J Mol Neurosci (1991) 3:39-45

Journal of Molecular Neurosc|ence © Birkh~iuser Boston 1991

Effects of Guanine Nucleotides on Kainic Acid Binding and on Adenylate Cyclase in Chick Optic Tectum and Cerebellum Diogo Onofre Souza 1 and Galo Ramfrez 2 lDepartamento de Bioqufmica, Instituto de Biocifncias, UFRGS, Rua Sarmento Leite 500, 90050 Porto Alegre, RS, Brasil 2Centro de Biologia Molecular (CSIC-UAM), Universidad Aut6noma, Cantoblanco, 28049 Madrid, Spain

Abstract. Adenylate cyclase activity and binding of neurotransmitters to some receptors can be modulated simultaneously by guanine nucleotides. Furthermore it has been shown, in different neurotransmitter systems, that the ability of GTP to inhibit agonist binding is related to the capacity of the transmitter to modulate adenylate cyclase activity. In the present report we show that in chick optic tectum and cerebellum the effects of guanine nucleotides on kainic acid binding and on adenylate cyclase activity can be dissociated. In lysed membrane preparations, GTP, GDP, and GMP, or their analogs, displace binding of kainic acid with the same efficiency, whereas only GTP stimulates adenylate cyclase. In vesicle preparations, all three nucleotides inhibit binding of kainic acid without modifying adenylate cyclase activity. The present results suggest that, if adenylate cyclase is indeed coupled to this particular type of excitatory amino acid receptor, the coupling mechanism would be probably different from those operating in other neurotransmitter systems and also that the displacement of kainic acid by GDP and GMP (and even perhaps by GTP) is not likely to depend on the interaction between the receptor and a Gs-protein-mediated effector system.

Transmembrane signaling systems, composed of receptor, G-protein, and effector, couple many extracellular stimuli to cellular responses (Gilman, 1987). The state of activity of the G-protein is controlled by guanine nucleotides: activation occurs when GTP binds to the G-protein; inactivation follows GTP hydrolysis to give bound GDP (Levitzki, 1987). In the active state, G-proteins interact with a receptor and an effector, simultaneously modulating the effector activity and decreasing the affinity of the transmitter for its own receptor. Adenylase cyclase is one of the effectors that may be modulated by guanine nucleotides through Offprint requests to: D.O. Souza

stimulatory (Gs) or inhibitory (Gi) G-proteins (Birnbaumer et al., 1985). The effect of GTP on adenylate cyclase activity is well documented (Birnbaumer et al., 1985; Gilman, 1987). However, perhaps due to the presence in membrane preparations of nucleotide-interconverting enzymes, the effect of GDP on agonist binding and cyclase activity is still controversial (Kimura and Shimada, 1983). It has been shown, however, that a stable GDP analog (GDP-S) antagonizes both the stimulatory (Eckestein, 1979) and the inhibitory (Hildebrandt and Birnbaumer, 1983) effects of GTP or analogs. No effect of GMP has been demonstrated so far. On the other hand, GTP (and analogs) and also GDP (and analogs) decrease the binding of several neurotransmitters to their receptors (Stryer, 1986). In these studies GMP, cGMP, or adenine derivatives had no effect on binding. The simultaneous effects of guanine nucleotides on neurotransmitter binding and on adenylate cyclase activity have been interpreted according to the above-mentioned ternary complex model, which proposes the formation of an agonistreceptor-G-protein complex (Birnbaumer et al., 1985; Gilman, 1987; Limbird, 1981). This ternary complex would modulate adenylate cyclase activity only when GTP is bound to the G-protein, this action being terminated by hydrolysis of GTD or GDP. However, not only bound GTP but also bound GDP decreases the affinity of the neurotransmitter for the ternary complex. It also has been shown that the capacity of GTP to inhibit binding of a neurotransmitter to its receptor is related to the ability of the transmitter to modulate adenylate cyclase activity (Birnbaumer et al., 1985; Limbird, 1981; Thomsen et al., 1988). Excitatory amino acids have been shown to stimulate adenylate cyclase activity (Baba et al., 1988; Bruns et al., 1980; Shimizu et al., 1974; Shonk et

