Holliday junction-ZMM protein feedback enables meiotic crossover assurance

Abstract

Holliday junctions (HJs) are branched four-way DNA structures that link recombining chromosomes during double-strand break repair¹. Despite posing a risk to chromosome segregation, HJs accumulate during meiotic prophase I as intermediates in the process of crossing-over^2,3. Whether HJs have additional regulatory functions remains unclear. Here we establish an experimental system in budding yeast that enables conditional nucleolytic resolution of HJs after the establishment of meiotic chromosome synapsis. We find that HJ resolution triggers complete disassembly of the synaptonemal complex without disrupting the axis-loop organization of chromosomes. Mechanistically, HJs mediate the continued association of ZMM proteins with recombination nodules that form at the axes interface of homologous chromosome pairs. ZMM proteins, in turn, promote polymerization of the synaptonemal complex while simultaneously protecting HJs from processing by non-crossover pathways. Thus, reciprocal feedback between ZMMs, which stabilize HJs, and HJs, which retain ZMM proteins at future crossover sites, maintains chromosome synapsis until HJ-resolving enzymes are activated during exit from prophase I. Notably, by polymerizing and maintaining the synaptonemal complex structure, the HJ-ZMM interplay suppresses de novo double-strand break formation and recombination reinitiation. In doing so, this interplay suppresses the DNA damage response, enabling meiotic progression without unrepaired breaks and supporting crossover assurance.

Fig. 1. HJs stabilize the SC during meiotic pachytene.

a, The experimental set-up for conditional nucleolytic resolution of HJs after meiotic chromosome synapsis is established. b, Representative images of meiotic chromosome spreads at the indicated times in SPM, immunostained for Zip1 (green); DNA was stained with 4′,6-diamidino-2-phenylindole (DAPI; grey). Yen1^ON was induced by addition of β-oestradiol (or methanol (MeOH) as a control) at 7 h. The insets show a magnification of the region indicated by a dashed box; Rec8 (magenta) marks chromosome axes. The arrowheads indicate Zip1 polycomplexes. c, Quantification of Zip1 synapsis and polycomplexes from b. n = 50 nuclei per timepoint, representative of three biological replicates. d, Quantification of Yen1^ON protein levels from Extended Data Fig. 1b and full Zip1 synapsis from c, normalized to peak values. e, Representative images of meiotic chromosome spreads at the indicated times in SPM as in b, but for P_GAL1-YEN1^ON-ND cells. f, Quantification of Zip1 synapsis and polycomplexes (as in c) from the experiment in e. g, Quantification of Yen1^ON-ND protein levels (as in d) from the experiment in Extended Data Fig. 1d and full Zip1 synapsis from the experiment in f. h, The experimental set-up to inhibit DSB formation by Rec104^FRB nuclear depletion and resolve HJs by Yen1^ON expression in pachytene-arrested ndt80∆ cells. i, Southern blot analysis of DSBs and DNA joint molecules (JMs) at the HIS4::LEU2 recombination hotspot from h. Gel source and biological replicate data are provided in Extended Data Fig. 3c. j, Quantification of DSBs and joint molecules from the experiment in i and a biological replicate, with or without Rec104^FRB nuclear depletion. Data are the mean and range of the percentage of total DNA. k, Quantification of Zip1 synapsis and polycomplex formation from the experiment in i, with or without Rec104^FRB nuclear depletion, including an additional timepoint (11 h in SPM). n = 50 nuclei per timepoint, representative of two biological replicates. l, Quantification of joint molecules as in j, after Rec104^FRB nuclear depletion and with or without Yen1^ON induction. m, Quantification of Zip1 synapsis and polycomplex formation as in k, after Rec104^FRB nuclear depletion and Yen1^ON induction. n = 50 nuclei per timepoint, representative of two biological replicates. For b and e, scale bars, 2 µm.

