J Biol Chem, Vol. 274, Issue 35, 25173-25180, August 27, 1999

Neuronal Sensitivity to Tetanus Toxin Requires Gangliosides^*

Lura C. Williamson, Karen E. Bateman, Julianne C. M. Clifford^§, and Elaine A. Neale^¶

From the  Laboratory of Developmental Neurobiology, NICHHD, National Institutes of Health, Bethesda, Maryland 20892-4480 and the ^§ Center for Biologics Evaluation and Research, Food and Drug Administration, Bethesda, Maryland 20892

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Tetanus toxin produces spastic paralysis in situ by blocking inhibitory neurotransmitter release in the spinal cord. Although di- and trisialogangliosides bind tetanus toxin, their role as productive toxin receptors remains unclear. We examined toxin binding and action in spinal cord cell cultures grown in the presence of fumonisin B₁, an inhibitor of ganglioside synthesis. Mouse spinal cord neurons grown for 3 weeks in culture in 20 µM fumonisin B₁ develop dendrites, axons, and synaptic terminals similar to untreated neurons, even though thin layer chromatography shows a greater than 90% inhibition of ganglioside synthesis. Absence of tetanus and cholera toxin binding by toxin-horseradish peroxidase conjugates or immunofluorescence further indicates loss of mono- and polysialogangliosides. In contrast to control cultures, tetanus toxin added to fumonisin B₁-treated cultures does not block potassium-stimulated glycine release, inhibit activity-dependent uptake of FM1-43, or abolish immunoreactivity for vesicle-associated membrane protein, the toxin substrate. Supplementing fumonisin B₁-treated cultures with mixed brain gangliosides completely restores the ability of tetanus toxin to bind to the neuronal surface and to block neurotransmitter release. These data demonstrate that fumonisin B₁ protects against toxin-induced synaptic blockade and that gangliosides are a necessary component of the receptor mechanism for tetanus toxin.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Tetanus toxin (TeNT)¹ blocks the release of inhibitory neurotransmitters in the spinal cord leading to hyperactivity of motor neurons and consequent spastic paralysis (for review, see Ref. 1). The toxin is synthesized as a single polypeptide (M_r = 150,000), released by bacterial lysis, and cleaved to form an active dichain toxin with the heavy and light chains linked through one disulfide bond. The carboxyl-terminal half of the heavy chain constitutes the receptor binding domain, the amino-terminal half of the heavy chain is responsible for membrane translocation of the toxin, and the light chain contains the catalytic domain (for review, see Ref. 2). Upon entering the synaptic cytosol, the toxin arrests neurotransmitter release by its action as a zinc endopeptidase, which cleaves vesicle-associated membrane protein (VAMP), a synaptic vesicle integral membrane protein believed to be critical for neuroexocytosis (3, 4). Whereas TeNT has been shown to bind to the surface of neurons and to act intracellularly (for reviews, see Refs. 2 and 5-7), its cell surface receptor(s) and mechanism of delivery to the cytosol are not well understood.

Since the demonstration of the affinity of TeNT for gangliosides present on the neuronal surface (8), there have been a number of studies (for review, see Ref. 6) characterizing the binding of TeNT to gangliosides in various in vitro and in vivo preparations. Although it is difficult to assign an affinity to this binding (7), there is general agreement that TeNT shows the highest affinity for GT1b and GD1b polysialogangliosides, although not nearly as high as the affinity of cholera toxin for its known GM1 ganglioside receptor. It has been suggested that gangliosides participate in pore formation (9), intracellular targeting (10), and/or presentation of the toxin to its specific substrate(s) (11). Incubation of chromaffin cells with exogenous GD1b and GT1b confers TeNT sensitivity (12), and differentiation of PC12 cells with nerve growth factor increases the expression of complex gangliosides and enhances TeNT binding (13, 14). Although for PC12 cells nerve growth factor-induced differentiation is critical, it is not clear that enhanced ganglioside synthesis is sufficient for conferring toxin sensitivity (15). Furthermore, degradation of polysialogangliosides on the neuronal surface with neuraminidase treatment does not diminish sensitivity to TeNT (16). Thus, the role of gangliosides in the delivery of TeNT to its substrate is not clear. Moreover, TeNT binding, assessed on various neuronal membrane preparations, decreases after protease treatment (17-21) suggesting protein involvement in toxin binding and internalization. These data and other observations on toxin transport and specificity for central neurons (discussed in Refs. 2 and 6) have led to the proposal of a dual receptor model consisting of both gangliosides and protein (6, 22).

