The Mechanism of Changes in Adrenoceptor - Mediated Responses 1 '

-Quantitative and qualitative changes in adrenoceptors under various conditions were studied by binding experiments. Chronic treatment with reserpine increased the level of a2-adrenoceptors in rat vas deferens and hypoxia increased the level of a1 -adrenoceptors in rat cardiomyocytes. Adenosine receptor agonists increased the affinity of the a2-adrenoceptor in rat vas deferens for the agonist with an increase in receptor-mediated responses. Thus two types of changes in receptor binding sites were observed. Next, changes in the GTP-binding (G) protein were studied. Activation of cyclic AMP-dependent protein kinase (PKA) decreased the ADP-ribosylation of Gi (41 K) protein by islet-activating protein (pertussis toxin, IAP). Purified Gi protein was phosphorylated by the enzyme. IAP-sensitive G protein-mediated coupling responses such as phosphatidylinositol turnover in differentiated HL-60 cells were also modified under this condition . These results indicated that phosphorylation of GI by PKA caused a qualitative change of Gi. Lithium ions also decreased the ADP-ribosylation of GI by IAP . Then it deter mined if the decrease was accompanied with a dissociation of the subunits of Gi. Phosphorylation of Gi by PKA impaired the dissociation of the subunits of GI caused by Mg2+ and GTPrS, wheareas lithium ions did not have any effect on their dis sociation. Thus some conditions caused a functional change in the so-called ""qualitative change" of Gi . Changes of adrenoceptor-mediated re sponses under various conditions have been observed many organs. These changes in re sponses are thought to be due to changes in the receptors in the cell membranes and/or to changes in the coupling responses to signals from the receptors. I studied the changes of adrenoceptors in several tissues under various conditions by binding experiments using radiolabeled ligands (1-3). The results showed two types of changes in adrenocep tors, quantitative and qualitative (Table 1). The latter type seemed to be due at least par tially to a change in the coupling between adrenoceptors and their GTP-binding regu lating (G) proteins. Then I studied the changes of G proteins. Since some G proteins coupled with adrenoceptors are ADP-ribosylated by islet-activating protein (pertussis toxin, IAP), I examined the ADP ribosylation of G proteins by IAP under various conditions. The results showed that some conditions caused a qualitative change of G proteins associated with the modulation of ADP-ribosylation of the proteins by IAP and impaired dissociation of the subunits of G proteins. [1] Binding studies on the changes of adrenoceptors (A) Quantitative changes of adrenoceptors: First, changes of the receptors were studied by binding experiments with radioisotope labeled ligands (1-17). Quantitative changes of receptors were observed in membranes or t This work was presented at the 62nd General Meeting of the Japanese Pharmacological Society , March 1989 when the author received the Award for Encouragement of Young Investigators from the Japanese Pharmacological Society. Table 1. Two types of changes of catecholamine receptors Quantitative changes of catecholamine receptors (1) Changes during aging (4-7) 1-a) Increase of (3-adrenoceptors in rat brain [3H-dihydroalprenolol] 1 -h) Increases of a, and a2-adrenoceptors in rat heart and vas deferens [3H-WB4101, 3H-clonidine] 1 -c) Increase of ~-adrenoceptors in human lymphocytes [3H-dihydroalprenol, 3H-hydrooxy benzylisoproterenol] 1 -d) Increases of a,-, a2 and (3-adrenoceptors in rat white fat cells [3H-bunazosin, 3H-yohimbine, 3H-CGP12177] (2) Effects of treatments with drugs (8, 9) 2-a) Increase of a2-adrenoceptors in rat vas deferens by chronic treatment with reserpine, chemical and mechanical denervation [3H-clonidine] 2-h) Increase of a2-adrenoceptors in rat vas deferens by glucocorticoids [3H-clonidine, 3H yohimbine] 2-c) Increase of (3-adrenoceptors in rat heart by forskolin or dibutyryl cyclic AMP [3H-CGP12177] (3) Pathological changes (10-12) 3-a) Increase of a,-adrenoceptors and decrease of (3-adrenoceptors in rat heart and hypothalamus [3H-WB4101, 3H-dihydroalprenolol] 3-b) Change of a,-adrenoceptors in experimental neuroma in rats [3H-WB4101] 3-c) Increases of a, and (3-adrenoceptors in hamsters in experimental cardiac myopathy [3H prazosin, 3H-CG P12177] 3-d) Increase of jS-adrenoceptors in human lymphocytes by chronic treatment with (3-blockers [3H-CGP12177] 3-e) Change of a,-adrenoceptors in human aorta of familial amyloidotic polyneuropathy [3H bunazosin] 3-f) Increase of a,-adrenoceptors in rat cardiac ventricular myocytes in hypoxia [3H-prazosin, 3H bunazosin] 3-g) Decrease of j9-adrenoceptors of cardiac myocytes in spontaneous hypertensive rats [1251 hydroxybebzylpindolol] 3-h) Increase of (3-adrenoceptors of microembolism in canine coronary vessels [3H-CGP12.177] 3-i) Change of a,-adrenoceptors in cardiac hypertrophy in pressure overload guinea pig hearts [3H-prazosin] 3-j) Increase of (3-adrenoceptors during reperfusion following hypothermic ischemia induced by cardiopleoic solution f3H-CG P12177. 