Syntaxin Is Required for Cell Division

Abstract

We recently identified a single family member homologue of syntaxin in the sea urchin. Syntaxin is present throughout development, and in rapidly dividing cleavage stage embryos it is present on numerous vesicles at the cell cortex. We hypothesized that syntaxin mediates essential membrane fusion events during early embryogenesis, reasoning that the vesicles and/or their contents are important for development. Here we show that functional inactivation of syntaxin with either Botulinum neurotoxin C1, which specifically proteolyzes syntaxin, or antibodies against syntaxin results in an inhibition of cell division. These observations suggest that syntaxin is essential for membrane fusion events critical for cell division.

INTRODUCTION

Cell division is a highly coordinated event requiring a variety of membrane fusion and fragmentation events. During mitosis in higher eukaryotes, for example, the nuclear envelope breaks down into nuclear membrane vesicles after chromosome condensation, and large cytoplasmic organelles such as the Golgi and endoplasmic reticulum (ER) are also believed to fragment (Lucocq and Warren, 1987; Warren, 1989). These fragmented organelle membranes then distribute equally into daughter cells and must refuse with each other to reconstitute their respective organelles. In addition to the breakdown and reformation of the nuclear envelope, Golgi, and ER during the cell cycle, the cell also increases its membrane surface area during cell division (for review, see Rappaport, 1996).

What proteins mediate these essential membrane fusion events during cell division? A highly conserved set of membrane proteins have been identified that are involved in many types of intracellular fusion (Rothman, 1994; Sudhof, 1995). These proteins localize to both vesicle and target membranes, known as v- and t-soluble NSF attachment protein (SNAP) receptors (SNAREs), respectively, and appear to function throughout the secretory pathway as the minimal machinery driving membrane fusion (Fasshauer et al., 1998; Weber et al., 1998). Recently, single family member homologues of syntaxin (t-SNARE), vesicle-associated membrane protein (VAMP; v-SNARE), and the monomeric GTP-binding protein rab3 were identified in the sea urchin egg in association with cortical granules, secretory vesicles whose contents give rise to the fertilization envelope (Conner et al., 1997). Syntaxin, VAMP, and rab3 are also present throughout embryogenesis enriched in cells with elevated levels of regulated secretion (Conner and Wessel, manuscript in preparation). During the cleavage stage of this embryo, a period of cell division every 45–60 min, we find enrichment of these molecules on vesicles accumulating at the cortex of cells, suggesting that these vesicles may play an important role in cell division. Thus, we hypothesized that these proteins not only mediate the complex array of membrane fusion events of secretion, as previously documented (for review, see Ferro-Novick and Jahn, 1994; Bock and Scheller, 1997; Rothman and Sollner, 1997), but also function in the contribution of new membrane to the cell surface during division. Using the sea urchin embryo, which has a single detectable syntaxin homologue in early embryos, we test this hypothesis by inactivating syntaxin with the microinjection of Botulinum neurotoxin C1, which specifically proteolyzes syntaxin family members (Blasi et al., 1994; Schiavo et al., 1995; Walch-Solimena et al., 1995), and affinity-purified antibodies against syntaxin. We find that disruption of syntaxin inhibits cell division, whereas cells injected with toxin or antibodies that have been heat inactivated develop as normal. Thus, we conclude that functional syntaxin is required to mediate membrane fusion events during cell division. This further suggests that the molecular models that describe protein-mediated membrane fusion events for regulated exocytosis are applicable to membrane fusion events required for basic processes of cell division.

MATERIALS AND METHODS

Animals

Adult Lytechinus variegatus were obtained from Scott Services (Miami, FL) and Mele Enterprises (Duke University Marine Lab, Beaufort, NC). Gametes were obtained as described (McClay, 1986).

