2023 Volume 98 Issue 1 Pages 45-52
Meiotic recombination between homologous chromosomes is promoted by the collaborative action of two RecA homologs, Rad51 and meiosis-specific Dmc1. The filament assembly of Dmc1 is promoted by meiosis-specific Mei5–Sae3 in budding yeast. Mei5–Sae3 shows sequence similarity to fission yeast Sfr1–Swi5, which stimulates DNA strand exchanges by Rad51 as well as Dmc1. Sae3 and Swi5 share a conserved motif with the amino acid sequence YNEI/LK/RD. In this study, we analyzed the role of the YNEL residues in the Sae3 sequence in meiotic recombination and found that these residues are critical for Sae3 function in Dmc1 assembly. L59 substitution in the Sae3 protein disrupts complex formation with Mei5, while Y56 and N57 substitutions do not. These observations reveal the differential contribution of conserved YNEL residues to Sae3 activities in meiotic recombination.
Meiosis is essential to produce haploid gametes by reducing the chromosome number to half of that in diploid germ cells (Marston, 2014). Meiotic recombination promotes DNA exchange between homologous chromosomes for physical linkages between the chromosomes, which is essential for accurate segregation of homologous chromosomes in meiosis I and creates genetic diversity in gametes (Cejka et al., 2005; Hunter, 2015).
The preference for interhomolog recombination over intersister recombination is one of the prominent features in meiotic recombination, as well as the programmed formation of DNA double-strand breaks (DSBs) and the regulated formation of reciprocal exchanges called crossovers (Pyatnitskaya et al., 2019). DNA exchanges between homologous chromosomes are mediated by the collaborative actions of two RecA homologs, Rad51 and Dmc1 (Bishop et al., 1992; Shinohara et al., 1992). Both proteins bind to single-stranded DNA (ssDNA) to form a helical filament for homology search and strand exchange between ssDNA and a homologous double-stranded DNA (dsDNA) (Ogawa et al., 1993; Luo et al., 2021; Xu et al., 2021). Rad51 plays a catalytic role in intersister recombination in mitotic cells and a structural role in interhomolog recombination by helping the assembly of Dmc1 in meiotic cells (Cloud et al., 2012; Lan et al., 2020). Dmc1 catalyzes the exchange between the DNAs in meiotic recombination (Li et al., 1997; Sehorn et al., 2004).
Both assembly and disassembly of Rad51 and Dmc1 filaments are tightly regulated by various factors. Rad51 assembly on ssDNA coated with an ssDNA binding protein, replication protein A (RPA), is facilitated by various Rad51 mediators: Rad52, Rad55–Rad57 and Psy3–Csm2–Shu1–Shu2 in the budding yeast Saccharomyces cerevisiae (Sung, 1997; Shinohara and Ogawa, 1998; Sasanuma et al., 2013). On the other hand, in addition to Rad51, Dmc1 assembly on RPA-coated ssDNA is promoted by the meiosis-specific complex Mei5–Sae3 (Hayase et al., 2004; Tsubouchi and Roeder, 2004; Ferrari et al., 2009; Chan et al., 2019). How Mei5–Sae3 promotes Dmc1 filament formation is still largely unknown.
Mei5–Sae3 belongs to a family of protein complexes, whose primordial form is Sfr1–Swi5 in the fission yeast Schizosaccharomyces pombe, which is involved in both mitotic and meiotic recombination (Akamatsu et al., 2003). Sfr1–Swi5 and Mei5–Sae3 complexes are conserved from yeast to mammals (Hayase et al., 2004; Akamatsu et al., 2007; Tsai et al., 2012). Fission yeast and mouse Sfr1–Swi5 promote both Rad51- and Dmc1-mediated strand exchange in vitro by stabilizing the active ATP-bound form of Rad51 filaments (Haruta et al., 2008; Chi et al., 2009). Consistent with this, a structure study showed that the C-terminal region of Sfr1 complexed with Swi5 forms a kinked rod, which might fit into a groove of the Rad51 filament (Kokabu et al., 2011; Kuwabara et al., 2012). At the amino acid sequence level, yeast Swi5/Sae3 is a short protein of around 100 amino acids (aa): 91 aa for budding yeast Sae3 and 85 aa for fission yeast Swi5 (Hayase et al., 2004; Akamatsu et al., 2007). On the other hand, Sfr1/Mei5 is highly divergent with limited sequence conservation (Hayase et al., 2004).