40

Souza & Ramirez: KainateReceptors and Adenylate Cyclase

al., 1987). However, the stimulation exerted by amino acids on cyclase activity differs in several aspects from the modulatory effects exerted by other neurotransmitters. Thus, these amino acidsonly stimulate enzyme activity at concentrations 100-10,000 times greater than those necessary for maximum binding (Baba et al., 1988; Bruns et al., 1980; Foster and Fagg, 1984; Ramirez et al., 1981; Sharif and Roberts, 1981; Shimizu et al., 1974; Shonk et al., 1987). Stimulation of the enzyme also depends on the integrity of the purinergic metabolic pathways yielding adenosine, which has no effect on binding, however (Bruns et al., 1980; Shimizu et al., 1974). Furthermore, while binding also can be demonstrated in cell-free preparations, stimulation of adenylate cyclase is observed only when cell integrity is maintained (Baba et al., 1980; Bruns et al., 1980; Foster and Fagg, 1984; Sharif and Roberts, 1981; Shimizu et al., 1974; Shonk et al., 1987). All this points to indirect effects as an explanation for any action of excitatory amino acids on cyclase activity. In the present study we have investigated the simultaneous effects of guanine nucleotides on kainic acid binding and on adenylate cyclase activity in two regions of the chick brain. Kainic acid is a rigid analog of glutamate that binds to receptors in membrane preparations from chick brain with high affinity (Ramirez et al., 1981). Our results show that the displacement of kainic acid binding by guanine nucleotides can be dissociated from their modulatory effects on the activity of adenylate cyclase in the same membrane preparations.

Materials and Methods Animals White-Leghorn young chicks (6-12 days old), kept on a 12-hour light-12-hour dark schedule, were used. Materials [3H]Adenosine-3',5 '-cyclic-monophosphate (cAMP: 23 Ci/mmol) and [3H]Kainic Acid (3.6 Ci/ mmol) were purchased from Amersham International, U.K.; adenylyl-imidodiphosphate (AMPPNP), guanosine-5'-O-(3-thiotriphosphate) (GMPPSP), guanosine-5'-O-(2-thiodiphosphate) (GDP-S), guanylyl-imidodiphosphate (GMP-PNP), guanylyl(13,~-methylene)-diphosphonate (GMPPCP), GTP, GDP, GMP, and ATP were all purchased from Boehringer Mannheim. Protein kinase for cAMP determination was purchased from Sigma

Company. All other chemicals of analytical grade were obtained from standard commercial suppliers.

Membrane preparations Lysed membranes (Ramfrez et al., 1981). Optic tectal lobes or cerebella were homogenized in 0.32 M sucrose containing 10 mM Tris HCI buffer, pH 7.4, and 1 mM MgCI2. The homogenate was centrifuged at 1000g, and the pellet was rehomogenized and spun again. The second pellet was discarded, and both supernatants were pooled and centrifuged at 100,000g for 30 minutes. The resulting pellet was lysed in 1 mM Tris HCI buffer, pH 7.4, for 30 minutes and centrifuged at 100,000g for 30 minutes. This pellet was washed three times in 10 mM Tris HCI buffer, pH 7.4, at 100,000g for 30 minutes. The final pellet was diluted in the same buffer and used for the determination of kainic acid binding and adenylate cyclase activity. Preparation of membrane vesicles (Daly et al., 1980). Optic tecta and cerebella were homogenized in the following medium (KRG): 122 mM NaCI, 3 mM KCI, 1.2 mM Mg2SO4, 1.3 mM CaCI2, 0.4 mM KH2PO4, 25 mM NaHCO 3, I0 IxM EDTA, and I0 mM glucose, pH 7.4, gassed with CO2/O2 (5%-95%). The homogenate was centrifuged at 1000g for I0 minutes, the supernatant was discarded, and the pellet washed three times as above. The final pellet was diluted in KRG and used for the determination of kainic acid binding and adenylate cyclase activity. Assay of kainic acid binding Lysed membrane preparations. Binding assays were performed at 30°C in small polycarbonate tubes containing, in a total volume of I ml, 0.5 mg of membrane protein, 10 mM Tris HCI, 10 mM MgCI2, 1 mM dithiothreitol, 1% albumin, and 40 nM radioactive ligand ([3H]kainic acid, 3.6 Ci/mmol), with or without the displacer (400 ~M nonradioactive kainic acid). The concentrations of guanine nucleotides are specified in each experiment. The incubation was started by addition of the radioligand and stopped after 16 minutes by centrifugation of the tubes at 100,000g for 30 minutes. The supernatant was discarded. The walls of the tubes and the surface of the pellets were quickly and carefully rinsed with cold distilled water. The pellets were processed for radioactivity and protein measurement as previously described (Ramfrez et al., 1981). Specific binding was defined as that part of total binding displaced by a concentration of nonlabeled ligand 10,000 times the radioligand concentration.