Antibody Purification

To affinity purify Fab fragment antibodies against syntaxin, a syntaxin-GST fusion protein was made using a nucleotide sequence representing amino acids 1–265 (MRDL … KKFY) of the syntaxin cDNA clone (Conner et al., 1997) ligated into a pGEX-3A vector for fusion with GST and transformed into BL21(DE3) cells for overexpression. Syntaxin-GST fusion protein-expressing BL21(DE3) cells were induced at 23°C with 0.1 mM isopropyl-β-d-thiogalactopyranoside for 3 h. Cells were then pelleted by centrifugation at 4000 rpm for 10 min, resuspended in PBS, lysed with high pressure using a French press, and solubilized with 1% Triton X-100 for 30 min. Cellular debris was then pelleted at 10,000 × g at 4°C for 20 min. The resulting supernatant was passed over a glutathione-agarose column (Sigma, St. Louis, MO), and the column was then washed with 10 column volumes of PBS. Syntaxin-GST fusion protein was specifically eluted with PBS containing 10 mM reduced glutathione (Sigma), and the purity of column elutant syntaxin-GST protein was verified by SDS-PAGE and immunoblot analysis. Affinity-purified syntaxin-GST protein was blotted to nitrocellulose in PBS and then blocked with preimmune sera for 10 min. The blot was then washed with PBS and incubated for 30 mins with syntaxin Fab fragment antiserum obtained using the Immunopure Fab preparation kit (Pierce, Rockford, IL), which previously had been conjugated to Oregon Green 488 using the FluoReporter labeling kit (Molecular Probes, Eugene, OR). The blots were then washed again with PBS, and the Oregon Green-labeled Fab fragment antibodies were eluted from the nitrocellulose with 100 mM glycine, pH 2.5, dialyzed extensively against PBS, and concentrated to 2 mg/ml using Ultrafree-4 centrifugal filters with a 10-kDa cutoff (Millipore, Bedford, MA). Protein concentration was determined using the Bradford method using BSA as a standard. Affinity-purified Fab fragment antibodies labeled with Oregon Green were tested by immunolocalization in thick sections of eggs (see below).

Injections

Eggs were fertilized and placed into a Kiehart chamber (Kiehart, 1982) in artificial seawater (ASW) (McClay, 1986). Fertilized eggs or a single blastomere of a two-cell-stage embryo was microinjected with various reagents. Botulinum neurotoxins A, C1, and E (BoNT-A, -C1, and -E; Wako Bioproducts, Richmond, VA) stock injection solutions were 1 mg/ml toxin in 200 mM NaCl and 50 mM sodium acetate, pH 6.0. BoNT-C1 was heat inactivated by incubation of the stock injection solution at 100°C for 10 min. BoNT-E was activated with 200 μg/ml trypsin at 37°C for 30 min. Trypsin was removed, selectively, by incubation with soybean trypsin inhibitor conjugated to agarose beads (Sigma) for 30 min at room temperature, trypsin-bound beads were then removed by centrifugation, and the supernatant was used subsequent to microinjection. Proteinase K (Sigma) stock injection solution was 5 mg/ml in deionized water. Affinity-purified fluorochrome-labeled Fab fragments against syntaxin and rab3 were resuspended in deionized water to ∼1.2 and ∼1.3 mg/ml, respectively. Affinity-purified antibodies were heat inactivated by incubation for 10 min at 100°C. Nonrelevant fluorochrome-labeled Fab fragment antibodies raised against rabbit immunoglobulin G (IgG; Sigma) were resuspended in deionized water to 1.5 mg/ml for injection. An oil droplet of dimethypolysiloxane (Sigma) was coinjected into cells as a marker. Injection volumes never exceeded 5% of the cell volume.

Immunolocalization Assays In Situ

Immunofluorescence localization was performed in whole mounts and on embryo sections that were fixed and processed as previously described (Laidlaw and Wessel, 1994). The polyclonal antibodies against the syntaxin were diluted 1:500 (∼1 μg/ml) and 1:200 (∼5 μg/ml) (Conner et al., 1997). The secondary antibodies (FITC-conjugated goat anti-mouse IgG [Cappel, West Chester, PA] or lissamine–rhodamine-conjugated affinity-purified Fab fragment goat anti-rabbit IgG [Jackson ImmunoResearch, West Grove, PA]) were diluted 1:100 (∼1 μg/ml). Signals were recorded by epifluorescence with a Zeiss (Thornwood, NY) Axioplan or a Zeiss LSM 410 laser scanning microscope.