Although the function of these complexes has been well studied, the contribution of conserved amino acids in these proteins to recombination has not yet been analyzed. In this study, we elucidated the role of conserved amino acid residues in Sae3 in meiotic recombination and found that a conserved YNEL sequence in Sae3 is critical for Dmc1 assembly. The possible contribution of these residues to Mei5–Sae3 function is discussed.
Amino acid sequence comparison reveals sequence conservation between budding yeast Sae3 and its fission yeast ortholog Swi5 (Akamatsu et al., 2003; Hayase et al., 2004) (Supplementary Fig. S1A). Sae3/Swi5 orthologs are found in most vertebrates (the rat genome encodes three orthologs), some invertebrates (e.g., fruit fly, sea urchin and sea squirt) and fungi including yeasts (Supplementary Fig. S1A), but seem to be absent in plants. While all orthologs contain a conserved C-terminal region of ~80 aa, some show a longer extension of the N-terminus; for example, human SWI5 protein contains 235 aa. Phylogenetic analysis showed that S. cerevisiae Sae3 belongs to a different branch of Swi5 orthologs from S. pombe Swi5 (Supplementary Fig. S1C). AlphaFold2 prediction shows a structure of Sae3 alone (Fig. 1A). The predicted Sae3 consists of two long α-helixes (α1 and α2) and a short α-helix (α3) near its C-terminus. Importantly, the predicted Sae3 structure is very similar to the structure of fission yeast Swi5 in the Swi5–Sfr1C (an N-terminally truncated version of Sfr1) complex (Kuwabara et al., 2012) (PDB, 3viq, Fig. 1B). Moreover, the predicted C-terminal region of human SWI5 is similar to both Swi5 and Sae3 (Supplementary Fig. S1D and S1E), and the N-terminal extension does not show any structure.
Predicted Sae3 structure and the relationship with Swi5–Sfr1. (A) Structure of budding yeast Sae3 as predicted by AlphaFold2 (https://alphafold.ebi.ac.uk); AF-P89114. (B) Structural comparison of predicted Sae3 (purple) with the Swi5 (blue)–Sfr1(green) complex. Structure alignment was performed by PyMOL. (C) Magnified view of the lower half of (B). Conserved Y56, N57, E58 and L59 residues are displayed in stick form. (D) Helical-wheel diagram of putative α2 of Sae3 (purple) as well as the pair of Swi5 and Sfr1 α2 helixes (Kuwabara et al., 2012). The conserved YNEL sequence is shown in red.
We compared amino acid sequences of Sae3/Swi5 proteins from various species from fungi to vertebrates (Supplementary Fig. S1A) and found that the YNEI/LK/RD sequence, comprising Sae3 residues 56–61, was highly conserved among Sae3/Swi5 orthologs (Supplementary Fig. S1B). This sequence is in the middle of the second α-helix (α2) of Sae3, and thus of Swi5, which forms the parallel α-helix (α2) of Sfr1 (Fig. 1C and 1D). This region in fission yeast Swi5 can interact not only with α2 of Sfr1 but also with the C-terminus of both Swi5 and Sfr1 (Fig. 1D), whereby four α-helixes fold with a unique packing (Kuwabara et al., 2012).
