RAD51 supports DMC1 by inhibiting the SMC5/6 complex during meiosis

Abstract

Meiosis is a fundamental process for sexual reproduction in most eukaryotes and the evolutionarily conserved recombinases RADiation sensitive51 (RAD51) and Disrupted Meiotic cDNA1 (DMC1) are essential for meiosis and thus fertility. The mitotic function of RAD51 is clear, but the meiotic function of RAD51 remains largely unknown. Here we show that RAD51 functions as an interacting protein to restrain the Structural Maintenance of Chromosomes5/6 (SMC5/6) complex from inhibiting DMC1. We unexpectedly found that loss of the SMC5/6 partially suppresses the rad51 knockout mutant in terms of sterility, pollen inviability, and meiotic chromosome fragmentation in a DMC1-dependent manner in Arabidopsis thaliana. Biochemical and cytological studies revealed that the DMC1 localization in meiotic chromosomes is inhibited by the SMC5/6 complex, which is attenuated by RAD51 through physical interactions. This study not only identified the long-sought-after function of RAD51 in meiosis but also discovered the inhibition of SMC5/6 on DMC1 as a control mechanism during meiotic recombination.

Figure 1

Loss of function of SMC5/6 partially suppresses fertility and pollen inviability of the rad51 mutant in a DMC1-dependent manner. A, The representative pictures of siliques of WT, dmc1, rad51, sni1-2, asap1, rad51 sni1-2, rad51 asap1, rad51 sni1-2 dmc1, and rad51 asap1 dmc1. Bar = 0.5 cm. A total of 30 siliques from different plants per genotype were examined. B, The length of siliques. C, The seed numbers per silique. The data in (B) and (C) are represented as means ± sd (n = 30). The statistical significance was determined by Student’s t test (two-tailed; ***P < 0.001). D, Representative images of Alexander staining of pollen. Red was viable, dark blue or purple were dead. Bar = 50 μm. A total of 20 siliques from different plants per genotype were examined. E, Quantification of viable pollen grains per anther. The data are represented as means ± sd (n = 20). The statistical significance was determined by Student’s t test (two-tailed; ns, not significant; ***P < 0.001)

Figure 2

Chromosomes morphologies stained with DAPI. Representative image showing meiotic chromosome feature of WT, rad51, sni1-2, asap1, rad51 sni1-2, rad51 asap1, rad51 sni1-2 dmc1, and rad51 asap1 dmc1 at specific stages labeled in each panel above. The arrows indicate chromosome fragments. Bar = 5 μm. At least 10 meiocytes from different plants per genotype were examined

Figure 3

Loss of function of SMC5/6 partially suppresses the chromosome fragmentation of the rad51 mutant in a DMC1-dependent manner. A, Representative meiotic chromosomes in anaphase I of WT, dmc1, rad51, sni1-2, asap1, rad51 sni1-2, rad51 asap1, rad51 sni1-2 dmc1, and rad51 asap1 dmc1 observed by chromosome spread with centromere FISH. Yellow Arabic number in each image showing the number of chromosome fragments. Blue color refers to the DAPI stained chromosomes. The red dots refer to the centromere signals. Bar = 5 μm. At least 23 meiocytes from different plants per genotype were examined. B, The percentage of cells with different number of chromosome fragments. Four levels of chromosome fragmentation (0, 1–5, 6–10, >10) were used according to the number of chromosome fragments per cell. The statistical significance was determined by Chi-squared test (two-tailed; ***P < 0.001)

Figure 4

Dual immunostaining of the ASY1 and ZYP1 of synaptonemal complex in zygotene and pachytene meiocytes of WT, rad51, sni1-2, rad51 sni1-2, asap1, and rad51 asap1. In both panels of zygotene and pachytene meiocytes, the left column in magenta refers to the ASY1 signals, the second column in green refers to the ZYP1 signals, the third column merges the left two columns, and the right column merges the ASY1, ZYP1, and DAPI stained chromosomes. Bar = 5 μm. At least 11 meiocytes from different plants per genotype were examined

