Adherence of botulinum and tetanus neurotoxins to synaptosomal proteins

0361-9230/92 $5.00 + .OO Copyright 0 1992 Pergamon Press Ltd.

Brain Research Bulletin, Vol. 29, pp. 917-924, 1992 Printed in the USA. All rights reserved.

Adherence of Botulinurn and Tetanus Neurotoxins to Synaptosomal Proteins CARA-LYNNE

SCHENGRUND,’

NANCY

J. RINGLER

AND

BIBHUTI

R. DASGUPTA*

Department of Biological Chemistry, The M. S. Hershey Medical Center, The Pennsylvania State University, Hershey, PA 17033 *Department of Food Microbiology and Toxicology, University of Wisconsin, 1925 Willow Drive, Madison, WI 53 706 Received

2 1 August

199 1; Accepted

12 May 1992

SCHENGRUND, C-L., N. J. RINGLER AND B. R. DASGUFTA. Adherenceofbotulinumandtetanusneurotoxinsto synaptosomal proteins.BRAIN RES BULL 29(6) 9 17-924, 1992.-The ability of %labeled botulinum type A and tetanus neurotoxins to adhere to blots of synaptosomal proteins separated by SDS-polyacrylamide gel electrophoresis was studied. Both neurotoxins appeared to adhere preferentially to an -80 kDa and to a lesser extent to an - I I6 kDa protein(s). Adherence of the neurotoxins to these proteins was enhanced by preincubation of the neurotoxins with GT I b. The - 100 kDa heavy chain segment of BTxA adhered to the same proteins. The carboxy terminal half of the heavy chain adhered primarily to the -80 kDa protein(s) while the amino terminal portion bound most intensely to the - I 16 kDa protein(s). The ability of the -80 and - 116 kDa proteins to stain positively with the periodic acid-Schiff reagent and to bind ‘251-labeledwheat germ lectin suggests that they are glycosylated. Both neurotoxins appear to adhere to the same -80 and - I I6 kDa proteins because tetanus neurotoxin preincubated with GT 1b was able to reduce binding of radiolabeled botulinum type A neurotoxin to both proteins. Neither neurotoxin adhered to blots of proteins from liver, spleen, or kidney, suggesting that the proteins adhered to are neural components. Botulinum neurotoxin

Tetanus neurotoxin

Synaptosomes

only partially inhibited by preincubation with crude gangliosides

BOTULINUM (BTx) and tetanus (TTx) neurotoxins produced by Clostridium botulinurn and Clostridium tetani, respectively, are selective for their sites of action. BTx binds to the presynaptic terminal ofthe neuromuscular junction (NMJ) and blocks release of acetylcholine. TTx binds to the NMJ and is retrogradely transported to the central nervous system (CNS) where it blocks release of the inhibitory neurotransmitters, glycine and GABA. The results are that BTx induces a flacid paralysis while TTx induces a spastic paralysis (19,44). Little is known about the binding sites present on the presynaptic membranes that determine the target selection of these neurotoxins. Numerous reports (22,24,48) indicate that TTx and BTx bind to gangliosides of the lb series (47). However, the affinity and specificity of these toxins for Gl b gangliosides is much lower than that of cholera toxin for GMl. In inhibition studies, we observed that 50% inhibition of BTx or TTx binding to GT 1b-coated plastic wells occurred at GT 1b concentrations in 3,000-fold molar excess (38). In contrast, analogous experiments showed that GM 1 inhibition of cholera toxin binding to GM l-coated plastic wells occurred at a less than IO-fold molar excess (39). Additional evidence suggesting that receptors other than gangliosides may exist includes (a) retrograde transport of cholera toxin can be completely inhibited by preincubation of the toxin with GMl, whereas retrograde transport of TTx is

or GTl b (46); (b) brief trypsinization or fixation with formaldehyde or glutaraldehyde of cells dissociated from 16-day-old fetal rat cerebra reduces TTx binding (33,5 1); (c) birds, which have high levels of G 1b gangliosides, are resistant to TTx (28).

This type of observation led Montecucco to suggest that in addition to the ganglioside binding site, the neurotoxins also contain a protein binding site. Recently, Schiavo et al. (41) found that a subclone of rat pheochromocytoma (PC12) cells that express high affinity receptors for TTx upon NGF-induced differentiation, contains an -20 kDa protein to which TTx can be cross-linked. They suggested that the protein is involved in the neurospecific binding of TTx. Because it is difficult to work with nerve termini at the NMJ, we initiated a search for protein (glycoprotein) components of synaptosomes, isolated from the gray matter of bovine brain, that might be selectively adhered to by BTxA and TTx. Our working hypothesis was that if the two neurotoxins adhered to a few synaptosomal proteins, the protein(s) bound might be (a) part of the low and/or high affinity binding sites, or (b) a component adhered to when the neurotoxin induces its physiological effects. The later hypothesis is a possibility because the assay monitored binding of the neurotoxins to blots of synaptosomal

’ To whom requests for reprints should be addressed,

917

918

SCHENGRUND.

proteins, hence, all proteins associated with the synaptosomal preparation were available as potential binding sites for the neurotoxins. EXPERIMENTAL

