- Email: [email protected]
Conantokin-G: A novel peptide antagonist to the N-methyl-d-aspartic acid (NMDA) receptor
Neuroscience Letters, 118 (1990) 241-244
241
Elsevier Scientific Publishers Ireland Ltd. NSL 07227
Conantokin-G: a novel peptide antagonist to the Nomethyl-D-aspartic acid (NMDA) receptor E. Edward M e n a l, M a r y F. Gullak l, Martin J. Pagnozzi 1, Karl E. Richter 1, Jean Rivier 2, Lourdes J. Cruz 3 and Baldomero M. Olivera 3 1Department of Neuroscience, Central Research Division, Pfizer Inc., Groton, CT 06340 (U.S.A.), 2Clayton Laboratories for Peptide Biology, Salk Institute, La Jolla, CA 92138 (U.S.A.) and aDepartment of Biology, University of Utah, Salt Lake City, UT84112 (U.S.A.) (Received 26 March 1990; Revised version received 22 June 1990; Accepted 27 June 1990)
Key words: NMDA antagonist; Excitatory amino acid antagonist; Snail venom; Peptide venom; Conantokin; Sleeper peptide Conantokin-G is a 17 amino acid peptide isolated from the venom of the fish-eating snail Conus geographus which produces hyperactivity when injected into the brains of adult mice. We show that this peptide is a selective N-methyl-D-aspartate (NMDA) antagonist based on its ability to block NMDA-induced elevation of cGMP in rat cerebellar slices in vitro (IC50 = 171 nM), but not kainic acid-induced elevations. This inhibition could not be overcome by increasing the NMDA concentration, indicating non-competitive inhibition. Conantokin-G displayed no affinity for binding sites for thienylcyclohexylpiperidine, various glutamate subclasses or those for several other neurotransmitters/neuromodulators. This peptide, however, enhanced [3H]glycine binding to rat forebrain membranes but not to spinal cord membranes. The activity profile of the peptide in various assays indicates that it is a novel type of non-competitive NMDA antagonist.
Several components isolated from the venom of the fish-hunting snail Conus geographus have been shown to have unique effects on vertebrate biological systems. Some of these components, such as the ~, a- and o>conotoxins have paralytic effects and block nicotinic acetylcholine receptors, muscle sodium channels and presynaptic Ca 2+ channels, respectively (reviewed in ref. 15). One component which has been recently isolated, chemically characterized and synthesized is a 17 amino acid peptide, conantokin-G (formerly called the 'sleeper peptide' and 'conotoxin GV') with the sequence Gly-GluGla-Gla-Leu-Gln-Gla-Asn-Gln-Gla-Leu-Ile-Arg-GlaLys-Ser-Asn-NH2) [16]. It contains 5 residues of the unusual amino acid 7-carboxyglutamate (Gla). This peptide produces unique behavioral effects when injected into the brains of mice. In young mice (less than two weeks old), it induces a profound sleep-like state while adult mice become hyperactive [13]. In many respects the hyperactivity induced by this peptide in adult mice resembles the behavioral effects induced by non-competitive glutamate (Glu) receptor antagonists of the Nmethyl-D-aspartate (NMDA) class. Thus, we examined conantokin-G for activity as a glutamate antagonist. The results reported here show that this peptide is a potent and selective N M D A antagonist whose mechaCorrespondence: E.E. Mena, Pfizer Central Research, Groton, CT 06340, U.S.A.
0304-3940/90/$ 03.50 © 1990 Elsevier Scientific Publishers Ireland Ltd.
nism of antagonism is distinct from either the competitive or previously described non-competitive types. We found that conantokin-G blocked the enhancement of cGMP induced by N M D A (100 aM) in the neonatal rat cerebellum in vitro with an IC50 of 171 + 28 nM (n = 7, Fig. 1A). This peptide had no effect on kainic acid (KA)-induced (100 aM) elevation of cGMP at concentrations of up to 3.2 a M (Fig. 1A). Comparison of the activity of this peptide with that of other known N M D A antagonists in this assay shows that this peptide is one of the most potent compounds described (Table I). We next analyzed N M D A concentration-response data obtained in the presence of various levels of the peptide in order to determine if this antagonism is competitive or non-competitive with NMDA. The results (Fig. 1B) show that there is a progressive flattening of the concentration-response curve as the concentration of conantokin-G increases. Furthermore, the apparent ICs0 of conantokin-G is unaffected by the NMDA concentration; both features are consistent with non-competitive antagonism. Other non-competitive antagonists such as phencyclidine (PCP) and MK-801 exhibit potent inhibition of [3H]TCP binding to brain membranes. In our hands, we find that non-competitive antagonists display similar potencies at blocking NMDA-induced elevations of cGMP (at 100 aM NMDA) and [3H]thienylcyclohexylpiperidine (TCP) binding (Table I). In con-
242 80-[3
/(3-- - O - - O
lOO - A / t-
.9 /
!
