Cytoplasmic domain of natriuretic peptide receptor-C inhibits adenylyl cyclase. Involvement of a pertussis toxin-sensitive G protein.

Natriuretic peptide receptor C (NPR-C) is a disulfide-linked homodimer with an approximately 440-amino acid extracellular domain and a 37-amino acid cytoplasmic domain; it functions in the internalization and degradation of bound ligand. The use of NPR-C-specific natriuretic peptide analogs has implicated this receptor in mediating the inhibition of adenylyl cyclase or activation of phospholipase C. In the present studies we have investigated the role of the cytoplasmic domain of NPR-C in signaling the inhibition of adenylyl cyclase. Polyclonal rabbit antisera were raised against a 37-amino acid synthetic peptide (R37A) corresponding to the cytoplasmic domain of NPR-C. Incubation of anti-R37A with rat heart particulate fractions blocked atrial natriuretic peptide-dependent inhibition of adenylyl cyclase. The cytoplasmic domain peptides R37A and TMC (10 residues of transmembrane domain appended on R37A) were equipotent in inhibiting adenylyl cyclase (Ki ~1 nM) in a GTP-dependent manner, whereas K37E (a scrambled peptide control for R37A) did not inhibit adenylyl cyclase activity. Prior incubation of membranes with pertussis toxin blocked R37A or TMC inhibition of cAMP production. Detergent solubilization of the rat heart particulate fraction destroyed natriuretic peptide inhibition of adenylyl cyclase, but TMC was able to inhibit cAMP production in a dose-dependent manner. Our results provide evidence that the 37-amino acid cytoplasmic domain of NPR-C is sufficient for signaling inhibition of adenylyl cyclase through a pertussis toxin-sensitive G protein.

Natriuretic peptide receptor C (NPR-C) is a disulfidelinked homodimer with an approximately 440-amino acid extracellular domain and a 37-amino acid cytoplasmic domain; it functions in the internalization and degradation of bound ligand. The use of NPR-C-specific natriuretic peptide analogs has implicated this receptor in mediating the inhibition of adenylyl cyclase or activation of phospholipase C. In the present studies we have investigated the role of the cytoplasmic domain of NPR-C in signaling the inhibition of adenylyl cyclase. Polyclonal rabbit antisera were raised against a 37-amino acid synthetic peptide (R37A) corresponding to the cytoplasmic domain of NPR-C. Incubation of anti-R37A with rat heart particulate fractions blocked atrial natriuretic peptide-dependent inhibition of adenylyl cyclase. The cytoplasmic domain peptides R37A and TMC (10 residues of transmembrane domain appended on R37A) were equipotent in inhibiting adenylyl cyclase (K i ϳ1 nM) in a GTP-dependent manner, whereas K37E (a scrambled peptide control for R37A) did not inhibit adenylyl cyclase activity. Prior incubation of membranes with pertussis toxin blocked R37A or TMC inhibition of cAMP production. Detergent solubilization of the rat heart particulate fraction destroyed natriuretic peptide inhibition of adenylyl cyclase, but TMC was able to inhibit cAMP production in a dose-dependent manner. Our results provide evidence that the 37-amino acid cytoplasmic domain of NPR-C is sufficient for signaling inhibition of adenylyl cyclase through a pertussis toxinsensitive G protein.
The natriuretic peptides are a family of three polypeptide hormones termed atrial natriuretic peptide (ANP), 1 brain natriuretic peptide (BNP), and C-type natriuretic peptide (CNP) (1)(2)(3). The role of ANP and BNP as endocrine hormones is apparently to be antagonists to vasopressin and endothelins and to the renin/angiotensin/aldosterone system (1,4). The role of CNP in vivo is less well defined. Although CNP may not be a significant modulator of diuresis or natriuresis (5,6), it is a vasodilator expressed by endothelial cells (3,7).