PROCEDURES

The following materials were purchased from the indicated sources: PD- 10 (Sephadex G-25M) columns, DEAE-Sephadex, SP-Sephadex, and QAE-Sephadex (Pharmacia LKB, Piscataway, NJ); polyvinylidene difluoride (PVDF) transfer membranes (Millipore, Bedford, MA); cholera toxin (Sigma Chemical Co., St. Louis, MO); Na”‘I (Amersham Corp., Arlington Heights, IL); “C-formaldehyde and Bolton Hunter reagent (New England Nuclear, Boston, MA); ‘251-wheat germ agglutinin (ICN, Irvine. CA); [ 1-‘4C]-acetic acid (American Radiolabeled Chemicals, Inc., St. Louis, MO); prestained molecular weight standards for sodium dodecylsulfate polyacrylamide gel electrophoresis (SDSPAGE, Bio-Rad Laboratories, Richmond, CA); Kodak XAR-2 X-ray film (Standard Medical Systems, Columbia, MD). TTx (Massachusetts Health Research Institute and Co., Boston, MA) was further purified by hydroxylapatite (Calbiochem, San Diego, CA) chromatography using the procedure described by Dr. William Habig (Food and Drug Administration, Bethesda, MD, personal communication). TTx applied to a 6 X 0.7 cm column in O.OOSM Na phosphate buffer (pH6.8) eluted over the range of 0.05-o. 10M Na phosphate; SDS-PAGE indicated its purity was > 98%. Gangliosides were isolated from the gray matter of bovine brains as previously described (39). 14C-labeled mixed bovine brain gangliosides (specific activity of-O.5 &i/mg) were prepared by reacetylating N-deacetylated samples with the sodium salt of [ 1“4C] acetic acid using the procedure described by Nores et al. (32). BTxA, BTxB, and BTxE were isolated from cultures of type A (strain Hall), B (strain Okra), and E (strain Alaska E43) as previously described (7,8,9,). The chromatographically isolated neurotoxins were pure; each preparation migrated in SDS-PAGE (Coomassie Blue stained) as an - 150 kDa single band as shown in the references cited for purification. Purity has been further verified based on amino acid sequence analysis (36). The -50 kDa light and - 100 kDa heavy chains of BTxA were separated by a dithiothreitol-induced reduction of the disulfide bond, followed by ion exchange chromatography. SDS-PAGE and amino acid sequence analysis (not shown here, see 36) indicated that the fractions were pure. The heavy chain of BTxA was cleaved with trypsin to generate - 50 kDa N- and C-terminal fragments which were separated using ion exchange chromatography. Purity and identity of the two separated fragments of the heavy chain have been established based on SDS-PAGE and amino acid sequence determinations (37). BTxA, its heavy chain, and the C-terminal fragment of the heavy chain, as well as BTxB and BTxE were labeled with “‘1 using the chloramine T procedure as modified by Williams et al. (50). Briefly, 200 gg of protein in 200 ~1 of O.lOM sodium phosphate buffer (pH 7.4) containing 0.15M NaCl was exposed to 2 mCi of Na’251 in the presence of 12 pg of chloramine-T for 1 min at 20°C. The reaction was quenched by adding 25 pg of tyrosine. The ‘251-labeled protein was isolated by chromatography on a PD- 10 column (pre-blocked with BSA) using 0.1 OM sodium phosphate buffer (pH 7.4) containing 0.15M NaCl as eluant. Specific activities were between 4 and 7 &i/pg. The LDSo (lowest concentration of neurotoxin, injected IP, needed to kill 50% of the mice within 4 days) of the pure and ‘251-labeled BTxA was determined. CD1 strain female mice, 19-21 gm, were injected IP with 0.5 ml of the serially diluted neurotoxin in 0.1 M sodium phosphate buffer containing 0.2% gelatin and time until death

RINGLER

AND

DASGIlPT.4

for the following 96 h. For cold BTxA the LD,,, was of BTxA, while that for “51-labeled BTxA was between 5 X 10’ and I X IO* mice/mg. TTx was labeled with ‘25I using either the chloramine-T procedure described which is milder (50) than the one used by Habermann (17: ?3 FCi/pg) or the Bolton-Hunter method (0.3 &i/fig) which has been found to have less effect on the toxicity of TTx (I ). SDSPAGE analysis of freshly iodinated preparations (5% stacking. 7.5% running gel, data not shown) indicated that the labeling procedures did not alter migration of the neurotoxins. Because the light chain of BTxA did not label well with lz51. all of the BTxA fragments were labeled with 14Cby reductive methylation of the amino groups ( 10). The specific activities of the 14C-labeled fragments were between 1-5 nCi/pg. Monitoring adherence of the heavy chain as well as its C-terminal fragment labeled with either 14C or lz51. provided an internal monitor of whether the labeling procedure or the amount of labeled fragment used had an effect on the binding patterns obtained. Synaptosomes were isolated from crude gray matter of bovine brain (obtained from a local abattoir) using the procedure described by Eichberg et al. (I 3). This procedure was selected because we had previously found (40) that the synaptosomal fraction was enriched in synaptosomes, a conclusion confirmed by transmission electron microscopic analysis of the recovered fraction (Fig. I). Synaptosomes were concentrated by reducing the concentration of the sucrose solution in which they were recovered to -0.4M and centrifuging at 100,000 X g for 60 min at 4°C. The pellet was then taken up in a small volume of water and aliquots used for protein analysis and assay. Membranes were isolated from the brain, liver. spleen, and kidneys of rats as follows: The tissue was cut into small pieces and the cells dissociated by sieving in cold erythrocyte lysis buffer (0. I55 M NH,CI containing 0.1 mM disodium EDTA and 0.0 I M KHCO+ Dissociated cells were recovered by centrifugation at I .OOOX g for 5 min at 4°C. Recovered cells were lysed in cold water and the membranes recovered by centrifuging at 100.000 X g for I h at 4°C. Protein concentration of the synaptosomes and membrane fractions was determined using the procedure of Lowry et al. (26) with bovine serum albumin (BSA) as the standard. monitored

