RAN-2^Ran-GAP and RAN-3^Ran-GEF Are Enriched on Anaphase Chromosomes

A previous study depleting Ran in C. elegans by ran-1(RNAi) showed female meiotic spindles still formed although with a reduced density of microtubules and showed relatively normal anaphase progression (Chuang et al, 2020). As RAN-1 may not have been fully depleted by RNAi, we created conditional knockdown worm strains of Ran-GEF and Ran-GAP by adding Auxin Induced Degron (AID) (Zhang et al, 2015) and HALO tag sequences to the endogenous ran-2^Ran-GAP and ran-3^Ran-GEF loci. RAN-2::AID::HALO and RAN-3::AID::HALO worms laid no eggs on auxin plates (Fig 1B). Previous depletion of RAN-3 or RAN-2 by RNAi caused >90% embryonic lethality but did not affect brood size (Askjaer et al, 2002). This suggests that depletion of RAN-2 or RAN-3 through the AID system results in more complete depletion than RNAi.

It has been proposed that chromosome associated RCC1(Ran-GEF) generates Ran-GTP near chromatin whereas cytoplasmic Ran-GAP generates Ran-GDP far from chromosomes. Ran-GTP proximal to chromosomes releases inactive SAFs from binding with importins (Fig 1A). Consistent with these ideas, we observed RAN-3::AID::HALO in the nucleoplasm and chromosomes before GVBD (Fig 1C). After GVBD, RAN-3 diffused from the nucleus and was not detected on chromosomes during spindle assembly in 12/12 embryos. RAN-3 only faintly associated with chromosomes at metaphase I and II. In contrast, RAN-3 strongly localized to chromosomes at anaphase I and anaphase II (Fig 1C). RAN-2::AID::HALO was strongly associated with the nuclear envelope before GVBD in 11/11 oocytes. Later, RAN-2 faintly labeled metaphase I and metaphase II spindles. At anaphase, RAN-2 localized to the spindle midzone, and its intensity increased as anaphase progressed. It is strongly associated with the inner side of separating chromosomes (Fig 1D). These results suggest that RAN-3 and RAN-2 might function primarily at anaphase.

Ran-GEF Is Required for Extrusion of the First Polar Body but not Metaphase I Spindle Formation

Ran regulators are also involved in nuclear transport, defects in which usually lead to small and leaky nuclei. By treating the worms with auxin for a brief period, we sought to only evaluate the function of RAN-3 and RAN-2 on meiotic spindle formation without disrupting meiotic prophase and nuclear transport. When treated with auxin for 4 h, expression of RAN-3::AID::HALO in the nucleoplasm and chromosomes in diakinesis oocytes was significantly reduced (Fig S1A and B). At 6-h auxin treatment, the expression level was comparable to control oocytes with no HALO expression. Moreover, −1 oocytes treated with auxin for 4 or 6 h were smaller than controls, whereas 36-h auxin resulted in much smaller nuclei (Fig S2A and B). Qualitatively, GFP:: tubulin partially leaked from the cytoplasm into nuclei after 4 h of auxin but GFP::tubulin leaking into the nuclei was much more severe at 36 h of auxin (Fig S2C). These results indicated that RAN-3 depletion after 4 h of auxin results in pre-NEBD defects and defects are more severe after 36 h of auxin. Expression of RAN-2::AID::HALO in the nucleoplasm and nuclear envelope in diakinesis oocytes was reduced to control levels after 4 h of auxin (Fig S1C and D). The sizes of oocyte nuclei were normal in RAN-2::AID::HALO worms after 4 or 6 h of auxin, and their nuclei were not leaky compared with controls (Fig S2A–C). RAN-2::AID::HALO oocytes treated with auxin for 36 h were not quantified as these oocytes were severely disorganized. We therefore analyzed meiotic spindle assembly by time-lapse imaging after 4–6 h of auxin because these timepoints showed strong depletion but no detectable effect on nuclear function for RAN-2 and only moderate early effects for RAN-3.