!
/
NMDA
Q.
0/ !
E
.=_ "E 5 0 -
/
o Q.. / O
/
/
/
! /O
10
::
50nM ~ ' ~ °
**
~
100~
'k
L"~ 50
\~E
40--
/
•
o
/ O9 *"
o
E
•
/
/
/
300
Conantokin-G (nM)
r~-"/ O" / / 20 - o' , = / / o . ii / 0 / "
500nM
...--e
t~O--O
I
I
I
102
103
104
Conantokin-G
/
e/
KA 0-
:
60--
ft.
/
¢,-.
Control .... .....D D°o°,°.°.°° o°°°°°~°
(nM)
L
200
I
400
I
600
NMDA (#M)
Fig. 1. A: effect of conantokin-G on agonist-induced elevation of cGMP in rat cerebellar slices in vitro. Slices were incubated as described in the legend to Table I with either NMDA or KA (100 pM) and varying concentrations of conantokin-G, The peptide completely antagonized responses to NMDA but had no effect on the KA-induced cGMP. B: concentration-response curves of the NMDA-induced (50-500 pM) elevation of cGMP in the neonatal rat cerebellum in the presence of the indicated concentration of conantokin-G. As the concentration of the conantokin-G is increased there is a clear flattening of the NMDA concentration-response curve. The inset shows the same data plotted as a conantokin-G concentrationresponse curve and shows that the apparent IC50 of the conantokin-G is unchanged by increasing NMDA concentrations. Both of these characteristics indicate non-competitive inhibition. A similar non-competitive profile was obtained with MK-801 in this assay, although this compound was 2-fold more potent than conantokin-G (see Table I). The results are from a representative experiment which was repeated three times. The ICs0's for the conantokin-G varied less than 10 % among these different experiments.
trast, conantokin-G has no significant effect on [3H]TCP binding at concentrations of up to 2 pM (Table I). The lack of activity in this assay clearly distinguishes this naturally occurring peptide from other 'non-competitive' NMDA antagonists. In order to determine if conantokin-G was interacting with a known binding site, we examined this peptide's had affinity for various Glu binding or uptakes sites, for non-strychnine-sensitive [3H]glycine (Gly) binding sites, or for binding sites for several other CNS neurotransmitters/neuromodulators. Conantokin-G (1 pM) had no effect on C1--dependent Glu binding (a putative anion dependent Glu uptake mechanism) [2], Cl--independent Glu binding (largely postsynaptic receptors) [12], Na +dependent Glu binding (binding to the Na+-dependent Glu carrier protein) [1] or on Na+-dependent glutamate uptake into rat brain synaptosomes. However, conantokin-G had the unexpected effect of enhancing strychnine-insensitive [3H]GIy binding to rat forebrain membranes in a concentration-dependent manner (Fig. 2). The enhancement was maximal at approximately 750 nM and had an apparent EDs0 of 152 _ 14 nM (n = 3). The binding was more than doubled in membranes prepared from forebrain but increased only 7.8 % in spinal cord membranes (Fig. 2), a result consistent with the distribution of NMDA-associated [3H]GIy receptors [3]. The peptide did not reveal a pharmacologically distinct [3H]GIy binding site since the ICs0's of Gly, D-Ser and
kynurenic acid were unchanged in this assay by conantokin-G (212 _ 16 nM, 269 -t- 55 nM and 8.3 _ 1.4 pM, respectively). Neither competitive nor non-competitive NMDA antagonists mimic this effect on [3H]Gly binding. Conantokin-G (at 1 pM) also displayed no affinity for the binding sites for KA, amino-3-hydroxy-5-methylisoxazole-4propionic acid (AMPA) benzodiazepine, ~q-adrenergic, ct2-adrenergic, 7-aminobutyric acid (GABA), serotonin 5-HTla, serotonin 5-HT2 or muscarinic nicotinic receptors (data not shown). The similarity between the behavioral effects of i.c.v. administered conantokin-G and i.p. administered PCP in mice led us to investigate the possible Glu antagonist activity of this peptide. Our results show that the 17 amino acid conantokin-G isolated from the venom of the predatory snail Conus geographus blocks NMDAinduced elevations of cGMP in the rat cerebellum in vitro. This blockade is not a non-specific effect on cGMP generation since it had no effect on KA-induced increases in cGMP. Thus, conantokin-G appears to be a potent NMDA antagonist selective for the NMDA class of receptors. This peptide does not act directly at the Glu receptor since it was unable to block [3H]GIu binding to postsynaptic receptors. Additionally, it had no effect on [3H]KA or [3H]AMPA binding. Conantokin-G also had no significant affinity for receptors of several other neurotransmitters.