Characterization of conserved residues in Sae3 functionTo elucidate the role of these conserved amino acid residues in Sae3, we constructed the sae3-Y56A, -N57A, -E58A and -L59A mutants in the background of the C-terminal Flag-tagged SAE3 gene. Flag tagging, which enabled us to detect the protein by western blotting and immunostaining, does not affect the function of the protein in vivo (Hayase et al., 2004). Meiotic time course analysis revealed that, like the sae3 deletion mutant, the sae3-Flag-Y56A, -N57A, -E58A, and -L59A mutants (for simplicity, “-Flag” is not shown hereafter) did not enter meiosis I, suggesting an arrest in meiotic prophase I (Fig. 2A). Western blotting showed normal expression of the Sae3-Flag proteins in the sae3-Y56A, -N57A and -L59A mutants after the induction of meiosis, like that of the control Sae3-Flag (Fig. 2B). On the other hand, the sae3-E58A cells showed a low abundance of the mutant Sae3-Flag protein (Fig. 2B), indicating that the substitution of E58 affects Sae3 protein stability. We note that Sae3-E58 is located near the conserved D61 in the YNEI/LK/RD sequence (Supplementary Fig. S1A) in the predicted Sae3 structure (Supplementary Fig. S2A). A similar arrangement of E52 and D55 is seen on Swi5 (Supplementary Fig. S2B). This suggests that Sae3-61D participates in the same activity as Sae3-E58.
Mutations in the Sae3 YNEL sequence cause defective meiosis progression. (A) Meiosis I progression in various strains. SAE3-Flag, sae3Δ, and sae3-Flag-Y56A, -N57A, -E58A and -L59A mutant cells underwent induced meiosis at 0 h. Meiosis I progression was verified by DAPI staining of each cell type at the indicated times. More than 200 cells were counted at each time point. A cell with more than two DAPI bodies was classified as a cell passing meiosis I. (B) Expression of Mei5 and Sae3-Flag protein in various strains. Cell lysates at each time were examined by western blotting with anti-Mei5, anti-Flag (for Sae3) and anti-tubulin (control) antibodies.
All four sae3 mutants expressed Mei5 protein normally during meiosis as seen in the control and the sae3 deletion cells (Fig. 2B).
Sae3 YNEL is necessary for Dmc1 assembly, but not for Rad51 assemblyWe characterized the localization of proteins involved in meiotic recombination on chromosome spreads in the sae3-Y56A, -N57A, and -L59A mutants. Immunostaining showed punctate staining (foci) of two RecA homologs, Rad51 and Dmc1 proteins, on meiotic chromosome spreads (Bishop, 1994; Hayase et al., 2004). In the control diploid cells (SAE3-Flag), Rad51 and Dmc1 foci appear at 3 h of incubation in sporulation medium (SPM), peak at 4 h, and gradually disappear by 6 h (Fig. 3A and 3B), when meiosis I begins (Fig. 2A). The sae3 deletion mutant is proficient in Rad51 focus formation but deficient in Dmc1 focus formation (Hayase et al., 2004; Tsubouchi and Roeder, 2004). Like the control, all three sae3 mutants showed normal appearance of Rad51 foci but accumulated the foci at late times (Fig. 3A and 3B), indicating that the three residues of Sae3 are critical for the turnover of Rad51 ensembles. Indeed, the sae3-N57A and -L59A mutants were almost completely defective in Dmc1 focus formation at 4 h (Fig. 3A and 3B). On the other hand, the sae3-Y56A mutant showed a delay in Dmc1 focus appearance: it did not show Dmc1 foci at 3 or 4 h but started to form Dmc1 foci at 5 h and later. In this mutant, Dmc1 foci did not show turnover by 12 h. The sae3-N57A mutant also did not show Dmc1 foci at early time points, but foci transiently formed at 8 h. These results suggest a critical role both for Sae3-N57 and -L59 in Dmc1 assembly and for Sae3-Y56 in efficient and timely Dmc1 assembly as well as disassembly. The Dmc1 disassembly defect in the sae3-Y56A mutant may be due to the formation of a defective Dmc1 complex.