Bipolar metaphase I spindles were observed in control worms (n = 15 no degron plus auxin; n = 10 ran-3::AID no auxin; n = 5 ran-2::AID no auxin) (Fig 2A; Video 1). Consistent with the dispersal of RAN-3 and RAN-2 at NEBD (Fig 1C and D), 4–6-h auxin treatment of RAN-2::AID::HALO (n = 33) or RAN-3::AID::HALO (n = 26) worms (Fig 2B and C) also resulted in bipolar metaphase I spindles. The mean metaphase I spindle GFP::tubulin pixel intensities were significantly decreased after RAN-3 depletion relative to both no degron controls or no auxin controls (Fig 2E and F). The length of metaphase I spindles was not significantly different than controls after depletion of RAN-2 or RAN-3 (Fig S1E) although a slight increase in metaphase II spindle length was observed after RAN-2 depletion (Fig S1F). The velocity of anaphase I (reported as the rate of increase in distance between separating chromosome masses) was significantly faster than no degron or no auxin controls after depletion of RAN-3 (Fig 2G). However, the velocity of anaphase II was not significantly different than controls (Fig S1G).

In 5/16 time-lapse sequences of RAN-3-depleted worms, chromosomes that moved toward the cortex at anaphase I merged into the metaphase II spindle. When viewed end-on, control metaphase II spindles have six univalents arranged in a pentagonal array (Fig 2A 810 s). In a side view of this pentagonal array during metaphase or early anaphase, three chromosome pairs are in focus (Fig 2C 1,020 s). Merging of polar body-bound anaphase I chromosomes into the metaphase II spindle was inferred from more than six chromosomes in an end-on view (Fig 2B 1,060 s) or from direct observation when merging occurred in a favorable focal plane (Video 2). However, this frequency of polar body failure was not significantly different than the 1/10 failures observed in no auxin controls of the same strain (Fig 2D; P = 0.35 Fisher’s exact test). Thus, failure to extrude the first polar body might be an artifact of time-lapse imaging or the genetic background. Conversely, the actual polar body failure rate might be much higher if the chromosomes that segregated toward the cortex at anaphase I do not move inward until later in development. To address both issues, we collected z-stacks of post-meiotic pronuclear to four-cell stage embryos dissected from worms at 5–7 h after placing worms on auxin. These embryos underwent meiosis 4–6 h after auxin treatment and the worms were not mounted for time-lapse imaging when these embryos underwent meiosis. 43/55 RAN-3 depleted embryos had a single polar body, whereas 29/29 no auxin controls of the same strain had two polar bodies (Fig 2D). This difference was significant (P = 0.0001 Fisher’s exact test). Because the chondroitin layer of the eggshell is secreted during anaphase I (Olson et al, 2012), the first polar body is embedded in the eggshell and remains stationary at the anterior tip of the oval embryo, whereas the second polar body is inside the eggshell and moves to an internal location over time. Consistent with this, 18/19 no auxin control one-cell embryos had two polar bodies at the anterior tip whereas 19/20 no auxin control four-cell embryos had one polar body at the tip and one polar body internal. The single polar body of 14/15 RAN-3 depleted one-cell embryos was at the anterior tip whereas the single polar body of 15/15 RAN-3 depleted four-cell embryos was internal. The four-cell result was significantly different than a random distribution of 7/14 (P = 0.002 Fisher’s exact test) and suggests that RAN-3 depletion causes resorption of chromosomes destined for the first polar body. Polar body extrusion defects after RAN-2 depletion compared with the same strain with no auxin were less definitive (P = 0.052 Fisher’s exact test), and chromatin was extremely condensed in RAN-2 depleted post meiotic embryos, making it difficult to distinguish polar bodies from micronuclei at the cortex. Overall, these results suggest that the Ran-GTP pathway is important for limiting anaphase I velocity and polar body formation. Notably, RAN-3 is concentrated on chromosomes during anaphase and polar body extrusion (Fig 1). RAN-3 may have a more limited role in promoting the density of spindle microtubules at metaphase I when its localization on chromosomes is not discernible (Fig 1).

