The guanylyl cyclase-A receptor transduces an atrial natriuretic peptide/ATP activation signal in the absence of other proteins.

Attempts to activate partially purified preparations of the guanylyl cyclase-A (GC-A) receptor with atrial natriuretic peptide (ANP) have previously failed, leading to speculation that essential cofactors are lost during purification procedures. The receptor was modified to contain the FLAG epitope (DYKDDDDK), expressed in Sf9 cells, and purified to apparent homogeneity (4.3 mumol cyclic GMP formed/min/mg protein; 5.8 mmol 125I-ANP binding site/mg protein) by a combination of immunoaffinity, Q-Sepharose FF, and wheat germ agglutinin batch chromatography. High initial protein/detergent ratios, the presence of glycerol (40%), and the inclusion of protein phosphatase inhibitors in all buffers resulted in the purification of a receptor that continued to transduce the ANP/ATP activation signal. Both native and purified GC-A contained a single class of high affinity ANP binding sites (Kd = 60 pM) and an equivalent EC50 for ATP (0.3 mM). Positive cooperativity as a function of MnGTP was retained during purification. Thus, GC-A is capable of transducing a ligand binding signal in the absence of other proteins.

The atrial natriuretic peptide receptor (NPR-A) coupled to guanylyl cyclase activity (GC-A) 1 is a prototype for a family of membrane-bound guanylyl cyclases currently numbering 6 in mammals (A-F). The B and C cyclases appear to be receptors for C-type natriuretic peptide and heat-stable enterotoxins/ guanylin, respectively, while D, E, and F remain orphan receptors (1)(2)(3)(4).
Members of the family contain a single apparent transmembrane domain that separates an amino-terminal extracellular, ligand binding domain from an intracellular region that contains both a cyclase catalytic domain and a protein kinase homology domain (KHD) (1). The function of the KHD is not fully understood, but since ATP is required for transduction of the ANP binding signal to activation of cyclase catalytic activity, and since various mutations within the KHD interrupt ANP/ATP signaling (5)(6)(7), it has been suggested that the protein kinase homology domain binds ATP (6). ATP appears to act without conversion to ADP or other metabolites since nonhydrolyzable analogues of ATP also substitute for ATP (8,9). The KHD of the cyclases resembles the kinase homology domain 2 of the JAKS; an invariant Asp found in the catalytic domain of known protein kinases is a different amino acid in either the cyclases or kinase homology domain 2 of the JAKS. Although the requirement of ATP and peptide ligand for activation of GC-A is somewhat analogous to the guanine nucleotide-dependent activation of adenylyl cyclase by hormones, non-hydrolyzable analogues of GTP activate adenylyl cyclase in the absence of hormone, and sub-micromolar as opposed to millimolar concentrations of nucleotide are effective (10).
During previous attempts to purify GC-A, the ANP/ATP activation signal has vanished, even though ANP has continued to bind to the receptor (11,12). Thus, it was suggested that additional factor(s), specifically ATP binding proteins (8), are required for transduction of a hormone binding signal to activation of the cyclase catalytic domain. Now we show, however, that after purification of detergent-solubilized GC-A to apparent homogeneity, the cyclase retains ability to be activated by ANP/ATP.

EXPERIMENTAL PROCEDURES
Materials-Sf9 cells (13) were obtained from the American Type Culture Collection. BaculoGold virus was from Pharmingen. cDNA encoding FLAG-GC-A was kindly provided by Dr. Michael Chinkers (Vollum Institute, Oregon Health Sciences Center). Lubrol 12A9 (from ICI) was a gift from Dr. Elliott Ross (University of Texas Southwestern Medical Center at Dallas). Polyoxyethylene-10-lauryl ether was from Sigma, M1 and M2 immunoaffinity columns were from IBI, wheat germ agglutinin-agarose was from Vector, and Q-Sepharose FF was from Pharmacia. 125 I-ANP, [␣-32 P]GTP, and [ 3 H]cGMP were from DuPont NEN, and rat ANP (99 -126) was from Peninsula. Reagents used for radioimmunoassay of cGMP have been previously described (14).
