interhomolog recombination. Altogether, this study underlines that multiple factors can counteract FIGL1 activity to promote accurate meiotic repair.

Keywords: DMC1; FIGL1; RAD51; chromosome axis; crossovers; meiotic DNA repair.

Fig. 1

Comparison of fertility in the figl1 mutant combined with mutations in genes regulating DMC1/RAD51 or the meiotic chromosome axis in Arabidopsis thaliana. Each colored dot represents one plant with the average number of seeds per fruit counted from at least 10 fruits. The mean and the SD are represented by the black bars for each genotype. Each figl1 double mutant combination is compared with wild‐type sister plants and respective single mutants that were cultivated together in a segregating population. Mean and error bars representing SD for each genotype are presented in black. The P values shown were computed using an unpaired Kruskal–Wallis test corrected with Dunn's test for multiple comparisons. ns, nonsignificant (P > 0.01). *, P < 0.01; **, P < 0.001; ***, P < 0.0001.

Fig. 2

Downregulation of the RAD51‐dependent pathway without FIDGETIN‐LIKE‐1 (FIGL1) alters meiotic break repair outcomes. (a) Representative metaphase I and anaphase II images of DAPI‐stained chromosome spreads of male meiocytes are shown in rad51b, figl1 rad51b, xrcc2, figl1 xrcc2, rad51 RAD51‐GFP, figl1 rad51 RAD51‐GFP, rad54, and figl1 rad54. Bars, 5 μm. (b) Quantification of bivalents and aberrant repair in metaphase I. The histograms show the percentage of metaphase cells exhibiting bivalents and chromosome fragmentation. The number of analyzed cells (n) and the mean number of bivalents per cell are indicated above each bar. (c) The number of HEI10‐MLH1 cofoci is reduced in figl1 rad54, but not in figl1 rad51 RAD51‐GFP. Representative images of HEI10 (red) and MLH1 (green) colocalization in wild‐type, figl1, rad54, rad51 RAD51‐GFP, figl1 rad51 RAD51‐GFP, and figl1 rad54. Bars, 2 μm. (d) Quantification of HEI10‐MLH1 cofoci. Each dot represents a HEI10‐MLH1 cofocus in an individual cell, and the red bars represent the mean and SD for each genotype. n, number of cells analyzed. The P values shown were calculated for unpaired Kruskal–Wallis tests corrected with Dunn's test for multiple comparisons. (e) 3D structure model of interaction between Arabidopsis XRCC2 (in green) and FIGL‐FRBD domain (in magenta) by Alphafold2 structural analysis. An enlarged 3D view of the binding interface between XRCC2 and the FIGL1‐FRBD domain shows the residues involved. The dark and light blue lines indicate the high confidence predicted aligned error (PAE) scores for residues with < 8 Å distances, as provided by Alphafold2. (f) Yeast two‐hybrid assays showing the interaction of Arabidopsis FIGL1 with XRCC2 and DMC1 (used as positive control). Proteins of interest were fused with the Gal4 DNA binding domain (BD, left) and with the Gal4 activation domain (AD, right), respectively, and co‐expressed in yeast cells with selection on SD/−LW (− Leu −Trp) for diploid strains and on SD/−LWH (− Leu −Trp −His) as well as SD/−LWHA (− Leu −Trp −His −Ade) for protein interaction.

Fig. 3

Analysis of the distribution of RAD51 or DMC1 in various figl1 mutants. (a) Dual immunolocalization of REC8 (magenta) and RAD51 (green) on male meiocytes from wild‐type (WT) (Col‐0), figl1, xrcc2, figl1 xrcc2, asy1, figl1 asy1, and WT, figl1, hop2‐2, figl1 hop2‐2 in the A. thaliana Col‐0/Ws hybrid background. (b) Dual immunolocalization of REC8 (magenta) and DMC1 (green) in male meiocytes from WT, figl1, hop2‐2, and figl1 hop2‐2 in the Col‐0/Ws hybrid background. (c) Quantification of RAD51 and DMC1 foci in different WTs and mutants. The mean and error bars representing SD for each genotype are presented in black. The P values shown were calculated using unpaired Kruskal–Wallis tests corrected with Dunn's test for multiple comparisons.