Free Tubulin Is Concentrated in the Nuclear Volume at GVBD

At the onset of mitosis in C. elegans and Drosophila, soluble tubulin concentrates in the nuclear/spindle volume relative to the surrounding cytoplasm (Hayashi et al, 2012; Yao et al, 2012; Schweizer et al, 2015; Métivier et al, 2021). Because of microtubule polymerization is concentration dependent, this concentration might facilitate spindle formation in the vicinity of chromosomes in meiotic oocytes. In DMSO-treated control −1 oocytes, mNG::TBB-2 (β-tubulin) labeled microtubules in the cytoplasm and was excluded from the nucleus (Fig 3A and B). Upon fenestration of the nuclear envelope (GVBD: germinal vesical breakdown), indicated by leakage of non-chromosomal mCherry::histone out of the nucleus, mNG::TBB-2 fluorescence increased within the nuclear volume in 10/10 time-lapse sequences (Fig 3B) as previously described (McNally et al, 2006; Mullen & Wignall, 2017). Tubulin fluorescence then transformed into a “microtubule cage” (Mullen & Wignall, 2017) and eventually a bipolar spindle in 10 out of 10 time-lapse sequences. In oocytes treated with nocodazole to depolymerize microtubules, tubulin was diffuse in the cytoplasm before GVBD (Fig 3C), and still concentrated in the nuclear volume during and after GVBD (Fig 3C–E). Simple diffusion from the cytoplasm into the nuclear volume should result in equal fluorescence intensities of mNG::TBB-2 in the nucleus and cytoplasm but fluorescence instead increased in the nuclear volume to 1.2-fold greater than the cytoplasm (Fig 3E). After ovulation, chromosomes in nocodazole-treated zygotes were dispersed and loosely wrapped by sparse short microtubules. No spindle formation or chromosome separation was observed before pronucleus formation in 12 out of 12 time-lapse sequences.

We noticed that before GVBD, fluorescence of mNG::TBB-2 was detectable in the nucleus, which contradicts with the general assumption that the nucleus is void of tubulin (Fig 3B top panel, Fig S2B, average of the ratio of tubulin fluorescence in the nucleus to the cytoplasm: 0.69). To determine whether there is tubulin in the nucleus or if this is because of pinhole crosstalk from the spinning disk confocal, we captured images on a Zeiss laser scanning confocal microscope which has a single pinhole and therefore removes out of focus light more efficiently. The ratio of tubulin fluorescence in the nucleus to the cytoplasm before GVBD was significantly reduced to 0.33 (Fig S3A and B) in images from the Zeiss LSM compared with 0.69 from the spinning disk confocal. The difference in nuclear versus cytoplasmic fluorescence was also greater on the Zeiss LSM for other probes (Fig S3C and D). This suggests that the actual concentration difference between nucleus and cytoplasm are likely greater than the ratios reported from spinning disk confocal images.

Similar results in previous studies of mitosis led to the interpretation that alpha/beta tubulin dimers concentrate in the nuclear volume of unperturbed cells during spindle assembly. However, it is possible that nocodazole does not completely block microtubule polymerization and the fluorescence accumulating in the nuclear volume of nocodazole-treated oocytes represents accumulation of short microtubules. It is also possible that this phenomenon is induced by nocodazole and does not occur in unperturbed cells.