Fig. 2. dHJ–ZMM protein interplay maintains chromosome synapsis and protects crossover precursors.

a, Representative chromosome spreads at the indicated times in SPM, after Yen1^ON induction by β-oestradiol addition (or methanol control) at 7 h, immunostained for Ecm11–Gmc2 (green) and Zip3 (magenta). The arrowheads mark Zip3 localizing to the Ecm11–Gmc2 polycomplex. b, Quantification of Zip3 focus number from a. Data are mean ± s.d. n = 30 nuclei per timepoint. Statistical analysis was performed using Kruskal–Wallis tests with Dunn’s multiple-comparison test (P ≤ 0.001). Representative of two biological replicates. c, The experimental set-up for conditional depletion of a ZMM protein in pachytene-arrested ndt80∆ cells. Os, Oryza sativa. d, Western blot analysis of Zip3^AID levels in cells at the indicated times in SPM, treated as described in c. Crm1 was used as the protein loading control. The asterisks indicate putative SUMOylated Zip3. Representative of two biological replicates. e, Representative images of chromosome spreads from d, immunostained for Zip1 (green) and Zip3^AID (magenta). ZIP3^AID cells exhibit frequent polycomplex formation but show normal spore viability (Supplementary Table 2). f, Quantification of Zip1 synapsis and polycomplexes from e. n = 50 nuclei per timepoint, representative of two biological replicates. g,h, Southern blot (g) and quantification of joint molecules and non-crossover (NCO1) and crossover (CO2) products (h) at the HIS4::LEU2 recombination hotspot from d. mcJM, multichromatid DNA joint molecule; P1, parental 1; P2, parental 2. i, The experimental set-up for conditional depletion of a ZMM protein and Sgs1 in pachytene-arrested ndt80∆ cells. j, Western blot analysis of Zip3^AID and Sgs1^AID levels as in d for ZIP3^AIDSGS1^AID cells treated as described in i. Representative of two biological replicates. k,l, Southern blot (k) and quantification of joint molecules and non-crossover and crossover products (l) as in g and h, respectively, for the experiment in j. m, Representative images of chromosome spreads as in e, for the experiment in j. Both Zip3^AID and Sgs1^AID carry the Myc epitope tag. n, Quantification of Zip1 synapsis and polycomplexes as in f, for the experiment in m. n = 50 nuclei per timepoint, representative of two biological replicates. For a, e and m, scale bars, 2 μm.

Fig. 3. dHJs enable reversible SC disassembly.

a, The experimental set-up for reversible SC disassembly in pachytene-arrested ndt80∆ cells through conditional protein deSUMOylation using the Smt3 isopeptidase mutant ulp1^∆N-FRB. b, Representative images of meiotic chromosome spreads from the experiment in a at the indicated times in SPM, immunostained for Zip1 (green) and Smt3 (magenta). c, Quantification of Zip1 synapsis from b. n = 50 nuclei per timepoint, representative of two biological replicates. d,e, Southern blot (d) and quantification of joint molecules (e) at the HIS4::LEU2 recombination hotspot. Ulp1^∆N was induced by β-oestradiol (β-oest.) addition (or methanol as a control) at 7 h in SPM. Data are the mean and range of two biological replicates. Gel source and biological replicate data are provided in Extended Data Fig. 6e. f, The experimental set-up to investigate whether dHJs are required for SC reassembly. g, Representative images of meiotic chromosome spreads from the experiment in f at the indicated times in SPM, immunostained for Zip1 (green). h, Quantification of Zip1 synapsis from g (n = 50 nuclei per timepoint). i, Time-lapse image montage of Zip1^GFP in a cell nucleus after β-oestradiol-induced Ulp1^∆N-FRB expression (t = 0 min; ~7 h in SPM) and rapamycin-induced nuclear depletion (t = 120 min). Corresponds to Supplementary Video 6. j, Quantification of structured Zip1^GFP signal from i. The arrow and dashed line indicate the timepoint of rapamycin addition. Data are the mean of two biological replicates. n = 40 cells each. k, Time-lapse image montage as in i, for the simultaneous expression of Ulp1^∆N-FRB and Yen1^ON by β-oestradiol addition. Corresponds to Supplementary Video 7. l, Quantification of structured Zip1^GFP signal as in j, for the experiment in k. Data are the mean of two biological replicates. n = 40 cells each. Scale bars, 2 μm (b and g) and 1 μm (i and k).