121 1-iodocvanopindololl Qualitative changes of catecholamire receptors (4) Change during aging (4-7) 4-a) Increase in affinity of receptors in rat brains for (3-adrenergic agonists [3H-dihydroalprenolol] 4-b) Increase in affinity of receptors in rat heart and vas deferens for a2-adrenergic agonists (increase of high affinity 3H-clonidine binding sites) [3H-clonidine] (5) The effects of treatments with drugs (8, 9, 13-17) 5-a) Increase of a2-adrenergic agonoist binding sites and of affinity of the receptors for a2-adrenergic agonists in rat vas deferens by chronic treatment with reserpine or chemical or mechanical denervation [3H-clonidine, 3H-yohimbine] 5-b) !ncrease of binding of a2-adrenergic agonists to receptors in rat vas deferens by glucocorticoids [3H-clonidine, 3H-yohimbine] 5-c) Increase of a2-adrenergic agonist binding sites and increase in affinity of receptors for a2 adrenergic agonists by adenosine receptor agonists in rat vas deferens [3H-clonidine, 3H yohimbine] 5-d)) Increase of a2-adrenergic agonist binding sites and of affinity of the receptors for a2-adrenergic agonists in rat vas deferens by treatment with phospholipase [3H-clonidine, 3H-yohimbine] (6) Pathological changes 6-a) Increase in "GTP-shift" of (9-adrenergic agonist binding to the receptors during reperfusion following hypothermic ischemia induced by cardioplegic solution [125-iodocyanopindolol]. [ ] shows the isotope-labeled ligands used in the studies. ( ) shows the number of the reference. Unpublished data were included in this table. homogenates of various tissues, mainly with a and 8-adrenoceptors hydrophobic an tagonists (Table 1). However, depending on the procedures used for preparation of the fraction, some receptors are present in the membrane fractions as sequestered receptors or receptosomes, which cannot interact with agonists. Therefore, it is better to study re ceptors on the surface of intact cells using hydrophilic ligands than to detect amounts of receptors in cell membranes and homogenates using hydrophobic ligands. The receptors detected with binding experiments using intact cells and hydrophilic ligands should interact with the agonists. So next, the bind ings of hydrophilic antagonists (3H-CGP 12177, 3H-bunazosin) to intact cells were ex amined. One example of such a study is to investigate the effect of hypoxia. Rat cardiac ventricular cells were isolated by collagenase treatment and cultured. Then ai -adrenocep tors in their surface were assayed by meas uring the binding of the hydrophilic ai adrenoceptor antagonist 3H-bunazosin to intact cardiac cells for 12 hours at 4°C. The number of binding sites was significantly higher in the cells that had been subjected to hypoxia than in the controls. In the pretreated cells, inositol trisphosphate formation induced by ai -adrenoceptor stimulation was in creased in parallel with the increase in the re ceptors (K. Iwakura et al., unpublished data). Thus quantitative changes of the receptor ob served using intact cells and hydrophilic ligands seem to reflect changes of receptor coupling responses better than the changes observed using membrane fractions and hy drophobic ligands. (B) Qualitative changes of adrenoceptors: Changes in the coupling of the receptors and their coupled G proteins under various con ditions such as treatment with SH-reagents affects the affinities of agonists for the re ceptors. In binding studies, changes in the interactions between receptors and their coupled G proteins are observed as a change of "GTP shifts" (2). In the variant cell line of mouse lymphoma cells, cyc S49 cells, in which the (3-adrenoceptor-adenylate cyclase system has no stimulatory G proteins, no ""GTP shift" is observed in binding experi ments on f3-adrenoceptors. Thus it is possible to observe a qualitative change associated with the ,"GTP shift" or with a change in af finity of agonist for the receptors, possibly caused by a change in coupling between the adrenoceptors and their G proteins or a con formational change of the receptor itself. In rat vas deferens, pretreatment of the tissue with