243 TABLE I
_
THE EFFECT OF CONANTOKIN-G AND STANDARD N M D A ANTAGONISTS ON cGMP PRODUCTION AND [3H]TCP BINDING.
I~ 200 ._c "O .¢_
The elevation of cGMP induced by excitatory amino acid agonists in neonatal rat cerebeUar slices was carried out essentially as described by Foster ~ind Roberts [5] and Garthwaite and Garthwaite [6]. Basal levels of cGMP were 0.54 pmol/mg protein, in the presence of either NMDA or KA (100 gM) the levels of cGMP varied from 65 to 80 pmol/mg protein among the different experiments. [3H]TCP (New England Nuclear, 48.9 Ci/mmol) binding was carried out with membranes prepared from rat forebrain [4] which were sonicated and rinsed 5 times with 5 mM Tris-HCl, pH 7.5. The incubation medium contained 1 mg membranes in 1 ml 5 mM Tris-HC1 and 2 nM [3H]TCP. Following 30 min at 30°C, the assay was terminated by filtration through GF/B filters (Whatman). Non-specific binding was determined in the presence of 100/zM PCP. All values are the mean + S.E.M. of at least 3 determinations.
c
Compound
NMDA-induced cGMP (IC50, nM)
KA-induced [3H]TCP binding cGMP (IC50, Ki(nM) nM)
MK-801 PCP AP7 Conantokin-G
78__+15 184___26 3,200 + 350 171 + 28
> 5,000 >5,000 > 100,000 > 3,200
64+ 16 212___48 5,500 + 290 > 2,000
Some additional results, obtained in collaboration with others, are consistent with conantokin-G functioning as an NMDA antagonist. The peptide reduces the increase in Fura-2 fluorescence elicited when NMDA and Gly are applied to cerebellar granule cells in culture [8]. In addition, conantokin-G antagonizes certain effects of NMDA or Glu in vivo. These include the ability of NMDA to induce seizures, and the effects of glutamate on frequency modulation of the pacemaker nucleus in the central nervous system of weakly electric fish (G. Rose, M. Mclntosh and B. Olivera, unpublished observations). Clearly, additional experiments, both biochemical and electrophysiological, are necessary to define further the mechanism by which this peptide exerts its NMDA antagonist activity. Efforts to classify this peptide into one of the established classes of Glu antagonists were unsuccessful. Conantokin-G blocked the NMDA-induced increases in cGMP in a non-competitive manner. However, in contrast to other non-competitive antagonists, it does not bind to the [3H]TCP site. The ability of this peptide to induce an apparent enhancement of [3H]GIy binding also underscores its uniqueness. Gly has been shown to be a positive modulator of NMDA activity in several systems [10]. It is unclear, however, how enhancement of [3H]GIy binding could result in NMDA antagonism. The enhancement is observed only in forebrain membranes
.~
.........
~ Forebrain
/
,D
150
,~
~s.ii,3._..___,.._._ili
- ] ; Spinal Cord
100~
I~ 5o l
1000 Conantokin-G
I
2000 (nM)
I
3000
Fig. 2. Conantokin-G enhances [3H]Gly binding to rat forebrain membranes but not to spinal cord membranes. Crude synaptosomes (P2) were prepared identically from each tissue and lysed. [3H]GIy binding was carried out essentially as described by Kessler et al. [11]. Membranes were incubated for 30 min at 0°C with 100 nM [3H]GIy. The assay was terminated by filtering and rinsing with ice cold buffer for 5 s. Non-specific binding was measured by including 100/iM nonradioactive Gly in the assay. Each point is the mean + S.E.M. of 2 4 experiments.