Accumulation of Tetrameric GFP and Un-Polymerizable Tubulin in the “Nuclear Volume” at GVBD

Previous investigators suggested that tubulin dimers concentrate either by binding to something in the nuclear volume (Hayashi et al, 2012; Métivier et al, 2021) or by being excluded by cytoplasmic organelles that are kept out of the spindle volume by the ER envelope that still envelopes the spindle after NEBD (Schweizer et al, 2015, Fig 4A). We analyzed the behavior of two fluorescent probes designed to address three major issues: (1) incomplete depolymerization by nocodazole; (2) concentration of alpha/beta dimers in the absence of nocodazole; (3) specific binding of tubulin to a nuclear binding site versus volume exclusion by cytoplasmic organelles. GFP::GCN4-pLI is a tetramerized GFP designed to have a native molecular weight (128 kD) similar to an alpha/beta tubulin dimer (101 kD without tag, 128 kD with mNeonGreen tag, see Fig 7A), but which should not bind to any tubulin-specific binding sites in the nuclear volume because GCN4-pLI is a well characterized synthetic four helix bundle (Mittl et al, 2000). GFP::TBA-2(T349E) is an alpha tubulin mutant that can dimerize with beta tubulin, cannot polymerize (Johnson et al, 2011), and should bind any specific tubulin binding sites in the nuclear volume. Both GFP::GCN4-pLI (Fig 4B; Video 3) and GFP::TBA-2(T349E) (Fig 4C; Video 4) concentrated in the nuclear volume at GVBD in the absence of nocodazole. These results (1) suggested that the concentration of mNG::TBB-2 in nocodazole was not because of incomplete depolymerization; (2) that tubulin dimers concentrate in the nuclear volume during unperturbed spindle assembly; and (3) that binding to a tubulin-specific binding site in the nuclear volume is not required for concentration.

Because GFP::GCN4-pLI and GFP::TBA-2(T349E) could be tracked in the absence of nocodazole, we could examine their behavior during normal meiotic divisions. Chromosome segregation was normal in GFP::TBA-2(T349E) oocytes and GFP::GCN4-pLI oocytes (12/12 and 10/10 filmed respectively) suggesting that expression of GFP::TBA-2(T349E) or GFP::GCN4-pLI did not severely disturb normal spindle function. Interestingly, their concentration at GVBD lasted through Metaphase I but diffused to a 1:1 spindle: cytoplasm ratio at Anaphase I, followed by re-accumulation at metaphase II and dispersion at Anaphase II (Fig 4B–D; Video 4 and Video 5). This raises the question of what barrier between spindle and cytoplasmic volume might change between anaphase I and metaphase II.

The ER Delimits the Accumulation of Free Tubulin during Meiosis and Early Mitosis

Re-accumulation of tubulin within the metaphase II spindle volume could be because of nuclear import if the nuclear envelope transiently reformed or the re-accumulation could be because of cell-cycle changes in the structure of the ER that surrounds the spindle (Fig 5A; Poteryaev et al, 2005). To test whether a functional nuclear envelope reforms transiently between anaphase I and metaphase II, we tracked GFP::NPP-6, a nuclear pore protein of the Y complex (Galy et al, 2003) (Fig 5B) and GFP::LMN-1, the C. elegans nuclear Lamin (Galy et al, 2003) (Fig 5C) by time-lapse imaging. Both proteins disappeared from the nuclear envelope at GVBD and did not re-locate to the nuclear membrane until pronuclear formation (Fig 5B and C). The absence of NPP-6 from a nuclear envelope during meiosis has been previously reported (Penfield et al, 2020). These results make it unlikely that tubulin concentrates in the spindle volume at metaphase II because of nuclear import. In contrast, the ER is contiguous with the outer nuclear membrane before GVBD (Fig 5A and D). After GVBD, the ER wraps around meiotic spindles with clustering at spindle poles during metaphase I and metaphase II (Fig 5D). In contrast, yolk granules (Fig 5D), maternal mitochondria (Fig 5E), and lipid droplets (Fig 5F) are excluded from the nuclear and spindle region. At the light microscope level (Fig 5D), the ER appears reticular during metaphase I and metaphase II and disperses during anaphase I and anaphase II as described previously (Poteryaev et al, 2005). To determine the ultrastructural changes in ER morphology, we manually segmented the ER in previously published electron tomograms (Lantzsch et al, 2021). During metaphase I and metaphase II, the sides of the spindle are enclosed by overlapping sheets of ER (Video 6, Fig 6A; Video 7) and the accumulation of ER at the spindle poles is a complex mixture of sheets and tubules (Fig 6B; Video 8). During anaphase I and anaphase II, the ER near the spindle consists entirely of tubules (Fig 6C and D; Video 9). By time-lapse imaging of worms expressing an ER marker as well as GFP::GCN4-pLI, we found that the dramatic morphological change of ER coincides with the concentration and dispersion of GFP::GCN4-pLI during meiosis (Video 5) and mitosis (Video 10). Thus, tubulin-sized proteins accumulate in the spindle volume when it is encased by sheet-like ER and disperse when the ER is tubular. These results favor a model in which a semi-permeable spindle envelope composed of sheet-like ER excludes a “crowding agent” from the spindle volume, so that tubulin-sized proteins concentrate in the spindle volume by volume exclusion from the surrounding cytoplasm (Fig 4A).

