Securin regulates the spatiotemporal dynamics of separase

During mitosis, the metaphase-to-anaphase transition is regulated by the spindle assembly checkpoint (SAC) to ensure proper chromosome segregation (McAinsh and Kops, 2023; Musacchio, 2015). Once chromosomes are properly aligned on the spindle, SAC signaling is silenced and anaphase commences. Entry into anaphase requires the anaphase-promoting complex/cyclosome (APC/C), an E3 ubiquitin ligase (Alfieri et al., 2017). The APC/C is kept inactive by the SAC until anaphase onset when it ubiquitinates multiple substrates including securin, an inhibitory chaperone of separase (Watson et al., 2019). Separase is a protease that cleaves a subunit of cohesin, which links sister chromatids together, allowing anaphase onset (Uhlmann et al., 2000).

The SAC pathway also controls chromosome segregation during meiosis to form proper gametes (Gorbsky, 2015). Oocyte meiosis consists of two asymmetric divisions that generate a single large haploid gamete (Mogessie et al., 2018). In many species, oocytes are arrested in metaphase II prior to fertilization, which involves APC/C inhibition (Tunquist and Maller, 2003). Egg activation is triggered by fertilization, which transforms the arrested oocyte into a rapidly dividing embryo. Egg activation events include the metaphase-to-anaphase transition and the exocytosis of cortical granule vesicles (Horner and Wolfner, 2008). Secreted cargo modifies the extracellular matrix of the oocyte to block polyspermy and protect the embryo (Liu, 2011; Liu et al., 2003; Wessel et al., 2001). Many fundamental cell cycle discoveries have been made by studying oocyte meiosis.

Caenorhabditis elegans is ideal for studying egg activation in vivo (Lui and Colaiácovo, 2013). Egg activation occurs during meiosis I in C. elegans oocytes (McCarter et al., 1999; McNally and McNally, 2005; Yang et al., 2005). Cortical granule exocytosis occurs during anaphase I, secreting eggshell material (Bembenek et al., 2007; Olson et al., 2012). Importantly, separase localizes to cortical granules in anaphase I and promotes exocytosis, linking cell cycle regulation to vesicle trafficking (Bai and Bembenek, 2017; Bembenek et al., 2007, 2010; Mitchell et al., 2014; Richie et al., 2011). In human cells, securin is known as the pituitary tumor transforming gene, is overexpressed in cancers, and affects secretion (Donangelo et al., 2006; Heaney et al., 1999; Yu et al., 2000). Separase also regulates membrane trafficking in mammalian and plant cells (Bacac et al., 2011; Moschou et al., 2013), suggesting wide conservation of this role. How separase function during exocytosis is regulated by the SAC pathway is unknown.

The activation of separase after securin degradation has been extensively characterized using biochemical methods. However, the dynamics of separase localization are not well characterized. In yeast, securin is required for nuclear and spindle localization of separase (Agarwal and Cohen-Fix, 2002; Jensen et al., 2001). Mammalian separase localization is poorly characterized and has only been observed on chromosomes by staining overexpressed separase on chromosome spreads (Chestukhin et al., 2003; Shindo et al., 2012; Sun et al., 2009). Separase activity has been detected with biosensors at specific subcellular compartments (Agircan and Schiebel, 2014; MacKenzie et al., 2023; Monen et al., 2015; Nam and van Deursen, 2014; Rosen et al., 2019; Shindo et al., 2012; Weber et al., 2020). In C. elegans meiosis, separase localization has been well characterized, making it an ideal context to investigate how separase localization is regulated (Bembenek et al., 2007).

In C. elegans meiosis I, kinetochore cup structures form on holocentric homologous chromosomes (Fig. 1 A, Schvarzstein et al., 2010). Separase localizes to kinetochore cups in prometaphase I, appears at the midbivalent, where cohesin is found at anaphase onset, and then to “linker structures” between separating chromosomes in anaphase (Bembenek et al., 2007; Dumont et al., 2010; Muscat et al., 2015; Schvarzstein et al., 2010; Severson and Meyer, 2014). In the cortex, separase localizes to poorly characterized “linear elements” with kinetochore proteins during prometaphase I and then appears on cortical granules in anaphase I to promote exocytosis (Howe et al., 2001; Monen et al., 2005; Bai and Bembenek, 2017; Bembenek et al., 2007). How separase localization is regulated is not well understood.

We used live imaging to investigate the precise timing and regulation of separase dynamics. We find that while securin is degraded over several minutes, separase relocalizes seconds before cohesin is removed from chromosomes. Depletion of securin and APC/C affect separase localization on chromosomes and vesicles. We also characterize the phenotypes caused by non-degradable securin expression. Our findings suggest that APC/C and securin control both separase protease activity and its localization.

To characterize separase dynamics, we generated worms with endogenously tagged separase (SEP-1::GFP) or securin (IFY-1::GFP) with the chromosome marker H2B::mCherry. SEP-1::GFP localization is dynamic during meiosis I (Fig. 1, B–E). During prophase I, SEP-1::GFP is cytoplasmic and excluded from the nucleus (Fig. 1 B). Just before nuclear envelope breakdown (NEBD), SEP-1::GFP accumulates in the nucleus and becomes enriched on kinetochore cups (Fig. 1 C). Simultaneously, SEP-1::GFP appears on cytoplasmic linear element structures (Fig. 1, C–D). In anaphase, SEP-1::GFP appears between chromosomes and is found on vesicles (Fig. 1 E). Therefore, separase moves from kinetochore structures to sites of action in anaphase.

