Journal of Neurochemistry
Raven Press, Ltd., New York
0 1989 International Society for Neurochemistry
Proteolytic Cleavage of Tetanus Toxin Increases Activity
*?Gregory K. Bergey, /William H. Habig, *Jeffrey I. Bennett, and /Clara S. Lin
Departments of *Neurology and ?Physiology, University of Maryland School of Medicine, Baltimore, and $Centerfor
Biologics Evaluation and Research, Laboratory of Bacterial Toxins, Division of Bacterial Products,
U.S. Food and Drug Administration, Bethesda, Maryland, U.S.A.
Abstract: Tetanus toxin is initially synthesized in the form
of a single polypeptide chain and then proteolytically
“nicked” by the bacteria to produce a two-chain structure
joined by a disulfide bond. This two-chain form of the toxin
is the form known to be biologically active. Whether such
nicking is necessary for activity, as it is for certain other bac-
terial toxins, has not been demonstrated previously. Single-
chain toxin preparations produced by salt extraction from
the bacteria are characterized and compared with pure two-
chain toxin obtained from extracellular filtrates. The ability
of these various toxin preparations to produce paroxysmal
activity in mouse spinal cord neurons grown in dissociated
cell culture is described. The pure two-chain toxin is dem-
onstrated to have greater activity than the single-chain toxin
preparations. Indeed the activity of the single-chain toxin
Tetanus toxin, one of the most toxic biological sub-
stances known, is initially produced by the Clostridium
tetani bacterium in a single polypeptide chain of about
150,000 MW. This single-chain toxin is then subse-
quently “nicked” or cleaved by endogenous bacterial
proteases to form a two-chain structure, a heavy chain
of about 100,000MW and a light chain of 50,000 MW,
joined by a disulfide bond (Matsuda and Yoneda, 1975;
Helting and Zwisler, 1977; Helting et al., 1979). This
is the form of the toxin normally found in culture su-
pernatant and the form known to produce the dramatic
effects of tetanus. Whether this proteolytic cleavage is
necessary for or enhances toxin activity is unknown.
Other bacterial toxins (e.g., cholera, diphtheria, Pseu-
domonas, Shigella, botulinum) have a somewhat sim-
ilar two-chain structure ( D r a i n et al., 1971; Gill, 1976;
Vasil et al., 1977; Olsnes et al., 1981; Simpson, 1981).
For these other toxins, proteolytic cleavage appears
necessary for activity (diphtheria toxin also requires
additional reduction for activity). Because of the com-
Received August 9, 1988; revised manuscript received December
5, 1988; accepted December 13, 1988.
Address correspondenceand reprint requests to Dr. G. K. Bergey
at Department of Neurology, University of Maryland School of
Medicine, Baltimore, MD 21201, U.S.A.
preparations can be explained by the small amounts of re-
sidual two-chain toxin present in these extracts. Using a pro-
tease from a toxin-minus strain of Clostridium tetani to con-
vert a single-chain toxin preparation to two-chain toxin in-
creases toxin activity. In vivo the single-chain toxin
preparation is also less toxic. These findings indicate that
proteolytic nicking of tetanus toxin increases activity. The
unnicked, single-chain form of tetanus toxin may be a rela-
tively nontoxic protoxin form of the toxin; this is a structure-
function relationship similar to that of other bacterial protein
toxins. Key Words: Tetanus toxin-Spinal cord-Proteoly-
sis-Tissue culture-Protoxin. Bergey G . K. et al. Proteolytic
cleavage of tetanus toxin increases activity. J. Neurochem.
53, 155-161 (1989).
mon two-chain structure of bacterial toxins it has been
postulated that the single-chain form of tetanus toxin
is also an inactive protoxin that is proteolytically
cleaved and activated as the two-chain form found ex-
tracellularly (Middlebrook and Dorland, 1984). No di-
rect evidence to support this relationship exists for tet-
anus toxin. Studies to date with tetanus toxin have
been limited by the unavailability of suitable prepa-
rations of single-chain tetanus toxin and the fact that
intact animal systems contain endogenous protease
activity.