Two classes of natriuretic peptide receptor (NPR) have been defined by molecular cloning. The first class includes the membrane guanylyl cyclases, NPR-A (8,9) and NPR-B (10,11). For this class of receptor ligand binding activates synthesis of the intracellular second messenger cGMP by the cytoplasmic catalytic domain. The second class of natriuretic peptide binding site is a non-guanylyl cyclase receptor (12) termed clearance, NPR-C or ANP-R 2 . This receptor is homologous in its extracellular domain to NPR-A and NPR-B, exists as a disulfide-linked homodimer, and has a 37-amino acid cytoplasmic domain (13,14). The role of NPR-C has originally been considered in terms of the clearance of bound ligand by internalization and degradation (15,16). However, work from a number of laboratories has implicated NPR-C in mediating signal transduction, either inhibition of adenylyl cyclase or activation of phospholipase C (for review, see Refs. 17 and 18).
The ligand specificity of the three NPRs has been well defined, with ANP and BNP as agonists for NPR-A, and CNP for NPR-B (19 -21). NPR-C has much broader specificity, binding all three natriuretic peptides as well as smaller peptide analogs that do not activate guanylyl cyclase (15). In vivo many of the responses to natriuretic peptides are thought to be mediated by cGMP, including vasorelaxation (22,23), increased glomerular filtration, and inhibition of tubular sodium reabsorption (24 -26). The involvement of NPR-C in mediating physiological responses to ANP has been inferred by pharmacology; ANP variants that are ineffective in binding and activating receptor guanylyl cyclases will bind to NPR-C and are efficacious in the inhibition of adenylyl cyclase or activation of phospholipase C. These responses are dependent on the presence of guanine nucleotides (27,28), suggesting a receptor-coupled phenomenon. Blocking of these ANP signal transduction pathways by pertussis toxin treatment is similar to the classical G proteincoupled adenylyl cyclase inhibitory pathways (29 -31) and suggests either the direct involvement of G i or indirect involvement of G o in the coupling of NPR-C to adenylyl cyclase (32)(33)(34)(35).
In this report we investigated whether the 37-amino acid cytoplasmic domain of NPR-C is involved in mediating pertussis toxin-sensitive signal transduction. Antibodies specific for the cytoplasmic domain inhibit signaling, and free cytoplasmic domain peptide inhibits adenylyl cyclase in a manner consistent with G protein coupling. These data strongly implicate NPR-C in signal transduction.

EXPERIMENTAL PROCEDURES
Materials-ATP, cAMP, and isoproterenol were purchased from Sigma. Creatine kinase ( , and GTP␥S were purchased from Boehringer Mannheim. [␣-32 P]ATP was from Amersham Corp. PT was from List Biochemicals (Campbell, CA). Rat ANP and the ring-deleted ANP analog cANP(4 -23) were from Peninsula Laboratories (Belmont, CA). The purified protein derivative of tuberculin was purchased from Statens Seruminstitut, Copenhagen, Denmark. Anti-rabbit horseradish peroxidase was from Sigma. Goat serum, biotinylated anti-rabbit antibody, the avidin/biotin blocking kit, and the ABC reagent were from Vector Laboratories (Burlingame, CA). Synthetic peptides R37A and TMC were synthesized and purified by standard solid phase techniques. K37E was custom synthesized by Peninsula Laboratories. The peptide was purified by reverse phase high performance liquid chromatography, analyzed by mass spectometry, and quantitative amino acid analysis was performed to confirm the amino acid composition.
R37A Antibody Production-The synthetic peptide termed R37A, corresponding to the 37-amino acid cytoplasmic domain of bovine NPR-C (see Table I) was conjugated to the purified protein derivative of tuberculin with glutaraldehyde and used to immunize New Zealand White rabbits as described previously (20). The IgG fraction from rabbit serum was purified by protein A affinity chromatography (20).