and not spinal cord membranes, which lack strychnineinsensitive Gly receptors [3]. Recently, we observed that another novel class of NMDA receptor antagonist, the arylamine toxins from spider venoms, also display this effect on [3H]GIy binding [7]. Detailed studies reveal that the increased binding observed in the assay results from a decrease of the association and dissociation rate of [3H]Gly with its receptor and does not alter the Kd of this interaction [7]. If conantokin-G and the arylamines from spider venoms have a similar mechanism, their apparent antagonism may result from an impaired ability of Gly to positively modulate the NMDA receptor. Why does this snail, which has a diet restricted exclusively to fish, have a peptide which functions as a Glu antagonist as a component of its venom? Although a target for this peptide in peripheral systems of fish is unknown, intramuscular injection of conantokin-G causes jerky and uncoordinated swimming which may slow the fish's escape (D. Yoshikami and B. Olivera, unpublished observations). Alternatively, the peptide may act in concert with other components of the venom to aid in the culinary pursuits of this snail. A final possibility is that some components of the venom are reserved for defensive measures. Conus geographus attempts to sting when handled and has, in some cases, administered fatal doses of toxin to unprotected humans [14]. Conus geographus is a thin-shelled mollusk which may be a tempting and vulnerable target for several species of crustacean. Since Glu is a neurotransmitter at the invertebrate neuromuscular junction, injection of a potent glutamate antagonist into an invertebrate predator may
244
be a very effective defense. While we can not distinguish among these and other possibilities at this time, we have found that this peptide from the venom of Conus geographus is a potent NMDA antagonist in the mammalian CNS. Furthermore, two other peptides homologous to conantokin-G have been purified from other Conus venoms which produce similar behavioral effects in mice. One of these, conantokin-T, blocks NMDA-induced Ca 2+ fluxes in cerebellar cells in vitro [9]. Together these define a family of ~-carboxyglutamate-containing neuropeptides in Conus venoms, the conantokins (from the Philippino word for sleepy, antokin), which may be useful ligands for novel sites on the NMDA receptor complex. This work was supported in part by GM22737. We would like to thank Dr. G.N. Woodruff, Merck and Co., for the gift of MK-801. 1 Baudry, M. and Lynch, G., Characterization of two [3H]glutamate sites in rat hippocampal membranes, J. Neurochem., 36 (1981) 811-820. 2 Bridges, R.J., Nieto-Sampedro, M., Kadri, M. and Cotman, C.W., A novel chloride-dependent L-[3H]glutamate binding site in astrocyte membranes, J. Neurochem., 48 (1987) 1709-1715. 3 Bristow, D.R., Bowery, N.G. and Woodruff, G.N., Light microscopic autoradiographic localization of 3H glycine and 3H strychnine binding sites in rat brain, Eur. J. Pharmacol., 126 (1986) 303307. 4 Cotman, C.W. and Taylor, D., Isolation and structural studies on synaptic complexes from rat brain, J. Cell Biol., 55 (1972) 696-71 I. 5 Foster, G.A. and Roberts, P.J., Pharmacology of excitatory amino acid receptors mediating the stimulation of rat cerebellar cyclic GMP levels in vitro, Life Sci., 27 (1980) 215-221.
6 Garthwaite, J. and Garthwaite, G., Cellular origins of cyclic GMP responses to excitatory amino acid receptor agonists in rat cerebellum in vitro, J. Neurochem., 48 (1987) 29-39. 7 Gullak, M.F., Pagnozzi, M.J., Richter, K.E. and Mena, E.E., Effects of potyamine spider venoms components on [3H]GIu and [3H]Gly binding, Soc. Neurosci. Abstr., 14 (1988) 168. 8 Haack, J., Mena, E., Parks, T.N., Rivier, J., Cruz, L. and Olivera, B.M., Novel peptide NMDA antagonists from the venom of Conus, Soc. Neurosci. Abstr., 15 (1989) 1167. 9 Haack, J.A., Rivier, J., Parks, T., Mena, E.E., Cruz, L.J. and Olivera, B.M., Conantokin-T: a gamma-carboxyglutamate containing peptide with NMDA antagonist activity, J. Biol. Chem., 265 (1990) 6025-6029. 10 Johnson, J.W. and Asher, P., Glycine potentiates the NMDA response in cultured mouse brain neurons, Nature, 325 (1987) 529531. 11 Kessler, M., Terrammani, T., Lynch, G. and Baudry, M., A glycine site associated with N-methyl-D-aspartic receptors: characterization and identification of a new class of antagonists, J. Neurochem., 52 (1989) 1319-1328. 12 Monaghan, D.T., Holets, V.R., Toy, D.W. and Cotman, C.W., Anatomical distribution of four pharmacologically distinct 3H-Lglutamate binding sites, Nature, 176 (1983) 176-179. 13 Olivera, B.M., Mclntosh, J.M., Clark, C., Middlemas, D., Gray, W.R. and Cruz, L.J., A sleeper toxin from Conus geographus venom, Toxicon, 23 (1985) 27%282. 14 Olivera, B.M., Gray, W.R. and Cruz, L.J., Marine snail venoms. In A.T. Tu (Ed.), Handbook of Natural Toxins, Vol. 3, Dekker, New York, 1988, pp. 327-352. 15 Olivera, B.M., Gray, W.R., Zeikus, R., Mclntosh, J.M., Varga, J., Rivier, J., de Santos, V. and Cruz, L.J., Peptide neurotoxins from fish-hunting cone snails, Science, 230 (1985) 1338-1343. 16 Rivier, J., Galyean, R., Simon, L., Cruz, L., Olivera, B.M. and Gray, W.R,, Total synthesis and further characterization of the gamma-carboxyglutamate-containing'sleeper' peptide from Conus geographus, Biochemistry, 26 (1987) 8508-8512.