In vivo Immunolocalization and Quantitation

To immunolocalize syntaxin and rab3 in vivo, affinity-purified fluorochrome-labeled Fab fragment antibodies were injected into fertilized eggs or single blastomeres. To quantitate immunolocalization or FM1-43 endocytosis, 15–20 confocal sections of each embryo were acquired with the Zeiss LSM 410 laser scanning microscope and analyzed using Adobe Photoshop (Adobe Systems, Mountain View, CA). For each confocal section, the average immunolabel or FM1-43 brightness and area were determined by using the histogram. Then using the color range option, highlighted pixels (bright saturated) were selected, and their number was determined by the histogram. The number of saturated pixels was subsequently divided by the area and the average brightness to obtain a standardized value. Values from 15–20 confocal sections were then averaged to obtain the immunolocalization or FM1-43 endocytosis value for each cell.

For rab3/syntaxin colocalization experiments, affinity-purified Fab-fragment antibodies against syntaxin were labeled with Oregon Green 488 (see Antibody Purification; Molecular Probes) and affinity-purified Fab fragment antibodies against Rab3 isolated as described (Conner and Wessel, 1998) labeled with Texas Red using the FluoReporter labeling kit (Molecular Probes) were mixed 1:1. This antibody mixture was then injected into a fertilized egg ∼30 mins after insemination. Immunolocalization was visualized at the appropriate channels for each fluorochrome with confocal microscopy using a Zeiss LSM 410 microscope.

Membrane Topology and Endocytosis

3,3′-dihexyloxacarbocyanine iodide [DiOC₆(3)] (Molecular Probes) was resuspended in methanol at 1 mg/ml and then transferred to Hollywood safflower oil (Big Daddy Wesley’s, Beaufort, NC) by mixing 500 μl of the methanol/DiOC₆(3) solution with the 500 μl of safflower oil. DiOC₆(3) resuspended in safflower oil was then used for microinjection into cells for membrane labeling. The volume of oil containing DiOC₆(3) did not exceed 5% of the cell volume. FM1-43 (Molecular Probes) was resuspended in methanol at 1 mg/ml. It was then diluted in ASW to give a working concentration of 1 μM. To evaluate endocytosis, experimentally manipulated embryos were transferred to the FM1-43 in ASW and visualized after 15–45 min incubation at room temperature using confocal microscopy using a Zeiss LSM 410 microscope.

Brefeldin A Treatment

Eggs were fertilized in ASW and after 10 min were transferred to ASW containing brefeldin A (BFA; Calbiochem, La Jolla, CA) at the indicated concentrations (stock solution was 4 mg/ml in methanol) or ASW containing identical concentrations of methanol as that of the BFA-treated embryos as a control. The methanol in the ASW of the experimental and control samples was given 30 min at room temperature to evaporate before embryo transfer.

RESULTS

Syntaxin Is Present in the Dividing Sea Urchin Embryo

Because only a single sea urchin syntaxin family member is detectable throughout sea urchin embryogenesis, we asked whether syntaxin localizes in vivo to a distinct intracellular compartment of the secretory pathway like other syntaxin family members in a variety of other systems (Bennett et al., 1993; Dascher et al., 1994; Bock et al., 1997). By microinjection of detection levels (∼200 nM, noninhibitory) of fluorochrome-labeled affinity-purified antibodies against sea urchin syntaxin, we find syntaxin on intracellular vesicles enriched at the cell cortex of the newly fertilized egg and cleavage stage embryo (Figure 1, A–F). When elevated levels of antibody are injected (∼2 μM), we also find immunolabeling on the ER (Figure 1H). Immunolocalization of syntaxin in fixed cells yields similar results, although we observe greater immunolabel at the cell cortex than that associated with the ER (Figure 1, I–K). These syntaxin distribution patterns led us to hypothesize that the syntaxin-positive vesicles might be involved in membrane fusion events during cell division.

Figure 1.

In vivo immunolocalization of syntaxin by injecting detection levels of fluorochrome-labeled polyclonal antibodies (∼200 nM) reveals syntaxin association with vesicles enriched at the cell cortex of the fertilized egg (A and B) and cells of the dividing embryo (C–F). By increasing antibody injection 10-fold (∼2 μM), syntaxin is found in association with ER, as seen in the four-cell-stage embryo (G and H). Immunolocalization in fixed sections confirms the pattern observed in vivo with syntaxin on vesicles at the cortex and in association with ER (I–K); however the syntaxin epitope at the cell cortex appears more accessible in fixed tissue, suggesting in vivo masking of the syntaxin epitope. Images visualized by indirect immunofluorescence using confocal microscopy. Oil droplets mark embryos injected with fluorochrome-labeled antibody. Bar, 50 μm.