We next investigated IFY-1::GFP localization during meiosis I. During prophase I, IFY-1::GFP is detected in the cytoplasm and is enriched in the nucleus while separase is excluded (Fig. 1 F). This nuclear pool of IFY-1::GFP is therefore likely not bound to separase. IFY-1::GFP colocalizes with SEP-1 at kinetochore cups and linear elements from NEBD through prometaphase I (Fig. 1, G, H, K, and L). Thus, separase is likely securin-bound and inactive when it localizes to kinetochore structures. IFY-1 signal begins to decline in the cytoplasm by late prometaphase I, ∼3 min before anaphase onset, as previously described in worms (Fig. 1 N, Wang et al., 2013). By anaphase onset, weak IFY-1 signal remains at the midbivalent, which rapidly disappears (Fig. 1, I and N). We did not detect IFY-1 on vesicles at any timepoint (Fig. 1, F–I). Endogenously GFP tagged and multiple, independent transgenic lines show similar degradation curves regardless of expression level (Fig. S1, C and C′). In the cortex, IFY-1::GFP and SEP-1::mScarlet colocalize on linear elements until anaphase I when SEP-1::mScarlet localizes to vesicles and IFY-1::GFP is degraded (Fig. 1 O and Fig. S1 B). Therefore, SEP-1 colocalizes with IFY-1 at kinetochore structures until anaphase when IFY-1 is degraded and SEP-1 localizes to sites of action. These observations suggest that IFY-1 degradation may be required for SEP-1 to relocalize in anaphase.

To better define the dynamics of anaphase onset, we imaged separase, securin, and cohesin at the metaphase-to-anaphase transition with high spatiotemporal resolution. We observed a rapid separase relocalization seconds before anaphase onset (Fig. 2, A–C and Video 1). During prometaphase I, separase localizes to kinetochore cups (Fig. 2 A) and is excluded from the midbivalent where the meiotic cohesin subunit, GFP::COH-3, localizes (Fig. 2, A and B). After spindle rotation, separase enriches at spindle poles ∼30 s before anaphase I onset (Fig. 2, A and B; and Fig. S1 D, E, and F) and invades the midbivalent region within 10 s of anaphase onset, briefly colocalizing with GFP::COH-3 (Fig. 2, A and B; and Videos 1 and 2). When chromosomes move poleward, separase is highly enriched at the midbivalent and GFP::COH-3 is lost (Fig. 2, A and B; Fig. S2 A; and Fig. S1 D). Therefore, separase is spatially restricted from cohesin until seconds prior to anaphase onset.

We next wanted to determine the timing of securin degradation relative to separase relocalization. To address this, we imaged GFP::IFY-1 with H2B::mCherry at high temporal resolution. During early metaphase I, the APC/C is required for meiotic spindle compaction, rotation, and translocation (Yang et al., 2003). During this time, spindle-associated GFP::IFY-1 gets redistributed and concentrated in the same pattern as separase, but the signal drops precipitously at spindle rotation, losing approximately a third of its average intensity by anaphase I onset (Fig. 2 C; Fig. 1, N and O; Fig. S1 A; and Video 1). When separase relocalizes to the midbivalent, securin levels rapidly decrease (Fig. 2 C; Fig. 1 O; and Fig. S1 A). At linear elements in the cortex, GFP::IFY-1 drops to near cytoplasmic levels before SEP-1::mScarlet relocalizes to vesicles (Fig. 1 O; Fig. S1 B; and Video 3). SEP-1::GFP is lost from linear elements and rapidly accumulates on cortical granules within 30 s after anaphase onset (Fig. 2 D and Video 3). Therefore, securin degradation rapidly occurs when separase relocalizes to sites of action at anaphase onset.

We next wanted to determine whether separase localization is controlled by APC/C-mediated securin degradation. Previously, cortical granule localization of separase was not observed after ify-1 RNAi in C. elegans (Kimura and Kimura, 2012). We repeated this experiment to observe separase localization to chromosomes and vesicles. In control animals, the cortical granule cargo protein mCherry::CPG-2 is observed in vesicles in oocytes and embryos before anaphase I and is in the eggshell of older embryos (Fig. S2 A). In contrast to the previous report, we detected apparent SEP-1::GFP vesicle localization in multiple embryos in the uterus of ify-1(RNAi)-treated animals (N = 16/30 animals). To validate this result, we characterized separase localization in animals with different RNAi penetrance expressing SEP-1::GFP together with H2B::mCherry and mCherry::CPG-2. In mild phenotype cases (14–48 h of feeding RNAi), vesicle localization of separase is only observed in one or fewer embryos and mCherry::CPG-2 is observed in the eggshell of most embryos (N = 31/101 animals). Animals with an intermediate phenotype (16–39 h of feeding) had two to three unicellular embryos with SEP-1::GFP localized to mCherry::CPG-2 vesicles, while older embryos had gradually increasing eggshell mCherry::CPG-2 signal, suggesting reduced and delayed exocytosis (N = 34/95, Fig. S2 B). In severely affected animals (24–48 h of feeding RNAi), all embryos were unicellular and lacked extracellular mCherry::CPG-2 signal, with cortical granules trapped in the cytoplasm (N = 31/61, Fig. S2 C). Cortical granule localized SEP-1::GFP signal was reduced but still observed (Fig. S2 C). Therefore, securin depletion does not prevent separase localization to vesicles but inhibits exocytosis.