I .25X IO'mice/mg

Sodium Dodecyl Sulfite-Pol~mcr~lumide Gel BkctrophorM~ ISDS-P.4GE) and Electrotrarzsfkr of Proteins Samples, containing - 1OOpg of bovine brain synaptosomal protein or 130-135 c(g of membrane protein isolated from the indicated rat organ were processed further by either: (a) incubating at room temperature in sample buffer (see below) minus P-mercaptoethanol [conditions analogous to those used by Lin et al. (25) for identifying the epidermal growth factor receptor] or (b) heating for 5-10 min at 100°C in sample buffer (0.063 M tris-HCl, pH 6.8, containing 2.3% SDS, 5% /?-mercaptoethanol, and 10% glycerol). When indicated, 15 wg of GTI b was added to the synaptosomal protein sample in order to determine whether neurotoxin, not preincubated with GT 1b, would adhere to protein and/or GTl b. It also allowed us to ascertain that conditions appropriate for TTx binding had been used because TTx adheres to GT 1b on protein blots (6,3 1). Samples were run on a 5% stacking, 7.5% running gel as described by Haas and Kennett (2 1) and modified by Ringler et al. (35). The electrode buffer was tris-glycine (pH 8.5) containing 24.8mM tris, 192mM glycine, and 0.1% SDS. Prestained standards (apparent molecular weights differ from those of the native protein due to the presence of the dye which can vary from one lot to the next, see Bio-Rad catalog) were included on each gel. Prior to blotting, gels were put in electrode buffer minus SDS for 10 min. Transfer to PVDF

NEUROTOXINS

BIND SYNAPTOSOMAL

919

PROTEINS

FIG. 1. Morphology of the fragments recovered in the synaptosomal fraction. Pelleted synaptosomes were fixed in formaldehyde-glutaraldehyde (23) postfixed in 1.33%osmium tetroxide and stained with many1 acetate-oxalate (30) prior to embedding. After sectioning and staining with lead hydroxide (27). samples were visualized using a RCA electron microscope.

membranes was carried out using a Semi-Phor TE70 semi-dry transfer unit (Hoefer Scientific Inst., San Francisco, CA) at a constant current ( 100 mA) for - 2.5 h in the presence of electrode buffer minus SDS. Prestained standards were used to monitor protein transfer. Transferred synaptosomal proteins were visualized by staining with Amido black. To ascertain the completeness of transfer each gel was stained with Coomassie blue R250. Essentially all of the synaptosomal proteins < 205 kDa were quantitatively transfered; transfer of >205 kDa proteins was incomplete. Blots were stained using the periodic acid-Schiff procedure ( 16) to identify glycoproteins. “‘I-wheat germ agglutinin (WGA) was employed to corroborate identification of glycoproteins (3). Binding Assay

In order to block nonspecific binding of the neurotoxin to the membrane, blots were preincubated either overnight at 4’C in 1OmM Tris-HCl buffer, pH 7.2, containing 0.1% BSA or for 1 h at room temperature in buffer containing 1% BSA. BSA was included to minimize nonspecific binding of labeled neurotoxin. BSA was included in the buffer throughout the overlay procedure and all incubations were carried out with gentle shaking. Increasing the concentration of BSA to 1% had no effect on the binding patterns obtained. Although the 10 mM Tris-HCl buffer permited optimal binding of GTl b by BTxA, effects of buffers of differing ionic strengths on the binding of BTxA to blots of synaptosomal proteins were also determined. Because prein-

cubation with GTI b enhanced binding of BTxA and TTx to synaptosomal proteins (see Results section), each neurotoxin was usually preincubated, in the buffer (containing 0.1% BSA) in which the assay was to be carried out, with the amount of GTlb found to completely inhibit BTxA or TTx binding to GTl b-coated plastic wells (20nM BTxA, BTxB or BTxE in 0.28mM GTlb and 20nM TTx in 0.84mM GTlb; 38). Samples, preincubated for 60 min, were diluted 120- or 60-fold (0.17 nM BTxA and 0.33 nM TTx, respectively) with buffer containing 0.1% BSA prior to overlaying blots. Concentrations of lZ51-BTxA as low as 2.4 pM could be used. However, the higher concentration was routinely used to shorten the time needed for autoradiography. i4C-labeled fragments of BTxA were preincubated with the same concentration of GTI b as was the intact toxin. However, it was necessary to increase the concentration of 14Clabeled fragment used to 0.2-0.3 PM due to their lower specific activity (1-5 nCi/pg). Blots were routinely overlayed for 2 h at 20°C and then rinsed 7 times, 5 min each with buffer minus BSA. Autoradiography was used to visualize synaptosomal proteins adhered to by the labeled neurotoxins. Air dried blots were overlaid with Kodak XAR-2 X-ray film and a “Wolf Plus” intensifying screen (Jersey Lab and Glove Supply, Livingston, NJ) was used to shorten exposure time; 16-18 h later, films were developed using an automatic processor. Specificity of neurotoxin adherence to synaptosomal proteins was examined as follows: (a) The ability of a lOO-fold excess of cold native neurotoxin to block adherence of the labeled neu-

920

SCHENGRLINII.