Botulinum Neurotoxin C1 Blocks Cell Division in a Concentration-dependent Manner

To test the function of syntaxin during cell division, we microinjected BoNT-C1 into single cells to specifically inactivate syntaxin by releasing the functional protein binding domains. cDNA sequence analysis and in vitro cleavage results indicate that the single sea urchin syntaxin family member contains the neurotoxin protease cleavage site (Schiavo et al., 1995; Conner et al., 1997; Coorssen et al., 1997). We find that both cytokinesis and karyokinesis are blocked in 22% of cells injected with 1.6 nM BoNT-C1 within one cell cycle after injection, whereas cell division is inhibited in 100% of cells injected with ≥5 nM BoNT-C1 (Figure 2, A–C; maximal BoNT-C1 activity is supported at 37°C; however, these embryos were incubated at 23°C to retain maximal viability). Cells injected with 0.5 nM BoNT-C1 or 53 nM heat-inactivated BoNT-C1 had no affect on cell division or embryonic development (Figure 2, D–F, and Table 1). The inhibitory BoNT-C1 concentrations seen here are consistent with those found to inhibit catecholamine release in chromaffin cells (100 nM; Foran et al., 1996) and synaptosome neurotransmitter release (150 nM; Blasi et al., 1993). However, because the toxin is internalized into vesicles in these cells, it is hard to accurately determine the intracellular toxin concentration.

Figure 2.

BoNT-C1, which specifically cleaves syntaxin, inhibits cell division in the sea urchin embryo. Injection of active BoNT-C1 blocks cells from dividing (5.3 nM; A–C), whereas uninjected cells or cells injected with either heat-inactivated BoNT-C1 (53 nM; G–I) or proteinase K (493 nM; G–I) develop normally. Cells injected with either BoNT-A (45 nM; J–L) or BoNT-E (58 nM; M–O), which specifically proteolyze SNAP-25, also divide as normal. Injected cells are marked with an oil droplet. Bar, 50 μm.

Table 1.

BoNT-C1 inhibits cell division in a concentration dependent fashion

Toxin	Concentration (nM)	n	% cleavage
			Normal	Blocked or delayed
BoNT-A	5–45	7	100	0
BoNT-E	27–58	8	100	0
Inactivated BoNT-C1	5.3	5	100	0
	53	7	100	0
Proteinase K	493	5	100	0
BoNT-C1	0.5	3	100	0
	1.6	9	78	22
	3.9	6	33	67
	5.3	21	0	100

Blastomeres injected with a general proteinase, proteinase K, or heat-inactivated BoNT-C1 have no affects on cell division. However, as concentrations of BoNT-C1 are increased to ≥1.6 nM, we begin to see an inhibition in cell division, whereas uninjected cells divide as normal. Single blastomeres were injected at the two-cell stage, and cell division was assayed before the 16-cell stage. 

To test the possibility that the block in cell division was the result of nonspecific proteolysis by the toxin, we injected a general protease, proteinase K, and found that cell division is unaffected in cells injected with up to 493 nM proteinase K (Table 1 and Figure 2, G–I). In addition to syntaxin, BoNT-C1 has also been shown to proteolyze SNAP-25 (Foran et al., 1996; Williamson et al., 1996); thus to test the possibility that the BoNT-C1-induced phenotypes seen here result from the proteolysis of both syntaxin and SNAP-25, we injected either BoNT-A or -E, proteases that specifically target SNAP-25, into single cells of two-cell-stage embryos. We find that injection of either BoNT-A or -E has no affect on cell division at concentrations greater than that required by BoNT-C1 to block cell division (Table 1 and Figure 2). These observations suggest that the BoNT-C1-induced block in cell division is the result of the syntaxin-targeted toxin activity.