The “crowding agent” might consist of membranous organelles like yolk granules (Fig 5D), or mitochondria (Fig 5E) as suggested by Schweizer et al (2015) or might consist of ribosomes as suggested by Delarue et al (2018). To test whether ribosomes are a viable candidate for excluding volume from the cytoplasm, we determined the density of ribosomes inside of the spindle and outside the ER envelope from previously published electron tomograms (Lantzsch et al, 2021) and found no significant difference in ribosome density inside versus outside the spindle (Fig 6E). As a complementary approach, we monitored a ribosomal protein, GFP::RPL-29, by time-lapse imaging before and after GVBD (Fig 6F). GFP::RPL-29 was excluded from the nucleus before GVBD, then rapidly rushed into the nuclear volume at GVBD in 9/9 time-lapse sequences, indicating that ribosomes could not be the cytoplasmic crowding agent causing concentration of tubulin-sized proteins that stay concentrated through metaphase.

To test the idea that yolk granules, mitochondria, and other organelles are the crowding agent that excludes soluble tubulin from the cytoplasm, we segmented all the membranous organelles visible in previously described electron micrographs (Howe et al, 2001; Bembenek et al, 2007; Fig S4A–D). In one thin section of a −1 oocyte, membranous organelles occupied 25% of total non-nuclear area. In two different sections covering an entire metaphase II embryo, organelles occupied 21% and 22% of total area. Membranous organelles occupied 28% and 31% of the non-spindle area in these sections. The reduced cytoplasmic volume outside the nucleus or metaphase spindle would thus be between 79% and 69%. The expected apparent concentration of molecules in the nuclear/spindle volume would thus be 1/.79–1/.69 or 1.3–1.4-fold. This is reasonably close to the observed 1.2-fold enrichment given the limitations of both measurements.

Notably, GFP::RPL-29 did not concentrate in the nuclear volume at GVBD but instead equilibrated to equal fluorescence intensity inside and outside the spindle envelope. This result indicated that there is specificity to the types of molecules that concentrate in the nuclear/spindle volume.

The Concentration of Molecules during GVBD Is Size Dependent

In immature starfish oocytes, fluorescent dextrans of 25 kD or larger are excluded from the nucleus (Lénárt et al, 2003), similar to our results with mNG::TBB-2, GFP::GCN4pLI, and GFP::TBA-2(T349E), presumably because they are too large to diffuse freely through NPCs, and/or because of nuclear export of tubulin (Schwarzerová et al, 2019). In contrast, 10 kD fluorescent dextrans accumulated in the nucleus of immature starfish oocytes at a concentration twice that of the cytoplasm (Lénárt et al, 2003). It was suggested that this is because the small dextrans diffuse freely through NPCs and because yolk granules occupy 50% of the cytoplasmic volume thus driving apparent concentration of small dextrans in the nucleus (Lénárt et al, 2003). We found that a 36 kD monomeric GFP concentrated in C. elegans oocyte nuclei before GVBD (Fig 7A and B) to a concentration twice that of the cytoplasm (Fig 7D), like 10 kD dextrans in starfish oocytes. To rule out the possibility that this might be mediated by a cryptic NLS on GFP, we expressed a 34 kD monomeric HALO tag in the C. elegans germline, which also concentrated to a twofold higher concentration in the nucleus relative to the cytoplasm in diakinesis oocytes before GVBD (Fig 7C–E).