Figure 5

Loss of function SMC5/6 restores the DMC1 foci in rad51 mutant. A, The colocalization of DMC1 and ZYP1 in zygotene meiocytes of WT, rad51, sni1-2, rad51 sni1-2, asap1, and rad51 asap1. Bar = 5 μm. At least 22 meiocytes from different plants per genotype were examined. B, Quantification of DMC1 foci per cell. The data are represented as means ± sd (n ≧ 22). The statistical significance was determined by Student’s t test (two-tailed; ***P < 0.001). C, The colocalization of γ-H2AX and ZYP1 in zygotene meiocytes of WT, rad51, sni1-2, rad51 sni1-2, asap1, and rad51 asap1. Bar = 5 μm. At least 25 meiocytes from different plants per genotype were examined. D, Quantification of γ-H2AX foci per cell. The data are represented as means ± sd (n ≧ 25). The statistical significance was determined by Student’s t test (two-tailed; ns, not significant)

Figure 6

DMC1, RAD51, ASAP1, and SNI1 interact with each other. A–C, In vitro pull-down assays. The recombinant GST, GST-SNI1, GST-ASAP1, or GST-RAD51 proteins were coupled to glutathione beads and incubated with the recombinant DMC1-His or RAD51-His proteins. After washing, the beads were subjected to western blotting using anti-His or anti-GST antibodies. The experiments were repeated three times using different batches of proteins with similar results. D, Split luciferase assays. The Agrobacterium bacteria carrying the indicated constructs were co-expressed in N. benthamiana leaves. The positive luminescence detected by a CCD camera indicates interaction. EV, empty vector. Nluc, N-terminal luciferase. Cluc, C-terminal luciferase. The experiments were repeated three times using leaves from different plants with similar results. E–G, CoIP assays. DMC1-HA or RAD51-HA was co-expressed with GFP, SNI1-GFP, ASAP1-GFP, or RAD51-GFP in Arabidopsis protoplasts. Immunoprecipitation was performed using GFP-Trap beads and western blotting was performed using anti-HA or anti-GFP antibodies. The experiments were repeated three times using proteins from independent transfection with similar results

Figure 7

RAD51 attenuates the interaction between DMC1 and SMC5/6. A, B, In vitro pull-down assays. The HisMBP-DMC1 proteins coupled with amylose resin beads were incubated with the same amount of GST-SNI1 or GST-ASAP1 proteins but different amounts of RAD51-His proteins. After washing, the beads were subjected to western blotting using anti-His, anti-GST, or anti-MBP antibodies. –, no GST-RAD51. +, ++, and +++ indicated 1, 2, or 3 volume of GST-RAD51. The experiments were repeated three times using different batches of proteins with similar results. C, D, Split luciferase assays. The Agrobacterium bacteria carrying the indicated constructs were co-expressed in N. benthamiana leaves. Nluc, N-terminal luciferase. Cluc, C-terminal luciferase. EV, empty vector. The density of luminescence was quantified using Image J. The relative density to EV group in each leaf (n = 3) was calculated. The data are represented as means ±sd (n = 3). The statistical significance was determined by Student’s t test (two-tailed; ***P < 0.001). The experiments were repeated three times using leaves from different plants with similar results

Figure 8

A simplified working model. SMC5/6, RAD51, and DMC1 can interact with each other. SMC5/6 interacts with DMC1 to inhibit its binding to DSBs. RAD51 attenuates the interaction between SMC5/6 and DMC1. In WT, the interactions between SMC5/6, RAD51, and DMC1 are balanced, which allows RAD51 and DMC1 to be loaded at the opposite end of DSB sites, thus resulting in successful strand invasion. In the rad51 mutant, SMC5/6 interacts with DMC1 to inhibit its binding to DSB, no strand invasion occurring. In the rad51 smc5/6 double mutant, the inhibition of SMC5/6 on DMC1 is abolished, which allows DMC1 to bind to both DSB ends, thereby resulting in successful strand invasion. In the rad51 smc5/6 dmc1 triple mutant, in the absence of recombinase, no strand invasion occurs