%E JOURNAL OF BIOLOGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 269, No. 39, Issue of September 30, pp, 24454-24458, 1994 Printed in U.S.A.

The Role of Transglutaminase in the Mechanism of Action of Tetanus Toxin*

(Received for publication, June 3, 1994)

Julie A. Coffield, Robert V. Considine, Janet Jeyapaul, Andrew B. Maksymowych, Ren-de Zhang, and Lance L. SimpsonS From the Departments of Medicine and Pharmacology, Jefferson Medical College, Philadelphia, Pennsylvania 19107

Tetanus toxin is a zinc-dependent metalloendopro- tease that cleaves synaptobrevin, a polypeptide found in the membranes of synaptic vesicles. This action is thought to account for toxin-induced blockade of trans- mitter release. However, Facchiano and Luini (Fac- chiano, F., and Luini, A. (1992) J. BioZ Chem. 267,13267- 13271) have proposed that tetanus toxin can stimulate transglutaminase, and Facchiano et aZ. (Facchiano, F., Benfenati, F., Valtorta, F., and Luini, A. (1993) J. BioZ Chem. 268, 4588-4591) have further proposed that the stimulated enzyme produces cross-linking of synapsin. These actions might also account for toxin-induced blockade of exocytosis. Therefore, a series of experi- ments were performed to evaluate the possibility that tetanus toxin exerts its effects via transglutaminase. The results indicated that clostridial neurotoxins were poor substrates for the cross-linking effects of transglu- taminase, and transglutaminase was a poor substrate for the proteolytic actions of tetanus toxin. In addition, at concentrations relevant to blockade of exocytosis, clostridial neurotoxins did not act on intact cells to stimulate transglutaminase, nor did they act on the iso- lated enzyme to stimulate cross-linking of putrescine and dimethylcasein. When used as competitive inhibi- tors of endogenous transglutaminase substrates, glycine methyl ester and monodansylcadaverine did not block toxin action. Furthermore, concentrations of calcium that were too low to support transglutaminase activity did not prevent toxin action. The data suggest that stimulation of transglutaminase is not the principal mechanism by which tetanus toxin blocks exocytosis in nerve cells.

Tetanus toxin is an unusually potent substance that acts inside vulnerable cells to block mediator release (1, 2) . In the recent past, two hypotheses have been advanced to explain the intracellular actions of the toxin. According to one hypothesis, tetanus toxin is a zinc-dependent metalloendopeptidase that cleaves synaptobrevin (3,4). This peptide is believed to partici- pate in the fusion process that regulates exocytosis, and thus cleavage of the substance might plausibly lead to blockade of neurotransmitter and hormone release.

According to a second hypothesis, tetanus toxin interacts with the enzyme transglutaminase (5). The toxin itself can

* This work was supported in part by NINCDS Grant NS-22153 and by United States Department of Army Contract DAMD17-90-C-0048, and National Research Service Award Fellowships l-F32-NS09472-01 and l-F32-DK08888-01. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisernent” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

ferson Medical College, 1020 Locust St., Philadelphia, PA 19107. Tel.: f To whom correspondence should be addressed: Rm. 314-JAH, Jef-

215-955-8381; F a : 215-955-2169.

serve as a substrate for transglutaminase, but more impor- tantly, the toxin can stimulate transglutaminase to act on other substrates. Thus, tetanus toxin has been shown to induce transglutaminase to cross-link synapsin, a polypeptide that has been implicated in neurotransmitter release (6). Cross- linking of this peptide might lead to immobilization of vesicles and blockade of exocytosis.

Although the synaptobrevin hypothesis and the transglu- taminase hypothesis are seemingly quite different, there is one possible area of commonality. Facchiano et al. (6) have specu- lated that tetanus toxin could activate transglutaminase by proteolytic cleavage that converts an inactive precursor to an active product. If this were true, the endoprotease activity of tetanus toxin would be directed against two substrates (uiz., synaptobrevin, transglutaminase), and in both cases proteoly- sis would contribute to blockade of exocytosis.

Tetanus toxin-induced proteolytic cleavage of synaptobrevin has now been well documented, and this phenomenon almost certainly has a role in blockade of mediator release. By con- trast, the purported role of transglutaminase has not been es- tablished. Therefore, experiments were done to address two related issues: (i) to determine whether transglutaminase plays a pathophysiological role in tetanus toxin action, and (ii) to determine whether transglutaminase plays a physiological role in the normal process of transmitter release.