YNEL motif in Sae3 is necessary for Dmc1 assembly. (A) Meiotic chromosome spreads were immunostained with Rad51 and Dmc1. Representative images of Rad51 (green) and Dmc1 (red) staining, with a merged image with DAPI (blue), at 4 h in each strain and at 8 h in the sae3-Y56A and -N57A mutant strains are shown. Bar, 1 μm. (B) Kinetic analysis of focus-positive cells for Rad51. Spreads containing more than five foci were counted. At each time point, more than 100 cells were counted. One representative result among two independent time courses is shown. (C) Focus number counting of Rad51. At each time point, focus-positive spreads (more than 5 foci) were selected, and the numbers of foci were counted manually. Average focus numbers are shown at the top. (D) Kinetic analysis of focus-positive cells for Dmc1. Spreads containing more than five foci were counted. At each time point, more than 100 cells were counted. One representative result among two independent time courses is shown. (E) Number counting of Dmc1 foci. At each time point, focus-positive spreads (more than five foci) were selected, and the numbers of foci were counted manually. Average focus numbers are shown at the top. P-values were calculated using the Mann–Whitney U-test; ***P < 0.001.
We confirmed the above cytological defects in the sae3 mutants by counting focus number on spreads (Fig. 3C and 3E). At 4 h in the control spread (SAE3-Flag), the Rad51 and Dmc1 focus number was 37.1 ± 1.8 and 33.7 ± 4.3 (mean ± S.D.). The sae3-Y56A, -N57A and -L59A mutants showed a similar number of Rad51 foci to the control at 4 h (Fig. 3C). All three mutants showed no Dmc1 foci at 4 h (Fig. 3E). While the sae3-L59A mutant also showed no Dmc1 foci at 8 h, the sae3-Y56A and -N57A mutants displayed increased Dmc1 focus number at 8 h compared to that at 4 h (12.0 ± 3.0 and 12.6 ± 3.0 among focus-positive spreads, respectively), albeit with fewer foci than in the wild-type at 4 h, supporting the idea that Sae3-Y56A and Sae3-N57A have residual activity.
We also studied the localization of Mei5 and Sae3-Flag proteins on the spreads. Like Rad51/Dmc1, both Mei5 and Sae3-Flag proteins exhibited focus staining with typical kinetics of appearance/disappearance (Fig. 4). The sae3-L59A mutant was almost completely deficient in focus formation involving both Mei5 and Sae3-Flag (Fig. 4A), indicating a critical role for L59 in the Mei5–Sae3 complex. Indeed, this residue is predicted to be at the interface of the putative Mei5 and Sae3 dimer (Fig. 1C and 1D). The sae3-Y56A and -N57A mutants showed delayed formation of Mei5 and Sae3 foci (Fig. 4B and 4D), which appeared later than Dmc1 foci, indicating an uncoupling of Dmc1 loading from Mei5–Sae3 loading. At 8 h, the focus number of Mei5 or Sae3 in the sae3-N57A mutant was still lower than that at 4 h in the wild-type control, while the number in the sae3-Y56A mutant was similar to that in the control at 4 h (Fig. 4C and 4F).
Sae3 L59 is critical for the interaction with Mei5. (A) Meiotic chromosome spreads were immunostained with Mei5 and Sae3. Representative images of Mei5 (green) and Sae3-Flag (red) staining, with a merged image with DAPI (blue), at 4 h in each strain and at 8 h in the sae3-Y56A and -N57A mutant strains are shown. Bar, 1 μm. (B) Kinetic analysis of focus-positive cells for Mei5. Spreads containing more than five foci were counted. At each time point, more than 100 cells were counted. One representative result among two independent time courses is shown. (C) Focus number counting of Mei5. At each time point, focus-positive spreads (more than five foci) were selected and the numbers of foci were counted manually. Average focus numbers are shown at the top. P-values were calculated using the Mann–Whitney U-test; *P < 0.05, ***P < 0.001. (D) Kinetic analysis of focus-positive cells for Sae3-Flag. Spreads containing more than five foci were counted. At each time point, more than 100 cells were counted. One representative result among two independent time courses is shown. (E) Number counting of Sae3-Flag foci. At each time point, focus-positive spreads (more than five foci) were selected, and the numbers of foci were counted manually. Average focus numbers are shown at the top. P-values were calculated using the Mann–Whitney U-test; *P < 0.05, ***P < 0.001. (F) Complex formation of Mei5–Sae3 was verified by immunoprecipitation (IP). Whole-cell extracts (WCE) from various cells after 4-h induction of meiosis were immunoprecipitated with anti-Mei5 antiserum and IP fractions were probed with either anti-Mei5 or anti-Flag (for Sae3).