Western Blot Analysis-Control samples were prepared from 293 cells expressing recombinant rat NPR-C using Triton X-100 non-ionic detergent as described previously (20). Rat heart membranes were prepared by homogenization in ice-cold 50 mM Hepes, 1 mM EDTA, 250 mM sucrose, 0.7 mg/ml pepstatin A, 0.5 mg/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, and 1 mM phosphoramidon, pH 7.4, with a Polytron on a setting of 5 twice for 10 s. One-tenth volume of 30 mM MgCl 2 ,100 mM NaCl, 1 mM phosphoramidon was added to the particulate fraction, and the solution was centrifuged at 400 ϫ g for 10 min at 4°C in an HB-4 rotor. The supernatant was transferred to another tube, centrifuged as before, decanted, and passed though four layers of sterile gauze, then centrifuged at 30,000 ϫ g for 20 min at 4°C in a 60 Ti rotor (Beckman). The resulting pellet was resuspended in 3-5 ml of 50 mM Hepes, 0.1 mM EDTA, 5 mM MgCl 2 , 100 mM NaCl, 0.7 mg/ml pepstatin A, 0.5 mg/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, 1 mM phosphoramidon, pH 7.4, using a 7-ml Wheaton Dounce homogenizer with a B pestle. Aliquots were stored at Ϫ80°C. For sodium dodecyl sulfate-polyacrylamide gel electrophoresis, approximately 100 g of rat heart membrane protein, 0.5 g of 293/rat NPR-C membrane protein, and 10 l of Rainbow colored protein molecular weight markers (Amersham) were suspended in nonreducing sample mix and fractionated by electrophoresis on a 7.5% gel. The proteins were then transferred onto nitrocellulose by electroblotting, probed with anti-R37A at a 1:2,500 dilution as described previously (20), and developed using goat anti-rabbit horseradish peroxidase at a 1:4,000 dilution with ECL Western blotting detection reagents (Amersham).
Preparation of Heart Particulate Fractions-Heart ventricles were dissected from Sprague-Dawley rats (200 -300 g), quickly frozen in liquid nitrogen, and stored at Ϫ70°C until used. Frozen hearts were pulverized to a fine powder with a mortar and pestle precooled in liquid nitrogen. The heart powder was homogenized using a Teflon-glass homogenizer (12 strokes) in a buffer containing 10 mM Tris-HCl and 1 mM EDTA, pH 7.5, then centrifuged at 1,000 ϫ g for 10 min. The supernatant was discarded, and the pellet was homogenized in the above buffer and centrifuged at 1,000 ϫ g for 10 min. The pellet was finally suspended by homogenization in 10 mM Tris-HCl and 1 mM EDTA, pH 7.5, and used directly for adenylyl cyclase specific activity determination. Protein was determined essentially as described (36), with crystalline bovine serum albumin as standard.
Pertussis Toxin Treatment-PT treatment was performed as described earlier (32,37,38). Briefly, heart particulate fractions were incubated in 25 mM glycylglycine buffer, pH 7.5, containing 1 mM NAD, 0.4 mM ATP, 0.4 mM GTP, 15 mM thymidine, 10 mM dithiothreitol, and ovalbumin (0.1 mg/ml) with and without PT (5 g/ml) for 30 min at 30°C. The particulate fraction was washed two to three times with 10 mM Tris, 1 mM EDTA buffer, pH 7.5, and finally suspended in the same buffer and used for adenylate cyclase activity determination. Preincubation of the membranes at 30°C for 30 min in the absence or presence of PT resulted in a significant loss of enzyme activity (40%) which was independent of the presence of PT in the incubation medium. However, the percent inhibition of adenylyl cyclase by R37A or TMC remained unchanged. 2 Adenylyl Cyclase Activity Determination-Adenylyl cyclase activity was determined by measuring [ 32 P]cAMP formation from [␣-32 P]ATP as described previously (33,37). The typical assay medium contained 50 mM glycylglycine, pH 7.5, 0.5 mM MgATP, [␣-32 P]ATP (1-1.5 ϫ 10 6 cpm), 5 mM MgCl 2 , 100 mM NaCl, 0.5 mM cAMP, 1 mM 3-isobutyl-1methylxanthine, 0.1 mM EGTA, 10 M GTP␥S, and an ATP-regenerating system consisting of 2 mM creatine phosphate, 0.1 mg of creatine kinase/ml, and 0.1 mg of myokinase/ml in a final volume of 200 l. Incubations were initiated with the addition of reaction mixture to the membranes (30 -70 g) which had been preincubated at 37°C for 10 min. The reactions, conducted in triplicate at 37°C for 10 min, were terminated by the addition of 0.6 ml of 120 mM zinc acetate, cAMP was purified by coprecipitation of other nucleotides with ZnCO 3 by the addition of 0.5 ml of 144 mM Na 2 CO 3 and by subsequent chromatography, using the double column system (39). Under these assay conditions, adenylyl cyclase activity was linear with respect to protein concentration and time of incubation.