BoNT-C1 blocks synaptic vesicle fusion in the neuron by cleaving syntaxin, resulting in an accumulation of synaptic vesicles at the active zone of the synapse (Marsal et al., 1997; O’Connor et al., 1997). However, we suspected that in the rapidly dividing sea urchin embryo with a single syntaxin homologue that injection of BoNT-C1 could result in major changes in membrane topology of the cell that would lead to the block in cell division. To test this possibility, we injected the lipophilic dye DiOC₆(3), which labels any contacting membrane (Terasaki, 1998), into fertilized eggs, allowed them to divide, and then injected BoNT-C1 into a single cell to ask whether gross morphological changes in cytoplasmic membrane could be detected. We find that although cells injected with BoNT-C1 are inhibited in cell division, there is no detectable difference in DiOC₆(3) membrane labeling patterns compared with toxin-free cells (Figure 3, F–H).

Figure 3.

Cell phenotypes resulting from BoNT-C1 treatment are not the result of gross membrane topological alterations or membrane flow blockage through the Golgi apparatus. Fertilized eggs, injected with the membrane marker DiOC₆(3) to reveal membrane topology, were allowed to develop and a single blastomere was then injected with BoNT-C1 (A and F). Cells injected with BoNT-C1 (5 nM) become inhibited in cell division (C and H); however, there is little difference in the topology of membranes between toxin-injected and uninjected cells (F–H). The contribution of Golgi-derived material for cell division was tested by treating fertilized eggs with BFA. Embryos treated with 10 μM BFA 10 min after insemination are unaffected in cell division (D and E) compared with control embryos (I and J). Bar, 50 μm.

We also suspected that the observed effects of BoNT-C1 on cell division might be the result of impeding general membrane flow through the Golgi apparatus leading to a depletion of membrane-targeted vesicles. To test this possibility, we treated fertilized eggs with 10 μM BFA, a concentration well known for its ability to disassemble the Golgi apparatus by preventing anterograde vesicle transport from the ER but not the retrograde pathway (Lippincott-Schwartz et al., 1989; Klausner et al., 1992; Sciaky et al., 1997) in a variety of tissue culture cells (Sciaky et al., 1997; Kok et al., 1998; Zhang et al., 1998) and cultured sea urchin embryonic cells (Hwang and Lennarz, 1993). Surprisingly, treatment of newly fertilized eggs with 10–100 μM BFA has no observable effect on the timing or ability of the embryo to undergo cell division (Figure 3, D and E) compared with control embryos (Figure 3, I and J). The efficacy of BFA on blocking vesicle transport through the Golgi apparatus was tested by incubating unhatched sea urchin embryos in BFA to ask what concentration prevents secretion of the hatching enzyme. Sea urchin embryos are surrounded by a fertilization envelope during early development until the blastula stage, at which time they begin translating and secreting the hatching enzyme, which digests the envelope and allows the ciliated embryo to freely swim (Lepage and Gache, 1989; Lepage et al., 1992). We find that as low as 10 μM BFA prevents embryos from hatching out of the fertilization envelope (our unpublished results). The above results suggest that BoNT-C1 inhibits cell division by a specific syntaxin-mediated vesicle fusion effect and not simply the result of obstructing membrane flow through the Golgi apparatus.

Additionally, some syntaxin family members have been shown to be involved in retrograde membrane traffic from the Golgi to the ER in yeast (Lewis and Pelham, 1996) and are also thought to participate in synaptic vesicle recycling (Walch-Solimena et al., 1995). Thus, we asked whether BoNT-C1 was in some way inhibiting the cell endocytic pathway, which might indirectly block cell division. To test this hypothesis, we asked whether neurotoxin-injected cells were still capable of endocytosing FM1-43, a membrane-impermient lipophilic dye that fluoresces only when associated with membranes and has been shown useful in studying membrane dynamics in this embryo (Whalley et al., 1995). Single blastomeres of a two-cell-stage embryo were injected with BoNT-C1 and allowed to develop until a phenotypic difference in cell division was observed between injected and uninjected blastomeres, within 45 min to 1 h. We then transferred the embryos to ASW water containing FM1-43 to assay for endocytosis by looking for FM1-43-labeled endocytic vesicles. We find that toxin-injected cells are active in endocytosis, as evidenced by the accumulation of fluorescent vesicles in the cell cytoplasm (Figure 4, D–F), and no significant differences in endocytosis were apparent when compared with uninjected cells (Figure 4G). Moreover, we subsequently find FM1-43 fluorescent labeling in ER surrounding the cell nucleus (Figure 4, D and E), presumably by retrograde membrane traffic through the endosome and Golgi. These observations strongly suggest that the endocytic pathway is generally unaffected by treatment with BoNT-C1 and that cells are still capable of other membrane fusion events. Thus, we conclude that the toxin treatment is not simply affecting general metabolic processes or global membrane trafficking in the cell.

