|Home | About | Journals | Submit | Contact Us | Français|
Soluble guanylate cyclase (sGC) serves as a receptor for the signaling agent nitric oxide (NO). sGC synthesis of cGMP is regulated by NO, GTP, ATP and allosteric activators such as YC-1. Guanylate cyclase and the adenylate cyclase activity of full-length sGC and the sGC catalytic domain constructs (α1catβ1cat) is reported here. ATP is a mixed-type inhibitor of cGMP production for both sGC and α1catβ1cat indicating that the C-terminus of sGC contains an allosteric nucleotide binding site. YC-1 did not activate α1catβ1cator compete with ATP inhibition of cGMP synthesis, which suggests that YC-1 and ATP bind to distinct sites. α1catβ1cat and NO-stimulated sGC also synthesize cAMP, but this activity is inhibited by ATP via non-competitive substrate inhibition, and by GTP via mixed-type inhibition. Additionally, the adenylate cyclase activity of purified sGC was inhibited by PC12 lysate, suggesting that an intracellular small molecule or protein regulates this activity in vivo.
Soluble guanlyate cyclase (sGC) is a heterodimeric hemoprotein that is essential to several physiological processes (1–3). sGC synthesis of cGMP from GTP is activated by the signaling agent nitric oxide (NO). NO activation of sGC is complex and is known to be influenced by the binding of GTP and ATP to an allosteric binding site on the protein (4, 5). In addition to nucleotides, several small molecules have been shown to influence sGC activity (6, 7). Despite several studies focused on understanding sGC regulation by nucleotides and allosteric activators, many questions pertaining to the mechanism of this regulation remain unresolved.
sGC consists of two homologous subunits, α1 and β1. Architectural information about sGC has been advanced by the expression and isolation of the minimal heme binding domain β1(1–194) 2 (8) and the catalytic domains α1(467–690) and β1(414–619) (9). The functional catalytic domains, termed α1catβ1cat, contain a pseudosymmetric active site. This pseudosymmetric site contains residues known to be involved in nucleotide binding, but lacks the amino acids required for catalysis (10). Although nucleotide binding to this pseudosymmetric site has not been directly observed, it is known that sGC binds two equivalents of substrate (11) and that nucleotide binding to an allosteric site influences activity (5, 12).
Pharmaceutical screens have also identified sGC allosteric activators including YC-1 (7), and its more soluble derivative BAY 41–2272 (13). These compounds activate the FeII-unligated sGC weakly (2- to 4-fold), but significantly increase sGC activity when a ligand is bound at the FeII heme (14, 15). This synergistic activation leads to a FeII-CO complex that is activated 100- to 400-fold and a FeII-NO complex that is activated 200- to 400-fold. While the precise mechanism of activation by these compounds is unknown, it has been proposed that they are GTP mimics that bind to the psuedosymmetric site (10, 11).
Detailed spectroscopic and kinetic studies on the effects of GTP, ATP, and YC-1 on sGC have been previously reported (4, 11, 16–19); however, it remains unclear where these molecules bind and how they function to regulate sGC activity. These questions were addressed here by examining guanylate cyclase (GC) and adenylate cyclase (AC) activity of both full-length sGC and α1catβ1cat. The kinetic study reported herein suggests that sGC and α1catβ1catcontain two nucleotide binding sites that are able to bind both GTP and ATP, thus localizing the allosteric GTP/ATP binding site to the C-terminus of sGC. We observed that GTP binding to this site leads to inhibition of ATP turnover, whereas ATP binding to this site leads to inhibition of both GTP and ATP turnover. Additionally, we observed that although α1catβ1catcontains a functional allosteric nucleotide binding site, it is unresponsive to YC-1 or BAY 41–2272. This work shows that sGC differentially responds to GTP and ATP at physiologically relevant nucleotide concentrations, and indicates that distinct conformational changes occur within the catalytic domain depending on the occupation of the allosteric nucleotide binding site.
sGC was expressed using a baculovirus/Sf9 expression system and purified as previously described (19). The sGC catalytic domains (α1cat and β1cat) were expressed and purified as previously described (9). Diethylammonium (Z)-1-(N,N-diethylamino) diazen-1-ium-1,2-diolate (DEA/NO) was from Cayman Chemical Co. All other reagents were from Sigma, unless otherwise noted.
sGC activity was examined in the presence of various allosteric regulators by performing duplicate end-point assays at 37 °C. The assay mixture contained 50 mM Hepes (pH 7.4), 10 mM MgCl2, and 1 mM DTT. When present, YC-1 and BAY 41–2272 (both in DMSO) were at 150 µM and 20 µM, respectively, and the final concentration of DMSO was 2% (v/v), which was shown not to affect enzyme activity. The GTP concentration ranged from 0.0025 to 3 mM and the ATP concentration ranged from 0 to 5 mM. In each assay, sGC (0.2 µg) was incubated with ATP for 1 min at 37 °C before initiating the reaction with GTP. All assays were in a final volume of 100 µL. Reactions were quenched after 2 minutes by the addition of 400 µL of 125 mM Zn(CH3CO2)2 and 500 µL of 125 mM Na2CO3. cGMP quantification was carried out using a cGMP enzyme immunoassay kit, Format B (Assay Designs), per the manufacturer’s instructions. The resulting data were fit to the Michaelis-Menten equation, v = Vmax[GTP]/(KM + [GTP]) to obtain KM and Vmax in the presence of ATP. The experiments were repeated 3 times to ensure reproducibility.