FIG. 2. Protein blots showing synaptosomal proteins (heated in the presprior to separation with SDS-PAGE) visualized with amido black in A and by autoradiography in B-F. For autoradiographs the protein blots had been overlaid with either I .7 X IO “M ‘*‘I-BTxA. B: 3.3 Y 10-rOM ‘251-labeledTTx. C-E; or “‘I-labeled cholera toxin. F. 15 pg of CiT I b was added to the synaptosomal preparation run in D. BTxA was preincubated with GTlb in IO mM Tris ence ofSDS and ij-mercaptoethanol

chloride buffer. pH 7.1. containing 0.19, BSA as was T.T‘x used for the overlay shown in E. TTx used for the overlays shown in C and D was not preincubated with CT I b. The cholera toxin was preincubated with GM I in phosphate buffered saline containing 0. I’? BS.A.

RlNGL,ER

ANI> I~ASGLIPTA

-80 kDa protein(s) (Fig. 3) suggesting that BTxA and TTx adhere to the same -80 kDa protein(s). The heavy chain of BTx.4, labeled with either “‘1 or 14C, showed a binding pattern similar to that of the intact neurotoxin (Fig. 4, A). The C-terminal portion of the heavy chain showed a preference for the -80 kDa protein(s) (Fig. 4. B) whereas the N-terminal portion adhered primarily to the - 116 kDa protein(s) (Fig. 4. C). Interestingly. the light chain also adhered to the -80 kDa protein(s) (Fig. 4. D), albeit to a lesser extent than the C-terminal fragment of the heavy chain. The band at the solvent front that is seen in most of the overlays (cut off in Fig. 4, A) is probably labeled neurotoxin adhering to ganglioside. This hypothesis is based on the ohservation that the addition of GT 1b to a synaptosomal preparation resulted in adherence of labeled TTx to a component migrating at the solvent front (Fig. 2. D). This observation is in agreement with that made by Nathan and Yavin (3 I). The effects of preincubation with GT 1b on the adherence of BTxA and TTx to synaptosomal proteins was studied. Binding patterns were compared for labeled neurotoxin. preincubated. or not preincubated with GT I b. Preincubation with ganglioside did not alter the binding pattern but did enhance binding of BTxA to the synaptosomal proteins. It also resulted in autoradiographs that had lighter backgrounds. TTx not preincubated with GT I b adhered very poorly to blots of synaptosomal proteins (Fig. 2 C). instead it appeared to adhere to GT I b when it was added to the synaptosomal preparation prior to its electrophoresis (Fig. 7 D). In contrast to the neurotoxins. ‘Wabeled I ‘rhrio c~lw/rru toxin did not bind to the -80 or -I I6 kDd proteins (Fig. 2. F), nor did a IOO-fold excess of cold cholera toxin inhibit adherence of labeled BTxA to synaptosomal proteins. Autoradiogrdphy of a blot of synaptosomal proteins overlaid with 14Clabeled bovine brain gangliosides indicated that they adhered to

rotoxin was determined; (b) Because cholera toxin is known to adhere to GM I, the effect of adding a 1OO-fold excess of Vibrio cholera toxin preincubated with a two-fold excess of GM 1 (39) was included as a negative control (expected not to affect neurotoxin binding); (c) Ability of BTxA to adhere to membrane proteins from rat liver, spleen, and kidneys was also tested. The -80 kDa and - I 16 kDa proteins from bovine brain synaptosomes were isolated using preparative gel electrophoresis, followed by electroelution of the appropriate components. Recovered proteins were analyzed using the binding assay just described for their ability to function as a ligand for the adherence of BTxA and TTx. RESULTS ‘251-labeled BTxA and TTx preincubated with GTl b were found to preferentially adhere to an -80 kDa synaptosomal protein(s) and an - 116 kDa protein(s) (Fig. 2, B & E). A small amount of adherence to other synaptosomal proteins was also noted. ‘251-labeled BTxB and BTxE exhibited the same binding patterns as BTxA. Adherence patterns similar to those presented in Fig. 2 were obtained for BTxA and TTx overlays of blots of synaptosomal proteins that were not heated or exposed to /3mercaptoethanol prior to SDS-PAGE. The same adherence pattern was obtained for TTx labeled either by the chloramine-T Preincubation of ‘*‘I-BTxA or the Bolton-Hunter procedure. with a lOO-fold excess of cold BTxA or TTx in the presence of GTI b, reduced the amount of labeled BTxA adhering to the

FIG. 3. Effect of IOO-fold excess cold BTxA or TTx on the binding of labeled BTxA to blots of synaptosomal proteins. Autoradiograph of synaptosomal proteins overlaid with 3.3 X 10-‘OM‘251-labeledBTxA preincubated with GTlb (A and C); 3.3 X 10-~‘OMi%labeled BTxA preincubated with I .2 mg CT 1b and 1OO-foldexcess of cold TTx (B); or 3.34 X 1O-‘OM‘251-labeledBTxA preincubated with I20 mg of GT I b in the presence of IOO-fold excess of cold BTxA (D). The buffer used for each was 10 mM tris-HCI (pH7.2) containing IO mM NaCl and O.I? BSA.

