Jun 3, 1994 - tetanus toxin exerts its effects via transglutaminase. The results indicated that clostridial neurotoxins were poor substrates for the cross-linking ...
%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 Mechanismof Action of Tetanus Toxin* (Received for publication, June 3, 1994)
Julie A. Coffield, RobertV. Considine, Janet Jeyapaul, AndrewB. Maksymowych, Ren-de Zhang, and Lance L. SimpsonS From the Departments of Medicine a n d Pharmacology, Jefferson Medical College, Philadelphia, Pennsylvania 19107
Tetanustoxin is azinc-dependentmetalloendoproserve as a substrate for transglutaminase, but more importease that cleaves synaptobrevin, a polypeptide found inthe toxin can stimulate transglutaminase to act on other tantly, themembranesofsynaptic vesicles. Thisaction is substrates.Thus,tetanustoxin has beenshowntoinduce of trans- transglutaminase t o cross-link synapsin, a polypeptide that thought to account for toxin-induced blockade mitterrelease. However,FacchianoandLuini(Fachas been implicated in neurotransmitter release (6). Crosschiano, F., and Luini,A. (1992)J. BioZ Chem. 267,13267- linking of this peptidemight lead to immobilizationof vesicles 13271) have proposed that tetanus toxin can stimulate a n d blockade of exocytosis. transglutaminase, and Facchiano et aZ. (Facchiano, F., Although the synaptobrevin hypothesis and the transgluA. (1993) J. BioZ taminase hypothesis are seemingly quite different, there Benfenati,F.,Valtorta,F.,andLuini, is one Chem. 268, 4588-4591) have further proposed that the possible area of commonality. Facchiano et al. (6) have specustimulated enzyme produces cross-linking of synapsin. lated that tetanus toxin could activate transglutaminase by Theseactionsmightalsoaccountfortoxin-induced proteolytic cleavage that converts an inactive precursor to an blockade of exocytosis. Therefore, a series of experi- active product. If this were true, the endoprotease activity of ments were performed to evaluate the possibility that tetanus toxin would be directed against two substrates (uiz., tetanustoxinexerts its effectsviatransglutaminase. synaptobrevin, transglutaminase), and in both cases proteolyThe results indicated that clostridial neurotoxins weresis would contribute to blockade of exocytosis. poor substrates for the cross-linking effects of transglu-Tetanus toxin-induced proteolytic cleavageof synaptobrevin taminase, and transglutaminase was a poor substrate has now been well documented, and this phenomenon almost for the proteolytic actions of tetanus toxin. In addition,certainly has a role in blockade of mediator release. By conat concentrations relevant to blockade of exocytosis, trast, the purported role of transglutaminase has not been esclostridialneurotoxinsdidnotactonintactcellsto tablished. Therefore, experiments were done to address two stimulate transglutaminase, nor did they act on the isorelated issues: (i) todeterminewhethertransglutaminase lated enzyme to stimulate cross-linking of putrescine plays a pathophysiological role in tetanus toxin action, and(ii) and dimethylcasein. When used as competitive inhibi- to determine whether transglutaminase plays a physiological tors of endogenous transglutaminase substrates, glycine role in the normal processof transmitter release. methyl ester and monodansylcadaverine did not block MATERIALSANDMETHODS toxin action. Furthermore, concentrations of calcium that were too low to support transglutaminase activity Toxins and Drugs-Tetanus toxin was purchased from Calbiochem. didnotpreventtoxinaction.Thedatasuggestthat Botulinum neurotoxin type B in the unactivated form was kindly provided by Dr. Y. Kamata (University of Osaka Prefecture). The neurostimulation of transglutaminase is nottheprincipal toxin was activated by adding it to N-tosyl-phenylalanine chloromethmechanism by which tetanus toxin blocks exocytosis in ylketone-treated trypsin that was coupled to agarose beads (trypsin: nerve cells. 