Souza & Ramirez:

Kainate Receptors and Adenylate Cyclase

Vesicle membrane preparations. Incubation was performed at 37°C in KRG medium. The incubation was stopped by centrifugation at 20,000g. Other conditions were as described for lysed membranes.

A Tecl'um

1.5

Lysed membrane preparations. The incubation medium was the same one used in the binding assays. Samples were processed according to Albano et al. (1974). The membranes were preincubated for 15 minutes at 30°C, and the incubation was started by the addition of ATP up to 1 mM, and stopped after 1 minute by the addition of two volumes of ethanol. The tubes were centrifuged at 3000g for 10 minutes, the pellets were discarded, and the supernatants were evaporated under a stream of N 2 ( o r air, with the same results). The residues were dissolved in 50 mM Tris HCI, pH 7.4, containing 4 mM EDTA. The cAMP content was measured by the protein binding method of Tovey et al. (1974). Vesicle membrane preparations. The vesicle preparations were gassed with CO2/O 2 for 30 minutes at 37°C in KRG medium (in order to stabilize the basal cAMP content), and the incubation was started by the addition of guanine nucleotides. The incubation was stopped by the addition of two volumes of ethanol, and the cAMP content was measured as described above for lysed membranes. Statistical analyses The effects of guanine nucleotides on kainic acid binding and on adenylate cyclase activity were analyzed by one-way analysis of variance followed by Duncan's multiple-range test when appropriate.

Results In addition to GTP and GDP, we have tested a number of analogs (GMP-PNP, GMP-PSP, GMP-PCP, and GDP-S), which have been shown to possess similar biochemical properties to the physiological nucleotides, while being resistant to enzymatic hydrolysis. Figure 1 shows that all the guanine nucleotides and analogs tested displace [3H]kainic acid, binding to a lysed membrane preparation from chick optic tectum or cerebellum to the same extent. Half-maximal inhibition of binding was seen at a nucleotide concentration of approximately 100 I~M. Therefore, only GMP-PNP, GDP-S, and GMP

II 0 ~. E]

.

GTP GMP'PCP GMP. PNP GMP. PSP

I~..~,~.~.

~S 1.0

Adenylate cyclase activity

41

[3 GOP" S

S. E ~05 0 E 123

z

0

I

;

t

B Cerebellum

o m 15

o 10

T

5

1

I

I

lO-S

10-5 10-4 I0-3 [NUCLEOTIDE] M Fig. 1. Effects of guanine nucleotides on kainic acid binding in lysed membrane preparations from optic tectum (A) and cerebellum (B). Standard deviation of each point was less than 10% of the mean for five (optic tectum) or three (cerebellum) independent experiments.

were used in further experiments dealing with the effects of nucleotides on either kainic acid binding or adenylate cyclase activity. In these lysed membrane preparations only GMP-PNP stimulated adenylate cyclase activity (Fig. 2). When vesicle membrane preparations from the same tissues were used, it was again found that all three nucleotides GMP-PNP, GDP-S, and GMP, decreased [3H]kainic acid binding (Fig. 3), while none of the guanine nucleotides modified the cyclase activity (Fig. 4). Since the above results suggested that the displacement of kainic acid by at least GDP-S and GMP was not mediated by a typical receptorGs-protein coupling, we did some preliminary experiments to explore the structural requirements for such displacement (Fig. 5). Besides GMP-PNP, GDP-S, and GMP only phosphoserine and phosphothreonine (at 400 p~M) were effective in displacing the binding of g3H]kainic acid to lysed membranes from the cerebellum. Guanosine, guanine, adenosine, cAMP, AMP-PNP, KH2PO 4, polyphosphate, fructose 1,6-diphosphate, phosphotyrosine, phospho-enol pyruvate, 3-phosphoglycerate and creatine phosphate were ineffective.