Recently fetal mouse spinal cord neurons in disso-
ciated cell culture have been utilized to characterize
the action of tetanus toxin (Bergey et al., 1983, 1987;
Habig et al., 1986). Using this system tetanus toxin has
been demonstrated to act at a presynaptic locus (Bergey
et al., 1987) to produce relative disinhibition following
a dose-dependent latent period (Bergey et al., 1983).
The relative activity of various toxin preparations can
be directly compared using this system to quantitate
Abbreviations used: H S , horse serum; MEM, minimum essential
medium; PDE, paroxysmal depolarizingevents; SDS, sodium dodecyl
sulfate.
156 G. K. BERGEY ET AL.
the time of onset of paroxysmal activity, a time marked
by reduced inhibitory activity that results in increased
polysynaptic excitation and prominent paroxysmal
depolarizing events.
We report here the results utilizing intracellular re-
cordings from mouse spinal cord neurons grown in
dissociated cell culture. T h e activity of a tetanus toxin
preparation containing predominantly the single-chain
form is compared with a pure two-chain toxin prepa-
ration to demonstrate that proteolytic cleavage is im-
portant for toxin activity. Portions of these data have
been reported in abstract form previously (Bergey et
al., 1986).
MATERIALS A N D METHODS
Culture techniques
Cultures of fetal mouse spinal cord neurons were prepared
as described in detail previously by Ransom et al. (1 977).
Spinal cords were removed from 124- to 14-day-old fetal
mice, trypsin treated, and mechanically dissociated and plated
on collagen-coated 35-mm plastic dishes. The culture me-
dium was Eagle's minimum essential medium (MEM) s u p
plemented with glucose (final concentration 30 mM) and
bicarbonate (final concentration 44 mM). Cultures were
grown and maintained at 35°C in 10%C 0 2 . During the first
24 h both 10% fetal calf serum and 10% horse serum (HS)
were included in the culture medium. After this time only
5% HS was included and 1% solution containing supplements
(Romijn et al., 1982) to reduce the amount of serum needed
was added. The antimetabolite 5-fluoro-2deoxyuridine
(concentration 54 pM with 140 p M uridine) was used for a
24-h period after day 7 to limit the growth of nonneuronal
cells. Cultures were maintained with biweekly subtotal
changes of medium for 4-8 weeks at which time they were
used for experiments.
Tetanus toxin
Homogeneous nicked tetanus toxin was prepared from
sterile culture filtrates of the Massachusetts C2 strain of Clos-
tridium tetani cultures. Unnicked toxin was extracted from
washed C. tetani cells with 1 M NaCI, 0.1 M sodium citrate
at pH 8.7 (Raynaud, 1951). The extraction solution also con-
tained 1 mM EDTA, I mM phenylmethylsulfonyl fluoride
(PMSF), and 10 mM benzamidine as potential protease in-
hibitors. Both toxin preparations were purified by (NH&S04
precipitation with 60% saturated ammonium sulfate, and gel
filtration on Sephadex G-100. The toxin fractions were
pooled, dialyzed against 5 mM potassium phosphate, pH 6.8,
and were applied to a hydroxyapatite column and eluted with
a gradient of 5 mM to 250 mM potassium phosphate, pH
6.8. After concentration by ultrafiltration, protein was mea-
sured by the Lowry protein assay using bovine serum albumin
as a standard (Lowry et al., 1951). Electrophoretic analysis
of the toxin preparations on sodium dodecyl sulfate (SDS)
gels in the presence and absence of mercaptoethanol deter-
mined the proportion of unnicked, single-chain toxin. Re-
duction of disulfide bonds in nicked toxin results in complete
dissociation into heavy and light chains whereas the unnicked,
single-chain toxin is not dissociable. Gels stained with Coo-
massie Brilliant Blue R-250 (Bio-Rad Laboratories) were used
for qualitative indications of the compositions, but for quan-
tification of the amounts of unnicked toxin in the various
preparations, toxin was labeled with ''1 using the Bolton-
Hunter reagent (Bolton and Hunter, 1973). SDS gel electro-
J. Neurochem., Vol. 53,No. I , 1989
phoresis was then performed in the presence of 2-mercap-
toethanol. Gels (10 cm length) were sliced into 2-mm fractions