Fig. 4. dHJ–ZMM protein interplay suppresses DSB formation and promotes meiotic progression.

a, Representative images of chromosome spreads at peak DSB formation (4 h), and before (7 h) and after (12 h) Yen1^ON induction by addition of β-oestradiol (or methanol control), immunostained for Zip1 (green) and Hop1 (magenta). b, The total Hop1 signal intensity per nucleus from a. Data are mean ± s.d. n = 50 nuclei per timepoint. Statistical analysis was performed using Kruskal–Wallis tests with Dunn’s multiple-comparison test (P ≤ 0.001). c, Western blot analysis of the levels of Hop1 and Hop1 phosphorylated at Thr318 from a. p-Hop1, phosphorylated Hop1. Pgk1 was used as the protein loading control. d,e, Southern blot analysis of DSBs at the CCT6 locus (d) and quantification at CCT6 and ERG1 (e). Yen1^ON was induced by β-oestradiol addition (or methanol control) at 7 h in SPM. Sae2^AID was simultaneously depleted by addition of 5-Ph-IAA (or DMSO control) to prevent DSB repair. Data are the mean and range of fold changes relative to the Sae2^AID depletion control from two biological replicates. Gel source and replicate data are provided in Extended Data Fig. 8c. P, parental. f, Western blot analysis as in c for Msh4^AID depletion at 7 h in SPM. Representative of two biological replicates. g,h, Southern blot analysis of DSBs (g) and quantification of CCT6 and ERG1 (h), as in d and e, for the experiment in f. Gel source and replicate data are provided in Extended Data Fig. 8e. i, Time-lapse montage of Zip1^GFP and Nup84^mCherry after Yen1^ON induction in early/mid-zygotene (t = 0 min, ~5 h in SPM). The arrowheads mark Zip1^GFP aggregates. Nup84^mCherry brightness was adjusted for signal visibility. Corresponds to Supplementary Video 9. j–l, Quantification of structured Zip1^GFP signal (j), meiosis I (k) and sporulation (l) from the experiment in i. The arrow and dashed line indicate β-oestradiol addition. PC, polycomplex. Data are the mean and range (error bars) of two biological replicates, n = 20 cells each. Scale bars, 2 μm (a) and 1 μm (i).

Fig. 5. Cdc5 promotes SC disassembly partly through the activation of HJ resolvases.

a, Representative chromosome spreads at the indicated times in SPM, with immunostaining for Zip1 (green). Cdc5 was induced by β-oestradiol addition at 7 h. b, Quantification of Zip1 synapsis and polycomplex formation from a. n = 50 nuclei per timepoint. c,d, Time-lapse montage of Zip1^GFP (c) and quantification of structured Zip1^GFP (d) in cells of the indicated genotypes after Cdc5 induction by β-oestradiol addition at around 7 h in SPM (t = 0 min). n = 40 cells. See Supplementary Videos 15 and 16. e, The time to complete loss of structured Zip1^GFP signal from d. Data are mean ± s.d. n = 40 cells per genotype. Statistical analysis was performed using two-tailed unpaired Mann–Whitney U-tests. f, Time-lapse montage of Zip1^GFP and Cnm67^tdTomato in cells of the indicated genotypes (left) and quantification of Zip1^GFP signal loss relative to SPB separation (right). Data are the median (dashed line), and first and third quartiles (dotted lines). n = 60 cells per genotype. Statistical analysis was performed using Kruskal–Wallis tests with Dunn’s multiple-comparison test (P ≤ 0.001). g,h, dHJ–ZMM protein interplay coordinates meiotic progression with crossover assurance. g, ZMMs promote the formation of dHJs, which maintain ZMM association at homologue axis interfaces. In turn, ZMMs protect dHJs from STR-mediated dissolution and promote SC assembly and maintenance. MutLγ resolves dHJs into crossovers, contributing to ZMM displacement and loss of SC maintenance. h, SC assembly induces chromosome-autonomous and, eventually, nucleus-wide DSB downregulation. DSB repair completion silences the DNA damage response, enabling Ndt80-mediated pachytene exit without unrepaired DSBs and at least one dHJ per homologue pair. Ndt80-induced Cdc5 activates HJ resolvases, displacing ZMMs and destabilizing the SC to promote rapid SC disassembly. Scale bars, 2 μm (a) and 1 μm (c and f).

Extended Data Fig. 1. HJs stabilize the SC during meiotic pachytene.