Expression of FLAG-GC-A in Sf9 Cells-The baculovirus expression vector PVL-FLAG-GC-A was cotransfected with BaculoGold virus into Sf9 cells, and the recombinant virus was then cloned according to Kitts et al. (15). Sf9 cells infected with the recombinant virus expressed ϳ0.7-1 pmol of 125 I-ANP binding site/mg of particulate protein, while no significant binding was detected in uninfected cells.
Solubilization of FLAG-GC-A-48 h post-infection with the recombinant virus, Sf9 cells expressing FLAG-GC-A were pelleted and frozen in liquid nitrogen. The frozen pellets were thawed in the presence of a solution containing 50 mM HEPES, pH 7.5, 100 mM NaCl, 40% glycerol, 1 mM EDTA, 10 g/ml leupeptin, 1 mM ␤-glycerol phosphate, and 5 mM sodium azide (buffer A). The thawed cells were broken in a Dounce homogenizer, and crude particulate fractions were obtained and washed by centrifugation (30,000 ϫ g) in the same buffer. The particulate fraction was resuspended in buffer A at a concentration of 20 -30 mg of protein/ml. FLAG-GC-A was then solubilized by rocking the solution for 60 min at 4°C in buffer A containing 1% Lubrol 12A9 or polyoxyethylene-10-lauryl ether. After 60 min, the detergent extract was diluted with buffer A to 0.5% of either non-ionic detergent and centrifuged at 100,000 ϫ g. The supernatant fluid was obtained and diluted to 0.1% detergent with buffer A. Approximately 20 -40% of the 125 I-ANP binding sites detected in the particulate fraction were solubilized. Higher recoveries (up to 70 -80% of 125 I-ANP binding sites could be obtained at higher detergent:protein ratios; however, enzyme solubilized under these conditions lost ANP and ATP sensitivity (not shown).
Purification of FLAG-GC-A-All purification procedures were performed at 4°C. Detergent extracts (200 ml) were rocked with 6 ml of M2 immunoaffinity resin for 2-3 h and then washed batchwise by centrif-* Supported by National Institutes of Health Grants 7R01-HL49588 (to S. K-F. W.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 The abbreviations used are: GC-A, guanylyl cyclase-linked atrial natriuretic peptide receptor; ANP, atrial natriuretic peptide; KHD, kinase homology domain. ugation (10 times, 30 ml) with buffer B (buffer A containing 0.05% detergent, 0.25 mM ATP, 150 mM NaCl, and 0.025% phosphatidylcholine). FLAG-GC-A was eluted from the resin by rocking with 6 ml of buffer B containing 0.1 mg/ml of FLAG peptide for 30 min. After centrifugation at 3000 ϫ g, FLAG-GC-A was collected in the supernatant fluid. The above elution procedure was repeated 10 times. The eluates were then pooled and filtered through a 0.45-m filter, and the filtrate was then mixed with 2 ml of Q-Sepharose FF for 30 min and centrifuged. FLAG-GC-A was collected in the supernatant fluid. The Q-Sepharose FF resin was then washed three times with 10 ml of buffer B by centrifugation, and additional FLAG-GC-A was also collected in the supernatant fluid. The supernatant fluids were then pooled and rocked with 10 ml of wheat germ agglutinin-agarose for 2 h and centrifuged. The pellet was washed batchwise (10 times, 40 ml of buffer B) by centrifugation. FLAG-GC-A was then eluted from the resin by rocking with 10 ml of buffer B containing 0.3 M N-acetylglucosamine for 20 min. After centrifugation, FLAG-GC-A was collected in the supernatant fluid. The procedure was repeated five times. The supernatant fluids were then pooled and filtered through a 0.45-m filter. The filtrate was then made 2.25 mM with respect to CaCl 2 and added to 2 ml of M1 immunoaffinity resin. The slurry was rocked for 2-3 h and washed batchwise by centrifugation (10 times, 10 ml of buffer B containing 2.25 mM CaCl 2 ), and then purified FLAG-GC-A was eluted from the resin by rocking 30 min with 1 ml of buffer B containing 2 mM EDTA. FLAG-GC-A was collected in the supernatant fluid after centrifugation. The elution procedure was repeated four times. Purified FLAG-GC-A was subsequently stored at Ϫ80°C. The purification procedures were routinely performed within 20 -24 h to minimize losses in ANP/ATP sensitivity.