MATERIALS AND METHODS Toxins and Drugs-Tetanus toxin was purchased from Calbiochem.

Botulinum neurotoxin type B in the unactivated form was kindly pro- vided by Dr. Y. Kamata (University of Osaka Prefecture). The neuro- toxin was activated by adding it to N-tosyl-phenylalanine chlorometh- ylketone-treated trypsin that was coupled to agarose beads (trypsin: toxin, 1:40 (w:w)). The mixture was incubated at 37 “C for 15 min in 0.02 M sodium phosphate buffer, pH 7.0. The reaction was terminated by centrifugation and aspiration of activated toxin. The homogeneity and molecular structure of the toxins were confirmed by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate (see below), and the biological activity of the toxins was measured on mouse phrenic nerve-hemidiaphragm preparations (see below). Guinea pig liver trans- glutaminase, glycine methyl ester, and monodansylcadaverine were purchased from Sigma.

Enzyme Assay-Transglutaminase activity was assayed as described by Facchiano and Luini (51, with two exceptions. First, enzyme and toxin were preincubated at the same concentrations at which they were assayed, rather than preincubated at a high concentration and then diluted. Second, tritiated putrescine (30 Cilmmol) was used in place of tritiated spermidine. The assay buffer was composed of 12 mM Tris-HC1, pH 7.8, 14 m~ dithiothreitol, 2 m~ MgCl,, 2 m~ CaCl,, and other ingredients as listed under “Results”. Depending on experimental pro- tocol, either dimethylcasein (20 p~), endogenous protein, or a clostridial neurotoxin was used as a substrate.

Neuromuscular Preparations-Mouse phrenic nerve-hemidiaphragm preparations were excised and suspended in physiological buffer that was bubbled with 95% 0,, 5% CO,. The physiological solution had the following composition (m): NaC1, 137; KCl, 5; CaCl,, 1.8; MgSO,, 1.0; NaHCO,, 24; Na,HPO,, 1.0; and D-glucose, 11. Gelatin (0.01%) was added as an auxiliary protein to diminish nonspecific inactivation of toxin.

24454

Tkansglutaminase and Tetanus Toxin 24455

Tissues were used to monitor stimulus-evoked muscle twitch or spon- taneous miniature end plate potentials. For experiments on evoked twitch, phrenic nerves were stimulated at 0.1 Hz, and muscle responses were monitored via a strain gauge transducer and a physiological re- corder. Toxin-induced paralysis was measured as a 90% reduction in twitch response to nerve stimulation. For experiments on end plate responses, preparations were pinned in a small Petri dish coated with sylgard and continuously perfused (3 mumin) with fresh physiological solution (34 "C) of the composition described above. Standard intracel- lular recordings were obtained using glass microelectrodes filled with 3 M KCl. Tip resistances ranged between 20 and 40 MIL Resting mem- brane potentials ranged between -60 and -80 mV.

Spontaneous miniature end plate potentials and evoked responses were monitored for a base-line period of 30-60 min before addition of toxin or drug. When toxin alone was added after the base-line period, evoked responses were monitored until onset of paralysis (see above). Spontaneous miniature end plate potentials were then recorded for an additional 60-90 min. When drug alone was added after the base-line period, spontaneous potentials were recorded for an additional 90-120 min. When toxin and drug were studied together there was a progres- sion of events, as follows: 30-min exposure to drug, recording of spon- taneous miniature end plate potentials for 30 to 60 min, addition of toxin and monitoring of responses until onset of paralysis, recording of spontaneous end plate potentials for 30-60 min.

The average number of end plates studied during each base-line period was five. The average number of end plates studied during drug or toxin exposure was nine. A minimum of three experiments was done for each paradigm.

Cell Culture-NG-108 neuroblastoma cells, which were kindly pro- vided by Dr. M. Nirenberg (National Institutes of Health), were main- tained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum, 100 p~ hypoxanthine, and 1 p~ aminopterin. Cells were differ- entiated by diminishing the serum content to 5%, then adding 1 mM N6,2'-O-dibutyryladenosine 3':5'-cyclic monophosphate. Cells were in- cubated in differentiation medium for 6 1 0 days prior to experiments.