I.AP attenuated the inhibitory ef fect of clonidine, an a2-adrenoceptor agonist, on the contractile responses to electric stimu lation, but had no effect on the contractile re sponse to exogenous norepinephrine (18). These results indicate that feedback inhibition mediated by presynaptic a2-adrenoceptors is partly mediated by a IAP-sensitive G protein coupled with presynaptic a2-adrenoceptors. The effect of the adenosine receptor agonist 2-chloroadenosine (2CA) on the binding of a2-adrenoceptor ligands to the receptor in the rat vas deferens was investigated (13-15, 17). In homogenates of vas deferens, 2CA (0.1 ,eM) increased the maximal number of 3H -clonidine binding sites . This effect was abolished by the addition of theophylline. The detectable binding sites for 3H-clonidine were considered to be high affinity binding sites, because lower affinity binding sites (K,>10 nM) were not detectable in binding studies using the membrane filter method. Therefore, the inhibitory effect of the a2-adrenoceptor agonists on 3H-yohimbine (a2 adrenoceptor antagonist) binding were examined. The in hibitions by agonists were potentiated by the treatment with 2CA and the maximal numbers of binding sites of 3H-yohimbine in the homo genates was unaffected by the treatment with 2CA (15, 17). The results suggested that stimulation of the adenosine receptor induced a change in its conformation, resulting in an increase in its affinity for agonists. On the other hand, the binding of 3H-prazosin, an a1-adrenoceptor antagonist, was not in fluenced by the presence of 2CA. The con tractions of isolated vas deferens induced a, adrenoceptor agonists were not changed by treatment with 2CA, but the inhibition by clonidine of contractions induced by electric stimulation was enhanced by preincubation with 2CA. The results showed that stimulation of adenosine receptors induced a conforma tional change of presynaptic a2-adrenoceptor, resulting in potentiation of feedback inhibi tion by the receptors in the tissue. Another example of a qualitative change is the change in (3-adrenoceptors induced by reperfusion following cardioplegic arrest of rat heart (N. Sakagoshi et al., unpublished data). In the study, the "GTP shift" observed in the competition experiment in 9-adreno ceptor antagonists (1251-iodocyanopindolol) binding by a-adrenoceptor agonists seemed to be potentiated by the pretreatment. In this case, the increase in the stimulation of adenylate cyclase mediated by 3-adreno ceptors was comparable with the potentiation in the "'GTP shift". [2] Studies on changes in GTP-binding regulating proteins I studied changes of G proteins coupled with adrenoceptors, especially IAP-sensitive G proteins. The changes of ADP-ribosylation of the proteins by IAP caused by several treat ments such as phosphorylation of the protein by cyclic AMP-dependent protein kinase (PKA) and addition of lithium ions were studied (Table 2). (A) Effects of protein kinase A activation on IAP-sensitive G protein (40-41 K): ADP ribosylation of a 40-41-K protein in rat cardiac myocyte membranes reached a maxi mum after 90 min under conditions used and then was not affected by further addition of LAP or 32P-NAD. Autoradiography showed that purified Gi and cardiac myocyte mem branes incubated with IAP each gave a 40 41-K band of ADP-ribosylated protein, and the intensities of these bands were reduced in the presence of 100 /,,M guanosine 5,-(3-0 thio)-triphosphate (GTPrS). This results suggests that the proteins ADP-ribosylated by IAP in cardiac myocyte membranes included an IAP-sensitive G protein, possible Gi (40 41 K). Intact cardiac ventricular myocytes were preincubated with I-isoproterenol, and then their ADP-ribosylation by IAP was ex amined. Results showed that this pretreatment significantly decreased ADP-ribosylation. The number of 3H-CGP 12177 (beta-adrenore ceptor hydrophilic antagonist) binding sites in the isoproterenol-treated cells was also de creased to about 60% that of the control, indicating that pretreatment of the cells with I-isoproterenol induced down-regulation of beta-adrenergic receptors. To determine whether the decrease of ADP-ribosylation by IAP resulted from down-regulation of the re ceptors directly or as the result of an increase of intracellular cyclic AMP, I examined the effects of pretreatment with db-cAMP (1 mM) and forskolin (100 ,uM ). Both pretreat ments also decreased the ADP-ribosylation of 40-41 -K protein by IAP. These results show that an increase of intracellular cyclic AMP had the similar effect as pretreatment with I isoproterenol. Then, I examined the effect of the addition of the