125 I-ANP Binding Assay-ANP binding activity was determined as described previously (12), with the exception that 1% polyethyleneimine-treated Whatman GF/C filters were used for filtration. Nonspecific binding was determined using 1 M non-radioactive ANP (99 -126), and data were analyzed by nonlinear curve fitting using the LIGAND computer program (16).
Guanylyl Cyclase Assay-Guanylyl cyclase activity was determined by incubating samples (0.5-1 fmol 125 I-ANP binding sites) at 37°C for the indicated time intervals in 0.1 ml of a reaction mixture consisting of 25 mM HEPES, pH 7.5, 50 mM NaCl, 0.1% bovine serum albumin, 0.25 mM 1-methyl-3-isobutylxanthine, 6 mM MgCl 2 , 5 mM creatine phosphate, 4 units of creatine phosphokinase, and 0.5 mM GTP. To determine the enzyme activity using MnGTP as substrate, 0.1% Triton X-100 and 6 mM MnCl 2 also were added. The reaction was initiated by the addition of enzyme samples and terminated by the addition of 0.45 ml of 0.11 M zinc acetate, followed by 0.45 ml of 0.11 M Na 2 CO 3 . After centrifugation at 3000 ϫ g, 5 or 45 l of the supernatant fluid were assayed by radioimmunoassay for cyclic GMP content (14). For the MgGTP dependence of the ANP/ATP experiment (Fig. 5), [␣-32 P]GTP was used as substrate, and the reaction was terminated by 0.1% SDS. The amount of [ 32 P]cGMP was determined as described previously for adenylyl cyclase assays (17), with the modification that only 0.2 ml of H 2 O was added in the initial wash of the Dowex.
Iodination of Purified Receptor-Purified FLAG-GC-A was incubated with 100 Ci of [ 125 I]NaI for 20 min at room temperature in the presence of 0.1 M Tris-Cl, pH 6.8, and 1 pellet of IODOBEAD (Pierce). After removing the supernatant fluid from the IODOBEAD, free [ 125 I]NaI was removed by a G-25 spin column. The sample was then electrophoresed using an 8% separating gel by the method of Laemmli (18). The gel was stained with silver (19), vacuum-dried, and subjected to autoradiography with HYPER-FILM-MP (Amersham).
Protein Determination-Protein content of the crude membrane fraction and the detergent extract was determined by amido black staining (20) using bovine serum albumin (Pierce, no. 23209) as a standard. The protein content of the purified FLAG-GC-A was determined by densitometry according to Merril and Pratt (21), as modified as follows. Purified FLAG-GC-A and bovine serum albumin standards (0.1-10 ng) were electrophoresed on SDS-polyacrylamide gels and then silver stained (19). After drying between transparent cellulose sheets (Bio-Rad), the protein content of the FLAG-GC-A sample was determined by comparing the intensity of staining with that of the bovine serum albumin. The intensity of staining was determined in a densitometric scanner (Molecular Dynamics model 300A) under conditions where the intensity was linearly proportional to the protein content.