Acetylcholine release was measured by a modification of the method of McGee et al. (7). NG-108 cells were grown in 60-mm tissue culture

cells were labeled for 24-36 h with 1.5 p~ [methyZ-14Clcholine chloride dishes and differentiated as described above. Prior to release studies,

(50 mCi/mmol) in differentiation medium containing dibutyryl cyclic adenosine monophosphate. Following the labeling period, the radioac- tive medium was removed, and cells were washed with Dulbecco's modi- fied Eagle's medium for 30 min. The wash was accomplished by remov- ing the culture medium and adding fresh medium every 5 min during the 30-min wash period. Following the washes the medium was changed to Dulbecco's modified Eagle's medium with 0.1 m~ eserine sulfate, plus 44 mM NaHCO, (control cells) or 44 mN KHCO, (depolar- ized cells). The cells were incubated in release medium for 10 min. Radioactive acetylcholine was separated from precursor choline by or- ganic extraction as described previously (7).

In Vitro 7kanslation of Rat VAMP-2 (Synaptobreviwj-Rat VAMP-2 cDNA (kindly provided by Dr. Scheller, Stanford University) inserted into Bluescript KSII' (Stratagene) was transcribed and translated us- ing an in vitro system (TNT", Promega). Synaptobrevin was synthe- sized a t 30 "C for 90 min in a 6O-pl reaction mixture containing 2 pg of DNA, 25 p1 of TNTTM rabbit reticulocyte lysate, 40 pCi of [%]methi- onine, and 1 pl of TNTTM T3 RNA polymerase. Following synthesis, the reaction mixture was portioned into 5-1.11 aliquots and stored at -20 "C.

pglpl) was diluted into 10 pl of phosphate-buffered saline supplemented Digestion of Synaptobreuin with Tetanus lbxin-Tetanus toxin 1 p l (1

with 10 m~ dithiothreitol. Toxin was reduced at 37 "C for 30 min, after which various concentrations of reduced toxin (see "Results") were added to 5 pl of synaptobrevin translation mix and incubated at 37 "C for 30 min. All digestion reactions were treated with 50 pg/ml RNase at 37 "C for an additional 15 min to remove tRNA background. Reaction products (7 PI) were mixed with 20 pl of SDS-sample buffer and resolved on 15% SDS-polyacrylamide gel electrophoresis gels. Gels were fixed, stained with Coomassie, and destained, with the final change of destain containing 10% glycerol. Gels were dried and exposed t o x-ray film (Kodak X-Omat A R ) overnight.

aminase (1 x M; Sigma) was incubated with tetanus toxin (1 x IO" Incubation of Tkansglutaminase with Tetanus Toxin-Transglut-

M ; Calbiochem) in assay buffer containing 12 nm "is-HC1, pH 7.8, 14 mM dithiothreitol, 2 mM MgCl,, and 2 mM NaCl at 0" or 37 "C for 1 h. Tetanus toxin and transglutaminase partially eclipse one another dur- ing SDS-polyacrylamide gel electrophoresis; therefore, the two mol- ecules were separated by two ultrafiltration steps (molecular weight 100,000 cutoff; molecular weight 30,000 cutoE,Amicon centricon tubes).

The molecular weight 100,000 flow-through, which represented toxin- free transglutaminase, was concentrated on the molecular weight 30,000 membrane. Transglutaminase fractions were then mixed with sample buffer and analyzed by SDS-polyacrylamide gel electrophoresis in 7.5% gels (and see below).

Polyacrylamide Gel Electrophoresis-The cross-linking of endoge- nous proteins in NG-108 cells was monitored by doing polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate as de- scribed by Laemmli (8). The underlying concept was that transglutami- nase-induced cross-linking would produce high molecular weight pro- teins that should be retained in the stacking gel (9).

NG-108 cells were homogenized in assay buffer (12.5 m~ Tris, pH 7.4, 20 m~ dithiothreitol) at 37 "C, then incubated for 180 min in homoge- nization buffer that contained different concentrations of calcium rang- ing from 0 to 10 mM. Cell lysates (50 pg of proteidane) were subjected to electrophoresis using a 4% stacking gel and a 10% separating gel. Gels were stained with Coomassie Brilliant Blue and destained with 10% acetic acid and 50% methanol.

Data Analysis-The data in the figures and tables are presented as the mean * S.E. For experiments on tissue preparations, each data point reflects an n of 3 or more. For enzyme assays, each experiment was done at least twice, and within each experiment samples were done in triplicate.

RESULTS

Sequence Homology and Substrate Activity-There are two domains in tetanus toxin that reportedly have sequence homol- ogy with transglutaminase substrates and thus are presum- ably sites of transglutaminase-induced cross-linking (5). It is interesting that these two domains do not have significant ho- mology with one another (Fig. 1).