NEUROTOXINS

BIND

SYNAPTOSOMAL

PROTEINS

FIG. 4. Autoradiograph of blots of synaptosomal proteins overlaid with “C-labeled BTxA heavy chain, A; the C-terminal portion of the heavy chain, B; the N-terminal portion of the heavy chain, C; or the light chain of BTxA, D. Each fragment at a concentration of 0.2-0.3pM was incubated with 0.28 mM GTlb in 10 mM tris-HCI (pH7.2) containing 0.1% BSA prior to being added to a blot.

a protein that had a mobility similar to that of the protein to which cholera toxin appeared to adhere (not shown). It is possible that the cholera toxin adhered to the ganglioside associated with the protein rather than to the protein per se. In addition to adhering to bovine brain synaptosomal proteins, BTxA was found to bind to rat brain membrane proteins. Membranes isolated from rat liver, spleen, and kidney did not appear to contain comparable proteins to which BTxA could adhere (Fig. 5, B, 2, 4, & 6). The -80 and - I 16 kDa proteins from bovine brain synaptosomes were each a mixture of two closely migrating bands. We resolved and recovered these by electrophoresis and electroelution; 79 and 84 kDa proteins were resolved from the slower and faster moving zones of the -80 kDa band (Fig. 6 A & B), 109 and 118 kDa proteins were resolved from the - 116 kDa band (Fig. 6 C & D). Staining with periodic acid-Schiff reagent (not shown) and binding of ‘2sI-WGA to the blotted isolated proteins (Fig. 6 F-I) suggested that the proteins adhered to by the neurotoxins are glycosylated. Both BtxA and TTx adhered to the recovered proteins. DISCUSSION

The overlay assay system used in this work, permitted identification of two synaptosomal glycoproteins that can be adhered to by ‘251-labeled BTxA, BTxB, BTxE, and TTx. Emphasis is placed on the adherence of the neurotoxins to the -80 and - 116 kDa proteins because they are the two that were consistently most pronounced upon autoradiography of blots of synaptosomal proteins overlaid with labeled neurotoxin. The lighter bands seen in the autoradiographs may reflect nonspecific binding, specific binding to very minor synaptosomal components,

921

or binding to proteolytic fragments of the -80 and/or - 116 kDa proteins. The observation that the same proteins were adhered to by BTxA and TTx regardless of whether the synaptosomal proteins had been either heated or kept at room temperature in the presence or absence of /3-mercaptoethanol prior to SDS-PAGE, suggests that the site recognized by the neurotoxins is relatively resistant to denaturation. If the site adhered to by the neurotoxins includes the oligosaccharide portion of the glycoproteins as suggested (14,15,19,49,50), it would explain its apparent resistance to denaturation. The proteins present in brain that were adhered to by ‘251labeled BTxA and TTx were not present at comparable levels in nonneural tissues such as liver, spleen, and kidney. The finding that the -80 kDa glycoprotein(s) bound by the neurotoxins is present in synaptosomes is appropriate because nerve endings are the sites of binding and action for BTxA and TTx. Solsona et al. (45) found that ‘251-BTxA adhered to an - 145 kDa glycoprotein present in protein blots of periodate treated presynaptic membranes of the electric organ of Torpedo marmoruta. Although the mammalian brain synaptosomal proteins recognized by BTxA in this study are smaller than that of the electric organ, they are also glycosylated. The heavy chain of BTxA appeared to bind to the same proteins adhered to by the intact neurotoxin. Interestingly, it appears that the N-terminal portion of the heavy chain conveys specificity for the - 116 kDa protein(s), while the C-terminal portion adheres preferentially to the -80 kDa protein(s). In experiments designed to identify the portion of BTxA and TTx that convey neuronal specificity, Poulain et al. (34) observed that the amino terminus of the heavy chains of both BTxA and TTx had two functional domains. One appeared to be unique to each neurotoxin and was hypothesized to convey neuronal specificity; the second was present in each neurotoxin and was postulated to function in the translocation of the light chain of the neurotoxins. It is possible that the common domain is that which recognizes the - 116 kDa protein(s). The light chain also adhered to the -80 kDa protein(s), although less binding was seen (lighter band on autoradiography) then occurred using similar molar concentrations (labeled to similar specific activities) of the Cterminal portion of the heavy chain (Fig. 4 B & D). It is possible that the adherence of the light chain to the -80 kDa protein(s) is the result of hydrophobic interactions analogous to those proposed for its adherence to GT 1b (29). BTxA did not show an absolute requirement for binding to GT I b prior to adherence to the synaptosomal proteins. This is consistent with the observations made by Williams et al (50). However, preincubation with GTl b did reduce the nonspecific association of labeled BTxA to the blot. The fact that preincubation of BTxA with GTlb did not prevent its binding to synaptosomal proteins confirms earlier observations (20,50). The ability of TTx to substantially but not completely inhibit the binding of BTxA (Fig. 3, A & B) is also consistent with earlier work (50). The adherence to synaptosomal proteins by the neurotoxins appears specific based on the following observations: (a) excess native BTxA reduced adherence of ‘251-labeled BTxA to synaptosomal proteins, (b) membranes isolated from nonneuronal sources (rat liver, spleen, or kidneys) lacked the -80 kDa and - 116 kDa proteins to which BTxA adhered, (c) Vibrio cholera toxin did not bind to either the -80 or - 116 kDa synaptosomal proteins and was unable to block the binding of ‘251-labeled BTxA, and (d) excess cold tetanus toxin in the presence of GTl b was able to reduce the binding of labeled BTxA. The last point is consistent with the finding that high concentrations of TTx

927

SCHENGRIJND.