Spinal cord neurons in TeNT-exposed cultures mimic the electrophysiology of clinical tetanus; i.e. toxin action is manifest initially as a loss of inhibitory postsynaptic potentials and the onset of convulsant activity (23, 24). In a previous study, we examined the effect of TeNT on potassium-evoked neurotransmitter release in murine spinal cord cell cultures (25) and demonstrated that inhibitory glycinergic neurons are particularly sensitive to the toxin.

In the present study, we assess the role of gangliosides as functional receptors for TeNT in spinal cord cultures grown in the presence of the mycotoxin fumonisin B₁ (FB₁), an inhibitor of ganglioside synthesis. Fumonisin B₁ inhibits ceramide synthase, an enzyme required for the formation of ceramide, a precursor of ganglioside production (26). Tetanus toxin binding was examined qualitatively in cultures grown in the presence or absence of FB₁ using the carboxyl terminus of the toxin heavy chain (the binding fragment) conjugated to horseradish peroxidase (HRP) or by immunohistochemistry using holotoxin and a monoclonal antibody. Toxin action was assayed by potassium-evoked release of the inhibitory neurotransmitter glycine, by activity-dependent uptake of the styryl dye FM1-43 into synaptic terminals, and by immunohistochemistry and immunoblot analysis of VAMP cleavage. This study demonstrates that chronic FB₁ treatment of spinal cord cell cultures virtually eliminates gangliosides and TeNT binding to the neuronal surface and essentially eliminates action of TeNT. Furthermore, replenishing ganglioside-depleted cultures with exogenous gangliosides fully restores sensitivity to TeNT.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- Tetanus toxin (2 × 10⁷ mouse lethal doses/mg of protein) and monoclonal antibody 18.2.12.6 were generously provided by Drs. William Habig and Jane Halpern (Center for Biologics Evaluation and Research, FDA, Bethesda, MD). Fumonisin B₁ was obtained from Sigma. FdUrd was a gift from Hoffman-LaRoche. Purified ganglioside standards, mixed gangliosides from bovine brain, GD1b, and GT1b were from Matreya Inc., Pleasant Gap, PA. Cholera toxin fragment B-HRP and TeNT fragment C-HRP were from Calbiochem. Affinity-purified rabbit antibody against synapsin 1a and the mouse monoclonal antibody against VAMP used for immunohistochemistry were obtained from Chemicon, Temecula, CA. Mouse monoclonal antibodies against synaptophysin and microtubule-associated protein 2 were from Roche Molecular Biochemicals, and monoclonal antibody against phosphorylated neurofilaments (SMI-31) was from Sternberger Monoclonals Inc., Baltimore, MD. For immunoblots, monoclonal antibody against syntaxin was obtained from Chemicon, and a guinea pig polyclonal antibody that recognizes VAMP 1 and VAMP 2 (27) was a generous gift from Dr. Clifford Shone (Center for Applied Microbiology and Research, Porton Down, United Kingdom). FM1-43 and Hoescht 33258 were from Molecular Probes, Eugene, OR.

Spinal Cord Cell Cultures-- Timed C57BL/6NCR pregnant mice were obtained from the Frederick Cancer Research and Development Center, Frederick, MD. Spinal cords from 13-day-old fetal mice were trypsin-dissociated (28, 29) and plated at 1 × 10⁶ cells onto a mouse cortical astrocyte feeder layer as described previously (30). Cultures were grown for 3 weeks in a humidified 10% CO₂ atmosphere at 35 °C, with half-changes of medium twice weekly. Culture medium was minimal Eagle's medium (MEM) containing N3 supplement (see Fitzgerald (29) for content) and 5% horse serum. The antimetabolite FdUrd (54 µM) was added with 140 µM uridine to cultures within 1-2 days of plating for 96 h to inhibit non-neuronal cell overgrowth. Fumonisin B₁ (20 µM final concentration) was added to cultures 48 h after plating spinal cord cells and was included in each medium change. In some experiments, culture medium was replaced with MEM or MEM containing gangliosides with or without FB₁ for 24 h prior to assay. Fumonisin B₁ and/or gangliosides were washed from the cultures prior to experiments to preclude interference with TeNT binding to the neuronal surface.

The HEPES-buffered isosmotic salt solution (HBSS) used for all incubations of living cultures outside of a CO₂ environment was the same solution used for electrophysiologic studies of spinal cord cell cultures (31) and contains 136 mM NaCl, 3 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, and 10 mM glucose. The pH is adjusted to 7.25 and the osmolarity is adjusted with sucrose to 325 ± 5 mmol/liter.