a, Representative fluorescence-activated cell sorting (FACS) analysis of DNA content at regular time intervals after induction of meiosis in SPM to monitor pre-meiotic DNA replication in cells with the indicated genotypes. Note that most cells have replicated their genomes after ~4 h in SPM. Corresponds to experiments in Fig. 1b–g and Extended Data Fig. 1b–g. b, c, Western blot analysis (b) and quantification of Yen1^ON and Zip1 protein levels normalized using Pgk1 as loading control (c) in cells at indicated times in SPM. Yen1^ON was induced by β-oestradiol addition (or MeOH as control) at 7 h. Corresponds to the experiment in Fig. 1b,c and Extended Data Fig. 1f. Yen1^ON protein levels after addition of β-oestradiol are also shown in Fig. 1d. d, e, As in (b) and (c), for P_GAL1-Yen1^ON-ND. Corresponds to the experiment in Fig. 1e,f and Extended Data Fig. 1g. Yen1^ON-ND protein levels after addition of β-oestradiol are also shown in Fig. 1g. f, Representative images of meiotic chromosome spreads at indicated times in SPM (left) and quantification of Zip1 synapsis and polycomplex formation (right) in the absence of Yen1^ON induction (MeOH-treated control) at 7 h (n = 50 nuclei per time point; representative of three biological replicates). Control for the experiment shown in Fig. 1b–d. g, As in (f), for P_GAL1-Yen1^ON-ND. Control for the experiment shown in Fig. 1e–g. h, i, Time-lapse image montage of Zip1^GFP in a cell nucleus (h) and quantification of structured Zip1^GFP signal (i) after Yen1^ON induction by β-oestradiol addition at ~7 h in SPM (t = 0 min). Arrowhead marks the emerging polycomplex. Mean of two biological replicates is plotted (n = 40 cells each). Corresponds to Supplementary Video 1. Scale bar = 1 µm. j, k, As in (h) and (i), for P_GAL1-Yen1^ON-ND. Mean of two biological replicates is plotted (n = 40 cells each). Corresponds to Supplementary Video 2.

Extended Data Fig. 2. HJs stabilize the SC central region but are dispensable for the axis-loop organization of pachytene chromosomes.

a, Representative images of meiotic chromosome spreads at indicated times in SPM, immunostained for Ecm11-Gmc2 (green) and Smt3 (blue). Yen1^ON expression was induced by β-oestradiol addition (or MeOH as control) at 7 h in SPM. Scale bars = 2 µm. b, c, Quantification of Ecm11-Gmc2 synapsis (b) and SC-associated Smt3 (c) from (a) (n = 50 nuclei per time point; representative of two biological replicates). d, Representative images of meiotic chromosome spreads at indicated times in SPM, immunostained for Zip1 (green) and Rec8 (magenta). Yen1^ON expression was induced by β-oestradiol addition (or MeOH as control) at 7 h in SPM. Scale bars = 2 µm. e, Quantification of chromosome axis morphology based on Rec8 immunostaining from (d) (n = 50 nuclei per time point; representative of two biological replicates). f, Representative STED microscopy images of meiotic chromosome spreads at indicated times in SPM, immunostained for Zip1 (green) and Rec8 (magenta). Yen1^ON expression was induced by β-oestradiol addition (or MeOH as control) at 7 h in SPM. Insets show magnified regions indicated by dashed lines. Arrowheads indicate example chromosome regions with split axes. Scale bars = 1 µm; scale bars in insets = 200 nm.

Extended Data Fig. 3. Recombination intermediates containing dHJs and SEIs are stable during pachytene.

a, Representative in situ immunofluorescence images of cells at indicated times in SPM after Rec104^FRB anchor-away by rapamycin addition (or DMSO as control) at 7 h, immunostained for the HA epitope to visualize the subcellular localization of Rec104^FRB. DNA is visualized with DAPI (grey). b, Quantification of cells with nuclear Rec104^FRB at indicated times in SPM from (a) (n = 100 cells per time point). c, Southern blot analysis of recombination at the HIS4::LEU2 recombination hotspot from the experiment described in Fig. 1h. Two biological replicates are shown; replicate 1 corresponds to the cropped blot in Fig. 1i. P1, parental 1; P2, parental 2; CO1, crossover 1; CO2, crossover 2; DSBs, double-strand breaks; JMs, DNA joint molecules. Asterisk indicates meiosis-specific recombinant band from gene conversion of the XhoI site closest to the DSB site. Dagger indicates ectopic recombination bands between HIS4::LEU2 and leu2::hisG at the native LEU2 locus. d, Native/native two-dimensional Southern blot analysis of branched JMs at the HIS4::LEU2 recombination hotspot from indicated time points of replicate 1 in (c). The different DNA JM species are depicted in the left blot image. SEI, single-end intermediate; IH-dHJ, interhomologue-dHJ; IS-dHJ, intersister-dHJ; mcJM, multichromatid DNA JM. e, Quantification of SEI, IH-dHJ, IS-dHJ and mcJM levels from (d), shown as the percentage of total DNA. f, Western blot analysis of Yen1^ON protein levels in cells from indicated times in SPM of replicate 1 in (c). Crm1 served as protein loading control. Representative of two biological replicates. g, Quantification of Yen1^ON protein levels from the Western blot in (f) and a biological replicate, normalized to the highest value (red line), along with mean JM levels from both replicates in (c) (grey line). Error bars represent the range.