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 Tetanus toxin is an unusually potent substance that acts molecular structure of the toxins were confirmedby polyacrylamide gel inside vulnerable cells to block mediator release (1, 2 ) . In t h e electrophoresis in the presence of sodium dodecyl sulfate (see below), recent past, two hypotheses have been advanced to explain and the the biological activity of the toxins was measured on mousephrenic intracellular actions of the toxin. According to one hypothesis, nerve-hemidiaphragm preparations (see below).Guinea pig liver transtetanus toxin is a zinc-dependent metalloendopeptidase that glutaminase, glycine methyl ester, and monodansylcadaverine were purchased from Sigma. cleaves synaptobrevin(3,4).This peptide is believed to partici- Enzyme Assay-Transglutaminase activity was assayed as described pate in the fusion process that regulates exocytosis, a n d thus by Facchiano and Luini (51, with two exceptions. First, enzyme and cleavage of the substance might plausibly lead to blockadeof toxin were preincubated at thesame concentrations at which they were assayed, rather than preincubated at a high concentration and then neurotransmitter and hormone release. diluted. Second, tritiated putrescine (30 Cilmmol) was used in place of According t o a second hypothesis, tetanus toxin interacts tritiated spermidine. The assay buffer wascomposed of 12 mM Tris-HC1, with the enzyme transglutaminase (5). The toxin itself can pH7.8, 14 m~ dithiothreitol, 2 m~ MgCl,, 2 m~ CaCl,, and other as listed under “Results”.Depending on experimental pro* This work was supported in part by NINCDS Grant NS-22153 and ingredients tocol, either dimethylcasein (20 p ~ )endogenous , protein, or a clostridial by United States Department of Army Contract DAMD17-90-C-0048, and National Research Service Award Fellowships l-F32-NS09472-01 neurotoxin was used as a substrate. Neuromuscular Preparations-Mouse phrenic nerve-hemidiaphragm 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 preparations were excised and suspended in physiological buffer that therefore be hereby marked “aduertisernent”in accordance with 18 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; U.S.C. Section 1734 solely to indicate this fact. NaHCO,,24;Na,HPO,, 1.0; and D-glucose, 11. Gelatin (0.01%) was f To whom correspondence should be addressed: Rm. 314-JAH, Jefferson Medical College, 1020 Locust St., Philadelphia, PA 19107. Tel.: added as an auxiliary protein to diminish nonspecific inactivation of 215-955-8381;F a : 215-955-2169. toxin.
Tetanus Tkansglutaminase and
Tissues were used tomonitor stimulus-evoked muscle twitch or spon- The molecular weight 100,000 flow-through, which represented toxinFor experiments on evoked free transglutaminase, was concentrated on the molecular weight taneous miniature end plate potentials. twitch, phrenic nerves were stimulated at 0.1Hz, and muscle responses 30,000 membrane. Transglutaminase fractions were then mixed with sample buffer and analyzedby SDS-polyacrylamide gel electrophoresis were monitored via a strain gauge transducer and a physiological rein 7.5% gels (and seebelow). corder. Toxin-induced paralysis was measured as a 90% reduction in Polyacrylamide Gel Electrophoresis-The cross-linking of endogetwitch response to nerve stimulation. For experiments on end plate nous proteins in NG-108 cells was monitored by doing polyacrylamide responses, preparations were pinned in a small Petri dish coated with sylgard and continuously perfused(3 mumin) with fresh physiological gel electrophoresis in the presence of sodium dodecyl sulfate as desolution (34"C) of the composition described above. Standard intracel- scribed by Laemmli (8).The underlyingconcept was that transglutamilular recordings were obtained using glassmicroelectrodes filled with 3 nase-induced cross-linking would produce high molecular weight proM KCl. Tip resistances ranged between 20 and 40 MIL Resting memteins that should be retained in the stackinggel (9). brane potentials ranged between -60 and -80 mV. NG-108 cells were homogenized in assaybuffer (12.5m~ Tris, pH 7.4, Spontaneous miniature end plate potentials and evoked responses 20 m~ dithiothreitol) at 37 "C,then incubated for 180 min in homogewere monitored for a base-line period of 30-60 min before addition of nization bufferthat contained different concentrations of calcium rangtoxin or drug. When toxin alone was added after the base-lineperiod, ing from 0 to 10 mM. Cell lysates (50 pg of proteidane) were subjected evoked responses were monitored until onset of paralysis (see above). to electrophoresis using a 4% stacking gel and a 10% separating gel. Spontaneous miniature end plate potentials were then recorded for a n Gels were stained with Coomassie Brilliant Blue and destained with additional 60-90 min. When drug alone was added after the base-line 10%acetic acid and 50% methanol. period, spontaneous potentials wererecorded for a n additional 90-120 Data Analysis-The data in thefigures and tables are presented as min. When toxin and drug were studied together there was progresa the mean * S.E. For experiments on tissue preparations, each data of spon- point reflects an n of 3 or more. For enzyme assays, each experiment sion of events, as follows: 30-min exposure to drug, recording taneous miniature end plate potentials for 30 to 60 min, addition of was done at least twice, and within each experiment samples were done toxin and monitoring of responses until onsetof paralysis, recordingof in triplicate. spontaneous end plate potentials for 30-60 min. The average number of end plates studied during each base-line RESULTS period was five. The average numberof end plates studied during drug or toxin exposure was nine. A minimum of three experiments wasdone are two Sequence Homology and Substrate Activity-There for each paradigm. domains in tetanus toxin that reportedly have sequence homolCell Culture-NG-108 neuroblastoma cells, which were kindly provided by Dr. M. Nirenberg (National Institutes of Health), were main- ogy with transglutaminase substrates and thus are presumably sites of transglutaminase-induced cross-linking (5). It is tained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum, 100 p~ hypoxanthine, and 1 p~ aminopterin. Cells werediffer- interesting that these two domains do not have significant hoentiated by diminishing the serum content to 5%, then adding 1 mM mology with one another (Fig. 1). N6,2'-O-dibutyryladenosine3':5'-cyclic monophosphate. Cells were inOne of the domains that reportedly has substrate homology cubated in differentiation medium for 6 1 0 days prior to experiments. is found in t h e light chain and the other is in the heavy chain Acetylcholine release was measuredby a modification of the method of the toxin. I t should be noted that only the light chain is of McGee et al. (7). NG-108 cells were grown in 60-mm tissue culture dishes and differentiated as described above. Prior to release studies, essential for the intracellular actionsthat culminate in blockcells were labeled for 24-36 h with 1.5 p~ [methyZ-14Clcholinechloride ade of exocytosis (10, 11). (50 mCi/mmol) in differentiation medium containing dibutyryl cyclic The primary sequenceof the light chain of tetanus toxin was adenosine monophosphate. Following the labeling period, the radioactive mediumwas removed, and cells were washed withDulbecco's modi- aligned with the primary sequencesof the light chains of botufied Eagle's medium for 30 min. The wash wasaccomplished by remov- linum neurotoxin types A t o E. This was done to determine ing the culture medium and adding fresh medium every 5 min during whether there was homology in the putative transglutaminase the 30-min wash period. Following thewashesthemediumwas substrate domain (Fig. 1).Of the five botulinum neurotoxins, changed to Dulbecco's modified Eagle's medium with 0.1 m~ eserine only type B possessed both sequence homology with tetanus sulfate, plus 44 mM NaHCO, (control cells) or 44 mN KHCO, (depolartoxin and an essential glutamine. Serotype A possessed weak ized cells). The cells were incubated in release medium for 10 min. homology and a misaligned glutamine. The other three seroRadioactive acetylcholine was separated from precursor choline by ortypes possessed variable homology and no reactive glutamine. ganic extraction as described previously (7). In Vitro 7kanslation of Rat VAMP-2 (Synaptobreviwj-Rat VAMP-2 T h e light chains of tetanus toxin and botulinum neurotoxin cDNA (kindly provided by Dr. Scheller, Stanford University) inserted types A to E were searched to determine whether there were into Bluescript KSII' (Stratagene) was transcribed and translatedus- o t h e r d o m a i n s in which the group shared sequence homology ing an in vitro system (TNT", Promega). Synaptobrevin was syntheand a reactive glutamine. The results, which are shown in Fig. sized a t 30 "C for 90 min in a 6O-pl reaction mixture containing2 pg of 1, indicate that there w a s a region in the carboxyl terminus of DNA, 25 p1of TNTTM rabbitreticulocyte lysate, 40 pCi of [%]methiT3 RNA polymerase. Following synthesis, the the molecules that possessed an aligned glutamine. However, onine, and 1 pl of TNTTM reaction mixture wasportioned into 5-1.11aliquots and storedat -20 "C. recent evidenceindicates that this portion of the toxin molecule Digestion of Synaptobreuin with Tetanus lbxin-Tetanus toxin 1 p l ( 1 is n o t essential for blockade of exocytosis(12). pglpl) was diluted into 10 pl of phosphate-buffered saline supplemented T h e alignment data indicate that only tetanus toxin and with 10 m~ dithiothreitol. Toxin was reduced