Figure 4.

BoNT-C1 does not disrupt the endocytic pathway. Single blastomeres, injected with BoNT-C1 (5 nM), were allowed to divide, and when cell division was inhibited (A–C), embryos were then exposed to the membrane marker FM1-43, which fluoresces when associated with lipids and only enters the cell through the endocytic pathway. FM1-43 is found in both toxin-injected and uninjected cells associated with vesicles throughout the cytoplasm (D–F) and eventually reaches the ER (arrowheads) after an ∼20 min incubation (D and E). Bar, 50 μm.

Finally, we also asked whether BoNT-C1 was simply blocking cell division indirectly by somehow preventing actin polymerization, thus preventing the formation of the contractile actin ring, which is required for cell division. However, phalloidin staining of toxin-injected embryos indicates this is not the case (our unpublished results).

Botulinum Neurotoxin C1 Specifically Removes Syntaxin from Intracellular Vesicles

BoNT-C1 cleaves syntaxin family members at an amino acid sequence specific site near the transmembrane domain at the C terminus (Blasi et al., 1993; Schiavo et al., 1995). Because the sea urchin syntaxin contains the conserved BoNT-C1 cleavage site, and BoNT-C1 cleaves sea urchin syntaxin in vitro (Coorssen et al., 1997), we hypothesized that antibodies to the N-terminal region of syntaxin should no longer localize to vesicles in toxin-injected cells in vivo (Conner et al., 1997), because BoNT-C1 cleavage would release the syntaxin N-terminal region from the vesicle membrane. To test this hypothesis we injected a single blastomere of a two-cell embryo with BoNT-C1 and waited (45 min to 1 h) for a phenotypic difference in cell division between toxin-injected and uninjected cells. We then asked whether syntaxin localized to vesicles at the cell cortex by injecting fluorochrome-labeled antibodies against syntaxin (∼200 nM). We find that in toxin-free cells, syntaxin localizes to vesicles enriched at the cell cortex, whereas in toxin-treated cells, vesicle-associated syntaxin signals are dramatically decreased (Figure 5). To quantify the effects of BoNT-C1 on syntaxin vesicle immunolocalization at the cortex, fluorescence measurements of at least 15 confocal sections for each embryo examined (embryos showing cell division delays; n = 3) indicate that as the concentration of BoNT-C1 is increased, syntaxin immunolocalization decreases compared with toxin-free blastomeres of the same embryo (Figure 6). Injection of 3.9 and 5.3 nM BoNT-C1 results in an ∼30 and 70% decrease in syntaxin immunolocalization, respectively, suggesting that proper cell division requires at least ∼70% intact syntaxin.

Figure 5.

BoNT-C1 removes syntaxin from intracellular vesicles. A single blastomere of a two-cell-stage embryo was injected with BoNT-C1 (5.3 nM). Once cell division was inhibited in the toxin-injected blastomere, daughter blastomeres originating from the uninjected cell and the toxin-injected blastomere were then injected with fluorescently labeled antibodies against syntaxin (E and F; injected blastomeres are marked with an oil droplet) to test syntaxin immunolocalization. Confocal sectioning of the embryo reveals a dramatic decrease in vesicle-associated syntaxin in the toxin-injected cell compared with control blastomeres (B–D). However, some vesicle-associated syntaxin can be seen in toxin-injected cells. Images were visualized by indirect immunofluorescence using confocal microscopy with a Zeiss LSM 410 microscope. Bar, 60 μm.

Figure 6.