These results suggested that the same mechanisms driving concentration of larger proteins during GVBD might be responsible for the concentration of smaller proteins before GVBD. However, monomeric HALO tag did not stay concentrated after GVBD and instead diffused to a 1:1 fluorescence ratio inside and outside the spindle envelope (Fig 7E and F). The monomeric HALO tag did not grossly perturb meiotic progression (Fig 5E) in 12/12 time-lapse sequences. These results suggested that there might be a size-dependence for concentration of proteins in the nuclear volume during meiotic spindle assembly.

Molecule Movement during GVBD Is Charge Dependent

Although the tubulin-sized molecules that concentrated in the nuclear volume at GVBD are larger than the smaller HALO tag that quickly dispersed to a 1:1 ratio between nuclear and cytoplasmic volumes at GVBD, these proteins also differ in net charge (Fig 7A), with the concentrating proteins more negative than the dispersing proteins. Single molecule diffusion studies in cytoplasm (Xiang et al, 2020) and inside organelles (Xiang et al, 2023) have revealed that net positive charge can slow or stop diffusion of small proteins. One possibility is that proteins with positive charge would interact transiently by ion exchange with negatively charged phosphatidyl serine-rich intracellular membranes, and this electrostatic interaction would slow diffusion of positively charged proteins into a fenestrated spindle envelope but would not affect diffusion of negatively charged proteins. To test this idea, we added arginines to GFP::GCN4-pLI to either neutralize net charge or add a net positive charge. Similar to GFP::GCN4-pLI (Fig 7A, net charge at pH 7.0: −28), GFP::GCN4-pLI (neutral), or GFP::GCN4-pLI (positive) (Fig 7A, net charge at pH 7.0: 0 or 56, respectively) were excluded from the nucleus before GVBD (Fig 8A–D; −10 min). GFP::GCN4-pLI (neutral) remained excluded from the nuclear volume at GVBD for a longer period of time after histone leakage out of the nucleus than GFP::GCN4-pLI and exhibited only a slight accumulation in the nuclear volume (∼7 min after GVBD onset) (Fig 8A and C), a significant delay compared with GFP::GCN4-pLI (∼2 min after GVBD onset, Fig 4C). GFP::GCN4-pLI (positive) remained excluded from the nuclear volume for an even longer period after GVBD (Fig 8B and D). This result suggested that proteins with net negative charge concentrate in the nuclear volume at GVBD whereas proteins with neutral or positive charges do not. Interestingly, GFP::GCN4-pLI (extra negative) (Fig 7A, net charge at PH 7.0: −56) did not accumulate in the nuclear volume as GFP::GCn4-pLI did (Fig S5), suggesting there might be a narrow range of charge that are not retained in the cytoplasm.