For neuronal cell counts, cultures were rinsed in HBSS containing 0.1% bovine serum albumin (BSA), permeabilized with 0.05% saponin, and incubated in Hoescht 33258 (20 ng/ml in HBSS/BSA) for 30 min at 35 °C to stain all nuclei (32). Cultures were examined using an inverted microscope with simultaneous phase-contrast and fluorescence imaging to discriminate neuronal nuclei. Stained neuronal nuclei were counted using a 20× lens in 32 fields selected by a programmed computer matrix (IP Lab Spectrum from Scanalytics, Vienna, VA).

Immunohistochemistry-- Cultures were fixed in 2% paraformaldehyde in HBSS for 30 min at room temperature. After rinsing in phosphate-buffered saline (PBS), free aldehyde groups were blocked by the addition of 0.1 M glycine for 15 min, and the cells were permeabilized with 0.05% saponin in PBS for 30 min. Cultures were incubated overnight at 4 °C in mouse monoclonal antibody against microtubule-associated protein 2 (1:1000) or against synaptophysin (1:70) in PBS containing 5% normal goat serum. For immunostaining phosphorylated neurofilaments, cultures were rinsed and incubated in the mouse monoclonal antibody SMI-31 (1:10,000) in HBSS/BSA (which contains no phosphate). After rinsing 4 times in PBS/normal goat serum, cultures were incubated in goat anti-mouse IgG conjugated to rhodamine (1:100) for 60 min at 35 °C.

For double-staining experiments, control and FB₁-treated cultures were incubated in either 0.06 nM (10 ng/ml) TeNT in serum-free culture medium or in serum-free medium alone for 24 h. Cultures were fixed as described above and incubated overnight at 4 °C in a mixture of mouse monoclonal anti-VAMP and affinity-purified rabbit antisynapsin 1a, each diluted 1:1000 in PBS/normal goat serum. After rinsing, cultures were incubated for 60 min at 35 °C in a mixture of goat anti-mouse IgG-fluorescein (1:50) and goat anti-rabbit IgG-rhodamine (1:100) in PBS/normal goat serum.

Thin Layer Chromatography-- Sphingolipids labeled by incubating cell cultures in [¹⁴C]galactose (2 µCi/ml) for 24 h were isolated and identified as described by van Echten et al. (33). In brief, after radiolabeling, cells were harvested, and lipids were extracted from the cell pellet with 3 ml of chloroform/methanol/water/pyridine (60:30:6:1, by volume) for 48 h at 50 °C. Lipid extracts were desalted by reverse-phase chromatography on silica gel LiChroprep RP18, applied to silica gel high performance thin layer chromatography plates, and chromatographed with chloroform/methanol/aqueous CaCl₂ (0.22%) (60:35:8, by volume). Ganglioside standards and sample spots were visualized with anisaldehyde spray reagent (Sigma) and quantified by scintillation spectrometry.

Toxin Binding-- Fragment C of TeNT or fragment B of cholera toxin (toxin-binding fragments) each conjugated to HRP (2.5 µg/ml) was added to control and FB₁-treated cultures for 30 min at room temperature. After rinsing in HBSS, cultures were fixed in 2% paraformaldehyde in HBSS. Cells were reacted for 15 min with 3,3'-diaminobenzidine tetrahydrochloride (0.75 mg/ml in 0.05 M Tris-HCl, pH 7.6) containing 0.01% hydrogen peroxide and treated for a few seconds with 0.1% osmium tetroxide to intensify the reaction product. After rinsing in Tris-HCl, cultures were stored in Tris-buffered saline containing glycerol for subsequent photography.

Alternatively, tetanus holotoxin (4 µg/ml) was premixed with the monoclonal antibody 18.2.12.6 (1:2000, 3.75 µg/ml) in HBSS/BSA for 30 min at room temperature (34). Living cultures were incubated for 30 min each at room temperature in toxin-antibody complex followed by goat anti-mouse IgG conjugated to rhodamine (1:100) with HBSS/BSA rinses between incubations. Cultures were rinsed, fixed in 2% paraformaldehyde, and stored in glycerol containing n-propyl gallate to prevent fluorescence photobleaching (35).

Evoked Release of Neurotransmitter-- Cell cultures were radiolabeled with [³H]glycine (3 µCi/ml) in HBSS containing 0.1% BSA for 30 min at 35 °C. After rinsing once in cold (4 °C) HBSS/BSA, control and FB₁-grown cultures were incubated in either 0.03 nM TeNT or HBSS/BSA alone for 30 min at 4 °C. Toxin was removed with a cold wash, and cultures were transferred to a water bath at 35 °C for 90 min to allow toxin internalization and trafficking to its site of action. Glycine release was evoked as described previously (25). Cultures were exposed sequentially to 1.25 ml of the following solutions for 5 min intervals at 35 °C: HBSS without calcium and with 0.5 mM EGTA; HBSS containing 56 mM KCl, 83 mM NaCl, and no CaCl₂; and HBSS containing 56 mM KCl, 83 mM NaCl, and 2 mM CaCl₂. Cell-associated radioactivity was assayed after dissolving the cells in 0.2 N NaOH and neutralizing with HCl. Calcium-dependent release is the amount of radiolabeled glycine released in the presence of calcium minus that released in the absence of calcium and expressed as a percent of the total radioactivity at the start of the release assay normalized to control values.

