A Role for Separase in the Regulation of RAB-11-positive Vesicles at the Cleavage Furrow and Midbody

Summary

Cell division requires coordinated regulation of chromosome segregation and cytokinesis. Although much is known about the function of the protease separase in promoting sister chromosome separation, the role of separase during cytokinesis is unclear. Using C. elegans, we show that separase localizes to the ingressing furrow and midbody during cytokinesis. Loss of separase function during the early mitotic divisions of the C. elegans embryo causes cytokinesis failure that is not due to eggshell defects or chromosome nondisjunction. Moreover, depletion of separase causes the accumulation of RAB-11-positive vesicles at the cleavage furrow and midbody that is not a result of chromosome nondisjunction, but is mimicked by depletion of vesicle fusion machinery. Collectively, these data indicate that separase is required for cytokinesis by regulating the incorporation of RAB-11-positive vesicles into the plasma membrane at the cleavage furrow and midbody.

Results and Discussion

Localization of separase to the cleavage furrow and midbody during cytokinesis

We previously found that separase localizes to vesicles and promotes cortical granule exocytosis after fertilization during the first meiotic anaphase. Degranulation is an essential prerequisite for functional eggshell formation in C. elegans. Furthermore, separase localizes to the base of the polar body as it is extruded [1]. These findings suggest that separase could directly regulate polar body cytokinesis by influencing membrane trafficking. To investigate whether separase has a general function during mitotic cytokineses in C. elegans, we analyzed the localization of GFP::SEP-1 in embryos during mitosis. The GFP::SEP-1 transgene can fully rescue separase mutation in C. elegans (Fig. 3H), and can therefore be used to report the localization of the endogenous separase. During metaphase, we find that GFP::SEP-1 localizes to chromosomes, centrosomes, and in a cloud surrounding the metaphase plate (Fig. 1A, SuppMov1). During anaphase, separase remains on chromosomes, centrosomes, and also spreads along the anaphase spindle (Fig. 1B). During cytokinesis, separase appears at the cleavage furrow (Fig. 1C) and midbody (Fig. 1D). We have confirmed the localization of endogenous separase by immunfluorescence to these same structures (Supplemental Fig. S1).

Figure 3.

Hypomorphic sep-1 mutant embryos fail cytokinesis. (A–C) DIC images of sep-1(e2406) mutant embryos failing cytokinesis. The nuclei fully separate (arrowheads, B) and the cleavage furrow completes, but later regresses (C). (D–F) sep-1(e2406) mutant embryos labeled with histone-GFP and FM2-10 dye were imaged. (D) Lagging chromosomes (arrowhead) are observed during anaphase when the furrow (arrow) ingresses. (E) The nuclei (arrowheads) fully separate when furrowing completes, and some membrane accumulation is observed at the midbody (arrow). (F) Subsequently, the furrow regresses and a large bolus of membrane material can be observed (arrow). (G) Quantitation of chromosome separation in sep-1(e2406) embryos shows less chromosomal bridging defects than RNAi depletion of separase. (H) Despite minimal chromosomal bridging, sep-1(e2406) embryos have a high rate of cytokinesis failure, which is rescued by the expression of GFP::SEP-1. Scale bar, 10µm.

Figure 1.

Separase localizes to the furrow and midbody during cytokinesis. (A) In metaphase, GFP::SEP-1 is observed on chromosomes (arrow), centrosomes (arrowhead) and on the spindle. (B) GFP::SEP-1 remains on chromosomes (arrow) and centrosomes (arrowhead), and spreads along the spindle during anaphase. (C) Image at a cortical plane shows a faint localization of GFP::SEP-1 at the furrow (arrow) during cytokinesis, and a faint signal is also observed at the midbody (D). Images in B–D were captured with elevated laser power and gain settings, as separase signal decreases during anaphase and would otherwise be difficult to illustrate. (E) GFP::SEP-1^PD localizes normally during metaphase. (F) During anaphase, GFP::SEP-1^PD localizes very prominently to two spots within the centrosome (arrowhead), likely centrioles, and the central spindle region (arrow) of the mitotic spindle. (G) During cytokinesis, GFP::SEP-1^PD is more prominently observed at the ingressing furrow and (H) midbody (arrows) as compared to GFP::SEP-1. Scale bar, 10µm.