activated protein kinase A catalytic subunit and MgATP on the ADP ribosylation by lAP. Their additions also de creased ADP-ribosylation on the 40-41-K protein by IAP. These results suggested that phosphorylation of Gi resulted in a decrease in its ADP-ribosylation by IAP. Next I examined the phosphorylations of membranes of rat cardiac myocytes and dif ferentiated HL-60 cells by the PKA catalytic subunit. In the membranes of both types of cells, the 40-41-K protein was the main protein phosphorylated. The band of this Table 2. Experimental conditions causing change in ADP-ribosylation of GTP-binding proteins by IAP 1) Phosphorylation of purified Gi proteins by cAMP-dependent protein kinase 2) Addition of lithium chloride to purified Gi proteins, rat heart, aorta and brain membranes 3) Chronic treatment with lithium chloride in rat heart and human platelet 4) Aging in rat fat cells protein was similar to that of the protein ADP ribosylated by IAP. Finally, phosphorylation of partially purified Gi (41 K) by the protein kinase A catalytic subunit was studied. The alpha-subunit (41 K) and beta-subunit (35 K) were both substrates for phosphorylation by the protein kinase. The phosphorylation was almost saturated in 60 min, and 0.3 mol 32P/ mol protein was incorporated into the alpha subunit of Gi (19, 20). The incorporation of 32P into Gi in the phosphorylation reaction was low, even compared with that of Gi by protein kinase C observed previously (21 ). This difference in phosphorylation seemed to be due to differences in experimental condi tions. In addition, protein kinase A activation resulted in functional modifications of the IAP-sensitive GTP protein. Therefore, im provement in the experimental conditions for phosphorylation should increase the incorpo ration of 32P into Gi protein. The Gi protein purified by the method of Katada et al. in cludes IAP-sensitive 41 and 40-K dalton G proteins (22). The former is Gi (Gil ), which seems to be coupled with adenylate cyclase inhibition (23, 24), but the coupling response of the latter is unknown. Possibly the 40-K G protein mediates receptor-stimulated phos pholipase C activation in an IAP-sensitive manner. In addition, three kinds of cDNA for IAP-sensitive Gi have been reported besides those of Go and Gt (24). Therefore, the Gi sample used in this study may also have con tained another type of Gi termed Gi3 (25). In rat cardiac cells, several receptors are coupled with adenylate cyclase inhibition through Gi, possibly Gil (41 K). So the G protein (40-41 K) which showed changes in ADP-ribosyla tion in this work was possibly mainly Gil, but as mentioned above, the proteins (40-41 K) in rat cardiac ventricular myocyte membranes that were ADP-ribosylated by IAP may also have included other kinds of IAP-sensitive G proteins. I examined whether the coupling responses of the cells mediated by the IAP-sensitive GTP-binding protein were affected by phos phorylation of the proteins by protein kinase A (20). First, I examined the phosphatidyl inositol (PI) response of differentiated HL-60 cells. The chemotactic peptide fMet-Leu-Phe (fM LP) induced PI turnover in these cells. This stimulation by fMLP was completely dependent on GTP and Ca2+ and was IAP sensitive (26, 27). The stimulation resulted in formation of inositol trisphosphates (IP3) in the membranes of 3H-inositol-labeled cells. When the membranes were treated with protein kinase A in the presence of MgATP and cAMP, the stimulation by GTPrS (0.1 10 p M) was significantly inhibited, but the stimulation with [Ca2+] (0.01-1 /LM) was not affected. Activation of phospholipase C by fMLP in these cells is reported to be mediated through IAP-susceptible 40-K G protein, possibly socalled Gi2. These results strongly suggest that protein kinase A phosphorylates IAP-sensitive G protein (Gi2, 40 K) and that the phosphorylation of this G protein impairs its function of mediation between the fM LP receptor and phos pholipase C. Next, arachidonic acid re lease was studied (20). Receptor-mediated arachidonic acid release was studied in rat cardiac ventricular myocytes. In the cells, arachidonic acid release stimulated by the a, -adrenergic receptor or muscarinic receptor was IAP-sensitive, while stimulation of PI turnover by the receptor was not affected by IAP treatment. Thus arachidonic acid release was mediated by an IAP-sensitive G protein, which might be Gi2 or Gi3. This release was increased by pretreatment of the cells with db cAMP (1 mM) or forskolin (10 NM). I could not determine whether activation of protein kinase A resulted in phosphorylation of only the G protein that