using an Aliquogel Fractionator (Gilson Medical Electronics)
and the gel fractions counted. The percentage of nicked toxin
was calculated from the total counts in the unnicked toxin,
heavy chain and light chain regions of the reduced SDS gels
and compared to gels obtained with stock (100%nicked, two-
chain) toxin. For all physiological experiments a preparation
that was 10% nicked and 90% unnicked was used. This was
the preparation most enriched for unnicked toxin. For ex-
periments assaying for endogenous protease activity in the
cultures another preparation that was 22% nicked, 78% un-
nicked was also used. Previous experiments have demon-
strated preservation of bioactivity of tetanus toxin following
labeling with the Bolton-Hunter reagent (Habig et al., 1986).
1251-labeledtoxin was added for either 10 min or 14 h to
spinal cord cultures that had been washed free of horse serum
and placed in MEM. The cultures were then washed and the
structure of the cell-associated toxin was analyzed by SDS
gel electrophoresis in the presence of 2-mercaptoethanol. A
total of three cultures were pooled for each gel analysis.
For experiments to test whether nicking of the 90% single-
chain toxin preparation increased activity, radiolabeled un-
nicked tetanus toxin (90%unnicked) was treated with partially
purified nicking protease from a toxin minus strain of C.
tetani (strain BT 101; Laird et al., 1980; Finn et al., 1984).
After treatment both the control and protease-treated toxin
samples were passed over a Sephadex G-100 column pn-
manly to remove any contaminating clostridial hemolysin
(tetanolysin) present in the protease preparation which could
damage cultured neuronal cells. SDS gels of 1251-labeledun-
treated and protease-treated toxins were obtained after the
respective toxin preparations were reduced with 1% 2-mer-
captoethanol. These labeled toxin preparations were used in
both neuronal cell physiologic assays and mouse toxicity
studies. In neuronal cell physiologic assays 50,000 counts of
either the unnicked preparation or the protease-treated toxin
preparation were added to each culture. Intracellular record-
ings (described below) determined the onset of convulsant
activity. To check for residual hemolysin activity additional
cultures were incubated with 50,000 counts of the protease-
treated toxin for 72 h and examined under phase optics for
gross morphological changes.
Bioassays of the unnicked and protease-nicked toxins in
mice were performed. Dilutions of 1.5-fold of the labeled
toxin preparations were counted and injected into groups of
four mice per dilution. The percentages of mortality were
calculated after 96 h.
Electroph ysiology
Cultures grown for 4-8 weeks were selected for physiolog-
ical studies. Because HS contains significant amounts of an-
titetanus toxin antibodies, cultures were washed three times
in MEM without HS. Following this washing procedure, no
detectable antitoxin antibodies remained as determined by
mouse bioassay (<0.001 U; 1 U is sufficient to neutralize
approximately 2 pg of toxin). Cultures were placed in a
HEPES-buffered MEM solution prior to physiologic studies.
Culture dishes were placed on a temperature-controlled stage
of an inverted phase microscope. Using a thermister probe
with a feedback heating loop (Medical Systems) the temper-
ature at the culture dish periphery was carefully maintained
at 36.6 + 0.2"C. Intracellular recordings were made under
direct vision using microelectrodes filled with 4 M potassium
acetate at neutral pH with resistances of 35-55 MOhm.
 CLEA VAGE OF TETANUS TOXIN INCREASES ACTIVITY
Physiological recordings were obtained using conventional
bridge circuitry in conjunction with a storage oscilloscope
and a continuous chart recorder.
RESULTS
Several preparations of toxin were produced from
extracts of washed C.tetani cells as described in Ma-
terials and Methods. SDS gel analyses (Fig. 1) dem-
onstrated the single- and two-chain structures of re-
spective unnicked and nicked toxins. Analysis of ra-
diolabeled toxin (SDS gel electrophoresis with gel
fractionation)allowed quantification of the percentage