We next tested whether securin depletion, which should prematurely reduce securin levels, might cause premature separase relocalization. To test this, we examined separase localization in embryos within the spermatheca, which are in early prometaphase I. In control animals, prometaphase I embryos have SEP-1::GFP at kinetochore cups (Fig. 3, A and A′; and Fig. S2, F and G) and linear elements (N = 5/5, Fig. 3 A′′ and Fig. S2 A), with no vesicle localization. In intermediate and severe ify-1(RNAi) embryos, separase signal was reduced and mislocalized on chromosomes and could be observed between homologs where cohesin resides in embryos within the spermatheca (N = 7/15 embryos, Fig. 3, B and B′; and Fig. S2 G). In addition, we observed SEP-1::GFP on cortical granules prematurely in ify-1(RNAi) embryos within the spermatheca (N = 26/35, Fig. 3 B′′). Time-lapse recordings of oocytes from ovulation confirm that no vesicle signal is observed in control prometaphase 1 oocytes and embryos, but can be detected shortly after ovulation after ify-1(RNAi) treatment (Fig. S2, A′ and B′). These results suggest that securin depletion causes premature relocalization of separase.

We also depleted the APC/C, which is required for securin degradation, and examined separase localization. APC/C was also found to be required for separase to localize to vesicles (Kimura and Kimura, 2012). Severely affected apc-2(RNAi) embryos arrest in prometaphase I with mCherry::CPG-2 trapped in cortical granules (Fig. S2 D). As expected, apc-2 RNAi blocked the degradation of IFY-1::GFP, which colocalized with SEP-1::mCherry on kinetochore structures in arrested embryos (Fig. S2 E). In apc-2(RNAi) embryos found in the spermatheca, SEP-1 is localized to kinetochore cups and linear elements like control embryos (Fig. 3, C–C′′; and Fig. S2, D and E). Therefore, depletion of APC/C prevents securin degradation and separase relocalization from kinetochore structures to sites of action.

Since separase relocalizes prematurely after securin depletion, we expected that it may also become proteolytically active prematurely. To test this, we examined embryos expressing GFP::COH-3 with chromosome marker H2B::mCherry (Fig. S3). Throughout prometaphase I, GFP::COH-3 appears as a crescent at the midbivalent, consistent with previous reports (McNally et al., 2022), which is likely a folded short arm axis of the homologous chromosomes (Fig. S3, B and B′). In WT embryos, GFP::COH-3 remains on chromosomes until seconds before chromosome segregation begins (Fig. S3 A and F). In prometaphase I embryos within the spermatheca, GFP::COH-3 signal remains high in control (N = 10) or apc-2 RNAi (N = 5, Fig. S3, C, D, and G). In older embryos, COH-3 signal remained on chromosomes in arrested apc-2(RNAi) embryos but not in control embryos (Fig. S3, C and D). In contrast, prometaphase I ify-1(RNAi) embryos had a significant decrease in chromosomal GFP::COH-3 levels, consistent with prematurely active separase (Fig. S3, A, E, and F, N = 13). As expected, COH-3 persisted on chromosomes after anaphase onset in sep-1(RNAi) embryos (Fig. S3 A, N = 7 spindles). Therefore, securin depletion causes premature cohesin removal from chromosomes.

APC/C and securin inactivation caused pleiotropic cell cycle defects. To directly test whether securin affects separase relocalization, we characterized a non-degradable mutant. In many systems, non-degradable securin expression causes cell division defects, (Cohen-Fix et al., 1996; Hagting et al., 2002; Herbert et al., 2003; Leismann et al., 2000; Zur and Brandeis, 2001), which has not been tested in worms. We developed methods for the expression of toxic transgenes in C. elegans using gfp(RNAi) and propagation of transgenes with female-biased promoters in the male germline to control expression (Mitchell et al., 2014). We made GFP::IFY-1 with mutations in the predicted N-terminal destruction box motif, required for APC/C recognition (Kitagawa et al., 2002; Fig. 4 A). We generated transgenic lines with similar expression levels between wild type and mutant securin (Fig. S1, C and C′) in a WT background. As expected, the expression of GFP::IFY-1^DM, but not GFP::IFY-1^WT, caused embryonic lethality (Fig. 4 B). The localization of GFP::IFY-1^DM was identical to wild type, except that it was not degraded in anaphase I (Fig. 4, C–F). During anaphase I, GFP::IFY-1^DM accumulated on the spindle, but was not observed on vesicles (Fig. 4 F). GFP::IFY-1^DM causes severe chromosome segregation and polar body extrusion defects (N = 28/31, Fig. 4 G). Older GFP::IFY-1^DM embryos shrink when dissected in high salt buffer (N = 31/32), which is never observed in wildtype (N = 0/22, Fig. 4 H), indicating an eggshell permeability defect. Therefore, GFP::IFY-1^DM is dominant negative and behaves as expected for a non-degradable mutant.