RINGLER

AND

DASGUPTA

FIG. 5. BTxA adheres to proteins from bovine synaptosomes and rat brarn hut not from liver. kidney, or spleen. Autoradiographs of blots of bovine brain synaptosomal proteins in A-l. and B-l. 3. and S; membranes isolated from rat brain in A-2: and membranes isolated from liver. kidney, and spleen in B-2, 4. and 6, respectively. All samples were overlaid with 3.34 i IO ‘OM ‘251-labeled BTxA, preincuhated with GTlh in IO mM tris-HCI buffer (pH7.2) containing 0.1% BSA. Although the - I 16 kDa band in A-l, appears to be as intense as the one at -80 kDa. it is the result of overexposure of the autoradiograph. Exposure for a shorter time indicated that the -80 kDa band was predominant. however. the hands in A-2 w’ere much lighter.

are ;ible to compete with BTx and bind to the NMJ (l&42,43). TTx initially binds to the presynaptic terminals of NMJs and is 1carried by retrograde axonal transport to the CNS. The 3heral action of TTx. observed at high doses, induces a lethal flacc :id paralysis similar to that induced by BTx (4.19). This im-

wi

plies that both neurotoxins may adhere to some of the same synaptosomal protein(s). Our finding that TTx adheres to synaptosomal prott :ins is consistent with independent observations (33,5 1) that a tr ypsinsensitive component on nerve cells may play a role in tc:tanus

of their glycoprotein nature. (Left) FIG. 6. Electrophoretic resolution of the -80 kDa and - I 16 kDa proteins and demonstration Coomassie blue stained gel showing in A, lower component of the -80 kDa protein; B, upper component of the -80 kDa protein: synaptosomal C, lower component of the - I 16 kDa protein; D, upper component of the - I 16 kDa protein; and E, unfractionated proteins. F-J show autoradiographs of these samples overlaid with - 5 fig of ‘251-labeled WGA (specific activity was 1 mCi/mg) in PBS; K indicates that the only molecular weight standard (Bio-Rad) adhered to by WGA was ovalbumin.

NEUROTOXINS

BIND SYNAPTOSOMAL

923

PROTEINS

binding. However, binding studies of TTx to blots of proteins from total cell extracts of rat cerebral neurons, PC1 2 cells (3 l), or rat brain membranes (6) indicated that TTx adhered only to GTl b. The differences between these and the present studies are probably due to (a) earlier, the TTx was not preincubated with GT 1b prior to protein binding studies, and (b) in this work synaptosomes were used, hence a higher concentration of synaptosomal proteins were present on the blot. Our observation that TTx bound poorly to synaptosomal proteins unless preincubated with GTl b (Fig. 2, C-E) supports the first point. In regard to the second point, we noted that although labeled BTxA adhered to an -80 kDa protein present in homogenates of rat brain membranes, the intensity of the band was less than that seen when a comparable amount of total synaptosomal protein was used (Fig. 5). Although BTx acts at the NMJ (1 l), acceptors for BTx have been demonstrated in rat brain (5) and BTx has been shown to inhibit the release of acetylcholine from rat brain synaptosomes ( 12). These observations suggest that brain synaptosomes are a valid substitute for NMJ preparations, as a source for searching for molecular target(s) of BTx (2).

We believe that we have shown that BTxA and TTx are able to adhere to synaptosomal proteins. Based on the intensity of the bands seen upon autoradiography of blots of synaptosomal proteins that were overlaid with labeled neurotoxin, the -80 and - 116 kDa glycoproteins (resolvable into 79 and 84 kDa bands and 109 and 118 kDa bands, respectively) appear to merit further investigation. Studies to determine (a) whether the molecular weight differences reflect differences in post-translational modification of the proteins or differences in the protein core, and (b) whether they serve as receptor/acceptor molecules or function at some stage after the initial adherence of the neurotoxin to the presynaptic terminus are currently in progress. ACKNOWLEDGEMENTS

We thank William Tepp (University of Wisconsin) for his expert technical assistance and William H. Habig for his helpful discussions (U.S. Food and Drug Administration, Bethesda, MD). This work was supported in part by Public Health Service grants AI 2372 1 awarded to C-L. S. and NS 17742 and NS 24545 to B. R. DG. A preliminary report of this work has appeared as an abstract (FASEB J. 4: 2725; 1990).

REFERENCES I. AnDerLan, B.; Habig, W. H.; Hardegree, M. C.; Chrambach, A.

2.

3.

4.

5.

6. 7.

8.

9. 10. II. 12.

13. 14.