Transgenes Generated for This Study

syb7781 and syb7819 were generated by inserting Auxin Induced Degron (AID) and HALO tag sequences to the endogenous ran-3 and ran-2 loci through CRISPR/Cas9-mediated genome editing by Suny Biotech. Sequences of these insertions are in Supplemental Data 1. duSi18, duSi20, duSi23, duSi24, and duSi25 were generated by inserting GFP::GCN4-pLI (or its charge variants), or GFP::tba-2(T349E) via Flp Recombinase-Mediated Cassette Exchange (RMCE) method (Nonet, 2020). The plasmids containing corresponding DNAs (GFP::GCN4-pLI, GFP::GCN4-pLI [extra negative], GFP::GCN4-pLI [neutral], GFP::GCN4-pLI [positive], GFP::tba-2[T349E]) were injected into jsSi1579 landing site in NM5402 by In Vivo Biosystems, then the NeonGreen + FLP transgene was removed by outcrossing. Plasmids for RMCE insertion are constructed as follows: DNA sequences of GCN4-pLI and tba-2(T349E) were obtained as gBlocks from IDT followed by restriction digestion and cloned into the vector by T4 DNA ligation. The vector is designed for RMCE insertion, and it contains GFP optimized to reduce germline silencing. For charge variants of GFP::GCN4-pLI, gblocks containing aspartic acid and glutamic acid were cloned into GFP::GCN4-pLI to add negative net charge. To neutralize negative charge of GFP::GCN4-pLI or add positive net charge, 7 arginine or 21 arginine were cloned onto GFP::GCN4-pLI. The strains were sequenced, and the sequences are listed in Supplemental Data 2.

Drug Treatment

The strain used for nocodazole experiments has a bus-17(e2800) mutation which makes the worms cuticle permeable to drugs. 5 mg/ml stock nocodazole solution (Sigma-Aldrich, dissolved in 100% DMSO) was diluted into tricaine/tetramisole anesthetics to 5 μg/ml just before adding to worms for live imaging. 100% DMSO without nocodazole was diluted in the same way for control treatment.

Live Imaging and Statistical Analysis

Worms were anesthetized with tricaine/tetramisole as described (Kirby et al, 1990; McCarter et al, 1999) and gently mounted between a coverslip and a thin 2% agarose pad on a slide. All time lapse images were captured with a Solamere spinning disk confocal microscope equipped with an Olympus IX-70 stand, Yokogawa CSU10, either Hamamatsu ORCA FLASH 4.0 CMOS (complementary metal oxide semiconductor) detector or Hamamatsu ORCA-Quest qCMOS (quantitative complementary metal oxide semiconductor) detector, Olympus 100x UPlanApo 1.35 objective, 100-mW Coherent Obis laser set at 30% power, and MicroManager software control. Pixel size was 65 nm for the ORCA FLASH 4.0 CMOS detector and 46 nm for the ORCA-Quest qCMOS detector. For measuring fluorescence intensity or nucleus sizes (in Figs 2A–C and 7B and C and S1A), z-stack images (except in Fig S2A) were taken on the same microscope, with 1-micron step size to capture the center of nucleus for measurements. Images in Fig S3A were captured with a Zeiss Laser Scanning microscope (LSM) 980, Zeiss Objective LD LCI Plan-Apochromat 40x/1.2 Imm Corr DIC M27 for water, silicon oil or glycerin. For time-lapse movies of GVBD and meiosis, images were captured every 30 s. The fluorescence intensity in the background (where there was no worm) was subtracted before the fluorescence in the nucleus or cytoplasm before quantification.

For segmentation of ER in electron tomograms, ER structures were manually assigned at 10 nm intervals using IMOD software (https://bio3d.colorado.edu/imod/). Ribosomes in 1 μm² sections of spindle and cytoplasm in both metaphase I and metaphase II tomograms were counted using ImageJ software. Threshold values for each section were set to half of the peak value and the number of particles was determined. The average particle size in metaphase I spindles and cytoplasm, and metaphase II spindles and cytoplasm was: 30.9 ± 2.0 nm, 28.0 ± 1.2 nm, 26.3 ± 1.0 nm, and 26.8 ± 1.7 nm, respectively. Anaphase velocities were reported as the increase in distance between chromosome masses divided by time.

Area occupied by membranous organelles was determined from previously reported electron micrographs (Howe et al, 2001; Bembenek et al, 2007). Organelles were manually segmented using the Labkit plugin in Image J. Resulting segmented layers were exported as bitmap tifs and total area was determined with the Measure function in Image J. Numerical values corresponding to data points in figures are included in Supplemental Data 3.