at 37 "C for 30 min, after which various concentrations of reduced toxin (see "Results") were botulinum neurotoxin typeB have domains that could account added to 5 pl of synaptobrevin translation mix and incubated at 37 "C for substrate activity. Therefore, these two toxinswere assayed as s u b s t r a t e s for transglutaminase at concentrations that a r e for 30 min. All digestion reactionswere treated with50 pg/ml RNase at 37 "C for an additional 15 min to remove tRNA background. Reaction of transmitter release at the neuromuscurelevant to blockade products (7PI) were mixed with 20 pl of SDS-sample buffer and resolved lar junction and in NG-108 cells M to M; see below). on 15% SDS-polyacrylamide gel electrophoresis gels. Gels were fixed, The results of these experiments were negative. Even when stained with Coomassie, and destained, with the final change of destain tested at a high concentration (1 x M), the toxins possessed containing 10% glycerol. Gels were dried and exposed t o x-ray film little if any ability toserve as substrates for transglutaminase(Kodak X-Omat A R ) overnight. induced cross-linking. This indicates that: ( a ) tetanus toxin Incubation of Tkansglutaminase with Tetanus Toxin-TransglutM; Sigma) was incubatedwith tetanus toxin (1 x IO" aminase (1 x and botulinum neurotoxintype B are not important substrates M ; Calbiochem) in assaybuffer containing 12 nm "is-HC1, pH 7.8, 14 for t r a n s g l u t a m i n a s e , a n d( h ) to the extent that the toxins are mM dithiothreitol, 2 mM MgCl,, and 2 mM NaCl at 0" or 37 "C for 1 h. substrates, this is n o t an i m p o r t a n t p a r tof the process of blockTetanus toxin and transglutaminase partiallyeclipse one another during SDS-polyacrylamide gel electrophoresis; therefore, the two mol- ing transmitter release. Stimulation of l+ansglutaminase Actiuity-Experiments ecules were separated by two ultrafiltration steps (molecular weight 100,000cutoff; molecular weight 30,000 cutoE,Amicon centricon tubes). were done on intactcells a n d on isolated enzyme preparations
ll-ansglutaminase and Tetanus Toxin Tetanus Toxin L i t Chain
Tetanus Toxin Heavy Chain
Botulinum Toxin A L i t Chain
Botulinum Toxin B L i t Chaii Botulinum Toxin C Light Chain
F G G I M Q M K F C P F G A L S I I S I S P
Botulinum Toxin D Light Chaii
Botulinum Toxin E Light Chain
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 ToxinLight Chain
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 correspondingregions 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 regionof the light chains of the six toxins in which there istrue alignment of glutamine residues. This region is inthe carboxylterminus of the light chains, and it is a portion of the molecule that is not required for blockadeof 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 type B would stimulate transglutaminase at meaningful concentrations. 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 acetylcholine release was measured. The results (Fig. 2) indicated that concentrations in therange of 10"' to 10"O M produced partial to complete blockadeof transmitter release. In thenext experiment, NG-108 cellswere incubated for 180 min with 1x lo-' M tetanus toxin. Cells werethen ruptured by sonication and exposed to varying concentrations of exogenous calcium (35 "C; 180 rnin). Transglutaminase activity was assayed 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 exocytosis, 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 concentrations (e.g. 10 m ~ ) . In the final experiment, tetanus toxin and botulinum neurotoxin type B were examined for their ability tostimulate transglutaminase-mediated incorporation of tritiated putrescine 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, 1x 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 increase in the amount of tritiated putrescine incorporated into dimethylcasein. Clostridial nxins, Bansglutaminase, and Neuromuscular Bansmission-Transglutaminase is a calcium-dependent enzyme. As gauged by incorporation of putrescine into dimethylcasein, the EC,, for calcium was in the range of10-100 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
Toxin [MI FIG.2. Blockade of acetylcholine release from NG-108cells. Various concentrations of tetanus toxin were incubated with cells that had been preloaded with [rnethyZ-'4C]choline chloride. After 180 min, cells were depolarized and the medium was assayed for radioactive acetylcholine. Complete blockade of exocytosis was obtained with 180 rnin exposure to 1o"O M toxin.