BoNT-C1 dramatically reduces syntaxin vesicle localization. Cells treated with 3.9 and 5.3 nM BoNT-C1 have an ∼30 and 70% decrease in syntaxin immunolocalization, respectively, when compared with untreated cells of the same embryo, whereas rab3 immunolocalization is unaffected by treatment with the toxin. Each bar represents the average relative immunolocalization of 15 confocal sections of three different embryos.

To assess the specificity of BoNT-C1 on syntaxin removal from vesicles, we tested its effects on another vesicle constituent, rab3. Rab3 is associated with vesicles enriched at the cortex of cleavage stage embryos (Conner and Wessel, manuscript in preparation), and to test its colocalization with syntaxin, early sea urchin embryos were injected with fluorochrome-labeled antibodies against both syntaxin and rab3. We find that rab3 associates with the same vesicles as syntaxin (Figure 7, A–C). However, when BoNT-C1-injected and uninjected blastomeres are compared, we find no significant difference in the immunolocalization of rab3 (Figure 6), arguing strongly that vesicle-associated syntaxin removal by BoNT-C1 is specific and that syntaxin removal does not stimulate rab3 loss from these same vesicles.

Figure 7.

Syntaxin and rab3 colocalize on intracellular vesicles in the sea urchin embryo. Fluorochrome-labeled Fab fragments affinity purified to syntaxin and rab3 were injected into fertilized eggs. A cell surface confocal section shows syntaxin (A) and rab3 (B) associated with vesicles at the cortex of the embryo. Image overlay reveals syntaxin (red) and rab3 (green) on the same vesicles (C, colocalization in yellow; boxed inset shows higher magnification), with corresponding bright-field image (D). Bar, 60 μm.

Syntaxin Antibodies Inhibit Cell Division

As an alternative approach to test the function of syntaxin in cell division, we injected affinity-purified antibodies against recombinant sea urchin syntaxin into single blastomeres of a two-cell-stage embryo. Injection of monovalent Fab antibody fragments against syntaxin at 480 nM blocks cell division in a similar manner as that of BoNT-C1 (Figure 8, D–F). Antibody injection, like BoNT-C1 treatment, inhibits both karyokinesis and cytokinesis, and once injected cells have been inhibited, their development is halted, whereas uninjected blastomeres develop as normal. Injection of heat-inactivated affinity-purified Fab fragments has no affect on cell division (Figure 8, G–I), nor are any affects observed when single blastomeres are injected with nonrelevant Fab fragment antibodies (anti-rabbit IgG molecules at 580 nM; Figure 8, A–C). These observations argue that it is the specific inactivation of the syntaxin by antibodies that results in inhibited cell division, adding further evidence that functional syntaxin is required for cell division.

Figure 8.

Syntaxin antibodies inhibit cell division. Cell division is unaffected in blastomeres injected with nonrelevant Fab fragment antibodies at 580 nM (A–C, marked by an oil droplet; arrowhead). Blastomeres are inhibited in cell division when injected with 480 nM Fab fragment antibodies against syntaxin (D–F); however, cells injected with up to 1.9 μM heat-inactivated syntaxin antibodies show no affect on cell division (G–I). Bar, 50 μm.

DISCUSSION

Does the syntaxin family of proteins have a general function during cell division? An essential role for syntaxin during embryogenesis has been recently implicated in Drosophila; female germ line mosaic mutants for syntaxin 1 do not cellularize after the syncytial blastoderm stage, a time when massive increases in membrane surface area are required, and syntaxin 1 nulls appear to be cell lethal (Burgess et al., 1997). This effect is likely the result of a failure to mediate the fusion of intracellular membrane vesicles with the cell surface during cellularization (Loncar and Singer, 1995). Moreover, the KNOLLE gene of Arabidopsis has been shown to be a cytokinesis-specific syntaxin. The KNOLLE protein is found in cells only during mitosis, localizing to the plane of cell division, and mutations in this gene result in incomplete cytokinesis thought to result from an impairment in vesicle fusion (Lauber et al., 1997). Consistent with findings presented here in the sea urchin embryo, these cumulative observations make a strong argument for an essential and general role of syntaxins during cell division and development. It is possible that syntaxin is required for a broad range of membrane fusion events during cell division, ranging from organelle reconstitution to the generation of membrane surface area.