To study the localization of separase further, we expressed protease-dead GFP-separase (GFP::SEP-1^PD) with a point mutation of the catalytic cysteine to serine (C1040S). Interestingly, the mutant protease exhibited a more intense localization at putative sites of separase action as compared to the GFP::SEP-1 during anaphase and cytokinesis. GFP::SEP-1^PD accumulated as two bright spots within the centrosome, which are likely centrioles. It also accumulated strongly in the central spindle region (Fig. 1E, SuppMov2). Separase has been implicated in licensing centrosome duplication in anaphase [2, 3] and also regulating spindle dynamics [4, 5]. The enhanced localization of GFP::SEP-1^PD to these sites may reflect persistent binding to substrates that are not cleaved. Remarkably, GFP::SEP-1^PD also accumulates more prominently on the ingressing furrow and midbody during cytokinesis than the wild-type protein (Fig. 1G, H), suggesting that separase may also act on substrates at these sites. Therefore, separase exhibits a dynamic localization pattern during mitosis, reminiscent of its localization during meiosis I [1]. The localization of separase to the furrow suggests that it may function during cytokinesis.

Separase directly regulates cytokinesis

Consistent with the idea that separase could regulate cytokinesis, previous reports have shown that inactivation of separase causes cytokinesis failure in C. elegans [6]. However, some studies have suggested that cytokinesis failures observed after inactivation of separase in C. elegans are indirect and can be rescued if the embryos are mounted in ideal conditions that do not apply osmotic or mechanical pressure during image analysis [6]. We sought to re-evaluate the role of separase in cytokinesis given our finding that separase localizes to the furrow and midbody. We analyzed a number of sep-1(RNAi) treated embryos by live-cell imaging with either differential interference contrast (DIC) or fluorescence microscopy to observe cell division. We also titrated the degree of RNAi depletion to match the phenotype presented in the previous study, which was at an intermediate level of separase depletion observed after 20–30hr of RNAi feeding. We observed embryos during mitosis using ideal buffer conditions and a hanging drop mount to reduce osmotic and mechanical pressures that might cause cytokinesis failures, as done previously by Siomos et al. These conditions allow for the culture of blastomeres completely stripped of the eggshell [7], and preserve viability of fragile meiotic embryos mounted prior to eggshell formation.

As expected, wild-type embryos divide normally under these conditions and do not display cytokinesis failures (n = 20, Fig. 2A). However, sep-1(RNAi) embryos exhibited a significant incidence of furrow regression resulting in cytokinesis failure, which occurred prior to the second mitotic division (n=6/22, Fig. 2B, E). Similar to WT embryos, we observed several known components properly localized to the midbody in sep-1(RNAi) embryos (Supplemental Fig. S2), confirming that these embryos form midbodies but then subsequently display furrow regression. Therefore, depletion of separase by RNAi that results in a similar phenotype as previously reported causes cytokinesis failures due to furrow regression under optimal mounting conditions.

Figure 2.

Depletion of separase causes cytokinesis failures. 4D multiphoton images of GFP-histone and -tubulin (green) were maximum z-projected and overlaid with brightfield images (grey). (A) WT embryos complete the mitotic divisions normally. (B) Depletion of separase causes chromosome bridges during mitotic anaphase and cytokinesis failures. (C) Embryos depleted of top-2 have extensive nondisjunction, leading to persistent chromosome bridges that do not cause furrow regression. (D) Quantitation of embryos showing normal chromosome segregation, lagging chromosomes or chromosomal bridging during anaphase. (E) Although sep-1(RNAi) and top-2(RNAi) share severe nondisjunction, 28% of sep-1(RNAi) embryos, but none of the the top-2(RNAi) embryos, display furrow regression. Scale bar, 10µm.