Heterogeneity of ‘251-labeledtetanus toxin in isoelectric focusing on polyacrylamide gel and polyacrylamide gel electrophoresis. Arch. Biochem. Bioohvs. 200: 206-2 15: 1980. Ashton, A. C.: Dolly, J. 0. Characterization of the inhibitory action of botulinum neurotoxin type A on the release of several transmitters from rat cerebrocortical synaptosomes. J. Neurochem. 50: 1808 1816; 1988. Battles, J. R.; Hubbard, A. L. ‘251-wheatgerm agglutinin blotting: increased sensitivity with polyvinylpyrrolidone quenching and periodate oxidation/reductive phenylamination. Analyt. Biochem. 140: 284-292; 1984. Bizzini, B.; Toth. P.: Fedinec, A. A. Defining a region on tetanus toxin responsible for neuromuscular blockade.Toxicon 26: 309-3 18; 1988. Black, J. D.; Dolly, J. 0. Selective location of acceptors for botulinum neurotoxin in the central and peripheral nervous system. Neurosci. 23: 767-779; 1981. Critchley, D. R.; Habig, W. H.; Fishman, P. H. Reevaluation of the role of gangliosides as receptors for tetanus toxin. J. Neurochem. 47: 213-222: 1986. DasGupta, B. R.; Rasmussen, S. Purification and amino acid composition of type E botulinum neurotoxin. Toxicon 2 1: 535-545; 1983. DasGupta, B. R.; Sathyamoorthy, V. Purification and amino acid composition of type A botulinum neurotoxin.Toxicon 22: 4 15-424; 1984. DasCupta, B. R.; Woody, M. Amino acid composition of Clostridium botulinum type B neurotoxin. Toxicon 22: 3 12-3 15; 1984. Dekleva, M. L.: DasGupta, B. R.; Sathyamoorthy, V. Botulinum neurotoxin type A radiolabeled at either the light or the heavy chain. Arch. Biochem. Biophys. 274: 235-240; 1989. Dolly, J. 0.; Black, J.; Williams, R. S.; Melling, J. Acceptors for botulinum neurotoxin reside on motor nerve terminals and mediate its internalization. Nature 307: 457-460; 1984. Dolly, J. 0.; Tse, C. K.; Black, J. D.; Williams, R. S.; Wray, D.; Gwilt, M.: Hambleton. P.: Mellina. J. Botulinum neurotoxin tvoe A as a probe for studying neurotransmitter release. In Lewis, G.,ed. Biomedical aspects of botulism. New York: Academic Press; 198I: 47-64. Eichberg, J.; Whittaker, V. P.; Dawson, R. M. C. Distribution of lipids in subcellular particles of guinea-pig brain. Biochem. J. 92: 91-100; 1964. Evans, D. M.; Richardson, P. J.; Fine, A.; Mason, W. T.; Dolly, J. 0. Relationship of acceptors for botulinum neurotoxins (types A

and B) in rat CNS with the cholinergic marker, Chol-1. Neurochem. Int. 13: 25-36; 1988. 15. Evans, D. M.; Williams, R. S.; Shone, C. C.; Hambleton, P.; Melling, J.; Dolly, J. 0. Botulinum neurotoxin type B. Its purification, radioiodination and interaction with rat brain synaptosomal membranes. Eur. J. Biochem. 154: 409-416; 1986. 16. Fairbanks, G.; Steck, T. L.; Wallach, D. F. H. Electrophoretic analysis of the major polypeptides of the human erythrocyte membrane. Biochemistry 10: 2606-26 17; 197 1. 17. Habermann, E. 1z51-labeledneurotoxin from Clostridium botulinum A: Preparation, binding to synaptosomes and ascent to the spinal cord. N-S. Arch. Pharmacol. 281: 47-56; 1974. 18. Habermann, E. Botulinum A and tetanus toxin-similar actions on transmitter systems in vitro. In Lewis, Jr., A. E., ed. Biomedical aspects of botulism. New York: Academic Press; 198 1: 129- 141. 19. Habermann, E.; Dreyer, F. Clostridial neurotoxins: Handling and action at the cellular and molecular level. Current Topics Microbial. Immunol. 129: 93-179; 1986. 20. Habermann, E.; Heller, I. Direct evidence for the specific fixation of Cl. botulinum A neurotoxin to brain matter. Naunyn-Schmiedeberg’s Arch. Pharmacol. 287: 97-106; 1975. 21. Hass, J. B.; Kennett, R. H. Characterization of hybridoma immunoglobulins by sodium dodecylsulfate-polyacrylamide gel electrophoresis. In Kennett, R. H.; McKearn, T. J.; Bechtol, K. B., eds. Monoclonal antibodies. New York: Plenum Press: 1979: 407-4 I 1. 22. Holmgren, J.; Elwing, H.; Fredman, P.; Svennerholm, L. Polystyrene adsorbed ganghosides for investigation of the structure of the tetanustoxin receptor. Eur. J. Biochem. 106: 371-379; 1980. 23. Karnovsky, M. J. A formaldehyde-glutaraldehyde fixative of high osmolality for use in electron microscopy. J. Cell Biol. 27: 137A; 1965. 24. Kitamura, M.; Iwamori, M.; Nagai, Y. Interaction between clostridium botulinum neurotoxin and gangliosides. Biochim. Biophys. Acta 628: 328-335; 1980. 25. Lin, P. H.; Selinfreund, R.; Wharton, W. Western blot detection of epidermal growth factor receptor from plasmalemma of culture cells using ‘251-labeledepidermal growth factor. Anal. B&hem. 167: 128139; 1987. 26. Lowry, 0. H.; Rosebrough, N. J.; Farr, A. L.; Randall, R. J. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193: 265-275; 195 1. 27. Millonig, G. A modified procedure for lead staining ofthin sections. J. Biophys. Biochem. Cytol. 11: 736-739; 1961. 28. Montecucco. C. How do tetanus and botulinum toxins bind to neuronal membranes? Trends Biol. Sci. 11: 314-317, 1986.