FM1-43 Uptake-- Living control and FB₁-treated cultures were exposed to 0.6 nM TeNT in HBSS/BSA for 30 min at 4 °C. After rinsing, cultures were incubated for 90 min at 35 °C as described above for the glycine release assay. Cultures were washed in HBSS/BSA and incubated in 2 µM FM1-43 in HBSS containing 56 mM KCl and 2 mM CaCl₂ for 5 min at room temperature. Finally, cultures were rinsed 3 times (10 min each) in 1 µM tetrodotoxin in HBSS/BSA to allow for the recycling of vesicle membrane labeled with FM1-43 without the loss of label through exocytosis of labeled vesicles. Cultures were maintained in 1 µM tetrodotoxin in HBSS/BSA for photography using a Zeiss Photomicroscope II and TMAX-3200 film.

Immunoblot Analysis-- After incubation with TeNT (0.06 nM for 24 h), spinal cord neurons from control and FB₁-treated cultures were detached from tissue culture dishes by trypsinization and rinsed once with PBS. Cells were dissolved by boiling for 5 min in electrophoresis sample buffer containing 2% SDS and dithiothreitol. Protein samples were run on 10-20% SDS-polyacrylamide gels (Bio-Rad) and transferred to nitrocellulose membranes. Membranes were blocked for 1 h in Tris-buffered saline (20 mM Tris, 500 mM NaCl, pH 7.5) containing 5% nonfat dry milk and 0.05% Tween 20 and incubated overnight with primary antibodies diluted 1:1000 with Tween 20 in Tris-buffered saline. Membranes were washed twice with Tween 20 in Tris-buffered saline prior to a 2 h incubation with the appropriate alkaline phosphatase-conjugated secondary antibody (Bio-Rad) followed by color development.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Mouse spinal cord cultures grown for 3 weeks, from day 2 after plating, in the continuous presence of 20 µM FB₁ show some decrease in neuron survival but no difference in overall morphology from untreated cultures. From neuronal cell counts (31 fields/dish, 3 dishes for each condition), we estimate the total number of neurons in control cultures at 2.86 × 10⁵ and in FB₁-treated cultures at 2.49 × 10⁵. Thus, cell counts indicate a 13% loss of neurons because of FB₁ treatment. Control and treated cultures appear similar, however, by phase contrast microscopy (Fig. 1). We further examined neuronal morphology by immunohistochemistry using monoclonal antibodies against microtubule-associated protein 2 as a marker for dendrites, against phosphorylated neurofilaments for axons, and against synaptophysin for synaptic terminals (Fig. 2). Control and FB₁-treated neurons exhibit similar immunostaining for dendrites (Fig. 2, A and B), axons (Fig. 2, C and D), and synaptic terminals (Fig. 2, E and F).

View larger version (165K): [in this window] [in a new window]   Fig. 1.   Phase-contrast photomicrographs of mouse spinal cord neurons in dissociated cell culture maintained in the absence (A) or presence (B) of 20 µM FB1. FB1 does not appear to be detrimental to neuronal development. Magnification bar, 100 µm.

View larger version (117K): [in this window] [in a new window]   Fig. 2.   Fumonisin B1 effects on neuronal morphology. Spinal cord neurons in culture grown for 3 weeks in the absence (A, C, and E) or presence (B, D, and F) of FB1 (20 µM) were immunostained with monoclonal antibodies against microtubule-associated protein 2 (A and B) to label dendrites, against phosphorylated neurofilaments (C and D) to label axons, and against the synaptic vesicle protein synaptophysin (E and F) to label synaptic terminals. Immunostaining of dendrites, axons, and terminals appears similar in control and FB1-treated cultures. Magnification bar, 25 µm.