One of the domains that reportedly has substrate homology is found in t h e light chain and the other is in the heavy chain of the toxin. I t should be noted that only the light chain is essential for the intracellular actions that culminate in block- ade of exocytosis (10, 11).

The primary sequence of the light chain of tetanus toxin was aligned with the primary sequences of the light chains of botu- linum neurotoxin types A to E. This was done to determine whether there was homology in the putative transglutaminase substrate domain (Fig. 1). Of the five botulinum neurotoxins, only type B possessed both sequence homology with tetanus toxin and an essential glutamine. Serotype A possessed weak homology and a misaligned glutamine. The other three sero- types possessed variable homology and no reactive glutamine.

The light chains of tetanus toxin and botulinum neurotoxin types A to E were searched to determine whether there were other domains in which the group shared sequence homology and a reactive glutamine. The results, which are shown in Fig. 1, indicate that there was a region in the carboxyl terminus of the molecules that possessed an aligned glutamine. However, recent evidence indicates that this portion of the toxin molecule is no t essential for blockade of exocytosis (12).

The alignment data indicate that only tetanus toxin and botulinum neurotoxin type B have domains that could account for substrate activity. Therefore, these two toxins were assayed as subs t ra tes for transglutaminase at concentrations that a r e relevant to blockade of transmitter release at the neuromuscu- lar junction and in NG-108 cells M to M; see below). The results of these experiments were negative. Even when tested at a high concentration (1 x M), the toxins possessed little if any abili ty to serve as substrates for transglutaminase- induced cross-linking. This indicates that: ( a ) tetanus toxin and botulinum neurotoxin type B are not important substrates for t ransglutaminase, and ( h ) to the extent that the toxins are substrates, this is not an important par t of the process of block- ing transmitter release.

Stimulation of l+ansglutaminase Actiuity-Experiments were done on intact cells and on isolated enzyme preparations

24456 ll-ansglutaminase and Tetanus Toxin

Tetanus Toxin L i t Chain

Tetanus Toxin Heavy Chain

F

L

G

E

S

F

I

D

A

K

F

N

C

I

P

L

Botulinum Toxin A L i t Chain Y G S T Q Y I R F S P Botulinum Toxin B L i t Chaii F G G I M Q M K F C P Botulinum Toxin C Light Chain F G A L S I I S I S P

Botulinum Toxin D Light Chaii F G T L S I L K V A P

Botulinum Toxin E Light Chain F G S I A I V T F S P

0

Botulinum Toxin A Light Chain

Botulinum Toxin B L i t Chain

Botulinum Toxin C Light Chain

Botulinum Toxin D Light Chain

Botulinum Toxin E Light Chain

Tetanus Toxin Light Chain

N G

R G M G

S G

R G

K G

N T

N K

N L

N I

N A

N M

FIG. 1. Alignment of the primary structures of tetanus toxin and botulinum neurotoxin. The upper part of the figure aligns the two domains of tetanus toxin that reportedly have sequence homology with known transglutaminase substrates (5). It is interesting that the only true homology between the two is the reactive glutamine (Q) that is characteristic of transglutaminase substrates. The middle part of the figure aligns the purported substrate domain of the light chain of tetanus toxin with the corresponding regions of the light chains of botulinum neurotoxin types A to E. As the figure illustrates, only the light chain of botulinum neurotoxin type B has significant sequence homology with the light chain of tetanus toxin. The lower part of the figure shows the only region of the light chains of the six toxins in which there is true alignment of glutamine residues. This region is in the carboxyl terminus of the light chains, and it is a portion of the molecule that is not required for blockade of exocytosis (12). The primary structures for the various toxins were obtained as follows: tetanus toxin (17, 18) and botulinum neurotoxin type A(19,20), type B (21), type C (22), type D (23), and type E (24, 25).

to determine whether tetanus toxin or botulinum neurotoxin r"---7 type B would stimulate transglutaminase at meaningful con- centrations. In the initial experiment various concentrations of tetanus toxin were incubated with NG-108 cells for 180 min, after which the extent of toxin-induced blockade of acetylcho- line release was measured. The results (Fig. 2) indicated that concentrations in the range of 10"' to 10"O M produced partial to complete blockade of transmitter release.