241
Elsevier Scientific Publishers Ireland Ltd. NSL 07227
Conantokin-G: a novel peptide antagonist to the Nomethyl-D-aspartic acid (NMDA) receptor E. Edward M e n a l, M a r y F. Gullak l, Martin J. Pagnozzi 1, Karl E. Richter 1, Jean Rivier 2, Lourdes J. Cruz 3 and Baldomero M. Olivera 3 1Department of Neuroscience, Central Research Division, Pfizer Inc., Groton, CT 06340 (U.S.A.), 2Clayton Laboratories for Peptide Biology, Salk Institute, La Jolla, CA 92138 (U.S.A.) and aDepartment of Biology, University of Utah, Salt Lake City, UT84112 (U.S.A.) (Received 26 March 1990; Revised version received 22 June 1990; Accepted 27 June 1990)
Key words: NMDA antagonist; Excitatory amino acid antagonist; Snail venom; Peptide venom; Conantokin; Sleeper peptide Conantokin-G is a 17 amino acid peptide isolated from the venom of the fish-eating snail Conus geographus which produces hyperactivity when injected into the brains of adult mice. We show that this peptide is a selective N-methyl-D-aspartate (NMDA) antagonist based on its ability to block NMDA-induced elevation of cGMP in rat cerebellar slices in vitro (IC50 = 171 nM), but not kainic acid-induced elevations. This inhibition could not be overcome by increasing the NMDA concentration, indicating non-competitive inhibition. Conantokin-G displayed no affinity for binding sites for thienylcyclohexylpiperidine, various glutamate subclasses or those for several other neurotransmitters/neuromodulators. This peptide, however, enhanced [3H]glycine binding to rat forebrain membranes but not to spinal cord membranes. The activity profile of the peptide in various assays indicates that it is a novel type of non-competitive NMDA antagonist.
Several components isolated from the venom of the fish-hunting snail Conus geographus have been shown to have unique effects on vertebrate biological systems. Some of these components, such as the ~, a- and o>conotoxins have paralytic effects and block nicotinic acetylcholine receptors, muscle sodium channels and presynaptic Ca 2+ channels, respectively (reviewed in ref. 15). One component which has been recently isolated, chemically characterized and synthesized is a 17 amino acid peptide, conantokin-G (formerly called the 'sleeper peptide' and 'conotoxin GV') with the sequence Gly-GluGla-Gla-Leu-Gln-Gla-Asn-Gln-Gla-Leu-Ile-Arg-GlaLys-Ser-Asn-NH2) [16]. It contains 5 residues of the unusual amino acid 7-carboxyglutamate (Gla). This peptide produces unique behavioral effects when injected into the brains of mice. In young mice (less than two weeks old), it induces a profound sleep-like state while adult mice become hyperactive [13]. In many respects the hyperactivity induced by this peptide in adult mice resembles the behavioral effects induced by non-competitive glutamate (Glu) receptor antagonists of the Nmethyl-D-aspartate (NMDA) class. Thus, we examined conantokin-G for activity as a glutamate antagonist. The results reported here show that this peptide is a potent and selective N M D A antagonist whose mechaCorrespondence: E.E. Mena, Pfizer Central Research, Groton, CT 06340, U.S.A.
0304-3940/90/$ 03.50 © 1990 Elsevier Scientific Publishers Ireland Ltd.
nism of antagonism is distinct from either the competitive or previously described non-competitive types. We found that conantokin-G blocked the enhancement of cGMP induced by N M D A (100 aM) in the neonatal rat cerebellum in vitro with an IC50 of 171 + 28 nM (n = 7, Fig. 1A). This peptide had no effect on kainic acid (KA)-induced (100 aM) elevation of cGMP at concentrations of up to 3.2 a M (Fig. 1A). Comparison of the activity of this peptide with that of other known N M D A antagonists in this assay shows that this peptide is one of the most potent compounds described (Table I). We next analyzed N M D A concentration-response data obtained in the presence of various levels of the peptide in order to determine if this antagonism is competitive or non-competitive with NMDA. The results (Fig. 1B) show that there is a progressive flattening of the concentration-response curve as the concentration of conantokin-G increases. Furthermore, the apparent ICs0 of conantokin-G is unaffected by the NMDA concentration; both features are consistent with non-competitive antagonism. Other non-competitive antagonists such as phencyclidine (PCP) and MK-801 exhibit potent inhibition of [3H]TCP binding to brain membranes. In our hands, we find that non-competitive antagonists display similar potencies at blocking NMDA-induced elevations of cGMP (at 100 aM NMDA) and [3H]thienylcyclohexylpiperidine (TCP) binding (Table I). In con-
242 80-[3
/(3-- - O - - O
lOO - A / t-
.9 /
!
!
/
NMDA
Q.
0/ !
E
.=_ "E 5 0 -
/
o Q.. / O
/
/
/
! /O
10
::
50nM ~ ' ~ °
**
~
100~
'k
L"~ 50
\~E
40--
/
•
o
/ O9 *"
o
E
•
/
/
/
300
Conantokin-G (nM)
r~-"/ O" / / 20 - o' , = / / o . ii / 0 / "
500nM
...--e
t~O--O
I
I
I
102
103
104
Conantokin-G
/
e/
KA 0-
:
60--
ft.
/
¢,-.
Control .... .....D D°o°,°.°.°° o°°°°°~°
(nM)
L
200
I
400
I
600
NMDA (#M)
Fig. 1. A: effect of conantokin-G on agonist-induced elevation of cGMP in rat cerebellar slices in vitro. Slices were incubated as described in the legend to Table I with either NMDA or KA (100 pM) and varying concentrations of conantokin-G, The peptide completely antagonized responses to NMDA but had no effect on the KA-induced cGMP. B: concentration-response curves of the NMDA-induced (50-500 pM) elevation of cGMP in the neonatal rat cerebellum in the presence of the indicated concentration of conantokin-G. As the concentration of the conantokin-G is increased there is a clear flattening of the NMDA concentration-response curve. The inset shows the same data plotted as a conantokin-G concentrationresponse curve and shows that the apparent IC50 of the conantokin-G is unchanged by increasing NMDA concentrations. Both of these characteristics indicate non-competitive inhibition. A similar non-competitive profile was obtained with MK-801 in this assay, although this compound was 2-fold more potent than conantokin-G (see Table I). The results are from a representative experiment which was repeated three times. The ICs0's for the conantokin-G varied less than 10 % among these different experiments.
trast, conantokin-G has no significant effect on [3H]TCP binding at concentrations of up to 2 pM (Table I). The lack of activity in this assay clearly distinguishes this naturally occurring peptide from other 'non-competitive' NMDA antagonists. In order to determine if conantokin-G was interacting with a known binding site, we examined this peptide's had affinity for various Glu binding or uptakes sites, for non-strychnine-sensitive [3H]glycine (Gly) binding sites, or for binding sites for several other CNS neurotransmitters/neuromodulators. Conantokin-G (1 pM) had no effect on C1--dependent Glu binding (a putative anion dependent Glu uptake mechanism) [2], Cl--independent Glu binding (largely postsynaptic receptors) [12], Na +dependent Glu binding (binding to the Na+-dependent Glu carrier protein) [1] or on Na+-dependent glutamate uptake into rat brain synaptosomes. However, conantokin-G had the unexpected effect of enhancing strychnine-insensitive [3H]GIy binding to rat forebrain membranes in a concentration-dependent manner (Fig. 2). The enhancement was maximal at approximately 750 nM and had an apparent EDs0 of 152 _ 14 nM (n = 3). The binding was more than doubled in membranes prepared from forebrain but increased only 7.8 % in spinal cord membranes (Fig. 2), a result consistent with the distribution of NMDA-associated [3H]GIy receptors [3]. The peptide did not reveal a pharmacologically distinct [3H]GIy binding site since the ICs0's of Gly, D-Ser and
kynurenic acid were unchanged in this assay by conantokin-G (212 _ 16 nM, 269 -t- 55 nM and 8.3 _ 1.4 pM, respectively). Neither competitive nor non-competitive NMDA antagonists mimic this effect on [3H]Gly binding. Conantokin-G (at 1 pM) also displayed no affinity for the binding sites for KA, amino-3-hydroxy-5-methylisoxazole-4propionic acid (AMPA) benzodiazepine, ~q-adrenergic, ct2-adrenergic, 7-aminobutyric acid (GABA), serotonin 5-HTla, serotonin 5-HT2 or muscarinic nicotinic receptors (data not shown). The similarity between the behavioral effects of i.c.v. administered conantokin-G and i.p. administered PCP in mice led us to investigate the possible Glu antagonist activity of this peptide. Our results show that the 17 amino acid conantokin-G isolated from the venom of the predatory snail Conus geographus blocks NMDAinduced elevations of cGMP in the rat cerebellum in vitro. This blockade is not a non-specific effect on cGMP generation since it had no effect on KA-induced increases in cGMP. Thus, conantokin-G appears to be a potent NMDA antagonist selective for the NMDA class of receptors. This peptide does not act directly at the Glu receptor since it was unable to block [3H]GIu binding to postsynaptic receptors. Additionally, it had no effect on [3H]KA or [3H]AMPA binding. Conantokin-G also had no significant affinity for receptors of several other neurotransmitters.