range of 100-300 nM (13), which is approximately 2 orders of magnitude below the EC,, for calciumsupported transglutaminase activity. Therefore, the actions of tetanus toxin and botulinum 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 calciumwas replaced with equimolar concentrations of magnesium or strontium. Interestingly, both toxins blocked transmission when added to tissues under conditions that would not be expected t o support transglutaminase activity ( n = 3 or more per group). When M) was added to unstimulated phrenic tetanus toxin (1 x
ToxinTetanus Dansglutaminase and
+ Cross-linked proteins
+ Heavy chain (-100KD)
* Light chain (- 50KD) toxin
24457 TABISI Effects of drugs on neuromuscular blockade Mouse phrenic nerve-hemidiaphragm preparations were excised and incubated inphysiological solution a t 35 “C. Phrenic nerves were stimulated (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
Tetanus Tetanus toxin Glycine methyl ester (3 Tetanus toxin Glycine methyl ester Tetanus toxin Monodansylcadaverine Botulinum Botulinum toxin Botulinum toxin Botulinum toxin
x (3 x
Glycine methyl ester (3 x Glycine methyl ester (3 x Monodansylcadaverine (1x
131* 12 124 11 129 2 13 126 * 8
*7 125 2 11 120 * 14 119 * 13
12 3 4 5 6 7 8 M) FIG.3. Effect of tetanus toxin on transglutaminase-mediated cross-linking of endogenous proteins in NG-108 cells. Two experimental paradigms were used, a s follows: (i) cells were homogenized and TABLEI1 assayed for enzyme-induced cross-linking of proteins in the absenceof Effects of false substrates on neuromuscular transmission toxin (lanes 1 4 ) ,or (ii) tetanustoxin was preincubated withcells (1 x Mouse phrenic nerve-hemidiaphragm preparations were excised from M; 180 min), after which cells were washed, homogenized, and animals and handled as described in Table I, except that miniature assayed for enzyme-induced cross-linking (lanes 5-8). In all cases, ly- endplate responses rather than muscle twitch responses recorded. were sates were incubated for 180 min before being submitted to polyacrylb Drug pretreatmenth,c Spontaneous miniature end amide gel electrophoresis. Assays were done in the presence of varying plate potentials concentrations of calcium, as follows: lanes 1 and 5 , no calcium; lanes 2 and 6 , 100 PMcalcium; lanes 3 and 7, 1 mM calcium; andlanes 4 and 8, min” 10 mM calcium. Note that pretreatment of cells with toxin did not 94 * 11 increase the amountof cross-linked protein inthe stacking gel, nor did + 11*2 it decrease the amount of free protein in the resolving gel. The location 135 -c 22 and approximate molecular weights of the heavy and light chains of Glycine methylester 1112 12 tetanus toxin are indicated. The location of synapsin I is between the + Glycine methylester 10 * 2 heavy and light chains of the toxin (doublet; S86KD and 80KD). 99 * 14 Monodansylcadaverine 85 * 6 nerve-hemidiaphragm preparations in physiological medium, + Monodansylcadaverine 16k 1
the average paralysis time was 172 2 9 min. Similarly, when botulinum neurotoxin type B (1x 10”’ M) was added t o preparations in physiological medium, the average paralysis time was 1912 12 min. The results for both toxins were not significantly 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 thepresence of glycine methyl esterand monodansylcadaverine.These agents can serve asfalse substrates for the enzyme, and thus they can be used to inhibit thecross-linking effects of endogenous substrates by transglutaminase (14-16). Glycine methyl ester and monodansylcadaverine were assayed 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 andmonodansylcadaverine 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 theIC,, values (glycine methyl ester, 3 x M; monodansylcadaverine, 1 x M), thedrugs 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 potentials 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 (1x M) the rate of a base-line spontaneous miniature end plate potentials during period was 99 2 14/min, and the rate during exposure to the drug was 85 2 6/min. These results show that transglutami-
Tetanus toxin was used a t a concentration of 1 x 10” M. involving drug pretreatment and subsequent additionof toxin is given under “Materials andMethods.” Glycine methyl ester was useda t 3 x lo-’$M, and monodansylcadavM. erinewasused a t 1x