RESULTS AND DISCUSSION
Expression of Recombinant FLAG-GC-A in Sf9 Cells-To facilitate rapid purification of the receptor, the FLAG epitope  (DYKDDDDK) was introduced just to the carboxyl side of the signal peptide, thereby representing the amino terminus of the processed cyclase. The insertion of the FLAG epitope did not appear to alter biological activity of the enzyme since the number of 125 I-ANP binding sites/mg protein or the relative cyclase activation in response to ANP/ATP in Sf9 membrane preparations was equivalent whether wild-type or FLAG-GC-A cyclase was expressed (data not shown). ANP/caged ATP maximally activated FLAG-GC-A at 48 h following infection with recombinant baculovirus, at which time 0.6 -1.0 pmol of 125 I-ANP binding sites/mg of protein in the crude particulate fraction were obtained (Table I). With the described protocol to grow Sf9 cells (2% fetal bovine serum in IPL-41 in suspension), 0.7-1.0 nmol of 125 I-ANP binding sites/ liter of culture were routinely obtained.
Solubilization of Epitope-tagged GC-A-Initial experiments using many types of detergents such as Triton X-100 and deoxycholate to solubilize GC-A resulted in complete losses of ANP/ATP sensitivity (not shown). Of the various detergents tested, Lubrol 12A9 or polyoxyethylene-10-lauryl ether resulted in the highest ANP/ATP-responsive cyclase. However, high detergent/protein ratios during the initial solubilization decreased responsiveness to ANP/ATP. The ANP/ATP response of GC-A also was very sensitive to the concentration of glycerol, and the response to ANP/ATP was largely attenuated when solubilization buffers contained 20% glycerol but was retained in 40% glycerol. For the highest recovery of ANP/ATP-respon-sive cyclase, FLAG-GC-A was routinely solubilized with 1% detergent at protein concentrations of greater than 20 mg/ml and 40% glycerol. Under these conditions, 41% of the initial 125 I-ANP binding activity and 7% of the particulate protein were solubilized (Table I). The K d for 125 I-ANP binding for the solubilized and the membrane-bound receptor was the same under these conditions ( Table I).
Purification of Ligand-sensitive FLAG-GC-A-The main objective of this study was to determine whether or not GC-A, serving as a prototype for the membrane guanylyl cyclases, could signal in the absence of other protein factors. Batch procedures using peptide competition (M2 immunoaffinity resins), N-acetylglucosamine competition (wheat germ agglutininagarose), and EDTA chelation (M1 immunoaffinity resin) were used to minimize the time required for purification. The capacity of the M2 immunoaffinity resin in buffer B restricted the amount of 125 I-ANP binding sites applied to 40 pmol/ml resin.
Purity of FLAG-GC-A-The silver-stained FLAG-GC-A recovered after the above purification steps was the only major band detected (Fig. 1), and densitometry of the silver-stained gel suggests that it is greater than 95% pure. The purity also was examined by the use of nonspecific radioiodination. FLAG-GC-A was the only major iodinated species, although longer exposures revealed a minor contaminant at about 70 kDa, a protein not detected by silver staining. The specific activity of the receptor preparation of 5.8 nmol of 125 I-ANP bound/mg protein is comparable with previously purified ANP receptor (6 -7 nmol/mg protein), suggestive of a highly purified cyclase preparation (12,22).
Ligand Binding Characteristics of FLAG-GC-A-With either the crude detergent extract or the purified preparation of cyclase, a single class of high affinity 125 I-ANP binding sites (K d ϭ 65 pM) was obtained (Fig. 2).