Extended Data Fig. 4. ZMMs are continuously required to maintain chromosome synapsis during prophase I.

a, Representative images of meiotic chromosome spreads at indicated times in SPM following Yen1^ON induction by β-oestradiol addition (or MeOH as control) at 7 h, immunostained for Ecm11-Gmc2 (green) and Msh5 (magenta). Scale bars = 2 µm. b, Quantification of Msh5 foci number from (a) (mean ± s.d.; n = 30 nuclei per time point; Kruskal-Wallis test (p ≤ 0.001) with Dunn’s multiple comparison test). Representative of two biological replicates. c, d, As in (a) and (b), for analysis of Zip4^9myc (mean ± s.d.; n = 30 nuclei per time point; Kruskal-Wallis test (p ≤ 0.001) with Dunn’s multiple comparison test). Representative of two biological replicates. e, Quantification of Zip3, Msh5 and Zip4^9myc aggregates associating with Ecm11-Gmc2 polycomplexes at indicated times in SPM. Yen1^ON was induced by β-oestradiol addition (or MeOH as control) at 7 h (n = 30 nuclei per time point; representative of two biological replicates). f, g, Western blot analysis of Msh4^AID (f) and Zip4^AID (g) protein levels in cells at indicated times in SPM. OsTir1^F74G expression was induced by addition of CuSO₄ at 6.5 h, and AID-dependent depletion was induced by 5-Ph-IAA addition (or DMSO as a control) at 7 h, as described in Fig. 2c. Crm1 served as protein loading control. Representative of two biological replicates. h, Representative images of meiotic chromosome spreads at indicated times in SPM from (f), immunostained for Zip1 (green) and Msh4^AID (magenta). Scale bars = 2 µm. i, As in (h), for ZIP4^AID in (g). j, k, Quantification of Zip1 synapsis and polycomplexes from (h) and (i) (n = 50 nuclei per time point; representative of two biological replicates). l, Representative images of meiotic chromosome spreads at indicated times in SPM, immunostained for Zip1 (green) and Rec8 (magenta). Zip3^AID depletion was induced by 5-Ph-IAA addition (or DMSO as control) at 7 h. Scale bars = 2 µm. m-o, Quantification of chromosome axis morphology based on Rec8 immunostaining in cells with indicated genotypes at specified times in SPM. Zip3^AID, Msh4^AID or Zip4^AID depletion was induced by 5-Ph-IAA addition (or DMSO as control) at 7 h (n = 50 nuclei per time point; representative of two biological replicates).

Extended Data Fig. 5. ZMMs protect dHJs from Sgs1-mediated dissolution throughout prophase I.

a, Western blot analysis of Msh4^AID protein levels in cells at indicated times in SPM. OsTir1^F74G expression was induced by addition of CuSO₄ at 6.5 h, and Msh4^AID depletion was induced by 5-Ph-IAA addition (or DMSO as a control) at 7 h, as described in Fig. 2c. Crm1 served as protein loading control. Representative of two biological replicates. b, c, Southern blot (b) and quantification of JMs, non-crossover (NCO1) and crossover (CO2) products (c) at the HIS4::LEU2 recombination hotspot from (a). d, As in (a), for MSH4^AID SGS1^AID. Representative of two biological replicates. e,f, As in (b) and (c), for MSH4^AID SGS1^AID from (d). g, As in (a), for SGS1^AID. h, i, As in (b) and (c), for SGS1^AID from (g). j, Representative images of meiotic chromosome spreads at indicated times in SPM from (d), immunostained for Zip1 (green) and Msh4^AID/Sgs1^AID (magenta). Scale bars = 2 µm. k, Quantification of Zip1 synapsis and polycomplexes from (j) (n = 50 nuclei per time point; representative of two biological replicates). l, Time-lapse image montage of Zip1^GFP in cell nuclei of indicated genotypes. OsTir1^F74G expression was induced by addition of CuSO₄ at 6.5 h in SPM, and Zip3^AID and Sgs1^AID depletion was induced by 5-Ph-IAA addition (or DMSO as a control) at ~7 h in SPM (t = 0 min). Arrowhead marks the emerging polycomplex. Corresponds to Supplementary Video 3 and 4. Scale bar = 1 µm. m, Quantification of structured Zip1^GFP signal in cells of indicated genotypes from (l) (n = 40 cells per genotype). n, Quantification of time to complete loss of structured Zip1^GFP signal in cells of indicated genotypes from (m) (mean ± s.d; n = 40 cells per genotype; two-tailed, unpaired Mann-Whitney U test). o, As in (a), for ZIP4^AID SGS1^AID. Representative of two biological replicates. p, As in (j), for (o). q, As in (k), for (p) (n = 50 nuclei per time point; representative of two biological replicates).