Antibody inhibition was performed by preincubating heart particulate fraction with anti-A-4 (20) or anti-R37A at 4°C for 2 h. Samples were centrifuged at 1,000 ϫ g for 10 min, the pellet washed twice in 10 mM Tris-HCl, 1 mM EDTA, pH 7.5, and resuspended in the same buffer for cyclase determination.
Detergent solubilization (40) was carried out by incubating heart particulate fraction with 0.3% (v/v) Lubrol PX-100 for 10 min at 25°C. The sample was centrifuged at 100,000 ϫ g for 1 h and the supernatant used for adenylyl cyclase activity measurements.

Antibodies Specific for the NPR-C Cytoplasmic Domain-The
binding of ANP to NPR-C leads to a PT-sensitive inhibition of adenylyl cyclase in several tissues (17,18); therefore, the 37amino acid cytoplasmic domain of NPR-C could mediate signal transduction. To address this, polyclonal rabbit antisera against the 37-amino acid sequence corresponding to the cytoplasmic domain of bovine NPR-C were raised (Table I). The purified protein derivative-coupled R37A peptide was very immunogenic, with serum titers of Ͼ1:500,000 when tested by enzyme-linked immunosorbent assay against unconjugated R37A. 2 The reactivity of anti-R37A antibody was tested by Western blot analysis (Fig. 1). Recombinant rat NPR-C expressed in 293 cells was detected as a nonreduced homodimer of approximately 150 kDa and comigrated with NPR-C homodimer in a rat heart membrane fraction (Fig. 1A). Blocking of immunoreactivity with R37A peptide (Fig. 1B) demonstrates the specificity of anti-R37A and reveals a nearly identical pattern of nonspecific background in the heart sample on un-  (Fig. 1A, lane 2) and blocked (Fig. 1B, lane 2) filters.
The reactivity of anti-R37A, raised against the bovine NPR-C cytoplasmic domain sequence, with rat ( Fig. 1) and human 2 NPR-C is consistent with the high degree of conservation of this 37-amino acid cytoplasmic domain (Table I).
Effect of Anti-R37A on ANP-mediated Inhibition of Adenylyl Cyclase-To extend the pharmacology implicating NPR-C in the inhibition of adenylyl cyclase, the NPR-C-specific anti-R37A antibodies were tested for blocking this pathway. Using anti-R37A, and the human NPR-A-specific antibody A-4 as a control (20), heart particulate fraction was preincubated with antibodies at a 1:100,000 dilution (ϳ0.1 g/ml) prior to assay of adenylyl cyclase activity (Fig. 2). In this ANP concentrationresponse assay for inhibition of adenylyl cyclase, there was 30 -40% maximal inhibition of enzyme activity in the presence of control antibody A-4. The NPR-C-specific anti-R37A resulted in marked attenuation of the ANP effect, with at most 10% maximal inhibition of the enzyme at 10 Ϫ6 M ANP. This experiment provides evidence for the involvement of the 37-amino acid cytoplasmic domain of NPR-C in mediating ANP-dependent inhibition of adenylyl cyclase. This result also suggests that NPR-C, and not another NPR with similar pharmacology, is mediating ANP signal transduction.