Ligand/ATP Responsiveness of FLAG-GC-A-In crude particulate fractions, ANP or caged ATP activated the cyclase by 1.3-or 1.7-fold, respectively, when added alone, but when added together, the cyclase was activated approximately 16fold. The requirement for both ANP and the nucleotide, ATP, for maximum stimulation of GC-A has been previously reported (5,8,9). After detergent extraction, ANP and ATP continued to stimulate minimally when added alone, but together stimulated the cyclase about 9-fold. However, the detergent-solubilized GC-A is highly labile. Considerable ANP/ATP stimulation (50%) was lost upon storage at 4°C for 24 h, and only about 20% remained after incubation at 25°C for 60 min. 2 Following an approximate 1300-fold purification of GC-A, the receptor remained responsive to ATP/ANP (Fig. 3). About 10% or 30% activations of FLAG-GC-A were obtained when ANP or caged ATP were added alone, respectively. When an equivalent number of 125 I-ANP binding sites (0.5-1 fmol ANP binding sites) were added per assay, the amount of cGMP produced upon incubation with ANP/ATP was approximately the same for the detergent extract or the purified enzyme preparation (see also Fig. 4, A and B). Therefore, the maximal responsiveness to ANP/ATP did not change significantly during the purification procedure. The basal activities were 2-3fold higher in the purified preparation, and consequently the -fold stimulation by ANP/ATP decreased from approximately 8-fold in the crude detergent extract to 3-4-fold in the purified GC-A preparation. The guanylyl cyclase reaction was linear for at least 5 min, and as shown in Fig. 5, ANP/ATP appears to alter V max as a function of MgGTP. ATP at millimolar concentration is required for maximal stimulation by ANP (Fig. 5). Although millimolar concentrations of ATP are quite high when compared to the binding of GTP to G proteins, which regulate the adenylyl cyclase system (EC 50 ϳ0.1 M), millimo- Kinetics of Activation by MnGTP-Membrane forms of guanylyl cyclases normally display positive cooperative kinetics as a function of MnGTP. FLAG-GC-A also yielded positive kinetics with respect to MnGTP in either the crude detergent extract or after purification (Fig. 6). This contrasts to the sea urchin guanylyl cyclase, which lost cooperative kinetics as a function of MnGTP after purification in initial work (23) but is consistent with later studies where the phosphorylated form of the sea urchin cyclase was purified and retained positive cooperativity (24).
Other Activities of GC-A-Could the KHD act as a protein kinase or as an ATPase? Using an ATPase assay (lower detection limit, 3 nmol/min/mg protein) (25), we have not observed any detectable ATPase activities from 150 ng of our purified preparation of FLAG-GC-A (not shown). Using nonspecific kinase substrates such as histone IIIs and VIIs, we have not observed detectable protein kinase activities in the purified preparations of FLAG-GC-A (not shown). This is consistent with the loss of protein kinase activity observed in the c-Kit receptor tyrosine kinase from the W 42 allele of the white spotting locus where a conserved Asp in the catalytic domain is mutated to an Asn (26). An Asn rather than an Asp is also found in the same relative position in GC-A. Also, the absence of a critical lysine residue known to be important for kinase activity (27) suggests that protein kinase activity may be absent in the cyclases. That GC-A contains the glycine loop implicated in nucleotide binding suggests that it can bind ATP. The failure to detect protein kinase activity with GC-A, how-ever, is consistent with its missing several critical residues found within the catalytic pocket of protein kinases (28).
GC-A represents a prototype of the particulate guanylyl cyclase family where there are at least six subtypes identified so far (1)(2)(3)(4). The results from this study initially show that the ligand/ATP are able to activate this class of particulate receptor guanylyl cyclases independent of other protein cofactors. In this unique signaling system, which utilizes ATP as the signal tranduction molecule, the domains for ligand binding, ATP regulation, and the effector guanylyl cyclase are located in one molecule. This is in contrast to the G protein-coupled receptor system, where the components for ligand binding, signal transducing, and effector are located in separate protein molecules. The mechanism of activation of GC-A is not yet understood, but the ANP/ATP activation is known to be regulated by phosphorylation (29 -31) and oligomerization (32,33). Although additional protein factors therefore are not required for signal transduction, it also seems likely that modulating factors will be discovered based on precedence with other cell surface receptors.