Extended Data Fig. 6. A molecular tool to reversibly induce SC disassembly independent of HJ resolution.

a, Ulp1 domain organization, adapted from. In the ulp1^∆N mutant, an in-frame GFP coding sequence replaces parts of the nuclear pore complex anchoring region (amino acids 172–340) in the N-terminal domain. b, Schematic representation of the ULP1 genomic locus. A diploid yeast strain heterozygous for the ulp1^∆N-FRB allele was used since ULP1 is an essential gene. The native promoter of ulp1^∆N-FRB is replaced by P_GAL1, which can be induced by the addition of β-oestradiol. In addition, the C-terminal FRB domain of ulp1^∆N-FRB enables the conditional nuclear depletion using the rapamycin-dependent anchor-away system. c, Western blot analysis of Ulp1^∆N-FRB, Smt3, Ecm11 and Zip1 protein levels in cells at indicated times in SPM. Ulp1^∆N-FRB expression was induced by β-oestradiol addition (or MeOH as control) at 8 h, and nuclear depletion was induced by rapamycin addition (or DMSO as control) at 9 h. Arrow indicates the band corresponding to the expected molecular weight of Ulp1^∆N-FRB. Asterisks indicate putative SUMOylated and polySUMOylated Ecm11 bands in the anti-Smt3 blot. Samples were run on separate gels due to similar molecular masses of the proteins, with Pgk1 as a sample processing control. Representative of two biological replicates. d, Quantification of SC-associated Smt3 on meiotic chromosome spreads from (c) (n = 50 nuclei per time point; representative of two biological replicates). e, Southern blot analysis of JMs and DSBs at the HIS4::LEU2 recombination hotspot. Ulp1^∆N was induced by β-oestradiol addition (or MeOH as control) at 7 h in SPM. Two biological replicates are shown, with replicate 2 including an additional time point (12 h in SPM). Replicate 1 corresponds to the cropped blot in Fig. 3d. f, Representative STED microscopy images of meiotic chromosome spreads at indicated times in SPM from (c), immunostained for Zip1 (green) and Rec8 (magenta). Insets show magnified regions indicated by dashed lines. Note that although Zip1 is lost from chromosomes and Rec8 localization appears to be more discontinuous after Ulp1^∆N expression (see also (g)), the co-alignment of homologous axes is maintained. Scale bars = 1 µm; scale bars in insets = 200 nm. g, Quantification of chromosome axis morphology based on Rec8 immunostaining on meiotic chromosome spreads at indicated times in SPM from (c) (n = 50 nuclei per time point; representative of two biological replicates). h, Experimental setup to determine whether SC reassembly upon Ulp1^∆N-FRB expression/nuclear depletion depends on the formation and repair of de novo DSBs. i, Western blot analysis of Ulp1^∆N-FRB, Rec104^AID, Smt3, Ecm11 and Zip1 protein levels in cells at indicated times in SPM, treated as described in (h). Arrow indicates the band corresponding to the expected molecular weight of Ulp1^∆N-FRB. Asterisks indicate putative SUMOylated and polySUMOylated Ecm11 bands in the anti-Smt3 blot. Samples were run on separate gels due to similar molecular masses of the proteins, with Pgk1 as a sample processing control. j, Quantification of Zip1 synapsis on meiotic chromosome spreads from (i) (n = 50 nuclei per time point).