We examined chromosome segregation during meiosis I in embryos overexpressing GFP::IFY-1^WT or GFP::IFY-1^DM together with the chromosome marker H2B::mCherry (Fig. S4, A–D). Spindle rotation in GFP::IFY-1^WT embryos, which depends on APC/C activity (Crowder et al., 2015; Ellefson and McNally, 2011), occurs 68 ± 7 s (N = 6) before anaphase onset. At anaphase onset, GFP::IFY-1^WT levels are rapidly decreasing and chromosomes move apart quickly (Fig. 2 C; and Fig. S4, A and C). In contrast, GFP::IFY-1^DM accumulates at the midbivalent 56 ± 6 s (N = 16) after spindle rotation (Fig. S4 B and Video 1). Chromosomes move apart 5 ± 2 s (N = 6) after midbivalent localization in GFP::IFY-1^WT embryos, but delay separation 115 ± 8.5 s (N = 23) after midbivalent localization in GFP::IFY-1^DM embryos (Fig. S4 A). After this extended delay, chromosomes move poleward at a significantly slower rate in the mutant relative to WT (Fig. S4, B and C). Chromosome segregation defects were observed in all embryos expressing GFP::IFY-1^DM (N = 18/18, Fig. S4, B and D), but not in GFP::IFY-1^WT embryos (N = 5/5, Fig. S4, A and D). Finally, GFP::IFY-1^DM labeled structures in between chromosomes sometimes drift away from the spindle in late anaphase I (Fig. S4 B), but not in WT. Therefore, GFP::IFY-1^DM inhibits chromosome segregation and causes a novel delay between the midbivalent localization of securin and poleward chromosome movement.

We next investigated whether GFP::IFY-1^DM inhibits cortical granule exocytosis. We imaged embryos expressing GFP::IFY-1^WT or GFP::IFY-1^DM together with the cortical granule cargo protein, mCherry::CPG-2 during meiosis I (Fig. S4, E–K and Video 4). In WT embryos, cortical granule exocytosis occurs shortly after anaphase onset, releasing mCherry::CPG-2 into the eggshell (Fig. S4, E–H). In embryos expressing GFP::IFY-1^DM, most of the mCherry::CPG-2 labeled cortical granules do not undergo exocytosis during anaphase I (Fig. S4, E, G, I, and K). Therefore, GFP::IFY-1^DM expression causes a severe block of cortical granule exocytosis.

Our results demonstrate that securin is a potent inhibitor of separase function during both chromosome segregation and exocytosis. While securin is known to inhibit separase protease activity, we wanted to investigate whether it also affects separase localization. Therefore, we examined endogenously tagged SEP-1::mScarlet in embryos expressing GFP::IFY-1^WT or GFP::IFY^DM during meiosis I (Fig. 5). From NEBD through prometaphase I, SEP-1::mScarlet colocalizes with GFP::IFY-1^WT and GFP::IFY-1^DM on kinetochore cups and linear elements (Fig. 5, A and B; and Video 5). After spindle rotation and shortening, SEP-1::mScarlet colocalizes with GFP::IFY-1^WT and GFP::IFY-1^DM at spindle poles, but GFP::IFY-1^DM accumulates at higher levels (Fig. 5, A and B). Separase relocalizes to the midbivalent in the presence of either the rapidly lost GFP::IFY-1^WT or the highly accumulated GFP::IFY-1^DM (Fig. 5, A and B). Therefore, expression of GFP::IFY-1^DM severely inhibits chromosome segregation without inhibiting the midbivalent or spindle localization of separase.

Next, we investigated whether securin affects separase localization to vesicles. We imaged SEP-1::mScarlet in the presence of overexpressed GFP::IFY-1^WT or GFP::IFY-1^DM (Fig. 5). Neither wild type nor mutant securin were detected on vesicles (Fig. 1, F–I and Fig. 4, C–F). SEP-1::mScarlet localized to vesicles normally in GFP::IFY-1^WT embryos (N = 11/11) but not in GFP::IFY-1^DM embryos (N = 11/12). To confirm vesicle localization, we co-expressed CAV-1::GFP (Sato et al., 2008) with GFP::IFY-1^WT or GFP::IFY-1^DM and examined SEP-1::mScarlet localization during anaphase I (Fig. 5, C–E). During prometaphase I, SEP-1 localizes to linear elements in GFP::IFY-1^WT or GFP::IFY-1^DM embryos (Fig. 5 C and Video 6). After anaphase I onset, SEP-1::mScarlet accumulates on vesicles in GFP::IFY-1^WT but not GFP::IFY-1^DM embryos (Fig. 5, C and D). In GFP::IFY-1^WT embryos, SEP-1::mScarlet rapidly accumulates on vesicles ∼30 s after anaphase I onset, while 11/12 GFP::IFY-1^DM embryos showed weak and partial vesicle localization after a significant delay (207 ± 37 s). In GFP::IFY-1^WT embryos, SEP-1::mScarlet localized to 40 ± 1 vesicles (N = 11 embryos), while GFP::IFY-1^DM embryos had only 4 ±1 SEP-1::mScarlet-positive vesicles (N = 12 embryos, Fig. 5 E). Therefore, GFP::IFY-1^DM interferes with cortical granule localization of SEP-1::mScarlet in anaphase I.