In the next experiment, NG-108 cells were incubated for 180 min with 1 x lo-' M tetanus toxin. Cells were then ruptured by sonication and exposed to varying concentrations of exogenous calcium (35 "C; 180 rnin). Transglutaminase activity was as- sayed by quantifying the amount of cross-linked protein in the stacking gel, as described by Barsigian et al. (9). The results indicated that, even at a concentration that totally blocks exo- cytosis, tetanus toxin did not alter the pattern or amount of cross-linked protein (Fig. 3). It did not induce the appearance of cross-linked protein at low calcium concentrations (e.g. 100 1" calcium) nor did it increase the amount at high calcium con- centrations (e.g. 10 m ~ ) .

In the final experiment, tetanus toxin and botulinum neuro- toxin type B were examined for their ability to stimulate transglutaminase-mediated incorporation of tritiated putres- cine into dimethylcasein. Toxins and transglutaminase were tested at various ratios (0.1 to 1.0; 1.0 to 1.0; 1.0 to 0.11, at various concentrations (maximum, 1 x lo-' M), and for various lengths of time (30, 60, and 120 rnin). There was no paradigm in which the toxins produced a statistically significant in- crease in the amount of tritiated putrescine incorporated into dimethylcasein.

Clostridial nxins, Bansglutaminase, and Neuromuscular Bansmission-Transglutaminase is a calcium-dependent en- zyme. As gauged by incorporation of putrescine into dimethyl- casein, the EC,, for calcium was in the range of 10-100 p~ (results not illustrated). Magnesium and strontium possessed no more than 10% of the activity of calcium in supporting transglutaminase activity.

The level of cytosolic calcium in quiescent nerves is in the

3 Y 0 > w

Toxin [MI FIG. 2. Blockade of acetylcholine release from NG-108 cells.

Various concentrations of tetanus toxin were incubated with cells that had been preloaded with [rnethyZ-'4C]choline chloride. After 180 min,

acetylcholine. Complete blockade of exocytosis was obtained with 180 cells were depolarized and the medium was assayed for radioactive

rnin exposure to 1o"O M toxin.

range of 100-300 nM (13), which is approximately 2 orders of magnitude below the EC,, for calcium supported transglutami- nase activity. Therefore, the actions of tetanus toxin and botu- linum neurotoxin type B were studied on quiescent nerves. To further ensure that cytosolic calcium levels did not rise, the experiments were repeated with quiescent nerves suspended in medium in which calcium was replaced with equimolar concen- trations of magnesium or strontium.

Interestingly, both toxins blocked transmission when added to tissues under conditions that would not be expected to sup- port transglutaminase activity ( n = 3 or more per group). When tetanus toxin (1 x M) was added to unstimulated phrenic

Dansglutaminase and Tetanus Toxin 24457

Stacking { gel

1 2 3 4 5 6 7 8

+ Cross-linked proteins

+ Heavy chain

* Light chain (-100KD)

(- 50KD)

FIG. 3. Effect of tetanus toxin on transglutaminase-mediated cross-linking of endogenous proteins in NG-108 cells. Two exper- imental paradigms were used, as follows: (i) cells were homogenized and assayed for enzyme-induced cross-linking of proteins in the absence of toxin (lanes 1 4 ) , or (ii) tetanus toxin was preincubated with cells (1 x

M; 180 min), after which cells were washed, homogenized, and assayed for enzyme-induced cross-linking (lanes 5-8). In all cases, ly- sates were incubated for 180 min before being submitted to polyacryl- amide gel electrophoresis. Assays were done in the presence of varying concentrations of calcium, as follows: lanes 1 and 5, no calcium; lanes 2 and 6 , 100 PM calcium; lanes 3 and 7, 1 mM calcium; and lanes 4 and 8, 10 mM calcium. Note that pretreatment of cells with toxin did not increase the amount of cross-linked protein in the stacking gel, nor did it decrease the amount of free protein in the resolving gel. The location and approximate molecular weights of the heavy and light chains of tetanus toxin are indicated. The location of synapsin I is between the heavy and light chains of the toxin (doublet; S86KD and 80KD).

nerve-hemidiaphragm preparations in physiological medium, the average paralysis time was 172 2 9 min. Similarly, when botulinum neurotoxin type B (1 x 10”’ M) was added to prepa- rations in physiological medium, the average paralysis time was 191 2 12 min. The results for both toxins were not signifi- cantly different when calcium was replaced by magnesium or strontium.