We anticipated that prolonged incubation in the presence of FB₁ would allow turnover of gangliosides synthesized during the 48 h prior to the addition of the inhibitor and would significantly deplete neuronal cell cultures of newly synthesized gangliosides. The loss of gangliosides from cells in FB₁-treated cultures was confirmed by thin layer chromatography and quantified by scintillation spectrometry (Fig. 3). Total radioactivity that comigrates with GT1b, GD1b, GD1a, and GM1 ganglioside standards is reduced more than 93% in FB₁-treated cultures compared with controls. Specifically, GD1b and GM1 are each depleted to 10% of control levels and GT1b and GD1a to 1 and 4%, respectively.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Thin layer chromatography of radiolabeled gangliosides extracted from 3-week-old spinal cord cultures grown in the absence (open symbols) or presence (closed symbols) of FB1 (20 µM). Cultures were radiolabeled with [14C]galactose (2 µCi/ml for 24 h). Gangliosides were extracted and identified by their comigration with ganglioside standards on thin layer chromatography plates. Newly synthesized gangliosides are reduced by 93% in FB1-treated cultures.

The role of gangliosides as putative receptors for TeNT was assessed in spinal cord cultures grown in the presence or absence of FB₁. We first examined the effect of FB₁ on toxin binding using fragment C, the binding domain of TeNT, conjugated to HRP. Essentially no peroxidase reaction product is seen in FB₁-treated cultures (Fig. 4B) in comparison to controls (Fig. 4A), suggesting very little fragment C binding to these neurons. We also failed to detect the binding of tetanus holotoxin to FB₁-treated cultures using indirect immunofluorescence with a monoclonal antibody that stabilizes toxin binding, enhancing it by greater than 2-fold when premixed with toxin (36). Intense immunostaining of TeNT to the surface of neuronal membranes is exhibited in control cultures (Fig. 4C) in marked contrast to the absence of immunoreactivity in FB₁-treated cultures (Fig. 4D). Thus, using saturating levels of fragment C-HRP (2.5 µg/ml) or tetanus holotoxin (4 µg/ml), we were not able to visualize TeNT bound to spinal cord neurons in FB₁-treated cultures. Moreover, the HRP-conjugated fragment B of cholera toxin, which binds GM1 gangliosides, labels the surface of control neurons (Fig. 4E) but is essentially absent from FB₁-treated neurons (Fig. 4F), providing further evidence for the loss of gangliosides from these cultures.

View larger version (120K): [in this window] [in a new window]   Fig. 4.   TeNT and cholera toxin binding to spinal cord neurons in control (A, C, and E) and FB1-treated (B, D, and F) cultures. Fragment C of TeNT conjugated to HRP (A and B), tetanus holotoxin premixed with the monoclonal antibody 18.2.12.6 (C and D), or fragment B of cholera toxin conjugated to HRP (E and F) was added to cultures for 30 min at room temperature. Very little peroxidase reaction product is seen on neuronal surfaces in FB1-treated cultures (20 µM for 3 weeks) incubated with TeNT fragment C (B) or cholera toxin fragment B (F) compared with their respective control cultures (A and E). Virtually no immunostaining for tetanus holotoxin was detected in cultures grown in the presence of FB1 (D) in comparison to control cultures (C). Magnification bar, 25 µm.

The most meaningful assessments of the requirement of TeNT for gangliosides would investigate the ability of TeNT to accomplish synaptic blockade by proteolytic cleavage of its substrate in neurons without gangliosides. We used activity-dependent uptake of the styryl dye FM1-43 (37) as a qualitative measure of TeNT-induced synaptic quiescence. In functional synaptic terminals, FM1-43 present during neuronal stimulation labels nerve terminal surface membrane that is endocytosed for synaptic vesicle recycling. In control and FB₁-treated neurons (without TeNT), FM1-43 labels synaptic terminals that have undergone synaptic vesicle exo- and endocytosis as a consequence of stimulation with 56 mM potassium and 2 mM calcium (Fig. 5, A and C). As expected, TeNT substantially inhibits the uptake of FM1-43 in control cultures (Fig. 5B). In contrast, the effect of TeNT in blocking FM1-43 uptake into synaptic vesicles in FB₁-grown cultures is minimal (Fig. 5D). FM1-43 uptake persists in FB₁-grown neurons exposed to blocking concentrations of TeNT, indicative of synaptic vesicle recycling in active terminals.

View larger version (97K): [in this window] [in a new window]   Fig. 5.   Fumonisin B1 effect on TeNT blockade of synaptic activity. Control and FB1-treated cultures were exposed to 0.6 nM TeNT for 30 min at 4 °C and then warmed for 90 min at 35 °C. Activity-dependent uptake of FM1-43 into synaptic terminals is seen as punctate fluorescence in control (A) and FB1-treated (20 µM for 3 weeks) cultures (C). FM1-43 uptake is inhibited in cultures exposed to TeNT (B), which is evidence of a block in synaptic vesicle recycling. However, in cultures grown in the presence of FB1, TeNT has little effect on FM1-43 uptake into synaptic terminals (D), indicating the persistence of synaptic activity in the presence of blocking concentrations of TeNT. Magnification bar, 25 µm.