Fig. 4

Chromosome spreads display the completion of meiotic break repair in A. thaliana wild‐type (WT), dmc1, figl1, figl1 dmc1, and haploid figl1 mutant plants. (a) Representative metaphase I images of DAPI‐stained chromosome spreads of male meiocytes are shown for WT, figl1, dmc1, and figl1 dmc1. b, bivalent or bivalent‐like; u, univalent. Bars, 5 μm. Figs 4(a) and 5(a) share the same metaphase image of figl1 genotype for a direct comparison with other genotypes. (b) Quantification of bivalents at metaphase I. The histogram shows the proportion of cells categorized based on the presence of bivalents. The mean number of bivalents per cell and the number of analyzed cells are indicated above each bar. (c) Metaphase I and II images of DAPI‐stained chromosome spreads from haploid WT and figl1 plants. Bars, 5 μm.

Fig. 5

Epistatic analysis of A. thaliana mnd1, hop2‐2, and figl1. (a) Representative metaphase I and anaphase II images of DAPI‐stained chromosome spreads of male meiocytes are shown in figl1, mnd1, hop2‐2, figl1 mnd1, and figl1 hop2‐2. Bars, 5 μm. Figs 4(a) and 5(a) share the same metaphase image of figl1 genotype for a direct comparison with other genotypes. (b) Quantification of bivalents and aberrant repair at metaphase I. The histogram shows the proportion of metaphase cells categorized based on the presence of bivalents and chromosome fragmentation. The mean number of bivalents per cell and the number of analyzed cells (n) are indicated above each bar. na, not applicable.

Fig. 6

Functional interaction of A. thaliana figl1 with asy1, asy3, and asy4. (a) Representative metaphase I and anaphase II images of DAPI‐stained chromosome spread of male meiocytes are shown in asy1, asy3, asy4, figl1 asy1, figl1/+ asy1, figl1 asy3, and figl1 asy4. Bars, 5 μm. (b) Quantification of bivalents and aberrant repair at metaphase I. The proportion of metaphase cells categorized based on the presence of bivalents and chromosome fragmentation is presented. The mean number of bivalents per cell and the number of analyzed cells (n) are indicated above each bar. na, not applied.

Fig. 7

Model for FIDGETIN‐LIKE‐1 (FIGL1) in vivo functions during meiotic break repair in A. thaliana. Blue and red lines represent homologous chromosomes, and four lines in the same color denote the two DNA strands of each sister chromatid. After the formation of double‐stranded breaks (DSB) and the resection of 5′ ends, RAD51 and DMC1 polymerize at 3′ single‐strand DNA (ssDNA) tails to form nucleoprotein homofilaments. DMC1 is proximal to the DSB site, but RAD51 is distal to the DSB site. Invasion of these homofilaments leads to the repair of meiotic breaks. FIGL1 can negatively regulate RAD51 and DMC1 activity at pre‐ and postinvasion steps. FIGL1 is not required for RAD51‐mediated intersister repair of meiotic breaks in the absence of DMC1 in plants. However, FIGL1 can antagonize RAD51‐mediated interhomolog invasions in the presence of DMC1. MND1 and HOP2 likely act upstream of FIGL1. The three meiotic axis proteins (ASY1, ASY3, and ASY4) act as positive modulators of interhomolog recombination by counteracting FIGL1 activity. In the absence of FIGL1, RAD51 is likely active for interhomolog invasions/repair in XRCC2 and RAD51B‐dependent manner. RAD54 and the catalytic activity of RAD51 are critical for ensuring the wild‐type level or distribution of meiotic CO/chiasmata formation when lacking FIGL1.