Separase RNAi also caused chromosome nondisjunction (Fig. 2D). Chromosome nondisjunction has been shown to cause cytokinesis failures in human cells [8]. Recent studies have shown that signaling pathways function in human cells and budding yeast to delay completion of cytokinesis when chromatin is trapped in the midbody [9, 10]. However, many mutations, including those affecting separase and securin, cause the “cut” phenotype in fission yeast where cytokinesis bisects the nucleus, suggesting that not all cell types fail cytokinesis with chromosome nondisjunction [11]. Nonetheless, it is possible that the furrow regression observed in the sep-1(RNAi) embryos could be an indirect effect of chromosome nondisjunction in C. elegans.

We sought to determine whether the cytokinesis failures we observed as described above are caused by chromosome nondisjunction. We analyzed embryos depleted of the topoisomerase II, top-2. As expected, top-2(RNAi) caused extensive chromosome nondisjunction (n=12/12), which persisted for multiple cell divisions in some cases (Fig. 2C, D). Despite the presence of extensive chromosome bridges, the cytokinetic furrow does not regress (n=0/12, Figure 2E). This shows that chromosome nondisjunction does not cause cytokinesis failures in the C. elegans embryo, consistent with previous studies [12]. As a control, we demonstrated that embryos depleted of both sep-1 and top-2 by RNAi display furrow regression (Supplemental Fig. 3). One caveat of these observations is that depletion of separase, unlike top-2(RNAi), is expected to cause the persistence of cohesin on chromatin, which could indirectly affect cytokinesis. However, the localization of separase to the furrow and midbody taken together with the absence of cytokinesis failure when severe chromosome bridging occurs suggests that separase activity is directly required for cytokinesis. Furthermore, if a checkpoint functions to delay abscission in the C. elegans embryo when chromatin is trapped in the midbody, it does not lead to the regression of the furrow within the time window we analyzed. Therefore, cytokinesis failures observed after sep-1(RNAi) due to furrow regression are likely a result of a direct function of separase during cytokinesis.

To further confirm that separase regulates cytokinesis directly, we analyzed embryos with a temperature-sensitive hypomorphic separase mutation, sep-1(e2406), under ideal mounting conditions. We previously demonstrated that this hypomorph is capable of promoting chromosome segregation, but not cortical granule exocytosis during anaphase of meiosis I [1]. We observed unlabeled embryos with DIC imaging at the restrictive temperature to observe mitosis (Fig 3A–C). Consistent with our previous observations, nuclei in sep-1(e2406) embryos separate normally as observed by DIC (n=12/17, Fig. 3B). We also observed sep-1(e2406) embryos expressing H2B::GFP at 25°C with fluorescent imaging and found that they rarely exhibit chromosome bridging during anaphase (n=2/20) but they do display lagging chromosomes (n=11/20, Fig. 3G). A significant fraction of the embryos have apparently normal chromosome segregation (n=6/20). Therefore, the sep-1(e2406) mutant does not display as severe chromosome nondisjunction as sep-1(RNAi) embryos.

Cytokinesis failures are observed in sep-1(e2406) embryos at a higher rate (n=13/26) than intermediate RNAi depletion, despite having less severe chromosome segregation defects (Fig. 3H). Importantly, chromosomal bridges were observed in all embryos (n=6/6) that display cytokinesis failure after sep-1(RNAi) treatment, whereas 38% (n=5/13) of the sep-1(e2406) embryos failing cytokinesis separated their nuclei normally. Therefore, the hypomorphic sep-1(e2406) mutation causes cytokinesis failures in the absence of chromosome nondisjunction. Together with the absence of cytokinesis failures after chromosomal bridges are induced by top-2(RNAi), we conclude that separase has a direct role in cytokinesis independent of chromosome segregation.