The role of gangliosides in TeNT internalization and subsequent action was quantified by measuring toxin-induced inhibition of glycine release in cultures grown in the presence or absence of FB₁ (Fig. 6). Cultures were incubated in 0.03 nM (5 ng/ml) TeNT as described for FM1-43 uptake. Under the conditions of the assay, TeNT blocks potassium-evoked calcium-dependent release of [³H]glycine by 40-45% in control cultures. [³H]Glycine release is depressed by about 15% in cultures treated chronically with FB₁, consistent with the decrease in neuron survival. However, when added to FB₁-treated cultures, TeNT is unable to produce any further reduction in release. Thus, the effectiveness of TeNT in blocking neurotransmitter release is abrogated in spinal cord neurons without gangliosides.

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Fumonisin B1 effect on TeNT blockade of evoked glycine release. Control (C) and FB1-treated cultures (20 µM for 3 weeks) were exposed to 0.03 nM TeNT for 30 min at 4 °C, rinsed, and then warmed for 90 min at 35 °C. Release of [3H]glycine was evoked with medium containing 56 mM potassium and 2 mM calcium. TeNT arrests glycine release by almost 45% in control cultures without FB1 treatment. In contrast, TeNT has no effect on glycine release in cultures grown in the presence of FB1. Control cultures in these experiments released about 31 ± 2% of total [3H]glycine with potassium depolarization. n = 10 for each condition, four experiments. T, tetanus toxin; F, fumonisin B1.

The proteolytic activity of TeNT on VAMP cleavage in control and FB₁-treated cultures was assessed using double-stain immunohistochemistry for VAMP and synapsin 1a. In control cultures, VAMP immunoreactivity is found in synaptic terminals (Fig. 7B) identified using antisynapsin 1a to label synaptic vesicles (Fig. 7A). In cultures exposed for 24 h to 0.06 nM TeNT, synapsin 1a immunoreactivity is similar to that seen in control cultures (Fig. 7C), whereas VAMP immunostaining is abolished (Fig. 7D). Chronic treatment with FB₁ prevents ganglioside production but does not alter either synapsin 1a (Fig. 7E) or VAMP (Fig. 7F) immunostaining. Finally, in cultures grown in FB₁, TeNT exposure has little discernible effect on VAMP immunoreactivity (Fig. 7H) as assessed qualitatively by fluorescence intensity and colocalization with synapsin 1a (Fig. 7G).

View larger version (126K): [in this window] [in a new window]   Fig. 7.   Fumonisin B1 effect on VAMP proteolysis by TeNT. Synapsin 1a (A, C, E, and G) and VAMP (B, D, F, and H) immunoreactivities colocalize in cultures grown for 3 weeks in the absence (A and B) or presence (E and F) of 20 µM FB1. In contrast, in cultures treated with TeNT (0.06 nM for 24 h), synapsin 1a marks synaptic terminals (C) in which VAMP immunostaining is abolished (D). In FB1-treated cultures, TeNT has little effect on VAMP immunostaining (H), which colocalizes with synapsin 1a (G). Magnification bar, 25 µm.

VAMP cleavage after a 24-h exposure to TeNT (same conditions as described above for immunohistochemistry) in control and FB₁-treated cultures was evaluated by immunoblot analysis of VAMP in relation to syntaxin (Fig. 8). Tetanus toxin completely cleaves VAMP in control cultures grown in the absence of FB₁. In contrast, VAMP is detected readily in cultures grown in the presence of FB₁ regardless of the presence of TeNT. These data indicate that ganglioside-depleted neurons show significant protection from TeNT effects, most likely because of the inability of TeNT to bind efficiently to the neuronal surface and undergo internalization into the neuronal cytosol.

View larger version (41K): [in this window] [in a new window]   Fig. 8.   Immunoblot analysis of VAMP cleavage by TeNT. VAMP proteolysis was evaluated relative to syntaxin, another synaptic protein, in cultures grown for 3 weeks in the absence or presence of 20 µM FB1. VAMP immunoreactivity is abolished in control cultures exposed to TeNT (0.06 nM for 24 h). In FB1-treated cultures, there appears to be little degradation of VAMP even after treatment with TeNT. Con, control; Fum, fumonisin B1.