modifies IAP-sensitive G protein mediated arachidonic acid release because activation of protein kinase A results in the opening of voltage-dependent Ca2+ channels in heart tissues (28). However, re ceptor-mediated arachidonic acid release is possibly modified by phosphorylation of G protein by protein kinase A, because ADP ribosylation of the G protein by IAP was modified by protein kinase A. Phosphoryla tion of Gi by protein kinase C impaired adenyl ate cyclase inhibition in human thrombocytes and S49 mouse lymphoma cells (21, 29-31), but it is uncertain whether phosphorylation of IAP-sensitive G protein by protein kinase A or C has similar affect on the function of the Gi protein that mediates adenylate cyclase inhibition. The changes in adenylate cyclase inhibition in the cells induced by phosphory lation by protein kinase A require study. In addition, phosphorylation of the catalytic subunit of adenylate cyclase by protein kinase C is reported to increase the response of the enzyme activity to agonist stimulation (32), indicating that phosphorylation of a protein does not always attenuate its function. Gi proteins were ADP-ribosylated by IA.P only when the three subunits were present as a trimer, and not when the subunits were dissociated (22). So the extents of ADP ribosylation of the G proteins by IAP seemed to reflect the extents of association of the three subunits of the proteins in various conditions. I examined the effects of phosphorylation of the G proteins by protein kinase A on the dissociation (33) of the three subunits in duced by Mg2+ (50 mM) and GTPrS (100 ,uM). As shown in Fig. 1, pretreatment of the proteins with protein kinase A inhibited the dissociation of the subunits induced in these conditions, whereas the pretreatment itself did not have any remarkable effect on the association of the subunits. These findings suggest that phosphoryla tion modified the function of the G protein and changed its ADP-ribosylation by IAP. Thus the adenylate cyclase activating pathway can crosstalk with other coupling pathways through phosphorylation of their coupled G proteins by activated protein kinase A and this can be one of the mechanisms causing "h eterologous desensitization" (34), in ad dition to other postulated mechanisms such as phosphorylation of Gi protein by protein kinase C (21, 29-31) or phosphorylation of some receptors by protein kinase A, protein kinase C (35) or C2+, cal modu! in -dependent protein kinase (Fig. 2). (B) Effects of lithium ion on IAP-sensitive G protein: I studied the effects of the addition of lithium ion on ADP-ribosylation of IAP sensitive G protein by IAP (H. Kawamoto et al., unpublished data). The ADP-ribosylation of 40-41 -K proteins in the membrane fraction of rat cardiac ventricular cells by IAP was re duced dose-dependently by the addition of LiCI (0.5-10 mM), and the effects was re versible. Of the monovalent ions tested, lithium ion was the strongest inhibitor of ADP-ribosylation of the protein by IAP. LiCI also decreased the ADP-ribosylation of par tially purified Gi proteins (41 K) by IAP. In addition, the inhibition by a2-adrenoceptor stimulation in forskolin-stimulated adenylate cyclase activity was impaired by lithium ion (2 mM) addition in human platelet mem branes. The results suggested that the ion also modulated the function of the Gi protein mediating adenylate cyclase inhibition. Lithi um ions had no appreciable effect on the association or dissociation of the three subunits of purified Gi proteins. From these results, IAP-sensitive G proteins are con cluded to be one of the targets for the therapeutic effect of lithium. (C) Qualitative change of inhibitory GTP binding protein: Because GTP-binding pro teins are supposed to be present at a higher Fig. 1. Hydrodynamic behavior of Gi: Effects of activation of protein kinase A on dissociation among the subunits of purified Gi protein. Purified Gi treated (40) or untreated (0, A) with PKA and Mg ATP for 30 min at 37'C was placed on top of 5-ml gradients of 5-20% (w/w) sucrose prepared in buffer containing either 50 pM GDP (0) or 50 ,uM GDP, 100 pM GTPTS and 50 MM MgCl2 (•, A). The gradients were centrifuged at 100,000 g for 60 min at room temperature and fractionated at 4'C into aliquots. The position of bovine serum albumin and cytochrome C in the gradients was determined by SDS-PAGE followed by Coomassie Blue staining and densitometric scanning. The position of Gi or its a-subunit was determined by first subjecting aliquots of the fractions to ADP-ribosylation by IAP in the presence of 32P-NAD and then analyzing them by SDS-PAGE followed by autoradiography