Effect of R37A and TMC on Adenylyl Cyclase Activity-The next series of experiments was designed to test the hypothesis that the cytoplasmic domain alone may mimic the ANP-dependent signaling of NPR-C. For this purpose, we used the R37A synthetic peptide and a 47-amino acid peptide, TMC (Table I), which consists of 10 additional residues from the transmembrane domain of NPR-C. The results shown in Fig. 3 demonstrate that R37A and TMC were equipotent and inhibited rat heart adenylyl cyclase in a concentration-dependent manner with an apparent K i of about 1 nM. In comparison, ANP was more potent (K i ϳ0.1 nM) and inhibited adenylyl cyclase activity by about 50 -55% compared with 35-45% for R37A and TMC. Thus, the free cytoplasmic domain of NPR-C can inhibit adenylyl cyclase activity, whereas the NPR-C homodimer inhibits adenylyl cyclase in a ligand-dependent manner. The inhibitory effect of these peptides was not due to the presence of positive charges per se, since the R37A scrambled peptide K37E has the same amino acid composition and did not exert any inhibitory effect on adenylyl cyclase (Fig. 4).
Dependence of R37A-and TMC-mediated Inhibition of Adenylyl Cyclase on Guanine Nucleotides-The inhibitory effect of cANP (4 -23) or ANP on adenylyl cyclase has been shown to be dependent on the presence of guanine nucleotides (32,33,37). These results are suggestive of a G protein-coupled signal transduction pathway. If the cytoplasmic domain of NPR-C mediates inhibition through the same or similar pathway as Effect of PT on R37A-and TMC-mediated Inhibition of Adenylyl Cyclase-Additional evidence for the role of a G proteincoupled pathway was obtained using PT. As shown in Fig. 6, R37A (panel A) and TMC (panel B) inhibited adenylyl cyclase activity in a concentration-dependent manner in heart particulate fractions, which was completely abolished by PT treatment. These results suggest that R37A and TMC, like cANP (4 -23) and ANP inhibited the enzyme activity through a PTsensitive G protein.
Effect of TMC on Adenylyl Cyclase Activity in Solubilized Membranes-To investigate if NPR-C or the cytoplasmic domain peptide TMC can inhibit adenylyl cyclase in solubilized membranes, heart particulate fraction was solubilized in nonionic detergent and used to assay for effects on adenylyl cyclase activity. 10 M GTP␥S stimulated enzyme activity in control and solubilized heart membranes by about 7-and 3-fold, respectively, whereas isoproterenol and c-ANP(4 -23), which stimulated or inhibited adenylyl cyclase activity by about 45 and 25% in control membranes, failed to stimulate or inhibit, respectively, cAMP production in the solubilized preparation (Table II). These results indicate a dissociation/uncoupling of the ␤-adrenergic receptor as well as NPR-C from the adenylyl cyclase system by detergent solubilization. In contrast, TMC was still able to inhibit GTP␥S-stimulated adenylyl cyclase activity by about 25%, and forskolin, which activates adenylyl cyclase by a receptor-independent mechanism, also stimulated the enzyme activity by about 9-fold. A concentration-response experiment in Fig. 7 shows the effect of cANP (4 -23) and TMC on adenylyl cyclase activity in detergent-solubilized heart particulate fraction. TMC inhibited adenylyl cyclase activity in a concentration-dependent manner in the solubilized fractions, whereas cANP(4 -23) was unable to elicit any inhibition of the enzyme activity. The simplest interpretation of the inhibition of adenylyl cyclase by TMC in the solubilized membrane fractions is that TMC may be (directly?) activating a G proteincoupled pathway.

DISCUSSION
Previous observations, from our group and others, on ANP inhibition of adenylyl cyclase through a receptor with NPR-Clike pharmacology (17,18), led us to design experiments to support further the role of this receptor. We raised polyclonal antibodies against the short 37-amino acid cytoplasmic domain of NPR-C (synthetic peptide R37A). The specificity of anti-R37A antisera against NPR-C was established by Western blotting against recombinant NPR-C and rat heart particulate fraction. Blocking of ANP signaling by anti-R37A in a rat heart particulate fraction indicates that NPR-C is involved in signaling, and another unidentified receptor need not be invoked to account for this ANP signal transduction pathway.