Extended Data Fig. 7. dHJ-ZMM protein interplay enables SC reassembly upon Ulp1^∆N-FRB depletion.

a, Western blot analysis of Ulp1^∆N, Yen1^ON and Ecm11 protein levels in cells with the indicated genotype at specified times in SPM, treated as described in Fig. 3f. Arrow indicates the band corresponding to the expected molecular weight of Ulp1^∆N-FRB. Samples were run on separate gels due to similar molecular masses of the proteins, with Pgk1 as a sample processing control. b, Representative images of meiotic chromosome spreads of the indicated genotype from (a) at specified times in SPM, immunostained for Zip1 (green). Control for the experiment shown in Fig. 3g. Scale bars = 2 µm. c, Quantification of Zip1 synapsis from (b) (n = 50 nuclei per time point). d, Time-lapse image montage of Zip1^GFP in a cell nucleus after Ulp1^∆N-FRB expression induced by β-oestradiol addition at 7 h in SPM (t = 0 min). Control for the experiment shown in Fig. 3i, with DMSO added after 120 min. Corresponds to Supplementary Video 5. Scale bar = 1 µm. e, Quantification of structured Zip1^GFP signal from (d). Mean of two biological replicates is plotted (n = 40 cells each). f, Representative images of meiotic chromosome spreads at indicated times in SPM, immunostained for Zip3 (green) and Smt3 (magenta). Ulp1^∆N-FRB expression was induced by β-oestradiol addition at 7 h in SPM, and nuclear depletion was initiated by rapamycin addition at 9 h in SPM. Arrowheads mark Zip3 aggregates. Scale bars = 2 µm. g, h, Quantification of Zip3 foci number (g) and fraction of nuclei exhibiting Zip3 aggregates (h) from (f) (mean ± s.d.; n = 30 nuclei per time point; Kruskal-Wallis test (p ≤ 0.001) with Dunn’s multiple comparison test). i, As in (f), for the simultaneous expression of Ulp1^∆N-FRB and Yen1^ON by β-oestradiol addition at 7 h in SPM. j, k, As in (g) and (h), for (i) (mean ± s.d.; n = 30 nuclei per time point; Kruskal-Wallis test (p ≤ 0.001) with Dunn’s multiple comparison test).

Extended Data Fig. 8. dHJ-ZMM protein interplay suppresses DSB formation.

a, Experimental setup to measure DSB formation by Southern blotting after Yen1^ON expression in ndt80∆ cells. Sae2^AID is depleted by 5-Ph-IAA addition (or DMSO as control) alongside Yen1^ON induction by β-oestradiol addition (or MeOH as control) to prevent endonucleolytic removal of Spo11 and subsequent DSB processing and repair. b, Western blot analysis of Yen1^ON, Sae2^AID and Hop1 protein levels in cells at indicated times in SPM, treated as described in (a). Hop1^P indicates phosphorylated Hop1. Crm1 served as protein loading control. Representative of two biological replicates. c, d, Southern blot analysis of DSB formation at the CCT6 (c) and ERG1 (d) hotspots from (b). Two biological replicates are shown side-by-side in the same blot for each hotspot; replicate 2 of the CCT6 blot corresponds to the cropped blot in Fig. 4d. e, f, Southern blot analysis of DSB formation at the CCT6 (e) and ERG1 (f) hotspots in cells with the indicated genotype, with Msh4^AID depletion by 5-Ph-IAA addition (or DMSO as control) at 7 h in SPM. Two biological replicates are shown per locus; replicate 1 of the CCT6 blot in (e) corresponds to the cropped blot in Fig. 4g. g, Western blot analysis of Zip3^AID, Hop1 and Hop1-pT318 protein levels in cells at indicated times in SPM. Zip3^AID depletion was induced by 5-Ph-IAA addition (or DMSO as control) at 7 h in SPM. Hop1^P indicates phosphorylated Hop1. Pgk1 served as protein loading control. h, As in (g), for ZIP4^AID.

Extended Data Fig. 9. Premature resolution of recombination intermediates disrupts the SC and causes meiosis I failure.