Previous studies have predominantly characterized separase regulation using in vitro assays (Rosen et al., 2019; Shindo et al., 2012; Yaakov et al., 2012). Separase is regulated by securin degradation (Cohen-Fix et al., 1996; Funabiki et al., 1996), autocleavage (Waizenegger et al., 2002), phosphorylation (Stemmann et al., 2001), isomerization (Hellmuth et al., 2015a), and CDK binding (Gorr et al., 2005; Yu et al., 2021). However, the localization of separase and its regulation at different subcellular locations has not been well characterized. Underscoring the importance of separase localization, human cancer cells show aberrant separase nuclear localization (Meyer et al., 2009). Given that separase regulates centriole duplication (Tsou et al., 2009), anaphase spindle dynamics (Jensen et al., 2001), and vesicle exocytosis (Bembenek et al., 2007, 2010), different control mechanisms might regulate these cellular processes. Our results show that the spatiotemporal control of separase localization is critical for the proper regulation of anaphase events.

Securin is a pseudo-substrate inhibitory chaperone of separase (Boland et al., 2017; Hellmuth et al., 2014, 2015a, 2015b; Hornig et al., 2002; Viadiu et al., 2005). As such it serves multiple regulatory roles including (1) positively regulating the folding of separase (Hellmuth et al., 2015a); (2) enabling its spindle and nuclear localization (Hornig et al., 2002); (3) inhibiting its protease domain; and (4) acting as a competitive APC/C substrate for proper timing of events (Kamenz et al., 2015; Lu et al., 2014; Thomas et al., 2021). Securin loss could cause separase to become unfolded and inactive, mislocalized, and/or prematurely active based on its known functions. The reduction in separase signal we observed after securin depletion could reflect mislocalization and/or unstable separase. In addition, we observe that securin depletion causes premature relocalization of separase to sites of action and causes premature loss of cohesin from chromosomes. Therefore, part of the securin depletion phenotype is due to a premature activation of separase.

Interestingly, GFP::IFY-1^DM does not inhibit the midbivalent and spindle localization of separase but significantly interferes with vesicle localization. This indicates that securin degradation is necessary but not sufficient to regulate separase localization in anaphase. The colocalization of separase and securin with kinetochore proteins on chromosomes and linear elements suggests that kinetochore proteins bind to the separase/securin complex. Kinetochore proteins relocalize to the midbivalent in anaphase (Dumont et al., 2010), where they may still be capable of binding separase and GFP::IFY-1^DM. However, linear elements disappear during anaphase, and kinetochore proteins are not known to localize to vesicles. Separase may bind directly to a substrate on vesicles, which would be inhibited by securin. We propose a model where separase is both kept inactive on kinetochore structures and away from substrates. APC/C-mediated securin degradation liberates separase protease activity and enables separase relocalization to sites of action.

We suspect that the dynamic relocalization of separase reflects an active transport mechanism. For example, spindle checkpoint inactivation at chromosomes involves dynein-mediated transport of proteins from the kinetochore to the centrosome (Griffis et al., 2007; Howell et al., 2001). In addition, APC/C regulates the microtubule motor activity required for spindle translocation (Yang et al., 2005). Finally, linear elements were shown to regulate cortical microtubules during polar body extrusion (Quiogue et al., 2023), which could also control separase localization. The spatiotemporal regulation of separase is an important facet of the metaphase-to-anaphase transition and may enable precise substrate cleavage by promoting a high local concentration of the enzyme near substrates.

Worm strains were maintained using standard protocols (Brenner, 1974; Mitchell et al., 2014). Some strains were obtained from the Caenorhabditis Genetics Center.

The ify-1 locus was PCR-amplified from genomic DNA to include restriction sites at the 5′ (SpeI) and 3′end (MluI) for integration into the pJK3 plasmid, following established protocol (Gibson et al., 2009). The pJK3 plasmid allows N-terminal GFP fusion protein expression through the pie-1 promoter. The highly conserved arginine (Arg38) and leucine (Leu41) within the conserved destruction box were mutagenized to alanine (RxxL -> AxxA) using the Quickchange mutagenesis kit (Stratagene). Worms were transformed with the pJK3 plasmid carrying gfp::ify-1^dm using microparticle bombardment as previously described (Praitis et al., 2001).

Cloning and mutagenesis primers are listed below. Restriction sites are underlined and mutagenized residues are boldened.

ify-1 Forward cloning prime with SpeI site:

5′-CGC​TCT​AGAACT​AGTATG​GAG​GAT​CTA​AAC-3′

ify-1 Reverse cloning prime with MluI:

5′-ACG​CGTTCA​CAG​GGG​AAG​GTT​GGC​TTC​TTC-3′

ify-1^dm Forward destruction box mutagenesis primer:

5′-GGTGCGGGG​CTG​GTT​GTA​AAC​TCG​TCA-3′

ify-1^dm Reverse destruction box mutagenesis primer:

5′-AGT​CGA​GTT​TAC​AAC​CAG​CCCCGCACC​GCCAGCAGA​AGG-3′.