of unnicked, single-chain toxin present. After repeated
efforts, a toxin preparation enriched to 90%unnicked
single-chain toxin was obtained and used for all phys-
iological experiments. No entirely pure unnicked
preparation could be obtained.
To determine whether the spinal cord cultures con-
tained proteases capable of cleaving the single-chain
form of tetanus toxin, cultures were incubated with an
enriched (78%) unnicked toxin preparation. Incubation
of radiolabeled tetanus toxin preparations enriched for
unnicked, single-chain toxin in culture medium for up
to 14 h did not result in demonstrable proteolytic
cleavage of the cell-associated toxin when the percent-
ages of nicked toxin were determined by gel analyses
(Fig. 2). This suggests that proteases capable of cleaving
the toxin were not present or active in these spinal cord
cultures during the 14-h incubation period. All phys-
iological experiments described below were of much
shorter duration than 14 h.
Intracellular recordings from spinal cord neurons
exposed to tetanus toxin reveal a transition from con-
trolled spontaneous activity to convulsant activity
FIG. 1. Electrophoretic analysis of nicked and unnicked toxin
preparations on 5% SDS gels. Unnicked toxin and nicked toxin
preparationswere preparedas described in Materials and Methods.
Gels were stained with Coomassie BrilliantBlue. Reduction of the
disulfide bonds in nicked toxin by 2-rnercaptoethanol results in
complete dissociation into heavy and light chains whereas the un-
nicked toxin is not dissociable by mercaptoethanol reduction be-
cause of its singlechainstructure. Analysis of radiolabeledunnicked
toxin preparation (not shown) indicatedthat it was 90% unnicked,
the other 10% being nicked toxin.
157
x)(xIo CONTROL (22%NICKED1 *I
10 MINUTES (18% NICKED)
500
0 1
0 10 20 30 40 50
FRACTION NUMBER
FIG. 2. Cultures of spinal cord neurons do not proteolyticallynick
single-chain tetanus toxin. Analyses of gel fractions from washed
mouse spinal cord cultures exposed to '251-labeledtetanus toxin
(22% nicked, 78% unnicked) for 10 min or 14 h are shown. SDS
gel electrophoresis was in the presence of 2-mercaptoethanol. A.
Gel analysis of control toxin. 8: Cell-associated toxin after 10 min
incubation with the spinal cord cultures. C Cell-associated toxin
incubated with cultures for 14 h. No increased percentages of
nicked toxin were present at either time.
characterized by paroxysmal depolarizing events (tet-
anus-PDE) and rhythmic depolarizing bursts of action
potentials resulting from decreased inhibition (Fig. 3).
The latent period prior to onset of convulsant activity
may reflect a complicated presynaptic membrane in-
teraction (Schmitt et al., 1981; Yavin et al., 1981;
Critchley et al., 1985).
Determinations of the time to onset of established
paroxysmal activity (appearance of PDE) following the
addition of tetanus toxin preparations were made by
intracellular recordings from spinal cord neurons as
described in Materials and Methods. After PDE was
noted in one neuron, additional recordings from 3-10
adjacent neurons were typically performed to confirm
the presence of paroxysmal activity in the neurons.
The onset of paroxysmal activity (PDE) following ad-
dition of toxin preparations containing 90%unnicked
toxin was delayed when compared with similar con-
centrations of pure nicked toxin (Figs. 4 and 5). A dose-
response plot for three concentrations of the two un-
labeled toxin preparations illustrates the significantly
increased activity of the pure nicked toxin (Fig. 4).
Furthermore the average times of PDE onset for the
90% unnicked preparation at and g/ml(53.5
and 75.0 min, respectively) were not significantly dif-
J. Neurochem., Vol. 53,No. I , 1989
158 G. K. BERGEY ET A L .
FIG. 3. Top: Fetal mouse spinal cord neurons grown in dissociated
cell culture for 5 weeks. The calibration bar equals 50 Fm. Intra-
cellular recordings from spinal cord under controlconditions typically
reveal prominent spontaneous postsynaptic potentials (excitatory
and inhibitory) and spontaneous action potentials as shown on the
chart recorder record under control conditions (A and B). After a