To establish the physiologic role of gangliosides in mediating the action on TeNT, we replenished FB₁-treated neurons with exogenous gangliosides. Experiments were performed on a 3-4-week-old control and FB₁ cultures in which the culture medium was replaced, 24 h prior to assay, with MEM with or without FB₁ or with MEM containing gangliosides with or without FB₁. Cultures were loaded with [³H]glycine and incubated in 0.03 nM TeNT (5 ng/ml, 150 mouse lethal doses/culture) as described. Glycine release under eight experimental conditions was compared in any given experiment. Fig. 9 illustrates results from two experiments in which cultures had been preincubated with bovine mixed brain gangliosides at 50 µg/ml. In these experiments, FB₁ affords complete protection against TeNT. Relative to control cultures preincubated with gangliosides, TeNT inhibits glycine release by almost 50%. Similarly, TeNT in FB₁-treated cultures preincubated with gangliosides produces a 52% block in release. Thus, exogenous gangliosides added back to ganglioside-depleted neurons fully restore the action of TeNT. As anticipated, the ganglioside supplement also restores TeNT association with the neuronal surface membrane (Fig. 9, lower panel). Because the added gangliosides insert into depleted glial membranes as well, these cells also exhibit TeNT binding not usually observed on glial cells in control cultures.

View larger version (79K): [in this window] [in a new window]   Fig. 9.   Exogenous gangliosides restore TeNT effect in FB1-treated cultures. Upper panel, control and FB1-treated cultures were incubated for 24 h in 50 µg/ml mixed brain gangliosides in MEM, loaded with [3H]glycine, exposed to 0.03 nM TeNT for 30 min at 4 °C, washed, and, after 90 min at 35 °C to allow toxin internalization, assayed for potassium-stimulated release of [3H]glycine. As shown in Fig. 6, [3H]glycine release is somewhat reduced from controls in cultures treated with FB1. TeNT at 0.03 nM inhibits release by about 40% in controls and by 48% in control cultures incubated with gangliosides. TeNT has no effect at this concentration in FB1-treated cultures. However, the addition of gangliosides to FB1-treated, ganglioside-depleted cultures completely restores TeNT action with the toxin blocking 51% of [3H]glycine release. In control cultures, 31 ± 1.4% of total [3H]glycine is released by potassium depolarization. n = 5 for each condition, two experiments. C; control; T, tetanus toxin; F, fumonisin B1; G, 50 µg/ml mixed brain gangliosides. Lower panel, tetanus toxin binding visualized using antibody 18.2.12.6 and rhodamine-conjugated secondary antibody. TeNT adheres to neuronal surface membranes in control cultures, whereas toxin binding in FB1-treated cultures is barely detectable. TeNT toxin binding to ganglioside-depleted neurons is restored by the addition of 50 µg/ml mixed brain gangliosides (MBG) or GD1b (7.5 µg/ml) and GT1b (5.0 µg/ml) to the culture medium for 24 h prior to staining. This treatment also caused the appearance of TeNT binding sites on the surface of glial cells. Magnification bar, 25 µm.

Similar experiments (data not shown) provide preliminary results that indicate that 30 µg/ml mixed brain gangliosides are somewhat less effective (42% toxin-induced block in controls plus gangliosides, 24% block in FB₁ cultures plus gangliosides). In one experiment, only GD1b and GT1b were used as the supplement at the concentration at which they were present in the 50 µg/ml ganglioside mixture, i.e. 7.5 µg/ml GD1b and 5 µg/ml GT1b. In this experiment, the TeNT block in glycine release was 62% in control cultures with gangliosides and 65% in FB₁ cultures with gangliosides. Thus, GD1b and GT1b appear to be as effective in restoring toxin sensitivity as the ganglioside mixture.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Spinal cord neurons in cultures treated chronically with FB₁ exhibit a greater than 90% decrease in ganglioside synthesis and significant protection against the effects of TeNT. Full sensitivity to TeNT is restored by a 24-h incubation, in the continued presence of FB₁, in medium containing exogenous gangliosides.

The survival of neurons grown in 20 µM FB₁ for 3 weeks is somewhat compromised, although the treated neurons develop dendrites, axons, and synaptic terminals similar to those of neurons with normal ganglioside synthesis. Comparable observations were reported for hippocampal neurons in culture (38); that is, long term incubation (12-14 days) with FB₁ failed to elicit changes in neuronal morphology.

We did observe that dividing glial cells were affected severely by FB₁ treatment, consistent with the report that interference with endogenous GM1 by the -subunit of cholera toxin leads to an inhibition of DNA synthesis and astroglial cell proliferation (39). The inability of astroglia to proliferate posed a technical problem for our studies. Dissociated spinal cord cells at plating contain both glia and neurons. Normally, glia attach to the culture substrate, proliferate, and spread; neurons adhere best to glial surfaces. Chronic FB₁ treatment renders such cultures fragile in that neurons without a glial substrate detach easily from the culture dish. Therefore, for most of the experiments reported here, we plated dissociated spinal cord cells onto a preformed, mitosis-arrested glial monolayer. The data in Fig. 3, however, were obtained from cultures without the additional glial monolayer to represent more closely the ganglioside content of neurons. In this regard, GD1b was the ganglioside that remained marginally detectable by thin layer chromatography of hippocampal neurons grown in FB₁, when it could not be detected on the neuronal surface by immunohistochemistry (40).