We also used fluorescent imaging of GFP-histone and membrane dye FM2-10-labeled embryos to more carefully observe chromosome segregation and plasma membrane dynamics during cytokinesis, as done previously [13] (Fig. 3D–F, SuppMov3). We observed an abnormal bolus of membrane material at the regressing furrow in sep-1(e2406) embryos, indicative of a defect in membrane dynamics during cytokinesis (Fig. 3F). Expression of SEP-1::GFP in sep-1(e2406) embryos rescued eggshell defects, chromosome nondisjunction, cytokinesis failures (Figure 3H) and viability. We also took advantage of the temperature-sensitivity of sep-1(e2406) allele to shift embryos to the non-permissive temperature after completion of the meiotic divisions and eggshell formation. We observed furrow regression causing cytokinesis failures in impermeable sep-1(e2406) embryos shifted just after eggshell formation (Supplemental Fig. 4). Together, the above studies indicate that inactivation of separase in C. elegans causes furrow regression leading to cytokinesis failures independently of meiotic division failures, eggshell permeability, or chromosome nondisjunction.

Separase regulates the RAB-11-positive vesicles during cytokinesis

To gain insight into the mechanism by which separase promotes cytokinesis, we investigated whether membrane trafficking during mitosis was affected given the abnormal accumulation of FM2-10 labeling at the furrows of sep-1(e2406) embryos during cytokinesis. We focused on studying RAB-11 because of its central role in regulating trafficking and exocytosis of vesicles to the plasma membrane during cytokinesis [14]. For this purpose, we analyzed embryos expressing GFP::RAB-11, previously reported to accurately reflect endogenous localization of RAB-11 [15]. In wild-type embryos, GFP::RAB-11 could be seen in transient patches on the ingressing furrow and weakly accumulated at the midbody during abscission (n = 10/10, Fig. 4A, E, F and SuppMov4). Similarly, embryos depleted of the top-2 topoisomerase (25–30hr feeding RNAi) had chromosome bridges, but showed no apparent defects in the trafficking of GFP::RAB-11 to the furrow and midbody (n = 8/8, Fig. 4B, E, F, and SuppMov5). By contrast, sep-1(RNAi) treatment (15–23hr feeding RNAi) caused prominent and persistent GFP::RAB-11 accumulation at the furrow and midbody during cytokinesis (n = 11/14, Fig. 4C, E, F, and SuppMov6). In contrast, we did not observe the accumulation of the early endosomal rab, RAB-5, at the ingressing furrow or midbody during cytokinesis in sep-1(RNAi) embryos (Supplemental Fig. 5). Depletion of the plasma membrane-localized t-SNARE (40–64hr feeding RNAi), syn-4, which is also required for cytokinesis in C. elegans [16], caused prominent, persistent accumulations of GFP::RAB-11 at the furrow and midbody (n = 6/9 Fig. 4D, E, F, and SuppMov7). These data show that regulation of RAB-11-positive vesicles at the ingressing furrow and midbody requires separase, in a similar way that a t-SNARE is required. Therefore, separase is required for cytokinesis and RAB-11 vesicle trafficking at the plasma membrane during cytokinesis in C. elegans.

Figure 4.