Multiple transformed worm lines were isolated and maintained using two methods we developed for the maintenance of worms harboring toxic transgenes (Mitchell et al., 2014). The first method exploits the transgenerational inheritance of siRNAs in C. elegans. Feeding animals gfp inhibits the expression of toxic GFP transgenes for an average of five generations in C. elegans. Using this first method, animals harboring gfp::ify-1^dm were grown on gfp RNAi plates for routine propagation. Animals carrying the GFP::IFY-1^DM transgene were then transferred onto OP50 plates and propagated for five generations. Fifth-generation hermaphrodites (F5) were selected, grown to adulthood, and their progeny were characterized. The second method involves mating male worms harboring our pie-1 promoter-driven toxic transgenes with hermaphrodites. Using this method, we can indefinitely propagate gfp::ify-1^dm transgenes through the male germline because the pie-1 promoter is not well expressed in the male germline. Male worms were mated with hermaphrodites and their F1 hermaphrodite progeny were characterized.

CRISPR/Cas9 was used to generate endogenously tag the wildtype ify-1 locus at the N-terminus with GFP and the wildtype sep-1 locus at the C-terminus with mScarlet, as previously described (Paix et al., 2015). The repair templates were amplified from the pDD282 (plasmid # 66823; Addgene) and pMS050 (plasmid # 91826, gifts from Bob Goldstein; Addgene). The primer sequences and repair templates used are listed below. Underlined amino acids denote flexible linker sequences.

ify-1::gfp Forward: 5′-ACG​ACC​TCC​TCG​CCG​AAG​AAG​CCA​ACC​TTC​CCC​TGGGA​GCA​TCG​GGA​GCCGGA​GCA​TCG​GGA​GCC-3′

ify-1::gfp Reverse: 5′-AAA​CAG​GTA​GAA​GAG​GCT​GAC​GTC​GTG​GGA​AAT​CAC​TTG​TAG​AGC​TCG​TCC​ATT​C-3′.

The ify-1::gfp guide RNA: 5′-GAC​GUC​GUG​GGA​AAU​CAC​AGG​UUU​UAG​AGC​UAU​GCU​GUU​UUG-3′

sep-1::gfp Forward:

5′-CAA​GTG​CCC​GAA​CTC​CAT​CAA​GAT​CCC​GAA​ATT​TGGGA​GCA​TCG​GGA​GCC​TCA​GGA​GCA​TCGATG​GTC​TCC​AAG​GG-3′

sep-1::gfp Reverse:

5′-ACG​ATC​CTT​AAG​ATC​CTT​CGG​GTC​AGA​TTA​TAT​TAC​TTG​TAG​AGC​TCG​TCC​ATT​C-3′.

The sep-1::gfp guide RNA:

5′-CAG​AUU​AUA​UUA​CAA​AUU​UCG​UUU​UAG​AGC​UAU​GCU​GUU​UUG-3′.

The coh-3 coding sequence and 3′ UTR were amplified from fosmid WRM068bC06 (Geneservice Ltd.) by PCR with primers AFS357 (5′-GGG​GAC​AGC​TTT​CTT​GTA​CAA​AGT​GGc​tAT​GGT​GAT​AAG​CAT​CGA​TGT​ACC-3′) and AFS358 (5′-GGG​GAC​AAC​TTT​GTA​TAA​TAA​AGT​TGG​CGC​CTT​TAA​AGC​TAC​CTG​TAA​C-3′). The PCR product was cloned into pDONRP2R-P3 via a Gateway BP Cloning reaction (Thermo Fisher Scientific). A single missense mutation identified in the resulting plasmid and the parental fosmid was repaired by site-directed mutagenesis using primers AFS436 (/5Phos/5′-TGAGTACTGAGAACTATGGTGTTTC-3′) and AFS437 (/5Phos/5′-AGTTGCTCGACTTCTTCGTAC-3′), yielding the error-free entry clone pAS139. pAS139 was used in a multisite Gateway LR reaction together with entry plasmids pCG142 (pie-1 intron:pie-1 promoter in PDONRP4P1R) and pCM1.53 (GFP with worm codon bias and synthetic introns in pDONR201) (plasmids # 17246 and # 17250, gifts from Geraldine Seydoux; Addgene) and destination vector pCFJ150 - pDESTttTi5605[R4-R3] (plasmid # 19329, gift from Erik Jorgensen; Addgene). The resulting targeting vector, pAS142, was inserted into the ttTi5605 Mos1 transposon site in strain EG4322 by MosSCI to create the single-copy integrated transgene ySi12[Ppie-1::GFP::coh-3] (Frøkjaer-Jensen et al., 2008; Merritt et al., 2008). ySi12 encodes a functional GFP::COH-3 fusion since it increases the embryonic viability of coh-4(tm1857) coh-3(gk112) double mutants from 3.1% (n = 1246) to 91.8% (n = 1522) and decreases male production from 36% (n = 25) to 3.8% (n = 1316). The low number of coh-4 coh-3 animals scored for male production is due to the low rate of survival of these worms to adulthood; this phenotype is also rescued by the ySi12 transgene.

Feeding RNAi was conducted as previously described using HT115 bacteria harboring the L4440 plasmid (Grishok et al., 2005). For apc-2 and ify-1 feeding RNAi, L4 hermaphrodites from WH416, JAB20, and JAB258 (Table S1) lines were plated onto the induced RNAi bacterial strains at 20°C or 25°C, and phenotype severity and penetrance were assessed after 14–48 h of treatment.