* The protocol for experiments
nase-induced cross-linkingof synaptic vesicle proteins does not play an importantrole in governing the normal process of neuromuscular transmission. Experiments were done to determine whether transglutaminaseinhibitors would alter clostridial neurotoxin-induced blockade of exocytosis. Both mechanical responses and electrophysiological responses were monitored. In studies on stimulus-evoked muscle twitch, neither monodansylcadaverine nor glycine methyl ester delayed the onsetof toxin-induced neuromuscular blockade (Table I). In studieson 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‘jPheI7 bond in synaptobrevin 2 (3). Thespecificity of this reaction is evident in the facts that: ( a )the toxin does not cleave the Gln-Lys or Gln-Ala bonds in synaptobrevin 1or synaptobrevin 2, and ( b ) it does not cleave the Val-Phe bond in synaptobrevin 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 deduce the importance of the doublet by attempting to produce proteolytic cleavage with tetanus toxin. Transglutaminase and toxin wereincubated under conditions similar to those described by Facchiano and Luini (5),after which they were submitted topolyacrylamide gel electrophoresis. The Gln-Phe dou-
5'Yansglutaminase and Tetanus Toxin
Experiments in the present study show that tetanus toxin does not produce obvious stimulation of the enzyme nor does it kDa produce obvious proteolysis of the enzyme in anisolated assay system. Furthermore, the toxin does not produce detectable stimulation of cross-linking in intact cells. It was also found 6620.1 that false substrates did not block toxin action on nerve endings, and ambientcalcium levels below those needed to support enzyme activity did not prevent paralysis. FIG.4. Effect of tetanus toxin on p u t a t i v e substrates. TransgluIn the aggregate, thisevidence weighs against the idea that taminase was incubated with, and then separated from, tetanus toxin toxin-induced stimulation of transglutaminase is theprincipal 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 reason for toxin-induced blockade of neurotransmitter release. 1h (control); lane 2 , 3 7 "C for 1 h. ""S-Synaptobrevin was also incubated Nevertheless, there are two cautionary notes that should be with tetanus toxin as described in the text (37 "C for 30 min). The added. First, the data do not rule out thepossibility that high mixtures were applied t o a 15% gel and then submitted to autoradiogconcentrations of, or lengthy exposure to, tetanus toxin could 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); alter transglutaminase activity. Second, the data do not clarify Lane 4, '"S-synaptobrevin plus toxin (1 x IO-" M). Note that thetoxin did the mechanism of toxin action on cells in which synaptobrevin not cleave either of the bands associated with transglutaminase, but it does not play a role in exocytosis. produced concentration-dependent cleavage of synaptobrevin.
- T I
blet typically exists one-quarter to one-third of the distance from the N terminus to theC terminus, and thusproteolysis at this site should be easily detected. As a positive control, tetanus toxin was incubated with synaptobrevin to demonstratecleavage of this molecule. The results,which are shown in Fig. 4, confirm the abilityof tetanus toxin to cleave synaptobrevin. However, the toxin produced 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 specific peptide bond in synaptobrevin 2, a vesicle-associated protein 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 tetanus toxin stimulates the cross-linking enzyme transglutaminase (5). Stimulation of the enzyme may be due to proteolytic cleavage that converts an inactive precursor toan active product (6). Therefore, aseries of experiments were doneto evaluate the possibility that tetanustoxin, or the structurally andfunctionally similar botulinum neurotoxin type B, exert their effects 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 sequence 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 reportedlystimulatestransglutaminase 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 cleavageof transglutaminase was not reported (6).
