Interaction of gephyrotoxin and indolizidine alkaloids with the nicotinic acetylcholine receptorion channel complex of torpedo electroplax

Neurochemical Research, Vol. 11, No. 8, pp. 1227-1240

INTERACTION OF GEPHYROTOXIN AND INDOLIZIDINE ALKALOIDS WITH THE

NICOTINIC ACETYLCHOLINE RECEPTOR- ION C H A N N E L COMPLEX OF TORPEDO

ELECTROPLAX

ROBERT S . ARONSTAM 1, JOHN W . D A L Y 2, THOMAS F .

SPANDE 3, TANJORE K . N A R A Y A N A N 1, AND ED S O N X .

A L B U Q U E R Q U E 3 1 Department of Pharmacology and Toxicology

Medical College of Georgia Augusta, GA 30912

2 Laboratory of Bioorganic Chemistry Bldg. 4, Room 212

National Institute of Arthritis, Diabetes and Digestive and Kidney Diseases National Institutes of Health, Bethesda, MD 20205

3 Department of Pharmacology and Experimental Therapeutics University of Maryland, School of Medicine

Baltimore, MD 20201

Accepted February 11, 1986

The interactions of eighteen natural and synthetic gephyrotoxin and indolizidine alkaloids with binding sites on nicotinic acetylcholine receptor channel (AChR) complex from Torpedo californica electric organ were investigated using two radiolabeled probes, [3H]perhydrohistrionicotoxin and [3H]phencyclidine. Both gephyrotoxins and indolizidines were moderately active inhibitors of the binding of these probes (K~'s = 0.1-20 ~M), but did not interact with the acetylcholine binding site. Structure-activity relationships indicate an important contribution of hydrophobic interactions to both gephyrotoxin and indolizidine binding. The stereoconfiguration of the alkaloids had little effect on binding. Carbamylcholine en- hanced the affinity of certain alkaloids up to 6 to 8-fold suggesting that interactions with open or desensitized conformations of the AChR complex are favored over interactions with resting conformations.

1227 0364-3190/86/0800-1227505.00/0 �9 1986 Plenum Publishing Corporatiol

1228 ARONSTAM ET AL.

INTRODUCTION

Gephyrotoxin is the parent member of a class of tricyclic perhydroben- zoindolizidine alkaloids isolated from the skin secretions of the Colombian poison frogs Dendrobates histrionicus and Dendrobates occultator (Fig- ure 1) (8-12). The name gephyrotoxin (Greek: gephyra meaning bridge) was derived in reference to the bridge presumably formed by the addition of the nitrogen function to one side chain of a postulated precursor 2,6- disubstituted piperidine to form a bicyclic indolizidine (see ref. 9). The Colombian poison frogs also produce a variety of indolizidines, some of which were formerly referred to as bicyclic gephyrotoxins (8, 9, 26), but will henceforth be referred to as indolizidines (11). The poison frogs of the family Dendrobatidae have elaborated a wide range of alkaloids for use in chemical defense against predators. Because of the large number of such dendrobatid alkaloids, a system of designation using boldface numbers corresponding to the nominal mass followed by identifying letter(s) for each alkaloid has been developed (8, 9). The dendrobatid alkaloids, which include batrachotoxins, histrionicotoxins, pumiliotoxins, and gephyrotoxins have remarkable effects on ion conductance of che- mosensitive and voltage-sensitive channels of nerve and muscle (see ref. 8).

Gephyrotoxin itself has weak antimuscarinic actions: it inhibits acetylcholine-, but not histamine-induced contractions of guinea pig ileum, and reduces the responses of guinea pig atria to acetylcholine (21). Re- cently, gephyrotoxin (GyTX, compound 1) has been shown to have complex effects on transmission at the neuromuscular junction (24, 25). GyTX blocked the nicotinic receptor complex in the open channel conformation, and stabilized a high affinity, desensitized conformation of the complex. GyTX inhibited ligand binding to sites associated with the receptor-gated ion channel without inhibiting ligand binding to the acetylcholine recognition site.

In the present study the interactions of eighteen natural and synthetic gephyrotoxin and indolizidine alkaloids with binding sites on the nicotinic acetylcholine receptor channel (AChR) complex from Torpedo californica electric organs were quantitated using two radiolabelled probes, [3H]perhydrohistrionicotoxin ([3H]H12-HTX; 4) and [3H]phencyclidine ([3H]PCP; 13). PCP and H12-HTX are noncompetitive (allosteric) inhibitors of nicotinic neurotransmission (2, 27). They bind with high affinity to sites that are believed to be located within the central hydrophilic depression of the receptor complex, which is probably the site of ion translocation gated by cholinergic agonists (17, 29). A similar investigation of structure-activity relationships in Torpedo californica electric organs

ALKALOIDS AND NICOTINIC RECEPTORS 1229

has been recently reported for histrionicotoxins (5). TorPedo electric organs are derived from striated muscle and possess high concentrations of nicotinic receptors whose biochemical properties closely resemble those found at the neuromuscular junction (3). The nature and mechanisms of noncompetitive neuromuscular blockade by drugs is currently under active investigation in many laboratories (1, 3, 7, 18), especially inlight of recently modeling of the tertiary structure of the AChR complex (see ref. 19 and references therein).

EXPERIMENTAL PROCEDURE

Materials. Natural/-gephyrotoxin,/-indolizidine 223AB, /-indolizidine 239An and/-indolizidine 239CD were isolated from skin extracts of the poison-frog Dendrobates histrionicus essentially as described (9, 11). Additional supplies of these alkaloids have been kindly supplied by Dr. T. Tokuyama (Osaka City University). Synthetic samples of alkaloids have been kindly provided as follows: d-gephyrotoxin (14) by Prof. Y. Kishi (Harvard University); d/-dihydrogephyrotoxin, d/-perhydrogephyrotoxin, dl-desamylgephyrotoxin, d/-(5E,9Z)-3- butyl-5-propylindolizidine, d/-(5Z,9E)-3-(2'-hydroxybutyl)-5-propylindolizidine and its 2'- epihydroxy isomer and d/-(5E,9Z)-3-(2'-hydroxyethyl)-5-propylindolizidine (15, 16) by Dr. D. J. Hart (Ohio State University); d/-perhydrogephyrotoxin (22) by Prof. L. E. Overman (University of California, Irvine); d/-indotizidine 223AB HC1 (see ref. 8) by Dr. O. E. Edwards (National Research Council, Ottawa, Canada) and d/-(5E,9E)-3-butyl-5-propylindolizine (28) by Dr. R. V. Stevens (University of California, Los Angeles). Syntheses of dl-(5E,9Z) and dl-(5Z,9Z)-3,5-dibutylindolizidines, d/-(5E,9Z) and d/-(5Z,9Z)-3-butyl-5-propylindolizidines, dl-(5E,9Z), d/-(5Z,9E) and dl-(5Z,9Z)-3-propyl-5-butyl-indolizidines and d/-5-propylindolizidine have been briefly described (10, 26). /-Indolizidine 223aB has recently been synthesized (23).

Tissue Preparation. Electric organs from Torpedo californica were purchased from Pacific Biomarine (Venice, CA) and stored at -70~ for up to 4 months before use. Electric organ was minced and then homogenized in a blender in 5 volumes of 50 mM Tris-C1, pH 7,4, containing 1 mM pbenylmethylsulfonyl fluoride to prevent proteolysis. The homogenate was filtered through 4 layers of cheese cloth to remove undisrupted material and then centrifuged at 20,000 g for 20 min. The pellets were resuspended in 50 mM Tris-C1 buffer and used without further treatment. Protein content was estimated by the method of Lowry et al. (20).

Binding Measurements. [a25I]-Alpha-bungarotoxin ([125I]BGT, 10-20 ~Ci/~zg, New Eng- land Nuclear) was used as a probe for receptor (i.e., ACh) binding sites. Torpedo membranes (5-15 ~g protein) were incubated with competing ligands (alkaloid or carbamylcholine) for 15 rain at room temperature. [125I]BGT (5 nM) was then added and the incubation con- tinued for 20 min. The reaction was then quenched by the addition of 0.5 volumes of a solution containing 5 mg/ml of methylated bovine serum albumin (Sigma). The suspensions were filtered on glass fiber filters (Whatman GF/B) which had been soaked in the albumin solution. The filters were washed once with 5 ml buffer, and their radioactivity content determined by liquid scintillation counting. Nonspecific binding was determined in the presence of 30 p~M d-tubocurarine.

Ion channel interactions were measured using two radiolabeled probes: [3H]Phencyclidine (48 Ci/mmol) was purchased from New England Nuclear, and [3H]perhydrohistrionicotoxin

1 2 3 0 ARONSTAM ET AL.

([3H]H12-HTX; 54.5 Ci/mmol) was prepared by reduction of 6.6 mg of dl-octa-hydrohistrion- icotoxin with 25 Ci of tritium gas by New England Nuclear. Torpedo membranes (20-50 I~g protein) were incubated with 2 nM [3H]H12-HTX or 3 nM [3H]PCP in the absence or presence of competing gephyrotoxins or indolizidines in 50 mM Tris-HCl, pH 7.4, in a final volume of 1 ml at room temperature for 60 min. The suspensions were filtered on glass fiber filters (Whatman GF/B), and the filters were washed with 5 ml buffer and their radioactivity content determined by liquid scintillation counting. The filters were pretreated with a 1% organo- silane solution (Sigmacote, Sigma) to eliminate [3H]HIz-HTX binding to the filters. Non- specific binding was determined in the presence of 100 IxM unlabeled histrionicotoxin,

Receptor and ion channel binding values were determined in triplicate in each experiment, and each experiment was performed 3 times using tissue prepared from different fish, In- hibition constants (Ki) for the gephyrotoxins and indolizidines were calculated from the concentrations which inhibited specific probe binding by 50% (i.e., the IC5o value) according to the relationship

Ki = IC5o/(1 + D/KD),

where D and KD are the concentration and dissociation constants of the radiolabeled probe ([3H]H1z-HTX or [3H]PCP). The KD'S of [3H]HI2-HTX were 0.089 and 0.135 txM in the presence and absence of I txM carbamylcholine. The corresponding values for [3H]PCP were 0.097 and 0.54 IxM.

RESULTS

Several binding functions were examined with each of the gephyrotoxins and indolizidines: 1) the ability to inhibit [ lzSI]BGT binding to the ACh binding site, 2) the ability to inhibit [3H]H12-HTX and [3H]PCP binding to sites associated with the receptor-gated ion channel, and 3) the susceptibility of their ion channel interactions to allosteric control by receptor agonists. This last function was examined by determining the influence of I p~M carbamylcholine on affinity of gephyrotoxins and indolizidines for binding sites on the AChR complex (4).

None of the gephyrotoxin or indolizidines inhibited [lzSI]BGT binding to the nicotinic receptor by more than 15% at the highest concentration tested (32-70 txM) (Figure 2). Under these same conditions, carbamylcholine inhibited [125I]BGT with an IC5o of 2.8 IxM.

The Ki values for inhibition of [3H]H12-HTX and [3H]PCP binding by gephyrotoxins and indolizidines are listed in Table I. Binding isotherms for gephyrotoxin interactions with [3H]Hlz-HTX binding sites are pre- sented in Figures 3 and 4. In most cases, the gephyrotoxins and indolizidines had a somewhat (up to 2-3 fold) higher affinity for [3H]PCP compared to [3H]H~z-HTX binding sites.

Among the tricyclic gephyrotoxins, the natural l and synthetic d-enantiomers of GyTX (compounds 1 and 2) did not differ in their affinity for binding sites (1Cso'S = 1.6-2.5 ixM), In the presence of 1 IxM car-


H, R

,,,H

CH2 (~H2OH

G E P H Y R O T O X l N S

1 ( I ) and 2{~d) a R : / / , ~ /

4 (d') R :

5 (all) R : ,~H

R 3 - ~ J

I N D O L I Z I D I N E S

R 1 R 2 R 3

e (o,) .... H .... \ A

(~,) .... H ~ A ~ .

8 (all) b " H .... ~ /

9 ( , , ) .... H .... v ~ .

1 0 I " ) " 'H .... ~ / ~

11 (d,) .... H ~

12 ~d,) .... H .... ~ ....

13 (d') .... H ~ - -H

14 (d' l .... H .... X / .... N 0 ~

15 (d') .... H , " X / .... " ~ OH

16 (d,) .... H ,"%/ "~",-..- O H

17 ( , ) -" H ,,"X/ ~ , / , ~ I O H

18 ( ' ) ' ' H .... ~ O H

FIG. 1. Structures of gephyrotoxin and indolizidine alkaloids, a The structure of naturally- occurring,/-enantiomer of GyTX (compound 1) was established by x-ray analysis (12). Com- pound 2 was later synthesized by Fujimoto and Kishi from L-pyroglutamic acid in a reaction sequence designed to arrive at the absolute configuration of compound 1 (14). Inexplicably, this synthesis gave the d-enantiomer. The structure depicted is that obtained by x-ray analysis, b The absolute configuration depicted for 8 is that of natural/-indolizidine 223AB (23) as are the configurations of natural indolizidines 239A~ and 239cD (17, 18).


1.0

~5

0.5 m

I I I I q

m

-7 -6 -5 -4

Log Concentration, M

Fro. 2. Influence of gephyrotoxin and indolizidine alkaloids on [ IaSI]BGT binding to Torpedo ACh receptors. [125I]BGT (5 nM) binding was measured in the presence of the indicated concentrations of/-gephyrotoxin (compound 1, A), dl-perhydrogephyrotoxin (compound 4, 0), (5E,9Z,)-3,5-dibutylindolizidine (compound 6, A), and carbamylcholine (0). Binding is expressed as fraction of control binding measured in the absence of gephyrotoxins or carbamylcholine. None of the 18 alkaloids depicted in Figure 1 inhibited specific [125I]BGT binding by more than 15% at the highest concentration tested (either 32 or 70 txM, depending on available supplies). Each point represents the mean from 3 experiments which varied by less than 15% of specific [125I]BGT binding measured in the absence of competing ligands.

bamylcholine, affinity for binding sites on the AChR complex was increased 2.7-6.3 fold. These were among the largest increases seen with any of the present alkaloids. An increase in affinity for both the [3H]Hle- HTX and [3H]PCP binding sites was associated with a decrease in the degree of unsaturation in the five-carbon side chain. Dihydrogephyrotoxin (compound 3) had 2-3 fold higher affinity for binding sites on the AChR complex measured in the absence of carbamylcholine. In the presence of carbamylcholine, the increases in affinity (compared to the parent compounds 1 and 2) were somewhat less. Complete saturation of the five- carbon side chain yielded the compound with the highest affinity for ion channel binding sites (perhydrogephyrotoxin, compound 4). Removal of the five-carbon side chain from the tricyclic nucleus (compound 5) re-


�9

>~ [..,

-z E- �9

" r [ j

~ k

m O

�9

�9 g <

[..., z

+

g

;>< [ -

o

o

o o o o - - o o o o o

§ 2 4 7 2 4 7 2 4 7 2 4 7 2 4 7 2 4 7 +1 +1 +1

§ +i +i +1 § § +1 +1 § §

~ ~ ~ �9 . . . . . . ~

+1 +1 +I +1 +1 +t § +t +1 +1 +1 +1 +1 +1 § +1 +1 +1

+1 +1 +l +1 +1 § § +l +1 +1 § § § +1 +1 +1 +1 +1

o~.~

' ~ .~ .~ .~ .~ ~ ,.~

. ~ . ~ ' ~ - ~ o - ~ ~ ~ ' e ,

'-'~ ~ ~ ~ . ~ ~ " ~, ~ , . o , . . o , . o ~, ~ ~ ~ ' ~ ~ --..-c: ' ~ ~ ' - - ' " ~ r ~ , , ' ~ 0 ~ " 0

�9 ~ ~ o ~ ~ ' , " ~ - - , ' ,.~' , ' ~ . ' . , ~

0 0 O ~ - _ . ,.~ . - , ~ , . . ~ ~ , . ~ ;,-~ r~ ,.o ,"~ ~., ~ , . ~ .~...~.~..~.~ ~

o ~ o ~ % o ~ , ~ . = - ~ ~ = o ~ . ~ ~ ~ ~ . ~ . ~

0~ ~0 , , , , , ~" , ~ T ' " T ' ~ T " , , , "T', "~" ~

~=

.o ~ .~

[J ' -6

' ~ ~ 0 " .~ c~ 00 0.~_~ ~

~ o

~1::~ ~ ~ 0 ~c~ o . ~. ~ , o . m

'~>~ ~ ~ . ~ ~


r ~

._= an

X I-- -r

1 . C m

0 . 5 -

I I I I

- 8 - 7 - 6 - 5

Log [Gephyrotox in] , M

Fro. 3. Binding of gephyrotoxins to ion channel sites of the nicotinic receptor complex from Torpedo californica electric organ. The binding of 2 nM [3H]H12-HTX was measured in the presence of the indicated concentrations of/-gephyrotoxin (compound 1; 0, 0), d/-dihydrogephyrotoxin (compound 3; D, II), d/-perhydrogephyrotoxin (compound 4; 5, A), or dl- desamytgephyrotoxin (compound 5; V, V). Binding was measured in the absence (open symbols) or presence (closed symbols) of 1 IxM carbamylcholine. Binding is expressed as fraction of specific 2 nM [3H]Hle-HTX binding measured in the absence of competing ligands. Each point represents the mean from 3 experiments which varied by less than 15%.

duced affinity at least 10-fold. The binding of desamylgephyrotoxin , how- ever, was st imulated by carbamylchol ine to the greatest extent of any of the compounds tested.

Among the indolizidines, little s tereoselect ivi ty was associated with ion channel interactions. Two s tereoisomers of 3,5-dibutylindolizidine (compounds 6 and 7) had virtually the same affinity for [3H]H12-HTX and [3H]PCP binding sites, as well as the same sensitivity to carbamylcholine. Similarly, the two s tereoisomers of 3-butyl-5-propylindolizidine (compounds 8, and 9) and three s tereoisomers of 3-propyl-5-butylindolizidine (compounds 10, l l , and 12) had similar affinity for [3H]H12-HTX binding sites and the same sensitivity to carbamylchol ine. Only the synthetic race- mate (compound 8) of the natural / - indol iz idine 223AB (3-butyl-5-propylindolizidine) was available, so the effect of d- and l-enantiomers has not been evaluated for the indolizidine alkaloids. It appears likely that as for histr ionicotoxins (5) and tricyclic gephyrotoxins (see above) there will be


re,

x

-r

1.0

0.5

r

\ - \

I - 7

I

\

- 6 - 5

Log ~lndolizidine], M

1.0

ifi x 0.5 T

[ I I

- 7 - 6 - 5

Log [Indolizidine], M

FIG. 4. Binding of indolizidines to ion channel sites of the nicotinic receptor complex from Torpedo californica electric organ. The binding of 2 nM [3H]H~z-HTX was measured in the presence of the indicated concentrations of the indolizidines. Binding was measured in the absence of carbamylcholine and is expressed as fraction of specific 2 nM [3H]H12-HTX binding measured in the absence of competing ligands. Each point represents the mean from 3 experiments which varied by less than 15%. Figure 4A: d/-(5E,9Z)-3,5-dibutylindolizidine (compound 6;0); d/-(5Z,9Z)-3,5-dibutylindolizidine (compound 7; O); d/-(5Z,9E)-3-(2'-hydroxybutyl)-5-propylindolizidine (compound 14; A); d/-(5Z,9E)-3-(2'-(epi) hydroxybutyl)-5- propylindolizidine (compound 15; A): (5E,9Z)-3-(2'-hydroxyethyl)-5-propyl indolizidine (compound 16; II). Fig. 4B: d/-(5E,9E)-3-butyl-5-propylindolizidine (compound 8; 0); dl- (5E,9Z)-3-butyl-5-propylindolizidine (compound 9; O); (5Z,9Z)-3-propyl-5-butylindolizidine (compound 11; A); d/-5-propylindolizidine (compound 13; A); /-indolizidine 239AB (compound 17; R);/-indolizidine 239CD (compound 18; II).

no effect. The precise nature of the alkyl substitutions (propyl or butyl) at C3 and C5 had little influence on affinity (compare compounds 8 and 9 to compounds 10, 11, and 12). Removal of one of the side chains (C3- butyl) resulted in a compound with diminished affinity (compound 13 compared to compounds 8 and 9).

The presence of an hydroxyl group on the side chain at C3 (compounds 14-17) or the side chain at C5 (compound 18) of the indolizidine was associated with diminished ion channel affinity. A greater decrease was associated with hydroxylation of the C5 side chain (compare compounds 17 and 18 with the parent compound 9). The stereo-configuration of a 2'- hydroxyl group on the C3 butyl side chain had only a slight affect on ion channel binding (compounds 14 and 15): The (epi)hydroxybutyl derivative was about twice as effective in blocking ion channel binding. Shortening the 2-hydroxylated C3 side chain from butyl to ethyl did not appreciably affect ion channel affinity or sensitivity to carbamylcholine (compare compounds 14 and 16). The two natural side chain hydroxylated alkaloids,


or"

I I

6 ] -

5 -

4 -

2 17

1 - / le

/ I I -7 -6

I -5

Log Ki, M

Fie. 5. Relationship of ion channel affinity of gephyrotoxins and indolizidines and carbamylcholine sensitivity. R is the ratio of inhibition constants for [3H]H1z-HTX ion channel binding sites measured in the presence and absence of l p,M carbamylcholine. Thus an R of 2 indicates that the ion channel affinity of the alkaloid is twice as great in the presence of 1 I~M carbamylcholine. R is plotted as a function of the log of the inhibition constant in the absence of carbamylcholine. The individual alkaloids are identified by number (Fig. 1 and Table I). The line is drawn from non-linear regression analysis for the tricyclic gephyrotoxins (compounds 1-5, closed circles). Among these compounds, there is a clear inverse relationship between ion channel affinity and susceptibility to stimulation by carbamylcholine (r = 0.95). Thus, the lower the affinity for the ion channel in the resting state, the greater the stimulation by carbamylcholine. No such relationship is observed with the indolizidines (compounds 6-18; open circles).

n a m e l y i ndo l i z i d ine s 239AB ( c o m p o u n d 17) and 239CB ( c o m p o u n d 18), w e r e o v e r 10-fold l ess a c t i ve t h a n the p a r e n t indo l i z id ine 223AB ( c o m p o u n d 10).

A m o n g the g e p h y r o t o x i n s , t h e r e was a c l ea r i n v e r s e r e l a t i onsh ip be- t w e e n the af f in i ty for the [3H]HI2 -HTX and [3H]PCP b ind ing s i tes and the d e g r e e o f b ind ing s t i m u l a t i o n b y c a r b a m y l c h o l i n e (F igure 5). Thus , in t e r m s o f a f f in i ty for b ind ing s i tes 4 > 3 > 2 /> 1 > 5. In t e r m s o f the deg ree o f i n c r e a s e in a f f in i ty b r o u g h t a b o u t b y c a r b a m y l c h o l i n e , 5 > 1 > 2 > 3 > 4 . Th is r e l a t i o n s h i p b e t w e e n af f in i ty and sens i t iv i ty to a l lo s t e r i c r egu la t i on b y the r e c e p t o r d id no t ho ld for the indo l i z id ines (F igure 5).


The affinity of the indolizidines were increased by approximately the same extent (1.2 to 2.2-fold) by carbamylcholine in all cases except for binding of compound of 14 to the [3H]PCP site, where carbamylcholine increased affinity by 3.6-fold. Carbamylcholine affected the affinity of 1A4 to a greater extent with respect to [3H]PCP binding than to [3H]Hlz-HTX binding (1.3- fold).

DISCUSSION

Interactions of eighteen gephyr0toxin and indolizidine alkaloids with the nicotinic acetylcholine receptor complex have been evaluated using two radiolabeled probes ([3H]PCP and [3H]Hlz-HTX) for sites associated with the ion translocation mechanism (channel) that is gated by cholinergic agonists. Antagonism by gephyr0toxins and indolizidines could be by direct competition or the result of allosteric alterations of channel structures. An alternate possibility is that the various classesof ligands bind to different domains of a single site, thereby modifying binding at the other domains. This last possibility is particularly attractive in light of the prevalence of differences in the ability of numerous compounds to inhibit [3H]Hlz-HTX vs [3H]PCP binding. Additionally, [3H]PCP binding, and the ability of other compounds to inhibit [3H]PCP binding, is influ- enced to a greater extent by receptor agonists than are [3H]Htz-HTX binding site interactions. There is certainly sufficient data, both in binding (5) and electrophysiological (1) studies, to Support the conclusion that [3H]lz-HTX and [3H]PCP binding sites are not identical. On the other hand, no compounds havebeen identified that interact with one site With- out affecting the other (2, 5, 13, 24).

The properties of drugs which block nicotinic transmission as a con- sequence of interference with receptor-gatedion'tl~ansport processes, as opposed to receptor ligand recognition functions, have been most clearly delineated in electrophysiological studies at the neuromuscular junction. These properties depend on whether the drug blocks the'receptor-channel complex in the open or closed conformations (1). PCP and H~2-HTX ap- pear to interact-with receptor ion channels in both the open and closed conformations (1, 27). On the other hand, the lack of voltage- and time2 dependent blockade of neuromuscular transmission by GyTX suggests that it does not interact with the receptor-channel complex in the closed conformation (25). GyTX decreases channel-lifetime (but-not conductance) and its actions are "use dependent"; i.e., they are potentiated by previous channel openings. This suggests that GyTX interacts with the receptor-channel complex in mainly the open conformation.


Another mechanism of some generality in the blockade of neuromuscular transmission is the stabilization of desensitized conformations of the receptor-channel complex. A number of noncompetitive blockers pro- mote receptor desensitization, including histrionicotoxins (1, 6), phencyclidine (2) and gephyrotoxin (24). The relationship between enhancement of desensitization and ion channel blockade is not yet clear. A comparison of structure activity relationships for gephyrotoxin blockade of [3H]H12-HTX and [3H]PCP binding and stabilization of desensitized receptor conformations may shed some light on this question.

In the present study, gephyrotoxin and structurally related indolizidines were found to be moderately active inhibitors of [3H]Hlz-HTX and [3H]PCP binding to sites on the AChR complex. The structure-activity relationships suggest an important contribution of hydrophobic interactions. For example, removal of an alkyl chain from either a gephyrotoxin or an indolizidine reduces affinity (compounds 5 vs l and 13 vs 8, respectively). Reduction of a side chain (compound 4 vs 1) increases affinity, and introduction of a hydroxyl moiety in a side chain (compounds 14-18) decreases affinity, again emphasizing the importance of hydrophobic interactions. The precise stereoconfiguration of the alkaloid had little influence on binding. In general, the alkaloids had a slightly higher affinity for [3H]PCP compared to [3H]Hlz-HTX binding sites, and their interactions with the [3H]PCP binding sites were more profoflndly affected by carbamylcholine. Indolizidines, as yet, have not been examined as noncompetitive blockers of nicotinic receptors, but the present data show- ing high affinity as ligands for AChR sites suggests that these relatively simple compounds may be useful probes.

Carbamylcholine increases the affinities of [3H]Hlz-HTX and [3H]PCP for their binding sites by up to 1.5 and 5.4 fold, respectively, without altering the number of binding sites (5). This stimulation has been inter- preted as indicating a removal of steric hindrances that restrict the access of channel ligands to their hydrophobic binding sites (4, 17). These steric hindrances are thought to be eliminated upon assumption of activated and/or desensitized conformations by the receptor complex. In the present study, desamylgephyrotoxin and d and l GyTX (compounds 5, 1, and 2) were among the most profoundly affected by carbamylcholine.

The present data on high activity for perhydrogephyrotoxin and for the indolizidines with two alkyl side chains suggests that synthesis of gephyrotoxin and indolizidine analogues with longer saturated alkyl chains may result in more potent noncompetitive blockers that are less sensitive to allosteric regulation by receptor agonists. If susceptibility of such ligands to allosteric regulation by receptor agonists is related to "use"


dependent effects on neurotransmission (25), such a compound may have a radically different spectrum of electrophysiological effects.

ACKNOWLEDGMENTS

This work is supported by NIH grants HL-31518 and DA-03303. A number of scientists (see EXPERIMENTAL PROCEDURES) have generously supplied natural and synthetic material and they are gratefully acknowledged.

REFERENCES

1. ALBUQUERQUE, E. X., ADLER, M., SPIVAK, C. E., and AGUAYO, L. 1980. Mechanism of nicotinic channel activation and blockade. Ann. N.Y. Acad. Sci. 358:204-238.

2. ALBUQUERQUE, E. X., TSAI, M.-C., ARONSTAM, R. S., ELDEFRAWl, A. T., and ELI)E- FRAWl, M. E. 1980. Sites of action of phencyclidine. II. Interaction with the ionic channel of the nicotinic receptor. Molec. Pharmacol. 18:167-178.

3. ARONSTAM, R. S. 1982. The Torpedo electromotor system. Pages 45-78, in HANIN, I., and GOLDBER~, A. M. (eds.), Progress in Cholinergic Biology: Model Cholinergic Syn- apses, Raven Press, New York.

4. ARONSTAM, R. S., ELDEFRAWI, A. T., PESSAH, I. N., DALY, J. W., ALBUQUERQUE, E. X., and ELDEFRAWl, M. E. 1981. Regulation of [3H]perhydrohistrionicotoxin binding to Torpedo ocellata electroplax by effectors of the acetylcholine receptor. J. Biol. Chem. 256:2843-2850.

5. ARONSTAM, R. S., KING, C. T., ALBUQUERQUE, E. X., DALY, J. W., and FEIGL, D. M. 1985. Binding of [3H]perhydrohistrionicotoxin and [3H]phencyclidine to the nicotinic receptor ion channel complex of Torpedo electroplax: Inhibition by histrionicotoxins and derivatives. Biochem. Pharmacol. 34:3037-3047.

6. BURGERMEISTER, W., CATTERALL, W. A., and WITKOP, B. 1977. Histrionicotoxin en- hances agonist-induced desensitization of acetylcholine receptor. Proc. Natl. Acad. Sci. USA. 74:5754-5758.

7. CHANGEUX, J.-P., DEVILLERS-THIERY, A., and CHEMOUILLI, P. 1984. Acetylcholine receptor: An allosteric protein. Science 225:1335-1345.

8. DALY, J. W. 1982. Alkaloids of neotropical poison frogs (Dendrobatidae). Prog. Chem. Org. Nat. Prod. 41:205-340.

9. DALY, J. W., BROWN, G. B., MENSAH-DWUMAH, M., and MYERS, C. W. 1978. Classi- fication of skin alkaloids from neotropical poison-dart frogs (Dendrobatidae). Toxicon 16:163-188, 1978.

t0. DALY, J. W., and SPA~DE, T. F. 1986. Chemistry, pharmacology and biology of alkaloids from amphibians. Pages 1-t29, in PELLETmR, S. W. (ed.), The alkaloids: Chemical and biological perspectives, Vol. 4, John Wiley & Sons, New York.

11. DALY, J. W., WHITTAKER, N., SPANDE, T. F., HIGHET, R. J., FEIGL, D., NISHIMORI, N., TOKUYUMA, T., and MYERS, C. W. 1986. Alkaloids from dendrobatid frogs: Struc- tures of two w-hydroxy congeners of 3-butyl-5-propylindolizidine and occurence of 2,5- disubstituted pyrrolidines and a 2,6-disubstituted piperidine. J. Nat. Products Chem. (Llyodia), in press.


12. DALY, J. W., WITKOP, B., TOKUYAMA, T., NISHIKAWA, T., and KARLE, I. L. 1977. Gephyrotoxins, histrionicotoxins and pumiliotoxins from the neotropical flog Dendro- bates histrionicus. Helv. Chim. Acta 60:1128-1140.

13. ELDEFRAWI, M. E., ELDEFRAWI, A. T., ARONSTAM, R. S., MALEQUE, M. A., WARNICK, J. E., and ALBUQUERQUE, E. X. 1980. [3H]Phencyclidine--a probe of the ionic channel of the nicotinic receptor. Proc. Natl. Acad. Sci. USA 77:7458-7462.

14. FUJIMOTO, R., and KISHI, Y. 1981. On the absolute configuration of gephyrotoxin. Te- trahedron Lett. 1981:4197-4198.

15. HART, D. J., and KANAI, K. 1983. Total synthesis of dl-gephyrotoxin and d/-dihydrogephyrotoxin. J. Am. Chem. Soc. 105:1255-1263.

16. HART, D. J., and TSAI, Y.-M. 1982. Stereoselective indolizidine synthesis: Preparation of stereoisomers of gephyrotoxin-223AB. J. Org. Chem. 47:4403-4409.

17. HEIDMANN, T., OSWALD, R. E., and CHANGEUX, J.-P. 1983. Multiple sites of action for noncompetitive blockers on acetylcholine receptor rich membrane fragments from Tor- pedo marmorata. Biochemistry 22:3112-3127.

18. KARLIN, A. 1983. The anatomy of a receptor. Neurosci. Commentaries 1:111-123. 19. KOSOWER, E. M. 1984. Revised assignments for the cx-, 13- and ~/-subunits of the ace-

tylcholine receptor structural model. FEBS Lett. 172:1-5. 20. LowRY, O. H., ROSEBROUGH, N. J., FARR, A. L., and RANDALL, R. J. 1951. Protein

measurement with the Folio phenol reagent. J. Biol. Chem. 193:265-275. 21. MENSAH-DWUMAH, M., and DALY, J. W. 1978. Pharmacological activity of alkaloids

from poison-dart frogs (Dendrobatidae). Toxicon 16:189-194. 22. OVERMAN, L. E., and FREERKS, R. L. 1981. Short total synthesis of (_~)-perhydroge-

phyrotoxin. J. Org. Chem. 13:2833-2835. 23. ROYER, J., and HUSSON, H.-P. 1985. Asymmetric synthesis III. Enantiospecific synthesis

of the natural DR, #5R, #9R, ( - ) gephyrotoxin-223 AB. Tetrahedron Lett. 26:1515- 1518.

24. SOUCCAR, C., VARANDA, W. A., ARONSTAM, R. S., DALY, J. W., and ALBUQUERQUE, E. X. 1984. Interactions of gephyrotoxin with the acetylcholine receptor-ionic channel complex. II. Enhancement of desensitization. Molec. Pharmacol. 25:395-400.

25. SOUCCAR, C., VARANDA, W. A., DALY, J. W., and ALBUQUERQUE, E. W. 1984. Inter- actions of gephyrotoxin with the acetylcholine receptor ion-channel complex. I. Block- ade of the ionic channel. Molec. Pharmacol. 25:384-394.

26. SVANBE, T. F., DALY, J. W., HART, D. J., TSAI, Y.-M., and MACDONALD, T. L. 1981. The structure of gephyrotoxin (GTX) 223AB. Experientia 37:1242-1245.

27. SPIVAK, C. E., MALEQUE, M. A., OLIVEIRA, A. C., MASUKAWA, L. M., TOKUYAMA, T., DALY, J. W., and ALBUQUERQUE, E. X. 1982. Actions of histrionicotoxin at the ion channel of the nicotinic acetylcholine receptor and at the voltage sensitive ion channel of muscle membranes. Molec. Pharmacol. 21:351-361.

28. STEVENS, R. V., and LEE, A. W. M. 1982. Studies on the stereochemistry ofnucleophilic additions to tetrahydropyridinium salts. A stereospecific total synthesis of one of the stereoisomers of gephyrotoxin 223. J. Chem. Soc. Chem. Commun. 103-104.

29. STROUD, R. M. 1983. Acetylcholine receptor structure. Neurosci. Commen. 1:124-138.

Documents

Interaction of gephyrotoxin and indolizidine alkaloids with the nicotinic acetylcholine receptorion channel complex of torpedo electroplax