As a further test of the involvement of transglutaminase in clostridial toxin action, experiments were done in the presence of glycine methyl ester and monodansylcadaverine. These agents can serve as false substrates for the enzyme, and thus they can be used to inhibit the cross-linking effects of endoge- nous substrates by transglutaminase (14-16).

Glycine methyl ester and monodansylcadaverine were as- sayed for their ability to inhibit transglutaminase-mediated incorporation of tritiated putrescine into dimethylcasein. The respective IC,, values were: glycine methyl ester, 2 x lo4 M; monodansylcadaverine, 1 x M. The concentrations of the drugs had to be incremented 2-4-fold to obtain comparable effects on intact NG-108 cells.

Glycine methyl ester and monodansylcadaverine were tested for their abilities to alter stimulus-evoked muscle twitch and spontaneous miniature end plate potentials (group n = 3 or more). At concentrations equal to or greater than the IC,, val- ues (glycine methyl ester, 3 x M; monodansylcadaverine, 1 x M), the drugs produced no observable effects on muscle twitch over a period of 120 min. The drugs similarly failed to produce an effect on the frequency of spontaneous miniature end plate potentials. For example, the rate of spontaneous po- tentials during a base-line period of 30-60 min was 135 2 22/min. When tissues were exposed to glycine methyl ester (3 x

M; 30 min), the frequency was 111 2 12/min. In a similar experiment with monodansylcadaverine (1 x M) the rate of spontaneous miniature end plate potentials during a base-line period was 99 2 14/min, and the rate during exposure to the drug was 85 2 6/min. These results show that transglutami-

TABIS I Effects of drugs on neuromuscular blockade

incubated in physiological solution a t 35 “C. Phrenic nerves were stimu- Mouse phrenic nerve-hemidiaphragm preparations were excised and

lated (0.1 Hz), and muscle twitch was recorded. Drugs were added to tissue baths 30 min before addition of tetanus toxin (1 x lo-“ M) or botulinum neurotoxin type B (1 x lo-” M). Each data point represents the mean 2 S.E. of four or more preparations.

Toxin Drug pretreatment Paralysis time

Tetanus toxin 131 * 12 Tetanus toxin Glycine methyl ester (3 x M ) 124 11 Tetanus toxin Glycine methyl ester (3 x M) 129 2 13 Tetanus toxin Monodansylcadaverine (1 x M) 126 * 8 Botulinum toxin 117 * 7 Botulinum toxin Glycine methyl ester (3 x M) 125 2 11 Botulinum toxin Glycine methyl ester (3 x M ) 120 * 14 Botulinum toxin Monodansylcadaverine (1 x M ) 119 * 13

TABLE I1 Effects of false substrates on neuromuscular transmission

animals and handled as described in Table I, except that miniature Mouse phrenic nerve-hemidiaphragm preparations were excised from

endplate responses rather than muscle twitch responses were recorded.

b Drug pretreatmenth,c Spontaneous miniature end plate potentials

min” - 94 * 11 + 11*2 - 135 -c 22 - Glycine methylester 111 2 12 + Glycine methylester 10 * 2 - 99 * 14 - Monodansylcadaverine 85 * 6 + Monodansylcadaverine 1 6 k 1

Tetanus toxin was used a t a concentration of 1 x 10” M. * The protocol for experiments involving drug pretreatment and sub-

Glycine methyl ester was used a t 3 x lo-’$ M, and monodansylcadav- sequent addition of toxin is given under “Materials and Methods.”

erine was used a t 1 x M.

nase-induced cross-linking of synaptic vesicle proteins does not play an important role in governing the normal process of neu- romuscular transmission.

Experiments were done to determine whether transglutami- nase inhibitors would alter clostridial neurotoxin-induced blockade of exocytosis. Both mechanical responses and electro- physiological responses were monitored. In studies on stimu- lus-evoked muscle twitch, neither monodansylcadaverine nor glycine methyl ester delayed the onset of toxin-induced neuro- muscular blockade (Table I). In studies on electrophysiological responses, the drugs similarly failed to protect tissues against toxin-induced effects (Table 11). These results are difficult to reconcile with the hypothesis that transglutaminase mediates the blocking effects of tetanus toxin.

Sequence Homology and Proteolytic Activity-Tetanus toxin has a highly selective proteolytic action; it cleaves the Gln7‘j- PheI7 bond in synaptobrevin 2 (3). The specificity of this reac- tion is evident in the facts that: ( a ) the toxin does not cleave the Gln-Lys or Gln-Ala bonds in synaptobrevin 1 or synaptobrevin 2, and ( b ) it does not cleave the Val-Phe bond in synapto- brevin 1.