Although it is known that TeNT binds with some specificity to di- and trisialogangliosides, the exact role of gangliosides in toxin uptake and delivery is not clear. TeNT binds preferentially to GT1b and GD1b rather than GM1 but the difference in affinities is small when compared with cholera toxin, whose binding affinity for GM1 gangliosides is at least three orders of magnitude over that for any other major brain ganglioside (for reviews, see Refs. 1, 6, and 41). If gangliosides comprise the receptor or essential components of the receptor for TeNT, loss of gangliosides should protect against the action of TeNT.

Bigalke et al. (16) reported that neuraminidase treatment of spinal cord neurons in culture did not protect against TeNT action. Neuraminidase eliminated surface membrane gangliosides of the GQ1, GT1, and GD1 series by degrading them to GM1. Although preferring GT1b and GD1b, TeNT binds to GM1 gangliosides, albeit with lower affinity, possibly explaining the persistence of a toxin effect in neuraminidase-treated cultures. The potency of TeNT arises from its efficient action as a zinc endopeptidase with very few molecules required to cleave VAMP and prevent neurotransmitter release (2). Furthermore, TeNT retains a long half-life in spinal cord neurons in cell culture (42). Consequently, it is not surprising that neuraminidase treatment does not confer protection against TeNT action.

At higher concentration (0.6 nM), TeNT is somewhat able to suppress neurotransmitter release from FB₁-treated neurons; i.e. when the toxin-induced block is almost complete in untreated cultures, the block in FB₁ cultures is about 25%. This effect most likely is related to persistent low level synthesis of GD1b (Fig. 3) together with a high concentration of toxin. However, from the data on neurotransmitter release depicted in Figs. 6 and 9, 0.03 nM (5 ng/ml) TeNT causes a 40-45% reduction in glycine release in control cultures in 2 h and yet produces no effect in FB₁ cultures. Toxin-induced synaptic blockade is fully restored by the addition of exogenous gangliosides to the culture medium prior to toxin exposure, and this establishes that gangliosides are required to mediate the action of TeNT. Gangliosides alone may not be sufficient but may function only in concert with a separate cooperating receptor, possibly a protein. Our data thus far are not incompatible with a dual receptor model (22). Fragment C of TeNT has been shown by x-ray crystallography to contain two domains; one is similar to carbohydrate-binding lectins and the other contains a -trefoil motif as observed in various proteins such as acidic and basic fibroblast growth factor (43). Thus, the structure of the binding subunit of toxin presumably could accommodate both ganglioside and protein receptors. Binding of botulinum neurotoxin B, another clostridial neurotoxin, to synaptotagmin, its reported protein receptor, is dependent on the presence of gangliosides GT1b and GD1a (44). If the receptor mechanism for TeNT is similar to that for botulinum neurotoxin B, the binding of TeNT to its protein receptor similarly might depend on the presence of gangliosides. Experiments to assess TeNT action in neurons whose surface membrane is depleted of both proteins and gangliosides would address this issue directly.

This study is the first to examine both binding and action of TeNT in a physiologically relevant neuronal culture system in which the majority of gangliosides have been removed from the neuronal surface by synthesis blockade. The data demonstrate that gangliosides are a necessary component of the mechanism for TeNT delivery to its substrate. Neurons grown in FB₁-treated cultures should prove useful in isolating protein receptors for TeNT and other clostridial neurotoxins, an endeavor that has been confounded by the abundance of gangliosides to which the toxins bind.

	ACKNOWLEDGEMENTS

We thank Veronica Dunlap for providing spinal cord cell cultures, Violet Aldaia for assistance with preliminary experiments, and Drs. William Habig and Jane Halpern for generous gifts of purified TeNT and antibodies and valuable criticism and suggestions over the years.

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^¶ To whom correspondence should be addressed. Tel.: (301) 496-6419; Fax: (301) 496-9939; E-mail: eneale@codon.nih.gov.

	ABBREVIATIONS

The abbreviations used are: TeNT, tetanus neurotoxin; FB₁, fumonisin B₁; VAMP, vesicle-associated membrane protein (also known as synaptobrevin); MEM, minimal Eagle's medium; HRP, horseradish peroxidase; HBSS, HEPES-buffered isosmotic salt solution; PBS, phosphate-buffered saline; BSA, bovine serum albumin; FdUrd, 5-fluoro-2'-deoxyuridine.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES