The mode of interaction of amiloride and some of its analogues with the adenosine A1 receptor

Neuroehem. Int. Vol. 20, No. 2, pp. 207-213, 1992 Printed in Great Britain. All rights reserved

0197-0186/92 $5.00 + 0.00 Copyright © 1992 Pergamon Press plc

THE MODE OF INTERACTION OF AMILORIDE A N D SOME OF ITS ANALOGUES WITH THE ADENOSINE AI RECEPTOR ANJA GARRITSEN,1 MARGOT W. BEUKERS,~ AD P. IJZERMAN, l* EDWARD J. CRAGOE JR 2 and WILLEM SOUDIJNI ~Division of Medicinal Chemistry, Center for Bio-Pharmaceutical Sciences, State University of Leiden, P.O. Box 9502, 2300 RA Leiden, The Netherlands 2P.O. Box 631548, Nacogdoches, TX, U.S.A. (Received 15 January 1991 ; accepted 31 July 1991)

A~traet--Amiloride, a potassium sparing diuretic, inhibits adenosine A ~receptor-radioligand binding in calf and rat brain membranes in the low micromolar range. The drug interacted with the A~ receptor in a manner different from classical A~ ligands, but structure-activity relationship studies indicated that this inhibitory effect is not related to the ion transport inhibiting properties of amiloride (Garritsen et al., 1990a,b) In the present study, the question is addressed how amiloride interacts with the adenosine A~ receptor. Amiloride and two of its analogues, in concentrations equivalent to their K~values in displacement studies, decrease the affinity of the A~ antagonist [3H]8-cyclopentyl-l,3-dipropylxanthine, but not the maximal binding capacity of the radioligand. Furthermore, the dissociation rate of the receptor ligand complex is unaltered in the presence of amiloride or its analogues in a concentration exceeding the K~ value 10-fold. These characteristics argue for a purely competitive mode of interaction. The functional consequences of the interaction between amiloride analogues and the At receptor were investigated at the level of cyclic adenosine 3',5'-monophosphate (cAMP) formation. The amiloride analogue 5-(N-butyl-N-methyl) amiloride (MBA) reversed A
The adenosine A 1 receptor is a prototypic example of a receptor that is coupled in an inhibitory manner to adenylate cyclase via a pertussis toxin-sensitive G protein (Londos et al., 1978). Inhibition of adenylate cyclase, however, is only one of the possible signal transduction mechanisms for this receptor (Fredholm and Dunwiddie, 1988). Recently, we demonstrated

* Author to whom all correspondence should be addressed. Abbreviations : Bmax,maximal binding capacity ; BSA, bovine

serum albumin; cAMP, cyclic adenosine 3',5'-monophosphate; CPDPX, 8-cyclopentyl-l,3-dipropylxanthine; ICs0, half maximal inhibiting concentration; KB, antagonist potency; KO, equilibrium dissociation constant; Ki, equilibrium inhibition constant; MBA, 5-(N-butyl-Nomethyl)amiloride; PIA, N6-R-l-phenyl2-propyladenosine. 207

that the diuretic amiloride inhibits binding of selective A~ receptor ligands to their binding sites (Garritsen et al., 1990a,b). This phenomenon is not limited to the A] receptor, as amiloride also interacts with the c~2-adrenoceptor, another G~ coupled receptor (Howard et al., 1987 ; Nunnari et al., 1987). This latter observation led to the hypothesis that the inhibition of binding might be related to a direct coupling of the ~2-adrenoceptor to the plasma membrane N a + / H + exchanger (Nunnari et al., 1987). Initially, by analogy with the ct2-adrenoceptor, we postulated that the interaction of amiloride with the At receptor might somehow be related to a coupling of the A1 receptor to a Na +/H + exchanger. However, several amiloride analogues with substituents on a terminal guanidino nitrogen atom have a relatively high affinity for the A ~ receptor, although this type of

208

ANJA GARRITSEN et al.

amiloride analogue is devoid of effect on the N a +/H + exchanger (Garritsen et al., 1990a), arguing against such a link. F u r t h e r characterization of the interaction revealed that arniloride is affected by G T P a n d protein modifiers in a m a n n e r similar to an adenosine antagonist (Garritsen et al., 1990a,b). In c o n t r a s t to the classic A~ antagonists however, the affinity of amiloride is largely d e p e n d e n t on the N a + a n d H + concentration o f the m e d i u m (Garritsen et al., 1990a). This differential regulation, c o m b i n e d with the fact that amiloride is structurally very different from classic A~ receptor antagonists, strengthened the suggestion that the recognition sites of these agents may be different. In the present study, the question whether amiloride acts at a distinct site is addressed by investigating the mode of interaction of amiloride a n d some of its analogues with the At receptor.

EXPERIMENTAL PROCEDURES

Materials Sources of most chemicals were described previously (Garritsen et al., 1990a ; Pirovano et al., 1989). Other compounds included: forskolin (Sigma, St Louis, U.S.A.); 8-phenyltheophylline (Janssen, Beerse, Belgium). Benzamil was synthesized as described previously (Cragoe et al., 1967). The following gifts are gratefully acknowledged: amiloride (Merck Sharp and Dohme, Haarlem, The Netherlands, USP grade) ; 5-(N-butyl-N-methyl)amiloride (MBA) (Dr G. Schmalzing, Max Planck Institute, Frankfurt, Fed. Rep. Germany) and rolipram (Dr N. Sprzagala, Schering AG, Fed. Rep. Germany). Membrane preparations Calf and rat brain membranes were prepared as described by Van Galen et al. (1987) and Lohse et al. (1984), respectively, and incubated with 2 IU/ml adenosine deaminase at 37 :~ for 30 min prior to storage in liquid N2. Rat fat cell membranes were prepared according to Pirovano et al. (1989). Protein concentrations were measured by the bicinchoninic acid method with bovine serum albumin (BSA) as standard (Smith et al., 1985). Measurement of [~H]8-cyclopentyl-l.3-dipropylxanthine ([3H]CPDPX) binding Saturation and displacement experiments were performed in 20 mM HEPES, adjusted to pH 7.4 with Tris as described previously (Garritsen et al., 1990a). Dissociation experiments were performed as follows: [3H]CPDPX (ca 0.12 nM) was incubated in bulk with calf brain membranes (28.5 #g protein/ml) for 60 min at 25. An excess of 8-phenyltheophylline was added at t = 0 (final concentration 5 pM). Test agents were added with 8-phenyltheophylline in a final concentration of 10 × the K, value (Garritsen et al., 1990a). At defined time intervals binding in two 400 #1 samples was measured. The experiments were performed in duplicate for at least three times with similar results.

Adenylate cyclase assay Adenylate cyclase activity was assayed according to Pirovano et al. (1989) in an incubation medium containing 0.1 mM ATP, 10 #M GTP, 5 mM creatine phosphate, 0.4 mg/ml phosphocreatine kinase, 2 mg/ml BSA, 0.1 mM rolipram, 0.2 IU/ml adenosine deaminase, 0.2 mM EGTA, I mM MgC12 and 150 mM NaC1 in 20 mM HEPES pH 7.4 with NaOH. cAMP production was stimulated by 10 #M forskolin (Lohse et al., 1987) for 15 min at 37'. Data analysis Data from saturation and displacement curves were analysed as described previously (IJzerman et al., 1984). The dissociation curves were analysed by a nonlinear curve-fitting program, according to the model described by Lohse et al. (1984). The data from the adenylate cyclase assays were expressed as a percentage of the value obtained with 10 #M forskolin alone and corrected for a time-dependent decrease in the stimulated production. The data from at least three independent experiments were pooled and fitted to the four parameter logistic equation (DeLean et al., 1978). All parameters were varied concomitantly. A KB value for MBA was estimated by calculating a pK~ value for each MBA concentration separately according to the formula pKB = log (CR- 1) - log [MBA], with CR being the ratio between the IC~0 values of N6-R-l-phenyl-2-propyladenosine (PIA) in the absence and presence of MBA. Subsequently, the mean of these pKB values and the corresponding standard error (SE*) were converted to a KB value+SE according to the formula K. = 10 pK, and SE = SE* × K,.

RESULTS The effect o f amiloride and its analogues on the binding parameters c~[[ - ~ H ] C P D P X

In the present experiments, the effects of amiloride and two of its analogues, M B A and benzamil, on adenosine A~ receptor binding were investigated. The drugs were selected to include different types of analogues with potentially different modes o f interaction. Amiloride is the parent c o m p o u n d of the drugs, whereas M B A and benzamil represent two classes o f analogues, namely those with substituents on the 5a m i n o nitrogen atom, which are selective for the plasma m e m b r a n e N a ~ / H + exchanger, a n d the guanidino-substituted derivatives, which inhibit the epithelial sodium channel, respectively ( K l e y m a n and Cragoe, 1988). C o n c e n t r a t i o n s tested were equivalent to the K~ values of the drugs in calf brain, i.e. 2, 0.65 a n d 0.07 /~M for amiloride, benzamil a n d M B A , respectively (Garritsen et al., 1990a,b). As a p p a r e n t from Table 1, the three drugs decreased the affinity o f [3H]CPDPX. Benzamil and amiloride were without effect on the B .... value, whereas M B A slightly increased this parameter. Similar experiments were performed with rat brain cortex membranes. In this tissue, M B A , in a c o n c e n t r a t i o n of 440 n M (the a p p a r e n t K~ value in

Interaction between amiloride and adenosine A, receptors Table 1. Binding parameters of [3H]CPDPX binding to calf brain membranes : influenceof amiloride and its analogues Addition None Amiloride MBA Benzamil

Conc. (/aM)

Kd_+SE (pM)

B,~x+ SE (fmol/mg)

2 0.07 0.65

72 q-3 139 + 6 200_+7 150_+5

800 _+33 808 _+42 901 _+38 832 + 69

Calf brain membranes (11 #g protein) were incubated for 60 rain at 25" with 0.02-1.5 nM [3H]CPDPX (total volume 400/A) with test agents. The values are presented as means+SE of three independent experiments.

rat b r a i n m e m b r a n e s , see Table 2), also reduced the affinity o f the radioligand b u t n o t its m a x i m a l binding capacity. Kd values were 0.18 a n d 0.51 n M in the absence a n d presence o f M B A , whereas the Bmax values in b o t h instances were 571 a n d 575 f m o l / m g protein, respectively.

The effect o f amiloride and its analo#ues on the dissociation rate o f [ 3 H ] C P D P X A l t e r a t i o n s in the dissociation rate m a y be indicative for allosteric interactions as exemplified by the effect o f local anaesthetics o n radioligand binding to the v o l t a g e - d e p e n d e n t sodium c h a n n e l ( P o s t m a a n d Catterall, 1984) a n d the effects o f amiloride analogues on ct2-adrenoceptor b i n d i n g ( N u n n a r i et al., 1987). The dissociation o f the [ 3 H ] C P D P X - c a l f b r a i n receptor complex was initiated by the addition o f 5/~M 8phenyltheophylline. The dissociation rate c o n s t a n t of [3H]CPDPX was 0 . 0 0 9 6 + 0 . 0 0 0 6 m i n ', corres p o n d i n g to a half-life o f the l i g a n d - r e c e p t o r complex o f 74__+4 min. Amiloride or its analogues were added in c o n c e n t r a t i o n s 10-fold higher t h a n their K~ values, i.e. 20, 6.5 a n d 0.7 p M for amiloride, benzamil Table 2. Displacement of [3H]CPDPX binding to rat and calf brain membranes by PIA and MBA : effect of the incubation medium Drug PIA PIA PIA MBA MBA MBA

Cocktail

-

+ +

Animal

Ki+ SE (nM)

Calf Rat Rat Calf Rat Rat

0.32_+0.01 7.9 + 0.8 72 + 7 70 -+4 440 + 20 3500+ 200

Rat brain membranes (23 #g protein) were incubated with ca 0.3 nM [3H]CPDPX for 1 h at 25° in HEPES buffer only or in HEPESbuffersupplementedwith the components of the adenylate cyclase cocktail (composition see ExperimentalProcedures). [3H]CPDPX binding in the presence of cocktail amounted to 70% of the value in the HEPES buffer only. K~values for rat brain were calculated in a one binding site-model with a Kd value of [3H]CPDPX of 0.28 nM (Pirovano et al., 1989). The Ki values in calf brain were calculated from previously published data (Garritsen et al., 1990a).

209

a n d M B A , respectively. N o n e of the c o m p o u n d s affected the 8-phenyltheophylline-induced dissociation rate.

Effect o f M B A on the inhibition o f adenylate cyclase by PIA To further define h o w amiloride interacts with the A . receptor, the effect of amiloride on the inhibition of adenylate cyclase activity by the A t agonist, PIA, was determined. The a d v a n t a g e o f measuring this biochemical response over m o r e d o w n - s t r e a m p h a r macological responses is t h a t it is feasible to directly c o m p a r e the effect of receptor stimulation or inhibition with receptor binding. F o r reasons o f simplicity, cyclic adenosine 3',5'm o n o p h o s p h a t e ( c A M P ) p r o d u c t i o n was m e a s u r e d in fat cell m e m b r a n e s , which have A , receptors, identical to those in rat b r a i n (Lohse et al., 1987; S h a m i m et al., 1989), b u t hardly any A2 receptors which m a y obscure the inhibitory effects o f P I A o n adenylate cyclase in b r a i n (Garcia-Sainz a n d Torner, 1985). As stated above the effects o f M B A o n the binding parameters of [3H]CPDPX in rat were a n a l o g o u s to those in calf, suggesting t h a t amiloride does not interact differently with rat a n d calf A l receptors. Preliminary experiments indicated t h a t amiloride inhibits c A M P p r o d u c t i o n in c o n c e n t r a t i o n s >~ 100 # M (data not shown). We excluded that this was due to interference of amiloride with the c A M P d e t e r m i n a t i o n , e.g. by binding to protein kinase A ( K l e y m a n a n d Cragoe, 1988). Since in adenosine receptor research the K~ a n d KB values are usually close, this interfering inhibitory effect would occur at c o n c e n t r a t i o n s near the antagonistic potency (Ka value) anticipated from the binding assays (ca 20 #M). Such mixed effects would be difficult to interpret a n d therefore amiloride was not the drug o f choice. Instead o f amiloride, M B A was used, a n analogue with higher potency towards the A , receptor (Garritsen et al., 1990a) a n d equal effects o n adenylate cyclase: like amiloride, M B A inhibited c A M P p r o d u c t i o n m a r k e d l y in concentrations ~> 100 # M but h a d n o significant effect in lower concentrations. c A M P f o r m a t i o n was m e a s u r e d in the presence o f graded c o n c e n t r a t i o n s o f P I A a n d four fixed conc e n t r a t i o n s of M B A . P I A inhibited the forskolinstimulated adenylate cyclase activity in a conc e n t r a t i o n - d e p e n d e n t m a n n e r , with a n IC50 value o f 23 ___4 n M a n d a m a x i m a l inhibition o f 35 _+ 1% (Fig. 1). The m a x i m a l forskolin-stimulated p r o d u c t i o n was n o t significantly affected by 3, 10 or 30/~M M B A , n o r was the m a x i m a l inhibition by PIA. The IC50 o f P I A was slightly increased in the presence o f 3 or 10 p M

210

ANJA GARRITSENet al. stituted derivatives were only 1-1.5 times lower (data not shown). Figure 2 and Table 2 illustrate that addition of the adenylate cyclase cocktail further decreased the apparent affinities of MBA and PIA. At the concentration radioligand used (0.3 nM), specific [3H]CPDPX binding in the absence of displacer was diminished by 30 + 3% by the adenylate cyclase cocktail.

110

100

90

D O EK Q_

80

.% 70 DISCUSSION

60

50

I

i

9

8 -LOG

I

I

I

7

6

5

[PIA ( M ) ]

Fig. l. The effect of 0 (O), 30), 10 (V), 30 (V) and 300 #M MBA (U]) on the inhibition of adenylate cyclase in rat fat cell membranes by PIA. The data are expressed as percentage of the cAMP production in the presence of 10 #M forskolin only. Data points are the average of at least three independent experiments. MBA from 23_+4 to 57_+ 14 and 50+_ 17 nM, respectively, whereas 30/~M MBA had a more pronounced effect and raised the ICs0 to 108_+57 nM (Fig. I). Inhibition by PIA in the presence of 300/~M MBA is apparent only in concentrations well over 100 nM. The inhibitory effects of 300 /~M MBA alone, however, thwarted the calculation of an ICs0 value for PIA. As only a limited number of drug concentrations, close to the inhibitory constant of the compound, could be used, analysis of a Schild plot was not feasible. On the assumption, based on the binding data, that competitive antagonism is involved, we calculated, as described in the experimental section, a Ka value of 5.3 -+ 1.9/~M. Bindin 9 studies with rat brain m e m b r a n e s

The potency of MBA in attenuating PIA-mediated inhibition is ca 70 times lower than its K~ value in calf brain membranes (Garritsen et al., 1990a). One possible explanation for this discrepancy is the difference in incubation conditions between the binding assays (performed with calf tissue in a HEPES-Tris buffer without further additions) and the adenylate cyclase assay (performed with rat tissue, in a HEPES buffer, containing an ATP regenerating system, rolipram, ATP, GTP and several ions). Table 2 shows that PIA and MBA are 25- and 6-fold less potent in rat than in calf brain membranes. In a single experiment, other 5-amino-substituted amiloride analogues (Garritsen et al., 1990a) were 3-5 times less potent in rat brain, whereas the affinities of the guanidino-sub-

Amiloride and its analogues interact with adenosine A, receptors in calf brain in a manner resembling that of an antagonist (Garritsen et al., 1990a,b). However, the interaction between amiloride and the A, receptor differs from that with classic A l antagonists with respect to sensitivity to pH and Na + concentration. The present study further explores the mode of interaction and the functional consequences of the amiloride A i receptor interaction. In radioligand binding experiments, the affinity of [3H]CPDPX is reduced by amiloride without an effect on the maximal binding capacity, which suggests that amiloride acts as a competitive inhibitor. The effects of benzamil and MBA are likewise, indicating that there are no obvious differences between the mode of interaction of amiloride and these analogues. This

x'°l

\ \\k \ \.\ \ 11

10

9 -

8

7

6

5

LOG [ DISPLACER ]

Fig. 2. The influence of the adenylate cyclase cocktail on the affinities of PIA and MBA in rat brain membranes. [3H]CPDPX binding was displaced by PIA (A, A) or MBA (O, 0 ) in HEPES buffer alone (A, O) or in HEPES buffer with adenylate cyclase cocktail (for composition see Experimental Procedures) (A, 0). The data points are from a representative experiment and were normalized for differences in binding in the absence of displacer.

Interaction between amiloride and adenosine A ~receptors decrease in the apparent affinities of the compounds is consistent with either of two interpretations : (1) CPDPX and amiloride (or its analogues) compete for the same binding site ; (2) the binding of amiloride (or its analogues) influences the conformation of the A ~ receptor via an allosteric site, such that the accessibility of the ligand binding site is diminished. As allosteric modulators of receptor binding will alter the dissociation rate of a receptor-ligand complex in the presence of a saturating concentration of a true competitor (Postma and Catterall, 1984), the effect of the three compounds on the dissociation rate of [3H]CPDPX was examined. None of the drugs influences the dissociation rate constant in concentrations of 10 x their K~value. Since amiloride and some of its analogues are not very specific agents, the interpretation of effects of higher concentrations is highly speculative and was therefore not determined. These results argue against an allosteric interaction. The data suggest that even if the binding sites of classic ligands and amiloride are not identical, binding of the drug to the receptor is mutually exclusive and the binding sites may overlap to some extent. This differs from the interaction with the ct2-adrenoceptor, where amiloride inhibits receptor binding at least partially in an aUosteric manner (Howard et al., 1987; Nunnari et al., 1987). The ct ~- and fl-adrenoceptor on the other hand are also blocked by amiloride in a competitive manner (Howard et al., 1987). Apparently, the interaction of amiloride with receptors for neurotransmitters is not uniform. Even though receptors may have similar amiloride binding sites, the interaction could be apparently competitive or allosteric in nature, depending on the relative location of the amiloride- and the ligand-binding domain within the receptor molecule. As for the a2-adrenoceptor, both the amiloride- and the ligand-binding site are located in the hydrophobic core of the receptor protein (Horstman et al., 1990). Information on the amiloride binding sites of other receptors is not yet available. The second part of this paper focusses on possible functional aspects of the amiloride-A ~receptor interaction. Stimulation of the At receptor in rat fat cell membranes with PIA results in inhibition of forskolinstimulated cAMP accumulation with an IC 50value for PIA of 23 nM, a value virtually identical to previously published values (Klotz et al., 1985; Londos et al., 1978 ; Ukena et al., 1986). Micromolar concentrations of MBA (compared to a K, value of 70 nM in calf brain membranes) reduce

211

the response to PIA, resulting in a small shift of the concentration response curve to higher concen, trations. The maximum and minimum of the curves are not significantlyaltered by MBA in concentrations up to 30 pM. Higher concentrations of MBA (/> 100 #M) could not be used due to their own inhibitory effects on adenylate cyclase. The latter inhibition is not related to the inhibition of A t receptor binding as amiloride and MBA are equipotent, whereas MBA is ca 30-fold more potent than amiloride in inhibiting A~ receptor binding. The fact that MBA could only be tested in a few concentrations that are close to the apparent KB value, where the error in the concentration ratio is large (Kenakin, 1987), made a Schild-analysis and thus a conclusive interpretation of the data difficult. On the assumption that competitive antagonism is involved (as the data from the binding studies indicate), the Ka value for MBA was estimated to be ca 5 # M . The potency of MBA in antagonizing the PIA response was much lower than in the previously performed binding experiments in calf brain (Garritsen et al., 1990a). One explanation for this discrepancy is that the two events are not related, another could be the fact that different assay conditions are employed. Our results confirm the latter : species differences and incubation conditions can fully account for the difference between K~ and KB values. Amiloride analogues with substituents on the 5-amino nitrogen atom have a higher affinity for calf A ~receptors than for rat A receptors, whereas the amiloride analogues with a substituent on a terminal guanidino nitrogen atom are less sensitive to this species difference. The mode of interaction did not appear to be different between calf brain and rat brain receptors. In the complex incubation medium of the adenylate cyclase assay the apparent affinity of PIA and MBA is decreased to 72 nM and 3.5 pM, respectively. Both values are close to the IC50 of PIA for inhibition of cyclase (23 nM) and the estimated KB value of MBA o f ca 5 pM. NaC1 and GTP may in part be responsible for these decreases in apparent affinity, given their reported effects on the affinities of PIA and MBA (Garritsen et al., 1990a; Pirovano et al., 1989). Anand-Srivastava (1989) described an attenuation of inhibitory hormone responses by amiloride in rat pituitary membranes, caused by an interaction with the guanine nucleotide-bindingregulatory protein, G~. Could such an interaction be an alternative explanation for the attenuation of the PIA response? Amiloride has only a stimulatory effect in the pituitary membranes, in contrast to its effects in the rat fat cell membranes, where amiloride inhibits adenylate

212

ANJA GARRITSEN el al.

cyclase in higher concentrations. A complete abolishment of inhibitory response is observed in the pituitary, whereas we find a rightward shift o f the conc e n t r a t i o n - r e s p o n s e curve. Finally, blockade of G~ c a n n o t be an explanation for the displacement o f [3H]CPDPX binding by amiloride, which is independent o f the coupling o f the A~ receptor to G~ (Garritsen et al., 1990a,b). Thus, the effects of amiloride o n the A ~ receptor a n d o n G~ seem to be distinct. The c o m b i n a t i o n o f adenosine a n t a g o n i s m and G~ inhibition is reminiscent of the dual effects of the cardiotonics sulmazole a n d milrinone a n d o f I B M X and X A C , as reported by Stiles a n d coworkers ( R a m k u m a r a n d Stiles, 1988a,b; Parsons et al., 1988a,b). These drugs are A ~receptor antagonists a n d concomitantly impair the function of G~. Amiloride could also have such a dual effect. This c o m b i n a t i o n of A ~ receptor a n t a g o n i s m a n d the functional blockade of G~ deserves attention in future investigations. Amiloride a n d its analogues a p p e a r to interact with a multitude of receptors (Garritsen et al., 1991). These interactions may contribute to the clinical effects of amiloride, a l t h o u g h the blockade of ion t r a n s p o r t by amiloride is p r o b a b l y the m a j o r m e c h a n i s m of action for its diuretic effect. The widespread amiloride sensitivity of n e u r o t r a n s m i t t e r receptors suggests that some structurally h o m o l o g o u s part of these receptors may be responsible for amiloride binding. In conclusion, the inhibiting effect o f M B A on A~ receptor binding is reflected in a reduction ofA~ receptor-mediated inhibition of adenylate cyclase. The binding experiments performed strongly argue for a purely competitive m o d e of interaction. This suggests t h a t if the binding sites of classical ligands a n d amiloride are not identical they may at least overlap. O t h e r n e u r o t r a n s m i t t e r receptors like the ~2-adrenoceptor might have similar amiloride binding sites with different degrees of overlap, resulting in apparently competitive or allosteric inhibition of receptor binding. A cknowled,qement--The skilful technical assistance of Anton van Weert is gratefully acknowledged.

REFERENCES

Anand-Srivastava M. B. (1989) Amiloride interacts with guanine nucleotide regulatory proteins and attenuates the hormonal inhibition of adenylate cyclase. J. biol. Chem. 264, 9491-9496. Cragoe E. J. Jr, Woltersdorf O. W. Jr, Bicking J. B., Kwong S. F. and Jones J. H. (1967) Pyrazine diuretics If. N-Amidino-3-amino-5-substituted 6-halo-pyrazinecarboxamides. J. Med. Chem. 10, 6{%75. DeLean A., Munson P. J. and Rodbard D. (1978) Simul-

taneous analysis of families of sigmoidal curves: application to bioassay, radioligand assay and physiological dose-response curves. Am. J. Physiol. 235, E97 102. Fredholm B. B. and Dunwiddie T. V. (1984) How does adenosine inhibit transmitter release? Trends Pharmac. Sci. 9, 13(~134. Garcia-Sainz J. A. and Torner M. L. (1985) Rat fat-cells have three types of adenosine receptors (R,, R~ and P). Biochern. J. 232, 439 443. Garritsen A., IJzerman A. P., Beukers M. W., Cragoe E. J. Jr and Soudijn W. (1990a) Interaction of amiloride and its analogues with adenosine A~ receptors in calf brain. Biochem. Pharmac. 40, 827 834. Garritsen A., IJzerman A. P., Beukers M. W. and Soudijn W. (1990b) Chemical modification of adenosine A~ receptors: implications for the interaction with R-PIA, DPCPX and amiloride. Biochem. Pharmac. 40, 835 842. Garritsen A., IJzerman A. P., Tulp M. Th. M., Cragoe E. J. Jr and Soudijn W. (1991) Receptor binding profiles of amiloride analogues provide no evidence for a link between receptors and the N a t / H + exchanger, but indicate a common structure on receptor proteins. J. Receptor Res. (in press). Horstman D., Brandon S., Wilson A. L., Guyer C. A., Cragoe E. J. Jr and Limbird L. E. (1990) An aspartate conserved among G-protein receptors confers allosteric regulation of alpha 2-adrenergic receptors by sodium. J. biol. Chem. 265, 21590-.21595. Howard M. J., Hughes R. J., Motulsky H. J., Mullen M. D. and Insel P. A. (1987) Interactions of amiloride with 7and fl-adrenergic receptors : amiloride reveals an allosteric site on 72-adrenergic receptors. Molec. Pharmac. 32, 53 58. IJzerman A. P., Bultsma T., Timmerman H. and Zaagsma J. (1984) The relation between ionization and affinity of /:~-adrenoceptor ligands. Naunyn-Schmiedeberg's Archs Pharmac. 327, 293 298. Kenakin T. P. (1987) Drug antagonism, in: Pharmacologic Ana(vsis o[" Dru9 Receptor Interaction (Kenakin T. P., ed.), pp. 205 244. Raven Press, New York. Kleyman T. R. and Cragoe E. J. Jr (1988) Amiloride and its analogs as tools in the study of ion transport. J. Membr. Biol. 105, 1 21. Klotz K.-N., Cristalli G., Grifantini M., Vittori S. and Lohsc M. J. (1985) Photoaffinity labeling of A t-adenosine receptors. J. biol. Chem. 260, 14659 14664. Lohse M. J., Klotz K.-N., Lindenbom-Fotinos J., Reddington M., Schwabe U. and Olsson R. A. (1987) 8-Cyclopentyl-l,3-dipropylxanthine (DPCPX)--a selective high affinity antagonist radioligand for A t adenosine receptors. Naunyn-Schmiedeber 9 ~ Archs Pharmac. 336, 204- 210. Lohse M. J., Lenschow V. and Schwabe U. (1984) Two affinity states of Ri adenosine receptors in brain membranes. Molec. Pharmac. 26, 1 9. Londos C., Cooper D. M. F., Schlegel W. and Rodbell M. (1978) Adenosine analogs inhibit adipocyte adenylate cyclase by a GTP-dependent process : basis for actions of adenosine and methylxanthines on cyclic AMP production and lipolysis. Proc. hath. Acad. Sci. U.S.A. 75, 5362 5366. Nunnari J. M., Repaske M. G., Brandon S., Cragoe E. J. Jr and Limbird L. E. (1987) Regulation of porcine brain 72adrenergic receptors by Na +, H ~ and inhibitors of Na+/H 4 exchange. J. biol. Chem. 262, 12387 12392.

Interaction between amiloride and adenosine A j receptors Parsons W. J., Ramkumar V. and Stiles G. L. (1988a) The new cardiotonic agent sulmazole is an At adenosine receptor antagonist and functionally blocks the inhibitory regulator, G~, Molec. Pharmac. 33, 441~448. Parsons W. J., Ramkumar V. and Stiles G. L. (1988b) Isobutylmethylxanthine stimulates adenylate cyclase by blocking the inhibitory regulatory protein, Gi. Molec. Pharmac. 34, 37~,1. Pirovano I. M., IJzerman A. P., Van Galen P. J. M. and Soudijn W. (1989) Influence of the molecular structure of N6-(co-aminoalkyl)adenosines on adenosine receptor affinity and intrinsic activity. Eur. J. Pharmac. Molec. Pharmac. Sect. 172, 185 193. Postma S. W. and Catterall W. A. (1984) Inhibition of binding of [3H]batrachotoxinin A 20-or-benzoate to sodium channels by local anesthetics. Molec. Pharmac. 25, 219 227. Ramkumar V. and Stiles G. L. (1988a) The new positive inotrope sulmazole inhibits the function of guanine nucleotide regulatory proteins by affecting GTP turnover. Molec. Pharmac. 34, 761-768. Ramkumar V. and Stiles G. L. (1988b) A novel site of action

213

of a high affinity A~ adenosine receptor antagonist. Biochem. biophys. Res. Commun. 153, 939 944. Shamim M. T., Ukena D., Padgett W. L. and Daly J. W. (I 989) Effects of 8-phenyl and 8-cycloalkyl substituents on the activity of mono-, di-, and trisubstituted alkylxanthines with substitution at the 1-, 3-, and 7-positions. J. Med. Chem. 32, 1231 1237. Smith P. K., Krohn R. I., Hermanson G. T., Mallia A. K., Gartner F. H., Provenzano M. D., Fujimoto E. K., Goeke N. M., Olson B. J. and Klenk D. C. (1985) Measurement of protein using bicinchoninic acid. Analyt. Biochem. 150, 76-85. Ukena D., Daly J. W., Kirk K. L. and Jacobson K. A. (1986) Functionalized congeners of 1,3-dipropyl-8-phenylxanthine: potent antagonists for adenosine receptors that modulate membrane adenylate cyclase in pheochromocytoma cells, platelets and fat cells. Life Sci. 38, 797 807. Van Galen P. J. M., IJzerman A. P. and Soudijn W. (1987) Adenosine derivatives with N6-alkyl, -alkylamine or -alkyladenosine substituents as probes for the Arreceptor. FEBS Left. 223, 197-201.