Degradation curve values are expressed as ratios reflecting the mean cytoplasmic GFP intensity in the newly fertilized oocyte relative to mean cytoplasmic GFP values in the −1 oocyte over time. Values for each timepoint correspond to an average of at least five independent movies for each IFY-1^WT or IFY-1^DM strain.

Lethality assays were performed as previously described (Mitchell et al., 2014). Lines expressing GFP::IFY-1^WT or GFP::IFY-1^DM were grown under identical conditions at 20°C or 25°C and embryo lethality was quantified. Lethality rates reflect the pooled average of embryonic lethality for each strain and condition after 24 h.

The polar body extrusion assay was performed using embryos dissected from mothers homozygous for H2B::mCherry and GFP::IFY-1^WT or GFP::IFY-1^DM five generations removed from gfp RNAi feeding. We quantified two cell–stage embryos to allow for the completion of meiosis and quantify polar bodies before the second polar body is internalized and degraded in older embryos (Fazeli et al., 2018).

The osmotic sensitivity assay was performed by dissecting wildtype (N2) or homozygous GFP::IFY-1^DM mutant embryos in a hypertonic solution of 300 mM KCl, as described (Knight et al., 2012). Animals were five generations removed from gfp RNAi feeding. Embryos were scored for normal appearance or obvious shrinkage.

Live cell imaging data was collected using spinning disk confocal systems using either an inverted Nikon Eclipse microscope with a 60× 1.40NA objective, a CSU-22 spinning disc system, and a Photometrics EM-CCD camera from Visitech International operated by MetaMorph software (Molecular Devices), or an inverted Nikon Eclipse Ti2-E with a 60× 1.42NA objective and 100× 1.45 NA objective, a CSU-X1 spinning disk system, and Andor iXon Life camera operated by NIS-Elements software (Nikon). Unless otherwise mentioned, live cell imaging was conducted at room temperature which was ∼20°C. Anaphase onset was determined using the movement of chromosomes or the presence of separase and/or securin at the midbivalent. Image analysis and manipulation were performed in Fiji (National Institutes of Health), Adobe Photoshop, and Adobe Illustrator (Adobe).

We used two immobilization methods to mount animals to image oocytes and embryos. The first method was an optimized nanoparticle-mediated immobilization technique based on a previously described protocol (Kim et al., 2013). This first strategy was used for Fig. 1, B–M; Fig. 3, A–C′′; Fig. 4, C–F; Fig. S1, C, E, and F; Fig. S2, C–G; and Fig. S3, A–E. We also used a chemical immobilization method by mounting worms in an M9 solution containing 5 mM levamisole on a 2% agarose pad following standard protocol (Bai and Bembenek, 2017; Bembenek et al., 2007; Mitchell et al., 2014). We used this second strategy for initially documenting control, apc-2, and ify-1 RNAi phenotypes, and for data presented in Fig. 3, A–C′′ and Fig. S3, A–E.

Before eggshell formation, meiotic embryos are especially fragile to osmotic and mechanical perturbations (Stein and Golden, 2018). To minimize perturbations ex utero, meiotic embryos were dissected from hermaphrodites in blastomere culture media using the hanging drop mounting technique (Edgar and Goldstein, 2012). Parental carcasses were removed from the media along with bacteria to prevent toxic effects associated with their presence when left in the media (Bai and Bembenek, 2017; Bembenek et al., 2007; Mitchell et al., 2014). Cortical granules have weak autofluorescence under 488-nm illumination that is quickly bleached after 25 exposures with standardized settings. The autofluorescent signal can be visually distinguished because it fills the vesicle, while our GFP fusion proteins appear to coat the membrane of the vesicle. Therefore, in conditions with weak GFP signal, we systematically performed an autofluorescence prebleach exposure before imaging GFP localization. For all ex utero imaging, L4 hermaphrodites were shifted from 20°C to 25°C for 18–24 h before imaging at either room temperature or 25°C. This approach was used for acquiring data for Fig. 1, E, N, and O; Fig. 2, A–D; Fig. 4 G; Fig. S1, A, B, and D; Fig. S2, A and B; Fig. S4, A, B, and F–K; and Videos 1, 2, 3, 4, 5, and 6.

Fluorescent values for degradation curves in Fig. 1 were acquired from single plane, ex utero movies of meiosis I embryos at approximately the same z-depth and using the same acquisition settings. The values represent the binned average signal from two to five independent movies, found in a three-pixel diameter circle at the spindle, filament, and cytoplasm. The values for each movie and each timepoint at these regions of interest are averages of between one and five independent measurements per movie minus the average background signal.

Fluorescent values for the securin degradation curves in Fig. S1 were acquired from single-plane, in utero movies of meiosis I. Worm age, imaging conditions (room temperature), and acquisition settings were identical for all data acquisition. Values represent the binned average of at least two to four independent movies, found in a 10-pixel diameter circle in the bulk cytoplasm. The values for each movie and each timepoint are averages of three independent cytoplasmic measurements minus the average background signal.

Quantification of prometaphase I GFP::COH-3 levels in Fig. S3 F were made from single-plane, in utero, movies after RNAi treatments (control and ify-1). One to three 3-pixel diameter circle region of GFP::COH-3 and H2B::mCherry signal were measured and averaged, subtracting the average background signal within a three-pixel diameter circle and expressed as a ratio. Curves reflect COH-3 levels over time for a single embryo for each condition.