a, Representative FACS analysis of DNA content in cells with the indicated genotype at regular time intervals after induction of meiosis in SPM. Corresponds to the experiment in Fig. 4i–l and Extended Data Fig. 9b–l. b, Time-lapse image montage of Zip1^GFP in a cell nucleus induced to enter meiosis in SPM, with imaging starting at 4.5 h in SPM (t = −30 min). The durations of the zygotene and pachytene stages of prophase I, and the time from pachytene exit to anaphase I are indicated (mean ± s.d of two biological replicates; n = 20 cells each). The brightness of Nup84^mCherry was adjusted to facilitate visibility of the signal. Asterisk indicates the gametogenesis uninherited nuclear compartment (GUNC). Ana, Anaphase. Corresponds to Supplementary Video 8. Scale bar = 1 µm. c, d, Western blot analysis (c) and quantification (d) of Yen1^ON protein levels, normalized to the highest value, in cells at indicated times after Yen1^ON induction by β-oestradiol addition (or MeOH as control) at 7 h in SPM. Error bars represent range of two biological replicates. e, Time-lapse image montage of Zip1^GFP in a cell nucleus after MeOH addition as a control in early to mid-zygotene (t = 0 min, ~5 h in SPM). The brightness of Nup84^mCherry was adjusted to facilitate visibility of the signal. Corresponds to Supplementary Video 10. Scale bar = 1 µm. f, Quantification of structured Zip1^GFP signal from (e). Mean of two biological replicates is plotted (n = 20 cells each). g-i, Quantification of structured Zip1^GFP signal (g), meiosis I entry (h), and sporulation (i) in pachytene cells at the time of β-oestradiol addition (t = 0 min, ~5 h in SPM). Mean of two biological replicates is plotted (n = 20 cells each; error bars represent range). Corresponds to Supplementary Video 11. j-l, Quantification of structured Zip1^GFP signal (j), meiosis I entry (k), and sporulation (l) in pre-leptotene/leptotene cells at the time of β-oestradiol addition (t = 0 min, ~5 h in SPM). Mean of two biological replicates is plotted (n = 20 cells each; error bars represent range). Corresponds to Supplementary Video 12. m, Time-lapse image montage of Zip1^GFP and Htb1^mCherry in a cell nucleus of the indicated genotype. H₂O was added as a control at 7 h in SPM (t = −65 min). Ndt80 expression was induced by β-oestradiol addition at ~8 h in SPM (t = 0 min). Corresponds to Supplementary Video 13. Scale bar = 1 µm. n, Quantification of structured Zip1^GFP signal and the fraction of cells undergoing meiosis I and II based on Htb1^mCherry labelled chromatin from (m) (n = 60 cells). o, As in (m), with CuSO₄ addition to induce expression of Yen1^ON at 7 h in SPM (t = −65 min). Corresponds to Supplementary Video 14. p, As in (n), for (o). q, Quantification of cells with completed or defective meiotic divisions from (n) and (p).

Extended Data Fig. 10. Cdc5 promotes SC disassembly in part through the activation of HJ resolvases.

a, Quantification of Zip1 synapsis and polycomplexes in the absence of Cdc5 induction (MeOH-treated control) at 7 h in SPM (n = 50 nuclei per time point). Controls for the experiments shown in Fig. 5a,b. b, Western blot analysis of Cdc5 and Zip1 protein levels in cells of indicated genotypes at specified times in SPM. Cdc5 was induced by β-oestradiol addition (or MeOH as control) at 7 h in SPM. Pgk1 served as protein loading control. Correspond to experiments in (a) and Fig. 5a,b. c, Quantification of Zip1 protein levels from (b), normalized to the highest value. d, Representative images of meiotic chromosome spreads at indicated times in SPM, immunostained for Zip1 (green). Cdc5 was induced by β-oestradiol addition (or MeOH as control) at 7 h in SPM. Scale bars = 2 µm. e, Quantification of Zip1 synapsis and polycomplexes from (d) (n = 50 nuclei per time point). f, As in (b), for cells of the indicated genotype. Corresponds to experiment in (d) and (e). g, As in (c), for (f). h, i, Time-lapse image montage of Zip1^GFP in cell nuclei of indicated genotypes (h) and quantification of structured Zip1^GFP signal (i) after Cdc5 induction by β-oestradiol addition at ~7 h in SPM (t = 0 min) (n = 50 cells per genotype). The bottom panel corresponds to Supplementary Video 17. Scale bars = 1 µm. j, Quantification of time to complete loss of structured Zip1^GFP signal in cells of indicated genotypes from (i) (mean ± s.d; n = 50 cells per genotype; two-tailed, unpaired Mann-Whitney U test). k, (Top) Representative images of meiotic chromosome spreads of the indicated genotypes at 6 h in SPM, immunostained for Zip1 (green), Rec8 (magenta) and the SPB component γ-tubulin/Tub4 (cyan). Insets show magnified, slightly separated SPBs. Scale bars = 2 µm. (Bottom) Quantification of nuclei with partial or full SPB separation containing Zip1 stretches (weighted mean ± weighted SD of two biological replicates; n = 152–174 nuclei per condition).