243 TABLE I
_
THE EFFECT OF CONANTOKIN-G AND STANDARD N M D A ANTAGONISTS ON cGMP PRODUCTION AND [3H]TCP BINDING.
I~ 200 ._c "O .¢_
The elevation of cGMP induced by excitatory amino acid agonists in neonatal rat cerebeUar slices was carried out essentially as described by Foster ~ind Roberts [5] and Garthwaite and Garthwaite [6]. Basal levels of cGMP were 0.54 pmol/mg protein, in the presence of either NMDA or KA (100 gM) the levels of cGMP varied from 65 to 80 pmol/mg protein among the different experiments. [3H]TCP (New England Nuclear, 48.9 Ci/mmol) binding was carried out with membranes prepared from rat forebrain [4] which were sonicated and rinsed 5 times with 5 mM Tris-HCl, pH 7.5. The incubation medium contained 1 mg membranes in 1 ml 5 mM Tris-HC1 and 2 nM [3H]TCP. Following 30 min at 30°C, the assay was terminated by filtration through GF/B filters (Whatman). Non-specific binding was determined in the presence of 100/zM PCP. All values are the mean + S.E.M. of at least 3 determinations.
c
Compound
NMDA-induced cGMP (IC50, nM)
KA-induced [3H]TCP binding cGMP (IC50, Ki(nM) nM)
MK-801 PCP AP7 Conantokin-G
78__+15 184___26 3,200 + 350 171 + 28
> 5,000 >5,000 > 100,000 > 3,200
64+ 16 212___48 5,500 + 290 > 2,000
Some additional results, obtained in collaboration with others, are consistent with conantokin-G functioning as an NMDA antagonist. The peptide reduces the increase in Fura-2 fluorescence elicited when NMDA and Gly are applied to cerebellar granule cells in culture [8]. In addition, conantokin-G antagonizes certain effects of NMDA or Glu in vivo. These include the ability of NMDA to induce seizures, and the effects of glutamate on frequency modulation of the pacemaker nucleus in the central nervous system of weakly electric fish (G. Rose, M. Mclntosh and B. Olivera, unpublished observations). Clearly, additional experiments, both biochemical and electrophysiological, are necessary to define further the mechanism by which this peptide exerts its NMDA antagonist activity. Efforts to classify this peptide into one of the established classes of Glu antagonists were unsuccessful. Conantokin-G blocked the NMDA-induced increases in cGMP in a non-competitive manner. However, in contrast to other non-competitive antagonists, it does not bind to the [3H]TCP site. The ability of this peptide to induce an apparent enhancement of [3H]GIy binding also underscores its uniqueness. Gly has been shown to be a positive modulator of NMDA activity in several systems [10]. It is unclear, however, how enhancement of [3H]GIy binding could result in NMDA antagonism. The enhancement is observed only in forebrain membranes
.~
.........
~ Forebrain
/
,D
150
,~
~s.ii,3._..___,.._._ili
- ] ; Spinal Cord
100~
I~ 5o l
1000 Conantokin-G
I
2000 (nM)
I
3000
Fig. 2. Conantokin-G enhances [3H]Gly binding to rat forebrain membranes but not to spinal cord membranes. Crude synaptosomes (P2) were prepared identically from each tissue and lysed. [3H]GIy binding was carried out essentially as described by Kessler et al. [11]. Membranes were incubated for 30 min at 0°C with 100 nM [3H]GIy. The assay was terminated by filtering and rinsing with ice cold buffer for 5 s. Non-specific binding was measured by including 100/iM nonradioactive Gly in the assay. Each point is the mean + S.E.M. of 2 4 experiments.