Quantification of prometaphase I GFP::COH-3 levels in Fig. S3 G were made from single-plane, in utero, movies after RNAi treatments (control, apc-2, and ify-1). The calculated signal ratio corresponds to the midbivalent signal of the −1 oocyte to a prometaphase I oocyte in the spermatheca acquired in the same plane in a single image. A three-pixel diameter circle region of GFP::COH-3 signal was measured, subtracting the average background signal within a three-pixel diameter circle. Each condition is an average of 5–13 independent worms.

Quantification of SEP-1::GFP kinetochore-to-cytoplasm ratios in Fig. S2, F and G were made from single-plane, in utero, movies and images after RNAi treatments (control, apc-2, and ify-1). Specifically for S2 F and SEP-1::GFP kinetochore values, two to five independent measurements were averaged from pixel values contained within a two-pixel diameter circle and subtracted from the average value of background from a 20-pixel diameter circle. For SEP-1::GFP cytoplasm values, a single measurement of a 20-pixel diameter circle was made with the background subtracted from a single measurement from the same-sized circle. Specifically for Fig. S2 G, we chose to analyze ideal cross sections of bivalent pairs per previously described parameters (Danlasky et al., 2020). To calculate SEP-1::GFP at the kinetochore and within the midbivalent region, GFP values were averages from three, single-pixel measurements from each structure on a per-bivalent basis, and the average background from GFP values within a 20-pixel diameter circle were subtracted out.

We were unable to generate viable animals with homozygous GFP::IFY-1^DM and homozygous SEP-1::mScarlet, but heterozygous GFP::IFY-1^DM was viable when combined with heterozygous SEP-1::mScarlet. Extensive troubleshooting was taken to avoid phototoxicity and photobleaching, while simultaneously being able to evaluate SEP-1 vesicle localization. To time anaphase I onset, we imaged GFP alone using a single plane until GFP::IFY-1^DM was detected at the midbivalent. We then quickly switched to multiplane imaging (5 µm steps × 3 planes) acquired every 45–120 s so that we could capture cortical planes over the course of an extended anaphase. Wildtype data was acquired only using multiplane ex utero imaging because embryos were not sensitive to phototoxicity or photobleaching problems within the normal duration of anaphase I. Quantification of SEP-1::GFP signal at vesicle structures was performed on movies with z planes that had a similar circumference to ensure a similar region of the cortex was quantified.

To quantitatively assess separase kinetochore localization after RNAi treatments, a one-pixel-wide line scan was generated from representative images along the prometaphase I spindle axis through a clearly resolved cross section of a homologous chromosome pair. Fluorescent intensities were generated using Fiji for both separase and H2B signal with the average background subtracted out. Average bulk cytoplasm was determined, with average background subtracted out. Background corrected values for the line scan were divided by background corrected values for the bulk cytoplasm for normalization.

Calculations for P values were done in Microsoft Excel using Student’s t test (two-tailed, assuming unequal variance) to determine statistical significance.

Fig. S1 shows separase and securin localization during meiosis I, quantification of securin expression lines, and separase colocalization with spindle markers (related to Figs. 1 and Fig. 2). Fig. S2 shows in utero images of separase localization in different conditions (related to Fig. 3). Fig. S3 describes cohesin dynamics, morphology, and regulation during meiosis I (related to Figs. 2 and 3). Fig. S4 documents the dominant negative phenotypes caused by the expression of non-degradable securin (related to Figs. 4 and 5). Table S1 lists the C. elegans strains used in this study. Video 1 shows the dynamics of separase, cohesin, and securin on chromosomes during the metaphase-to-anaphase transition (related to Figs. 1 and 2). Video 2 shows separase dynamics relative to chromosomes, cohesin, and securin at anaphase I onset (related to Fig. 2). Video 3 shows securin and separase in the cortex during metaphase–anaphase I (related to Fig. 2). Video 4 shows cortical granule exocytosis during meiosis I in embryos expressing WT or non-degradable securin (related to Figs. 4 and 5). Video 5 shows that non-degradable securin does not affect separase relocalization to the midbivalent in anaphase I (related to Figs. 4 and 5). Video 6 shows that non-degradable securin prevents separase from localizing to cortical granules during anaphase I (Related to Figs. 4 and 5).

The data are available from the corresponding author upon reasonable request.

We thank Julie Ahringer, Arshad Desai, Barth Grant, Tony Hyman, Karen Oegema, Martha Soto, and Asako Sugimoto for sharing strains and members of the Bembenek laboratory for comments and support.

CGC and Wormbase provided C. elegans strains and information, funded by the National Institutes of Health (NIH) (P40 OD010440) and the National Human Genome Research Institute (U41 HG002223). Funding was provided by the NIH grants R15 GM143731 to A.F. Severson and R01 GM114471 to J.N. Bembenek.

Author contributions: C.G. Sorensen Turpin: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization, Writing - original draft, Writing - review & editing, D. Sloan: Resources, M. LaForest: Data curation, Investigation, L. Klebanow: Investigation, D. Mitchell: Investigation, Methodology, Resources, Writing - review & editing, A.F. Severson: Conceptualization, Resources, Writing - review & editing, J.N. Bembenek: Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing - original draft, Writing - review & editing.