Numerous forms of transglutaminase have been sequenced in several laboratories, and many of these molecules possess a single Gln-Phe doublet. Therefore, an effort was made to de- duce the importance of the doublet by attempting to produce proteolytic cleavage with tetanus toxin. Transglutaminase and toxin were incubated under conditions similar to those de- scribed by Facchiano and Luini (5), after which they were sub- mitted to polyacrylamide gel electrophoresis. The Gln-Phe dou-

24458 5'Yansglutaminase and Tetanus Toxin

A 6 kDa kDa

66- 20.1 - T I FIG. 4. Effect of tetanus toxin on puta t ive substrates. Transglu-

taminase was incubated with, and then separated from, tetanus toxin as described in the text. The isolated enzyme ( -3 pg of protein) was applied to a 7.5% gel and stained with Coomassie (A ). La?7e I , 0 "C for 1 h (control); lane 2,37 "C for 1 h. ""S-Synaptobrevin was also incubated with tetanus toxin as described in the text (37 "C for 30 min). The mixtures were applied to a 15% gel and then submitted to autoradiog- raphy (panel B). Lane 1, ""S-synaptobrevin; Lane 2, "'S-synaptobrevin plus toxin ( 1 x 10" at); Lane 3, "S-synaptobrevin plus toxin ( 1 x IO" M); Lane 4, '"S-synaptobrevin plus toxin (1 x IO-" M). Note that the toxin did not cleave either of the bands associated with transglutaminase, but i t produced concentration-dependent cleavage of synaptobrevin.

blet typically exists one-quarter to one-third of the distance from the N terminus to the C terminus, and thus proteolysis at this site should be easily detected. As a positive control, tetanus toxin was incubated with synaptobrevin to demonstrate cleav- age of this molecule.

The results, which are shown in Fig. 4, confirm the ability of tetanus toxin to cleave synaptobrevin. However, the toxin pro- duced no obvious cleavage of transglutaminase, even when tested a t lo-' M.

DISCUSSION

Tetanus toxin is a zinc-binding protein that possesses the properties of a metalloendoprotease. The toxin cleaves a spe- cific peptide bond in synaptobrevin 2, a vesicle-associated pro- tein thought to play a role in exocytosis. This action almost certainly contributes to toxin-induced blockade of exocytosis (3, 4).

An additional action for tetanus toxin has been proposed by Facchiano, Luini, and their colleagues, who reported that tet- anus toxin stimulates the cross-linking enzyme transglutami- nase (5). Stimulation of the enzyme may be due to proteolytic cleavage that converts an inactive precursor to an active prod- uct (6). Therefore, a series of experiments were done to evaluate the possibility that tetanus toxin, or the structurally and func- tionally similar botulinum neurotoxin type B, exert their ef- fects via transglutaminase.

Sequence analysis data and enzyme-substrate experiments have been interpreted to mean that tetanus toxin is a substrate for transglutaminase (5). However, closer analysis of the se- quence data reveals that evidence for substrate homology is not compelling (Fig. 1). In addition, enzyme-substrate experiments a t meaningful concentrations of toxin demonstrated that the latter is not an important substrate.

Tetanus toxin reportedly stimulates transglutaminase to cross-link endogenous proteins, and this stimulation may be due to proteolytic activation of the cross-linking enzyme (6). I t is noteworthy that toxin-induced cross-linking was not shown on intact cells, and toxin-induced cleavage of transglutaminase was not reported (6).

Experiments in the present study show that tetanus toxin does not produce obvious stimulation of the enzyme nor does it produce obvious proteolysis of the enzyme in an isolated assay system. Furthermore, the toxin does not produce detectable stimulation of cross-linking in intact cells. It was also found that false substrates did not block toxin action on nerve end- ings, and ambient calcium levels below those needed to support enzyme activity did not prevent paralysis.

In the aggregate, this evidence weighs against the idea that toxin-induced stimulation of transglutaminase is the principal reason for toxin-induced blockade of neurotransmitter release. Nevertheless, there are two cautionary notes that should be added. First, the data do not rule out the possibility that high concentrations of, or lengthy exposure to, tetanus toxin could alter transglutaminase activity. Second, the data do not clarify the mechanism of toxin action on cells in which synaptobrevin does not play a role in exocytosis.