dose-dependentlatent period during which time there may be little
change in spontaneous activity, tetanus toxin produces prominent
effects characterized by paroxysmaldepolarizing events (tetanus-
PDE), abrupt depolarizing shifts of 5-20 mV with associatedtrig-
gered action potentials. Inhibitorypostsynaptic potentials dramat-
ically diminish as the tetanus-PDE become established. The
recording shown in B illustratesan example of a neuron with prom-
inent spontaneous inhibitory activity under control conditions. The
onset of activity produced by tetanus toxin is concomitant with
the disappearance of these prominent inhibitory postsynaptic po-
tentials. In these and other chart records the amplitudes of the
action potentials are attenuated by the frequency response of the
penwriter.
ferent from the average times to PDE onset for the
pure nicked toxin preparation at lo-’ g/ml and lo-’
g/ml(54.8 and 75.8 min, respectively).
Protease derived from a non-toxin-producing strain
of C. tetani was used to treat the 90% unnicked toxin
preparation. After protease treatment little unnicked
toxin was detectable by radioautography of SDS gels
of the mercaptoethanol-reduced toxin (Fig. 6). Al-
though most of the detectable label was in gel regions
corresponding to heavy and light toxin chains, some
small amounts of other cleavage products were de-
tectable. Cultures incubated with the protease-treated
toxin for 72 h showed no gross morphological changes
when examined under phase microscopic optics, in-
dicating that most if not all of the hemolysin was re-
moved during preparation.
Treatment of the 90% unnicked toxin preparation
with bacterial protease significantly shortened the time
to onset of tetanus-PDE. Following the addition of
50,000 counts of each preparation the average time to
onset of tetanus-PDE was 190 min (SEM 9.25; n = 5)
J. Neurachern., Val. S3,No. 1. 1989
with the unnicked preparation. With protease treat-
ment this was shortened to an average of 134 min (SEM
8.27; n = 4, p < 0.01).
Injection of protease-treated (previously 90% un-
nicked) toxin into whole animals also indicated that
protease treatment increased the toxicity of the single-
chain toxin preparation (Fig. 7). With pure nicked
toxin, 250 cpm produced 100%mortality within 4 days,
whereas no mice died when injected with 250 cpm of
the 90% unnicked, single-chain preparation. One
hundred percent mortality was not achieved with the
90% unnicked toxin preparation unless 600 cpm or
more were administered. Analyses of all results indicate
an approximate threefold increase in toxicity on nick-
ing of the unnicked preparation. Similar relationships
were observed with purified, unlabeled, non-protease-
treated 100% nicked and 90% unnicked toxin prepa-
rations. Protease treatment had no effect on the mea-
sured toxicity of nicked toxin preparations (data not
shown).
DISCUSSION
Because tetanus toxin is not cytotoxic per se, assays
of toxicity require physiological effects to be measured
and quantitated. Intracellular recordings from spinal
cord neurons grown in dissociated tissue culture allow
the demonstration of action and potency of prepara-
tions of tetanus toxin by allowing quantification of the
time of onset of paroxysmal activity (Bergey et al., 1983,
1987). Since the onset of this activity has been dem-
onstrated to parallel the reduction of inhibition pro-
duced by tetanus toxin (Bergey et al., 1983, 1987) this
is a specific measure of toxin action, more readily
quantifiable than limb rigidity in whole animals and
more closely related to the specific onset of toxin action
100-
80-
-
-.-
E
W
n
0
60-
eg 40-
-t-
.
o S O % “NNlCKFDi10% N l C I t D TOXlN P R E P A R A T I O N
i O O % NICKED TOXlN P R E P A R A T I O N
*0 10-8 O10’7 10-6 L
TOXIN CONCENTRATION (g/ml)
FIG. 4. Average times to the onset of PDE produced by three
concentrations of the 100% nicked toxin (0)and 90% unnicked
toxin (0)preparations. Averages shown represent three to five
experiments per point (fSEM). Analysis of covariance produced
a nicked-unnicked difference of 21.8 f 5.5 ( p < 0.001).
 CLEA VAGE OF TETANUS TOXIN INCREASES ACTIVITY 159