and not spinal cord membranes, which lack strychnineinsensitive Gly receptors [3]. Recently, we observed that another novel class of NMDA receptor antagonist, the arylamine toxins from spider venoms, also display this effect on [3H]GIy binding [7]. Detailed studies reveal that the increased binding observed in the assay results from a decrease of the association and dissociation rate of [3H]Gly with its receptor and does not alter the Kd of this interaction [7]. If conantokin-G and the arylamines from spider venoms have a similar mechanism, their apparent antagonism may result from an impaired ability of Gly to positively modulate the NMDA receptor. Why does this snail, which has a diet restricted exclusively to fish, have a peptide which functions as a Glu antagonist as a component of its venom? Although a target for this peptide in peripheral systems of fish is unknown, intramuscular injection of conantokin-G causes jerky and uncoordinated swimming which may slow the fish's escape (D. Yoshikami and B. Olivera, unpublished observations). Alternatively, the peptide may act in concert with other components of the venom to aid in the culinary pursuits of this snail. A final possibility is that some components of the venom are reserved for defensive measures. Conus geographus attempts to sting when handled and has, in some cases, administered fatal doses of toxin to unprotected humans [14]. Conus geographus is a thin-shelled mollusk which may be a tempting and vulnerable target for several species of crustacean. Since Glu is a neurotransmitter at the invertebrate neuromuscular junction, injection of a potent glutamate antagonist into an invertebrate predator may
244
be a very effective defense. While we can not distinguish among these and other possibilities at this time, we have found that this peptide from the venom of Conus geographus is a potent NMDA antagonist in the mammalian CNS. Furthermore, two other peptides homologous to conantokin-G have been purified from other Conus venoms which produce similar behavioral effects in mice. One of these, conantokin-T, blocks NMDA-induced Ca 2+ fluxes in cerebellar cells in vitro [9]. Together these define a family of ~-carboxyglutamate-containing neuropeptides in Conus venoms, the conantokins (from the Philippino word for sleepy, antokin), which may be useful ligands for novel sites on the NMDA receptor complex. This work was supported in part by GM22737. We would like to thank Dr. G.N. Woodruff, Merck and Co., for the gift of MK-801. 1 Baudry, M. and Lynch, G., Characterization of two [3H]glutamate sites in rat hippocampal membranes, J. Neurochem., 36 (1981) 811-820. 2 Bridges, R.J., Nieto-Sampedro, M., Kadri, M. and Cotman, C.W., A novel chloride-dependent L-[3H]glutamate binding site in astrocyte membranes, J. Neurochem., 48 (1987) 1709-1715. 3 Bristow, D.R., Bowery, N.G. and Woodruff, G.N., Light microscopic autoradiographic localization of 3H glycine and 3H strychnine binding sites in rat brain, Eur. J. Pharmacol., 126 (1986) 303307. 4 Cotman, C.W. and Taylor, D., Isolation and structural studies on synaptic complexes from rat brain, J. Cell Biol., 55 (1972) 696-71 I. 5 Foster, G.A. and Roberts, P.J., Pharmacology of excitatory amino acid receptors mediating the stimulation of rat cerebellar cyclic GMP levels in vitro, Life Sci., 27 (1980) 215-221.
6 Garthwaite, J. and Garthwaite, G., Cellular origins of cyclic GMP responses to excitatory amino acid receptor agonists in rat cerebellum in vitro, J. Neurochem., 48 (1987) 29-39. 7 Gullak, M.F., Pagnozzi, M.J., Richter, K.E. and Mena, E.E., Effects of potyamine spider venoms components on [3H]GIu and [3H]Gly binding, Soc. Neurosci. Abstr., 14 (1988) 168. 8 Haack, J., Mena, E., Parks, T.N., Rivier, J., Cruz, L. and Olivera, B.M., Novel peptide NMDA antagonists from the venom of Conus, Soc. Neurosci. Abstr., 15 (1989) 1167. 9 Haack, J.A., Rivier, J., Parks, T., Mena, E.E., Cruz, L.J. and Olivera, B.M., Conantokin-T: a gamma-carboxyglutamate containing peptide with NMDA antagonist activity, J. Biol. Chem., 265 (1990) 6025-6029. 10 Johnson, J.W. and Asher, P., Glycine potentiates the NMDA response in cultured mouse brain neurons, Nature, 325 (1987) 529531. 11 Kessler, M., Terrammani, T., Lynch, G. and Baudry, M., A glycine site associated with N-methyl-D-aspartic receptors: characterization and identification of a new class of antagonists, J. Neurochem., 52 (1989) 1319-1328. 12 Monaghan, D.T., Holets, V.R., Toy, D.W. and Cotman, C.W., Anatomical distribution of four pharmacologically distinct 3H-Lglutamate binding sites, Nature, 176 (1983) 176-179. 13 Olivera, B.M., Mclntosh, J.M., Clark, C., Middlemas, D., Gray, W.R. and Cruz, L.J., A sleeper toxin from Conus geographus venom, Toxicon, 23 (1985) 27%282. 14 Olivera, B.M., Gray, W.R. and Cruz, L.J., Marine snail venoms. In A.T. Tu (Ed.), Handbook of Natural Toxins, Vol. 3, Dekker, New York, 1988, pp. 327-352. 15 Olivera, B.M., Gray, W.R., Zeikus, R., Mclntosh, J.M., Varga, J., Rivier, J., de Santos, V. and Cruz, L.J., Peptide neurotoxins from fish-hunting cone snails, Science, 230 (1985) 1338-1343. 16 Rivier, J., Galyean, R., Simon, L., Cruz, L., Olivera, B.M. and Gray, W.R,, Total synthesis and further characterization of the gamma-carboxyglutamate-containing'sleeper' peptide from Conus geographus, Biochemistry, 26 (1987) 8508-8512.