Acknowledgments-We are grateful to Dr. Jose Martinez and Grace Gherovici for assistance and helpful suggestions on the transglutami- nase assay.

1.

2.

3.

4.

5. 6.

7.

8. 9.

10. 11.

12.

13. 14.

15. 16.

17.

18 19

20

21

REFERENCES Habermann, E., and Dreyer, F. (1986) Cum Top. Microhiol. Zmmunol. 129,

Simpson, L. L. (1989) Botulinum Neurotoxin and Tetanus Toxin, Academic 93-179

Schiavo, G., Benfenati, F., Poulain, B., Rossetto, 0.. de Laureto. P. P., Das- Press, San Diego, CA

Link, E., Edelmann, L., Chou, J. H., Binz, T.,Yamasaki, S., Eisel, U., Baumert, Gupta, B. R., and Montecucco, C. (1992) Nature 359,832-835

M., Siidhof, T. C., Niemann, H., and Jahn. R. (1992)Riochem. Riophys. Res. Commun. 189, 1017-1023

Facchiano. F., and Luini. A. (1992) J. Biol. Chem. 267, 13267-13271 Facchiano, F., Benfenati. F., Valtorta. F.. and Luini, A. (1993) J. Biol. Chem.

McGee, R., Simpson, P., Christian, C., Mata, M., Nelson, P., and Nirenberg, M.

Laemmli, U. K. (1970) Nature 227,680485 Barsigian, C., Fellin, F. M., Jain, A,, and Martinez, J. (1988) J . Bid. Chem.

Ahnert-Hilger, G., Weller, U., Dauzenroth, M. E., Habermann, E., and Gratzl, Bittner, M.A., Habig, W. H., and Holz, R. W. (1989) J. Neurochem. 53,966-968

Kurazono, H., Mochida, S., Binz, T., Eisel, U., Quanz, M., Grebenstein, 0.. M. (1989) FEBS I A t . 242, 245-248

Wernars, K., Poulain, B., Tauc, L., and Niemann, H. (1992) J . Bid. Chem. 267,14721-14729

268,4588-4591

(1978) Proc. Natl. Acod. Sci. U. S. A. 75, 1314-1318

263, 14015-14022

Blaustein, M. P. (1988) Handh. Exp. Pharmacol. 83, 275-304 Pastuszko, A,, Wilson, D. F., and Erecinska, M. (1986) J. Neurochem. 46,

Bungay, P. J., Potter, J. M., and Griffin, M. (1984) Biochem. J. 219,819-827 Senner, A,, Dunlop, M. E., Gomis, R., Mathias, P. C. F., Melaisse-Lagae, F., and

Malaise, W. J. (1985) Endocrinology 117, 237-242 Eisel. U., Jarausch. W., Goretzki, K., Henschen, A,, Engels. J., Weller, U.,

Hudel, M., Habermann, E., and Niemann, H. (1986) EMRO J. 5,249.5-2502 Fairweather, N. F., and Lyness, V. A. (1986) Nucleic Acids Res. 14,7809-7812 Binz, T., Kurazono, H., Wille, M., Frevert, J., Wernars, K., and Niemann, H.

Thompson, D. E., Brehm, J. K., Oultram, J. D., Swinfield, T. J., Shone, C. C., (1990) J. Riol. Chem. 265,9153-9168

Atkinson, T., Melling, J., and Minton, N. P. (1990) Eur: J. Riochrm. 189, 73-81

Whelm. S. M.. Elmore, M. J., Bodsworth, N. J., Brehm, J. K., and Atkinson, T.

499-508

(1992) Appl. Enuiron. Microhiol. 58, 2345-2354 22. Hauser, D., Eklund, M. W., Kurazono, H., Binz, T., Niemann, H., Gill, D. M.,

Boquet, P., and Popoff, M. R. (1990) Nucleic Acids Res. 18, 4924 23. Binz, T., Kurazono, H., Popoff, M. R., Eklund, M. W.. Sakaguchi, G., Kozaki, S.,

Acids Res. 18, 5556 Krieglstein, K., Henschen, A., Gill, D. M., and Niemann, H. (1990) Nucleic

24. Poulet, S., Hauser, D., Quanz, M., Niemann, H., and Popoff, M. R. (1992)

25. Whelm, S. M., Elmore, M. J., Bodsworth, N. J., Atkinson, T., and Minton, N. Biochem. Biophys. Res. Commun. 183, 107-113

P. (1992) Eur: J . Riochem. 204,657-667