100% NICKED TOXIN PREPARATION 90% UNNICKED/lO% NICKED TOXIN PREPARATION

10 minutes

3
30 minutes

60 minutes

FIG. 5. Representative penwriter records of intracellular recordings from spinal cord neurons at 10, 30, and 60 min following the addition
of pure nicked, double-chain toxin and 90% unnicked toxin (each g/ml) reveal the earlier onset of action with the pure nicked
preparation. In each instance, intracellular recordings at 10 min reveal a combination of spontaneous action potentials and postsynaptic
potentials (inhibitory and excitatory). After 30 min, the culture exposed to pure nicked toxin shows the early onset of paroxysmal activity
that is even more prominent at 60 min. lntracellular recordings from neurons in a culture exposed to the toxin enriched for the unnicked
form continue to show essentially a control pattern of activity; recordings at 60 min reveal paroxysmal activity (tetanus-PDE).

than death of animals. Using this neuronal culture sys-
tem the results above demonstrate that proteolytic
nicking of single-chain tetanus toxin into the two-chain
structure found extracellularly increases activity of the
toxin. This is illustrated above by the dose-response
curves of activity of the nicked toxin versus the activity
of a preparation highly enriched for unnicked toxin.
Not all strains of Clostridium tetani produce tetanus
toxin. Toxin production is plasmid related (Laird et
al., 1980). Using protease from a toxin minus strain of
C. tetani allowed an endogenous protease to be used
for the cleavage. Protease from a toxin producing strain
would have resulted in the addition of tetanus toxin
that would have confounded the results of the subse-
quent tissue culture and mouse assays of toxicity. This
experimental nicking of the unnicked toxin with a
clostridial protease increased toxin activity. Gel anal-
yses revealed that this experimental nicking produced
a two-chain toxin form similar to that naturally pro-
duced by the bacteria.
Despite considerable effort we have thus far been
unable to obtain toxin that was greater than 90% un-
nicked. This reflects the difficulty of purifying the in-
tracellular toxin without exposure to clostridial pro-
teases. Several protease inhibitors, including benzam-
idine (as recommended by Helting et al., 1979), have
been tried, but did not increase the purity of single-
chain toxin. Exactly where in the bacteria the toxin is
nicked is not known but the extracellular toxin is in
the nicked two-chain form. A mutant organism lacking
the nicking protease would be useful in producing a
pure single-chain toxin preparation.

- -- -
/-T
m
I

- /

/
- i
I
0 Protease Treated

II 0 Control
I
I
I I I

500 lo00 1500

INJECTED TOXIN (cprnl

FIG. 7. Protease treatment of previously singlechaintetanus toxin

FIG. 6. Radioautography of SDS gels of unnicked toxin before
(control) and after protease treatment, using protease derived from
a toxin-minus strain of C. tetani. Control toxin was 90% unnicked,
single-chain toxin. After protease treatment little unnicked toxin
remained.
increases in vivo toxicity. A bioassay of the unnicked (0)and pro-
tease-nicked (0)toxin in mice is shown. Dilutions of 1.5-fold of
the labeled toxin preparations were counted and injected into
groups of four mice per dilution. The percentages of mortality after
96 h in each group are shown.