924

29 Montecucco, C.; Schiavo, G.; DasGupta. 8. R. Effect ofpH on the interaction of botulinum neurotoxins A. B and E with liposomes. Biochem. J. 259: 47-53: 1989. an en bloc 30. Mumaw, V. R.; Munger, B. L. Uranyl acetate-oxalate stain as well as a fixative for lipids associated with mitochondria (abstract). Anat. Rec. 169: 383; 197 1. 31. Nathan. A.: Yavin, E. Periodate-modified ganghosides enhance surface binding of tetanus toxin to PC12 pheochromocytoma cells. J. Neurochem. 53: 88-94; 1989. 32. Nores. G. A.; Hanai. N.; Levery, S. B.: Eaton, H. L.; Salyan. M. E. K.; Hakomori. S. Synthesis and characterization of lyso-GM3 (I13NeuSAc lactosyl sphingosine). de-N-acetyl-GM3 (II’NeuNH, lactosyl cer). and related compounds. Carbohydr. Res. 179: 393410; 1988. 33. Pierce. E. J.; Davison, M. D.; Patton, R. G.; Habig, W. H.: Critchley. D. R. Characterization of tetanus toxin binding to rat brain membranes. Biochem. J. 236: 845-852; 1986. 34. Poulain. B.; Mochida, S.: Weller, U.: Hogy, B.; Habermann, E.; Wadsworth. J. D. F.: Shone. C. C.: Dollv. J. 0.: Taut. L. Heteroloaous combinations of heavy and light chains from botulinurn neurotoxin A and tetanus toxin inhibit neurotransmitter release in Aplysia. J. Biol. Chem. 266: 9580-9585; 199 I, 35. Ringler. N. J.: Selvakumar, R.; Woodward, H. D.; Simet, I. M.: Bhavanandan, V. P.; Davidson. E. A. Structure of canine tracheobronchial mucin glycoprotein. Biochem. 26: 5322-5328; 1987. 36. Sathyamoorthy, V.; DasGupta, B. R. Separation, purification, partial characterization and comparison of the heavy and light chains of botulinum neurotoxin types A. B. and E. J. Biol. Chem. 260: 1046110466: 1985. 31. Sathyamoorthy, V.; DasGupta, B. R.: Foley, J.: Niece, R. L. Botulinum neurotoxin type A: Cleavage of the heavy chain into two halves and their partial sequences. Arch. Biochem. Biophys. 266: 142-151; 1988. 38. Schengrund, C-L.; DasGupta. B. R.: Ringler, N. J. Binding of botulinum and tetanus neurotoxins to ganghoside CT I b and derivatives thereof. J. Neurochem. 57: 1024-1032: 1991. 39. Schengrund. C-L.; Ringler, N. J. Binding of vibrio cholera toxin and the heat-labile enterotoxin of Escherichia coli to GM I. derivatives

SCHENGRUND.

40. 41.

42.

43. 44. 45.

46.

47.

RINGLER

AND

DASGUPTA

of GM I. and nonlipid oiigosaccharide polyvalent ligands. J. Biol. Chem. 264: 13233-I 3237; 1989. Schengrund. C-L.: Rosenberg. A. Intracellular location and properties of bovine brain sialidase. J. Biol. Chem. 245: 6196-6200; 1970. Schiavo. G.: Ferrari. G.; Rossetto. 0.; Montecucco, C. Specific crosslinking of tetanus toxin to a protein of NGF-differentiated PC12 cells. FEBS 290: 227-230; 199 I Simpson. I_. I.. The binding fragment from tetanus toxin antagonizes the neuromuscular blocking actions of botulinum toxin. J. Pharmacol. Exp. Ther. 229: 182- 187: 1984. Simpson. L. L. Botulinum toxin and tetanus toxin recognize similar membrane determinants. Brain Res. 305: 177-l 80: 1984. Simpson. L. L. Molecular pharmacology of botulinum toxin and tetanus toxin. Ann. Rev. Pharmacol. Toxicol. 26: 427-453: 1986. Solsona, C.; Egea. G.: Blasi. J.: Casanova. C.: Marsal, J. The action of botulinum toxin on cholinergic nerve terminals isolated from the electric organ of Torpedo marmorata. Detection of a putative toxin receptor. J. Physiol. (Paris) 84: 174-179; 1990. Stoeckel. K.: Schwab. M.: Thoenen, H. Role of ganghosides m the uptake and retrograde axonal transport of cholera and tetanus toxin as compared to nerve growth factor and wheat germ agglutinin. Brain Res. 132:273-285. 1977. Svennerholm. L. Ganglioside designation. Adv. Exp. Med. Biol. 125:

I I: 1980. W. E. Gangliosides as membrane receptors for tet48. vanHeyningen. anus toxin. cholera toxin and serotonin. Nature. Lond. 249: 4 IS417: 1974. 49. Wadsworth. J. D. F.; De&, M.: Tranter, H. S.; King, H. J.; Hambleton. P.; Melling, J.: Dolly. J. 0.; Shone. C. C. Botulinum type F neurotoxin. Large-scale purification and characterization of its binding to rat cerebrocortical synaptosomes. Biochem. J. 268: I23128: 1990. P.; Melling, J. 50. Williams, R. S.; Tse, C-K.: Dolly. J. 0.; Hambleton, Radioiodination of botulinurn neurotoxin type A with retention of biological activity and its binding to brain synaptosomes. Eur. J. Biochem. 131: 437-445: 1983. 51. Yavin, E.; Nathan, A. Tetanus toxin receptors on nerve cells contain a trypsin-sensitive component. Eur. J. Biochem. 154: 403-407; 1986.