42

Souza & Ramirez: I

2OO

Kainate Receptors and Adenylate Cyclase

I

I

1

I

1

A Tectum

A Teclum

zx GMP- PNP

[]GOP.s

Q2 - ~ ~ - - ~

O -

tO u

~ loo

~ 0.1 -6 E ea Q

-*" GMP. PNP 0 GDP' S • GMP

w

o

.J

I B Cerebellum

I

o>- 200

t (.)

0

O 6

B Cerebellum

1"1

UJ p-

z 100 W El C

1

10-6 O

I

10-7

10-6

I

10-5

I

10-4

[NUCLEOTIDE]" M

Fig. 2. Effects of GMP-PNP, GDP-S, and GMP on adenylate cyclase activity in lysed membrane preparations from optic tectum (A) or cerebellum (B). Values are expressed as percentage of controls (100%). The control mean values were 3.0 pmol cAMP/mg protein/min in the optic tectum and 1.2 pmol cAMP/mg protein/min in the cerebellum. The standard deviation of each point was less than 10% of the mean for three independent experiments. *Significantly different from control at P < 0.05. Discussion

It has been known for a long time that both GTP and GDP inhibit the binding of neurotransmitters to a wide range of receptors, such as dopaminergic (Grigoriadis and Seeman, 1985), adrenergic (Hoffman and Lefkovitz, 1980), serotoninergic (Lyon et al. 1987), A l purinergic (Stiles, 1988), 13-GABAergic (Asano et al., 1985), muscarinic (Haga et al., 1986), H l histaminergic (Chang and Snyder, 1980) and opioid (Selley et al., 1988) receptors. In all these studies GMP, cGMP, and adenine derivatives had no effect on binding. While the displacing effect of GTP on agonist binding could have been predicted from the ternary model proposed for the complex interaction among receptor, G-protein, and effector systems, the mechanism underlying the effect of GDP was rather less obvious. It was soon discovered, in this context, that membrane preparations contain enzymatic activities that interconvert nucleotides (Kimura and Shimada, 1983) and com-

10-5

[

10-4

I

10-3

rNUCLEOT,DE].M Fig. 3. Effects of GMP-PNP, GDP-S, and GMP on kainic acid binding in vesicle membrane preparations from optic tectum (A) and cerebellum (B). The standard deviation of

each point was less than 10% of the mean for three independent experiments. plicate the interpretation of the effects of nucleotides (especially GDP) on receptor and effector systems. The use of nonhydrolyzable, fully active GTP and GDP analogs has helped us to overcome some of the uncertainties in the interpretation of the results. Neither the analogs used nor GMP were hydrolyzed by our membrane preparations (data not shown). Our experiments show quite clearly that whereas in lysed membrane preparations GMPPNP, GDP-S, and GMP inhibit [3H]kainic acid binding, only GMP-PNP modulates adenylate cyclase activity (Figs. I and 2). In vesicle membrane preparations, all three guanine nucleotides displace kainic acid binding, but none modifies the cyclase activity (Figs. 3 and 4). Thus, the effect on enzyme activity is specific for GTP and seems to be exerted from the inside of the membrane, since it occurs in lysed membrane preparations but not in occluded vesicle preparations. On the other hand the effects on kainic acid binding are presumably exerted from the outside or from both sides of the membrane. This characteristic of the effects of guanine nucleotides on the binding of excitatory amino acids is in agreement with other studies (Sharif and Roberts, 1981; Monahan et al., 1988). Our results indicate that the modulatory effects

Souza & Ramirez: I

A

Kainate Receptors and Adenylate Cyclase

I

43

I

Tectum

O

loo

- -

50

I-A GMP" PNP

o GDP. S

I-

• GMP


-

0 I-< ._I )Z W r~