J. Neurochem.. Vol. 53,No. 1. 1989

160 G. K. BERGEY ET A L
The fact that some nicked two-chain toxin was pres-
ent in all preparations guarantees that all preparations
will have some physiologic activity. Therefore it cannot
be definitely determined whether the unnicked single-
chain form of the toxin is atoxic or merely less toxic
than the two-chain form of toxin. Results from the
dose-response curves for the two preparations, how-
ever, indicate that the activity of the enriched unnicked
preparation could all be accounted for by the 10%
nicked toxin present. This suggests that the single chain
form of toxin is of very low potency and may in fact
be atoxic.
Injections into whole animals of proteolytically
nicked and 90% unnicked preparations demonstrated
about a threefold difference in toxicity. This is less than
the 10-fold difference that would be predicted and was
observed in the in vitro preparations. This reduced dif-
ference probably resulted from protelolytic nicking of
the single-chain preparation by endogenous trypsin-
like proteases active in the whole animal, effectively
reducing the percentage of unnicked, single-chain
toxin. Analysis of radiolabelled preparations of un-
nicked toxin have shown that, 24 h after injection into
mice, a large portion of the toxin remaining in the
injected muscle site as well as that which has undergone
intraaxonal transport has been nicked by mouse pro-
teases (Habig et al., 1983). As shown in Fig. 2 the
equivalent proteases appear to be absent in the spinal
cord cultures. Indeed the half-life of tetanus toxin in
spinal cord neurons in culture is rather long (estimated
to be 5-6 days, Habig et al., 1986), consistent with the
lack of protease activity seen here.
From the results described above, cleavage or nicking
of the toxin molecule is necessary for maximal toxicity.
Thus tetanus toxin appears to be similar to certain other
bacterial toxins in having an inactive or relatively in-
active single-chain protoxin form that is then activated
by proteolytic cleavage by the bacteria. Studies have
been conducted on the proteolytic activation of bot-
ulinum toxin, types A-G (Simpson, 1981). Activity
can be increased 100-foldor more by treatment of some
unnicked botulinum toxins with trypsin (Yokosawa et
al., 1986). Such cleavage may be exposing the active
site of the toxin, a site that resides on the nonbinding
portion of the molecule, fragment B. Interestingly, nei-
ther the heavy chain nor the light chain of tetanus toxin
alone is toxic (Matsuda and Yoneda, 1974, 1977), so
although nicking is necessary for maximum toxicity,
apparently portions of both chains are necessary for
maximum toxicity.
The benefits to the bacterium of such a relationship
requiring proteolytic nicking for activation are unclear.
Certainly there is no neurotransmission to be per-
turbed, but recent reports have indicated that tetanus
toxin may affect vesicle secretion in nonneuronal cells
such as macrophages (Ho and Klempner, 1986) and
adrenal chromaffin cells (Penner et al., 1986). Deter-
mining the structure-function relationships of this ex-
quisitely potent biological toxin will allow for a better
J. Neurochem., Vol 53,No 1. 1989
understanding of its site of action and broader use of
the toxin as a probe for investigating mechanisms of
neurotransmission and vesicular release.
Note added in proof: Weller et al. (1 988) recently reported
increased activity of bichain tetanus toxin compared to single
chain toxin utilizing assays in intact animals and phrenic
nerve-hemidiaphragms as well as measures of noradrenaline
release from isolated brain cell preparations. The differences
in vivo were only about twofold; in vitro the bichain toxin
was five to twelve times more potent than single chain toxin.
Acknowledgment: This work was supported by grants from
the J o h n A. Hartford Foundation, the Epilepsy Foundation
of America and U.S. Army Contract DAMD17-86-C-6056
to G.K.B. W e wish to express our appreciation to Ms. Yvonne
Logan for her expert preparation and maintenance of the
neuronal cultures used in these experiments. Dr. Thomas
Permutt kindly provided statistical analyses of the data.
REFERENCES
Bergey G. K., Macdonald R. L., Habig W. H., Hardegree M. C., and
Nelson P. G. (1983)Tetanus toxin: convulsant action on mouse
spinal cord neurons in culture. J. Neurosci. 3, 2310-2323.
Bergey G. K., Habig W. H., and Lin C. (1986) Nicking of tetanus
toxin increases activity. Ann Neurol. 20, 137-138.
Bergey G. K., Bigalke H., and Nelson P. G. (1987) Differential effects
of tetanus toxin on inhibitory and excitatory synaptic transmis-
sion in mammalian spinal cord neurons in culture: a presynaptic
locus of action for tetanus toxin. J. Neurophysiol. 57, 121-131.
Bolton A. E. and Hunter W. M. (1973) The labelling of proteins to