Depletion of separase affects RAB-11 vesicle trafficking. GFP::RAB-11 expressing embryos were imaged during cytokinesis to observe vesicle trafficking at the furrow and midbody (arrows). (A) During cytokinesis in WT embryos, GFP::RAB-11 puncta can be observed transiently on the ingressing furrow and at the midbody for a short time during early abscission, consistent with its role in regulating vesicle trafficking during cytokinesis. (B) Depletion of top-2 causes severe chromosome nondisjunction but does not affect the trafficking of GFP::RAB-11-vesicles to the furrow and midbody during cytokinesis as compared to wild-type embryos. (C) During furrow ingression in sep-1(RNAi) embryos, persistent GFP::RAB-11 localization is observed at the furrow and prominently accumuates at the midbody during abscission. (D) Depletion of the plasma membrane syntaxin, SYN-4, also causes GFP::RAB-11 to accumulate on the furrow and midbody. (E) The duration of time that GFP::RAB-11 was observed on the furrow was quantified in embryos showing persistent accumulations during furrow ingression for each condition. Whereas GFP::RAB-11 was observed transiently in WT (37 +/− 6s, n = 5) and top-2(RNAi) (26 +/− 5s, n = 4) embryos, it was observed for much longer time in sep-1(RNAi) (140 +/− 22s, n = 5, p<0.01) and syn-4(RNAi) (166 +/− 34s, n = 4, p<0.01) embryos. (F) Quantification of the RAB-11::GFP signal in embryos that displayed accumulations of RAB-11 during cytokinesis. To control for variations in signal intensity due to z-depth, we calculated the ratio of the intensity of the RAB-11 signal observed at the furrow (blue) and midbody (red) in embryos relative to the average intensity of three equal areas of cytoplasm. The ratio of RAB-11 signal in WT embryos at the furrow (1.4 +/− 0.05, n=8) or midbody (1.5 +/− .04, n=8) was similar to the ratios observed for top-2(RNAi) embryos (1.5 +/− ). Error bars indicated standard error of the mean. (G) Proposed model for trafficking of RAB-11-vesicles (purple) during cytokinesis (centrosome and spindle depicted in green). We suggest that at least three steps occur in RAB-11-vesicle trafficking after leaving the recycling endosome: (1) vesicle delivery, (2) tethering and (3) exocytosis at the plasma membrane. Separase and syntaxin localize to the furrow and cause accumulations of RAB-11-vesicles at the furrow and midbody when depleted, suggesting that vesicles are properly delivered and tethered. We hypothesize that RAB-11- vesicles fail to undergo exocytosis, but their endocytosis (dashed arrow) might be impaired. Scale bar, 10µm.

Our observations suggest that separase has a direct role in cytokinesis and regulates the trafficking of RAB-11-positive vesicles during cytokinesis. RAB-11-positive vesicles are known to traffic to the cleavage furrow from a centrosomal pool of recycling endosomes during cytokinesis where they fuse with the plasma membrane to provide the additional membrane necessary for cytokinesis (Fig 4F) [17]. The observation of RAB-11-positive vesicle accumulation at the plasma membrane during cytokinesis is a novel phenotype with respect to RAB-11 trafficking. Other regulators of RAB-11 trafficking have been shown to be important in proper delivery or tethering of vesicles to the cleavage furrow, causing a reduction of vesicles at the plasma membrane when inactivated [18–20]. Our observations that separase localizes to the furrow and midbody and that loss of separase causes an accumulation of RAB-11-positive vesicles at the cleavage furrow and midbody, similar to syntaxin depletion, shows that separase is required for a step in RAB-11-positive vesicle trafficking after delivery to the plasma membrane. Given that separase promotes cortical granule exocytosis during meiosis I in the C. elegans embryo [1], we favor the possibility that separase promotes exocytosis of RAB-11 vesicles during cytokinesis. However, our data cannot rule out the possibility that separase regulates the endocytic recycling of RAB-11 from the plasma membrane after vesicle fusion during cytokinesis. The function of separase in regulation of RAB-11 trafficking may partly explain the furrow regression and failure of cytokinesis we observed in separase RNAi-depleted and mutant embryos. Recently, RAB-11 was shown to localize to cortical granules and is required for their exocytosis during anaphase of meiosis I [21]. Therefore, these studies indicate that separase and RAB-11 function together in a pathway that coordinates chromosome segregation with membrane trafficking during both meiosis and mitosis in C. elegans.

Our findings suggest that separase directly regulates both chromosome separation and cytokinesis during cell division. We show that during mitotic cell divisions, separase first localizes to chromosomes, spreads along the anaphase spindle and later appears on the cleavage furrow during cytokinesis. This dynamic localization of separase may allow it to coordinate chromosome segregation with membrane trafficking events essential for the success of cytokinesis. The protease activity of separase is also involved in licensing centriole duplication [3] and regulating the spindle during anaphase [4]. Since protease-dead separase accumulates at the centriole, spindle midzone, cleavage furrow, and midbody more than the wild-type protein (Fig. 1D–E), we speculate that separase may proteolyze substrates at each of these cellular locations to coordinate chromosome separation, anaphase spindle dynamics, centrosome duplication, and membrane trafficking during cell division.