Sea urchin syntaxin associates with vesicles enriched at the cortex of the cleaving sea urchin embryo in addition to apparent ER labeling. Although in mammalian cells there appear to be distinct syntaxins that mediate membrane fusion events in discrete secretory compartments (Bennett et al., 1993; Bock et al., 1997; Tang et al., 1998; Wong et al., 1998), extensive PCR screening of sea urchin cDNAs has revealed only a single syntaxin homologue (Conner et al., 1997). Thus it is possible that a single syntaxin family member may be functioning in various secretory compartments in the sea urchin embryo, and its specific inactivation leads to membrane limiting steps and cessation of cell division. However, this is unlikely because cell division is unaffected by BFA treatment, arguing that surface membrane addition and the targeting of plasma membrane proteins early in embryogenesis come from a Golgi-independent membrane source or post-Golgi vesicles from maternally derived vesicle stores.

Syntaxin, in cooperation with VAMP and SNAP-25, has been shown to be involved in the formation of the minimal core membrane fusion machinery (Weber et al., 1998). Its role in neurotransmitter vesicle fusion has been extensively studied by taking advantage of BoNT-C1 (Foran et al., 1996; Marsal et al., 1997; O’Connor et al., 1997; Williamson and Neale, 1998), which specifically proteolyzes syntaxin family members possessing the appropriate cleavage site (Schiavo et al., 1995). Sea urchin syntaxin cDNA analysis indicates that it possesses the BoNT-C1 cleavage site (Conner et al., 1997), the protein can be cleaved in vitro by BoNT-C1 (Coorssen et al., 1997), and here we have shown that the toxin specifically removes syntaxin from vesicles enriched at the cortex and that syntaxin-specific antibodies block cell division. However, we are currently unable to test whether these vesicles are blocked in their fusion ability by either treatment, because we have no markers for the contents of these vesicles. It is feasible that it is the vesicle contents in addition to the inherent vesicle membrane proteins that are vital to cell division, and thus we are interested in their identification.

Although BoNT-C1 specificity for some syntaxin family members has been demonstrated (Schiavo et al., 1995), reports exist that BoNT-C1 can proteolyze both SNAP-25 and syntaxin in permeabilized chromaffin cells (Foran et al., 1996) and intact cultured neurons (Williamson et al., 1996) with equal efficiency. SNAP-25 cleavage by BoNT-C1 appears to occur at the C terminus, and although the exact site of protease cleavage is unknown, it is suspected that the protease recognizes a conserved conformation. A highly conserved SNAP-25 family member has recently been cloned in the sea urchin sperm (Schulz et al., 1998). Although we have been unable to detect SNAP-25 in eggs with antibodies against sperm SNAP-25, it is possible that the observed inhibition in cell division may be the cumulative affects of BoNT-C1 proteolysis of both syntaxin and SNAP-25. However, because cells injected with either BoNT-A or -E develop normally, we conclude that the BoNT-C1-induced phenotypes are specific for syntaxin proteolysis.

In this study we find that syntaxin inhibition blocks both cytokinesis and karyokinesis. However, it has been appreciated for some time that cytokinesis is separable from karyokinesis. For example, in the starfish, microinjection of antibodies against myosin results in blocking cytokinesis by preventing cleavage furrow formation, even though karyokinesis continues, as evidenced by the appearance of multiple daughter nuclei (Mabuchi and Okuno, 1977). More recently, selective inhibition of cytokinesis is observed when embryos are exposed to the natural marine toxins stypoldione from alga (O’Brien et al., 1989) and pseudopterolide from soft coral (Grace et al., 1992). These toxins are thought to target sulfhydryl-containing proteins involved in the formation of the contractile ring, yet karyokinesis continues in the cells. These studies focused on disruption of the cytoskeleton in cell division in contrast to the present study, which examines membrane dynamics. It is possible that if BoNT-C1 has targets on the ER, Golgi, or nuclear envelope, the introduction of the toxin could be disrupting homotypic membrane fusion events necessary for the reformation, fragmentation, or stability of these organelles during or after cell division. Thus, we hypothesize that treatment with syntaxin antibodies or BoNT-C1 could halt cell progression through the cell cycle at a checkpoint that monitors membrane status within the cell.