high specific activities by conjugation to a '251-containingacyl-
ating agent. Biochem. J. 133, 529-539.
Critchley D. R., Nelson P. G.. Habig W. H., and Fishman P. H.
(1 985) Fate of tetanus toxin bound to the surface of primary
neurons in culture: evidence for rapid internalization. J. Cell
Biol. 100, 1499-1507.
Drazin R., Kandel J., and Collier R. J. ( I 97 1) Structure and activity
of diphtheria toxin. 11. Attack by trypsin at a specific site within
the intact toxin molecule. J. Biol. Chem. 246, 1504-1510.
Finn C. W., Silver R. P., Habig W. H., Hardegree M. C., Zon G.,
and Garon C. F. (1984) The structural gene for tetanus neuro-
toxin is on a plasmid. Science 224, 881-884.
Gill D. M. (1976) The arrangement of subunits in cholera toxin.
Biochemistry 15, 1242-1248.
Habig W. H., Kenimer J. G., and Hardegree M. C. (1983) Retrograde
transport of tetanus toxin: toxin mediated antibody transport,
in Frontiers in Biochemical and Biophysical Studies of Proteins
and Membranes (Liu T. Y., Sakakibara S., Schecter A. N., Yagi
K., Yajima H., and Yasunobu B. T., eds), pp. 463-473. Elsevier,
New York.
Habig W. H., Bigalke H., Bergey G. K., Neale E. A,, Hardegree
M. C.. and Nelson P. G. (1986) Tetanus toxin in dissociated
spinal cord cultures: long-term characterization of form and ac-
tion. J. Neurochem. 47, 930-937.
Helting T. B. and Zwisler 0. (1977) Structure of tetanus toxin. I.
Breakdown of the toxin molecule and discrimination between
polypeptide fragments. J . Biol. Chem. 252, 187-193.
Helting T. B., Parschat S., and Engelhardt H. (1979) Structure of
tetanus toxin: demonstration and separation of a specificenzyme
converting intracellular tetanus toxin to the extracellular form.
J. Biol. Chem. 254, 10728-10733.
Ho J. L. and KIempner M. S . (1986) Inhibition of macrophage se-
cretion by tetanus toxin is not directly linked to cytosolic calcium
homeostasis. Biochem. Biophys. Res. Cornrnun. 135, 16-24.
Laird W. J., Aaronson W., Silver R. P., Habig W. H., and Hardegree
M. C. ( 1980) Plasmid-associated toxigenicity in Clostridium tet-
ani. J. infect. Dis. 142, 623.
Lowry 0. H., Rosebrough N. J., Farr A. L., and Randall R. J. (1951)
 CLEAVAGE OF TETANUS TOXIN INCREASES ACTIVITY
Protein determination with the Folin phenol reagent. J. Biol.
Chem. 193,265-275.
Matsuda M. and Yoneda M. (1974) Dissociation of tetanus neuro-
toxin into two polypeptide fragments. Biochem. Biophys. Res.
Commun. 57, 1257-1262.
Matsuda M. and Yoneda M. (1975) Isolation and purification oftwo
antigenically active “complementary” polypeptide fragments of
tetanus neurotoxin. Infect. Immun. 12, 1147-1 153.
Matsuda M. and Yoneda M. (1977) Antigenic substructure of tetanus
neurotoxin. Biochem. Biophys. Res. Commun. 77, 268-274.
Middlebrook J. L. and Dorland R. B. (1984) Bacterial toxins: cellular
mechanisms of action. Microbiol. Rev. 48, 199-22 1.
Olsnes S., Reisbig R., and Eiklid K. (1981) Subunit structure ofShihi-
gella cytotoxin. J. Biol. Chem. 256, 8732-8738.
Penner R., Neher E., and Dreyer F. (1986) Intracellular injected tet-
anus toxin inhibits exocytosis in bovine adrenal chromaffin cells.
Nalure 324, 76-78.
Ransom B. R., Neale E. A., Henkert M., Bullock P. N., and Nelson
P. G . (1977) Mouse spinal cord in cell culture. I. Morphology
and intrinsic neuronal electrophysiologic properties. J. Neuro-
physiol. 40, 1 132- 1 150.
Raynaud M. (1951) Extraction de la toxine tetanique a partir des
corps microbiens. Ann. Inst. Pasteur 80, 356-377.
Romijn H. J., Habets A. M. M. C., Mud M. T., and Wolters P. S.
161
( 1982) Nerve outgrowth, synaptogenesis and bioelectric activity
in fetal rat cerebral cortex tissue cultured in serum-free, chem-
ically defined medium. Dev. Brain Res. 2, 583-589.
Schmitt A., Dreyer F., and John C. (1981) At least three sequential
steps are involved in the tetanus toxin-induced block of neu-
romuscular transmission. Naunyn Schmiedebergs Arch. Phar-
macol. 317, 326-330.
Simpson L. (1981) The origin, structure and pharmacologic activity
of botulinum toxin. Pharrnacol. Rev. 33, 155-188.
Vasil M. L., Kabat D., and Iglewski B. H. (1977) Structure-activity
relationships of an exotoxin of Pseudomonas aeruginosa. Infect.
Immun. 16,353-361.
Weller U., Mauler F., and Habermann E. (1988) Tetanus toxin: bio-
chemical and pharmacological comparison between its portoxin
and some isotoxins obtained by limited proteolysis. Nauyn-
Schmiedebergs Arch. Pharmacol. 338,99- 106.
Yavin E. Z., Yavin A., Habig W. H., Hardegree M. C., and Kohn
L. D. (1 98 1) Tetanus toxin associated with developing neuronal
cell cultures: kinetic parameters and evidence for ganglioside-
mediated internalization. J. Biol. Chem. 256, 70 14-7022.
Yokosawa N., Tsuzuki K., Syuto B., and Oguma K. (1986) Activation
of Clostridium botulinum type E toxin purified by two different
procedures. J. Gen. Microbiol. 132, 1981-1988.
J. Neurochem., Vol. 53,No. I , 1989