Acknowledgments-We are grateful to Dr. Jose Martinez and Grace Gherovici for assistance and helpful suggestions on the transglutaminase assay. REFERENCES 1. Habermann, E., and Dreyer, F. (1986) Cum Top. Microhiol. Zmmunol. 129, 93-179 2. Simpson, L.L. (1989) Botulinum Neurotoxin and Tetanus Toxin, Academic
Press, San Diego, CA 3. Schiavo, G., Benfenati, F., Poulain, B., Rossetto, 0.. de Laureto. P.P., DasGupta, B. R., and Montecucco, C. (1992)Nature 359,832-835 4. Link, E., Edelmann,L., Chou, J. H., Binz, T.,Yamasaki, S., Eisel, U., Baumert, M., Siidhof, T. C., Niemann, H., andJahn. R. (1992)Riochem.Riophys. Res. Commun. 189, 1017-1023 5. Facchiano. F., and Luini. A. (1992)J. Biol. Chem. 267, 13267-13271 6. Facchiano, F., Benfenati. F., Valtorta. F.. and Luini, A. (1993)J. Biol. Chem. 268,4588-4591 7. McGee, R., Simpson, P., Christian, C., Mata, M., Nelson, P., and Nirenberg, M. (1978)Proc. Natl. Acod. Sci. U. S.A. 75, 1314-1318 8. Laemmli, U. K. (1970)Nature 227,680485 9. Barsigian, C., Fellin, F. M., Jain, A,, and Martinez, J. (1988)J . B i d . Chem. 263, 14015-14022 10. Bittner, M.A., Habig,W. H., and Holz, R. W. (1989)J. Neurochem. 53,966-968 11. Ahnert-Hilger, G., Weller, U., Dauzenroth, M. E., Habermann, E., and Gratzl, M. (1989) FEBS I A t . 242, 245-248 12. Kurazono, H., Mochida, S., Binz, T., Eisel, U., Quanz, M., Grebenstein, 0.. Wernars, K., Poulain, B., Tauc, L., and Niemann, H. (1992) J . Bid. Chem. 267,14721-14729 13. Blaustein, M. P. (1988)Handh. Exp. Pharmacol. 83, 275-304 14. Pastuszko, A,, Wilson, D.F., andErecinska, M. (1986) J. Neurochem. 46, 499-508 15. Bungay, P. J., Potter, J. M., and Griffin, M. (1984)Biochem. J. 219,819-827 16. Senner, A,, Dunlop, M. E., Gomis, R., Mathias, P. C. F., Melaisse-Lagae, F., and Malaise, W. J. (1985) Endocrinology 117, 237-242 17. 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 18 Fairweather, N. F., and Lyness, V.A. (1986) Nucleic Acids Res.14,7809-7812 19 Binz, T., Kurazono, H., Wille, M., Frevert, J., Wernars, K., and Niemann, H. (1990)J. Riol. Chem. 265,9153-9168 20 Thompson, D. E., Brehm, J. K., Oultram, J. D., Swinfield, T. J., Shone, C. C., Atkinson, T., Melling, J., and Minton, N. P. (1990) Eur: J. Riochrm. 189, 73-81 21 Whelm. S. M.. Elmore, M. J., Bodsworth, N. J., Brehm, J. K., and Atkinson,T. (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., Krieglstein, K., Henschen, A., Gill, D. M., and Niemann, H. (1990)Nucleic Acids Res. 18, 5556 24. Poulet, S., Hauser, D.,Quanz, M., Niemann, H., and Popoff, M. R. (1992) Biochem. Biophys. Res. Commun. 183, 107-113 25. Whelm, S. M., Elmore, M. J., Bodsworth, N. J., Atkinson, T., and Minton, N. P. (1992)Eur: J . Riochem. 204,657-667