We checked the formation of Mei5–Sae3 complexes in the sae3 mutants by immunoprecipitation (IP) using an anti-Mei5 antibody. Consistent with a previous report (Hayase et al., 2004), IP fractions of meiotic cell lysates from the control (SAE3-Flag diploid cells incubated in SPM for 4 h) contained Sae3-Flag protein (Fig. 4F, upper panels). Like the wild-type protein, Sae3-Flag-Y56A and -N57A proteins, which are defective in vivo, could bind to Mei5 in cell lysates. On the other hand, Sae3-Flag-L59A, which is completely defective in vivo, could not interact with Mei5 (Fig. 4F, lower panels), showing a critical role for L59 in complex formation with Mei5. Consistent with this, in fission yeast Swi5, I53 (L59 in Sae3) in α2 showed a hydrophobic interaction with the helix of Sfr1 (α2, Fig. 1D, right). This supports the idea that complex formation by Mei5 and Sae3 is critical for the function of these proteins. On the other hand, although Y50 in Swi5 (Y56 in Sae3) also exhibits similar hydrophobic interaction with Sfr1, the cognate amino acid substitution in Sae3 (Y56) did not affect the interaction with Mei5, but abolished the function, suggesting that the Y56A substitution affects the function without influencing the interaction between Mei5 and Sae3. Since N57 is probably located on the surface of the Mei5–Sae3 complex, the substitution of this residue may affect the function of the complex in relation to other interacting proteins such as RPA, Rad51 and Dmc1 (Hayase et al., 2004; Ferrari et al., 2009; Chan et al., 2014). Further characterization is necessary to understand the contribution of these conserved residues in Sae3/Swi5 functions.
All strains described here are derivatives of S. cerevisiae SK1 diploids, NKY1551 (MATα/MATa, lys2/’’, ura3/’’, leu2::hisG/’’, his4X-LEU2-URA3/his4B-LEU2, arg4-nsp/arg4-bgl). The genotypes of each strain used in this study are described in Supplementary Table S1.
Strain constructionsae3 mutant genes were constructed by the pop-in/pop-out method. PCR-based site-directed mutagenesis was carried out using the yIPlac195 plasmid (Gietz and Sugino, 1988) with the wild-type SAE3 gene as a template. The sequences of primer DNAs used in site-directed mutagenesis are provided in Supplementary Table S2. DNA changes were confirmed by DNA sequencing. Plasmids with the mutant SAE3 gene were introduced in the yeast strain MSY832 by transformation and yeast cells were selected for uracil prototrophy. The presence of the mutations was confirmed by restriction digestion of PCR products. Candidate yeast cells were grown overnight in YPAD liquid culture and were plated on a selection medium containing 5-FOA (5-fluoroorotic acid) for uracil auxotrophy. The mutations were reconfirmed by restriction digestion of PCR products and sequencing.
Antiserum and antibodiesAnti-Flag (anti-DYKDDDDK tag; Wako 012-22384) and anti-tubulin (MCA77G, Bio-Rad/Serotec) were used for western blotting. Guinea pig anti-Rad51 (Shinohara et al., 2000), rabbit anti-Dmc1 and rabbit anti-Mei5 serum (Hayase et al., 2004) were used for staining. When background staining was high, IgG was purified using the IgG purification kit (APK-10A, Cosmo Bio). The secondary antibodies for staining were Alexa Fluor 488 (Goat) and 594 (Goat) IgG used at a 1/2000 dilution (Thermo Fisher Scientific).