Our studies exploring the effect of the two NPR-C cytoplasmic domain peptides R37A and TMC on signal transduction are consistent with previous reports on ANP signaling via a PTsensitive G protein (17,18). The inhibitory effect of these peptides on adenylyl cyclase was not due to the net positive charge present (i.e. amino acid composition) since the scrambled peptide K37E with the same composition as R37A but a different sequence did not inhibit adenylyl cyclase activity. These data also suggest that there is structural specificity of the ANP-C cytoplasmic domain peptides to exert the inhibitory effects on adenylyl cyclase. The inability of cANP(4 -23) and isoproterenol to inhibit or stimulate adenylyl cyclase, respectively, in detergent-solubilized particulate fraction suggests that the NPR-C and ␤-adrenergic receptors are dissociated or uncoupled from adenylyl cyclase by detergent treatment. However, the inhibition of adenylyl cyclase by TMC in these detergent-solubilized membrane preparation lends itself to the interpretation that there may be a direct peptide/G protein interaction.
Receptor coupling to G proteins in hormonal inhibition of adenylyl cyclase is typically with seven transmembrane domain receptors such as the angiotensin II, ␣ 2 -adrenergic, endothelin B, muscarinic 2 and 4 subtypes, or the dopamine D2 receptor, in which the third cytoplasmic loop plays a role in determining G protein coupling and specificity (41)(42)(43). However, not all pathways for G protein activation need necessarily involve a seven-transmembrane domain receptor. The neuronal protein GAP-43 stimulates GTP binding to G o , suggesting that an intracellular protein can activate a G protein pathway (44). In addition, mastoparan, a 14-amino acid wasp venom peptide, activates G proteins directly (45). The single transmembrane domain insulin and insulin-like growth factor II/ mannose 6-phosphate receptors have also been reported to FIG. 7. Inhibition of adenylyl cyclase in detergent-solubilized rat heart fraction. Adenylyl cyclase activity was determined in the presence of the NPR-C cytoplasmic domain peptide TMC or the NPR-C specific ANP analog cANP (4 -23). Values are the means Ϯ S.E. of three separate experiments. The basal enzyme activity was 524 Ϯ 9 pmol of cAMP/mg of protein/10 min. activate protein-coupled pathways (46 -48). The relationship of these activation pathways to that reported here remains to be explored. Regulation of cellular cAMP production is subject to an interplay of positive and negative signals of at least six isoforms of adenylyl cyclase (30,31). Inhibition of cAMP production is through the inhibitory G proteins (G i ␣1, G i ␣2, and G i ␣3) and G␤␥ released by G o and G i ; however, not all forms of cyclase are sensitive to this pathway (49). Of the six completely cloned forms of adenylyl cyclase (50), type I adenylyl cyclase is a brain-specific enzyme activated by calmodulin and inhibited by ␤␥ (51) types V and VI expressed in heart are subject to inhibition by G i ␣1, G i ␣2, and G i ␣3, but not by G o ␣ and G␤␥ (51); (G i ␣1 is absent in heart (52)); whereas type IV also expressed in heart may not be sensitive to inhibitory G proteins (49). Taken together, it may thus be suggested that NPR-C-mediated inhibition of rat heart adenylyl cyclase by cytoplasmic domain peptides may be through their interaction with G i ␣2 or/and G i ␣3 and not through G o ␣ or G␤␥.
In conclusion, our data provide evidence for the involvement of NPR-C in signal transduction and extend previous pharmacological studies on G protein-mediated inhibition of adenylyl cyclase with NPR-C-specific ANP variants. A short 37-amino acid cytoplasmic domain appears to be sufficient to activate this response. The role of this signal transduction pathway in the in vivo physiology of the natriuretic peptides ANP, BNP, or CNP remains to be explored fully. The use of guanylyl cyclasespecific natriuretic peptide variants (53) or the application of gene targeting techniques to eliminate receptor expression may help address this question.