Experimental Procedures

Strains

We maintained C. elegans according to standard protocols [22]. Temperature-sensitive strains were maintained at 16°C and shifted to 25°C for 2–6 hours or as indicated, all other strains were maintained at 20–C. WH0520 was propagated on GFP(RNAi) to eliminate transgene expression and restore viability. Animals 5 generations off GFP(RNAi) expressed the transgene and were analyzed for localization. The following strains were used in this study: N2, WH0438 {unc-119(ed3) III, ojEx64[GFP::SEP-1 unc119(+)]}, WH0520 {unc-119(ed3) III, ojIs71[GFP::SEP-1^PD unc119(+)]}, TY3558 {unc-119(ed3) III, ruIs32[pie-1::GFP::his-11] III; ojIs1[tbb-2::GFP]}, WH0408 {sep-1(e2406) I /hT2[bli-4(e937) let-?(q782) qls48]}, WH0468 {sep-1(e2406) I ; unc-119(ed3) III, ruIs32 III[GFP::His-58 unc119(+)]/hT2[bli-4(e937) let-?(q782) qls48]}, WH0485 {sep-1(e2406) I ; unc-119(ed3) III, ojIs58 [GFP::SEP-1 unc119(+)]/hT2[bli-4(e937) let-?(q782) qls48]}, WH0347 {unc-119(ed3) III; ojIs35 [GFP::rab-11.1 unc-119(+)]}, MG170 {zen-4(or153ts) IV; xsEx6[ZEN-4::GFP]}, WH0370 {unc-119(ed3) III, ojIs50[GFP::AIR-2 unc119(+)]}, WH0433 {unc-119(ed3) III, ojIs26[GFP::NMY-2 unc119(+)]}, and RT1043 {unc-119(ed3) III, pwIs403[GFP::RAB-5 unc119(+)]}.

Molecular Biology

Separase feeding RNAi was performed using a previously described construct [6], and all other constructs were obtained from the Ahringer library [23]. L4 animals were fed at 20°C for the indicated time. The protease-dead separase mutant was generated using Quickchange mutagenesis (Stratagene, La Jolla, CA) using the following primers: CAGTTTGCTGATGGGTTCTGGAAGGTTAATATA ATTTATTTTG, and CAAAATAAATTATATTAACCTTCCAGAACCCATCAGCAAA CTG. Mutagenesis was performed on the wild-type sep-1 genomic sequence in the pJK3 plasmid that allows for pie-1-driven expression of N-terminal mGFP fusions. Biolistic transformation was done as previously described [24].

Immunohistochemistry and Staining

Staining methods that preserve membrane structures were performed as previously described [15]. The following dilutions were used: rabbit α-SEP-1 antibody [1], 1:200; mouse tubulin antibody DM1α (Sigma, St. Louis, MO USA), 1:100; alexa-conjugated secondary antibodies (Invitrogen, Carlsbad, CA USA), 1:200; TO-PRO-3 Iodide (Invitrogen, Carlsbad, CA USA), 1:500 and DAPI, 1:1,000.

Microscopy

Mounting Embryos

Embryos were mounted as previously described [1]. Live cell imaging was performed with a multiphoton system at LOCI (www.loci.wisc.edu) as previously described [1] or with a Yokogawa CSU-10 confocal (Newnan, GA USA). Fixed samples were imaged using a Bio-Rad MRC 1024 confocal microscope (Hercules, CA USA) or a Leica SP5 system (Wetzlar, Germany). 4D movies were analyzed in ImageJ v1.37p [25] using the Bio-Formats plugin from LOCI (www.loci.wisc.edu). Images were processed in Photoshop v8.0 (Adobe Systems) using Gaussian blur to reduce noise and levels to best display structures of interest. To quantitate the time that GFP::RAB-11 persisted on the furrow, movies captured at 5 second time intervals for the entire duration of cytokinesis were compared. Each frame showing RAB-11 localization at the furrow was counted as 5 seconds of persistent localization. We used metamorph to quantify GFP::RAB-11 at the furrow and midbody during cytokinesis. To do this, we determined the ratio between the total signal observed in a region at the furrow and midbody and that of an average of three equal regions of the cytoplasm after subtracting out the background signal.