Meiotic time courseSaccharomyces cerevisiae SK1 strains were patched onto YPG plates (2% bacteriological peptone, 1% yeast extract, 3% glycerol, 2% bacteriological agar) and incubated at 30 ℃ for 12 h. Cells were inoculated onto YPD plates (2% bacteriological peptone, 1% yeast extract, 2% glucose, 2% bacteriological agar) and grown for two days to obtain isolated colonies. A single colony was inoculated into 3 ml of YPD liquid medium and grown to saturation at 30 ℃ overnight. To synchronize cell cultures, the overnight cultures were transferred to pre-warmed SPS medium (1% potassium acetate, 1% bacteriological peptone, 0.5% yeast extract, 0.17% yeast nitrogen base with ammonium sulfate and without amino acids, 0.5% ammonium sulfate, 0.05 M potassium biphthalate) and grown for 16–17 h. Meiosis was induced by transferring the SPS-cultured cells to pre-warmed SPM (1% potassium acetate, 0.02% raffinose). The SPM-cultured cells were harvested at various times after transfer.
CytologyChromosome spreads were immunostained as described previously (Shinohara et al., 2000, 2003). Spheroplasts were burst in the presence of 1% paraformaldehyde and 0.1% lipsol. Stained samples were observed using a fluorescence microscope (BX51; Olympus/Evident) with a 100× objective (NA1.3). Images were captured by CCD camera (CoolSNAP; Roper), and then processed using IP lab and/or iVision (Bio Vision Technologies) and Photoshop (Adobe) software. For focus counting, ~30 nuclei were counted at each time point.
Western blottingWestern blotting was performed as described previously (Hayase et al., 2004; Shinohara et al., 2008), using cell lysates extracted by the TCA method. After being harvested and washed twice with 20% TCA, cells were roughly disrupted with zirconia beads by the Multi-beads shocker (Yasui Kikai). The protein precipitate recovered by centrifugation at 5,000 rpm for 1 min was suspended in SDS-PAGE sample buffer adjusted to pH 8.8 and then incubated at 95 ℃ for 10 min. After electrophoresis, the proteins were transferred onto a nylon membrane (Immobilon-P, Millipore) and incubated with primary antibodies in a blocking buffer (1× PBS, 0.5% BSA) and then with alkaline phosphatase-conjugated secondary antibody (Promega). The color reaction was developed with NBT/BCIP solution (Nacalai Tesque).
ImmunoprecipitationYeast cell lysates were prepared by the glass bead disruption method. The cells were resuspended in lysis buffer (50 mM HEPES-NaOH [pH 7.5], 140 mM NaCl, 10% glycerol, 1 mM EDTA, 5% NP-40). An equal amount of glass beads (Zircona Y2B05) was added along with a protease inhibitor (10×; Roche cOmplete, Mini, EDTA-free Protease Inhibitor Cocktail; 4693159001). The cells were disrupted using the Multi-beads shocker (2,300 rpm; 60 s on, 60 s off cycle, four times). The lysates were incubated with magnetic beads (Dynal M260 Protein-A conjugated; GE Healthcare) coated with anti-Mei5 antibody (6 μl serum in 100 μl beads) at 4 ℃ for 12 h and washed extensively (Sasanuma et al., 2013). Bound proteins were eluted by adding SDS sample buffer and were analyzed on an SDS-PAGE gel, transferred to a nylon membrane (Millipore) and probed with specific antibodies.
Software and statisticsFigures for protein structure analysis were generated by PyMOL. Mean ± S.D. values are shown. Datasets (focus number) were compared using the Mann–Whitney U-test (Prism, GraphPad).
The authors declare no competing financial interest.
A. S. conceived and designed the experiments. P. S. and S. M. carried out experiments. P. S., Y. F., M. I., A. F. and A. S. analyzed the data. A. S. wrote the manuscript with input from coauthors.
We are grateful to members of the Shinohara lab. P. S. was supported by a Japanese government scholarship and by the Institute for Protein Research. This work was supported by a Grant-in-Aid from the JSPS KAKENHI, Grant Numbers 22125001 and 22125002, to A. S.