Adenylate cyclase assays were preformed as described above but with the following differences. In each assay sGC (0.2 µg) was incubated with or without GTP (0.5 mM) for 1 min at 37 °C before initiating the reaction with ATP (0–10 mM). Reactions were quenched after 5 minutes and cAMP quantification was carried out using a cAMP enzyme immunoassay kit, Format B (Assay Designs), per the manufacturer’s instructions. The resulting data were fit to the equation for non-competitive substrate inhibition to obtain KM, KI and Vmax, where α is the maximum velocity in the presence of saturating substrate (20).
The experiments were repeated 3 times to ensure reproducibility.
The sGC catalytic domains were assayed as described above but with the following exceptions. α1catβ1cat (10 µg heterodimer) was assayed in 50 mM Hepes (pH 7.4), 6 mM MnCl2, 2 mM DTT and varying amounts of GTP or ATP as indicated. Reactions were quenched after 16 minutes and cGMP or cAMP quantification was carried out as previously described.
Semi-adherent PC12 cells were obtained from the American Type Culture Collection (ATCC) and were maintained in a humidified atmosphere of 95% air and 5% CO2 at 37 °C in Dulbecco’s modified Eagle’s medium (Invitrogen) supplemented with 10% fetal bovine serum (FBS), l-glutamine (200 µg/mL), penicillin (100 U/mL), and streptomycin (100 µg/mL). Cells were counted with a hemocytometer and viability was routinely assessed by trypan blue staining and found to be >90%. PC12 cells in 150 × 15 mm petri dishes were washed extensively with ice-cold assay buffer (50 mM Hepes (pH 7.4), 50 mM NaCl) and harvested with a rubber policeman. Cells were collected by centrifugation at 4 °C and gently resuspended in ice-cold assay buffer plus 1 mM 3-isobutyl-1-methylxanthine (IBMX). Whole cell assays were initiated by adding 1 µM PROLI/NO (Cayman Chemicals) to 100 µL of cells (50–100 µg of total protein) at 37 °C. Reactions were quenched 5 seconds after NO addition with 1 volume of 133 mM HCl. For lysate assays cells were resuspended in 50 mM Hepes (pH 7.4), 50 mM NaCl, 5 mM DTT, 1 mM benzamidine, 1 mM Pefabloc (Roche), 1 mM IBMX, plus Complete EDTA-free protease inhibitor cocktail (Roche) and lysed by sonication. Cell lysate was equilibrated at 37 °C for 10 seconds before initiating with 1 mM GTP or ATP/3 mM MgCl2. Assays contained 100 µM DEA/NO and/or 13 nM purified sGC where indicated, and were quenched after 5 minutes by addition of 133 mM HCl. Protein concentration was measured in each sample using the Bradford Microassay (Bio-Rad Laboratories) against a bovine serum albumin standard and cGMP or cAMP was quantified using an enzyme immunoassay kit, Format A (Assay Designs), per the manufacturer’s instructions. The detection limit of cAMP was 0.078 pmol/mL and the detection limit of cGMP was 0.08 pmol/mL, and all the data reported in this work was above these limits. Each assay point was done in duplicate and all experiments were repeated 3 times to ensure reproducibility.
The effect of ATP on sGC activity was examined with both full-length sGC (Figure 1A) and α1catβ1cat (Figure 1B). Figure 1A shows that ATP is a mixed-type inhibitor of FeII-unligated sGC as indicated by an increase in the apparent KM and a decrease in the Vmax. This is in agreement with previously published reports (5, 12) and indicates that ATP inhibits full-length sGC by binding to both the catalytic site as well as an allosteric nucleotide binding site. Kinetic data from previous reports show that this mechanism of ATP inhibition is observed with NO-stimulated sGC (5, 12, 21). This indicates that both nucleotide binding sites are accessible in the stimulated and basal sGC states. Figure 1B shows that ATP is also a mixed-type inhibitor of α1catβ1cat. α1catβ1catactivity was measured in the presence of Mn2+ because the complex is 50-fold more active in the presence of Mn2+ when compared to Mg2+ (9). α1catβ1catconsists of the C-terminus of both α1 and β1 (residues 467–690 of α1 and 414–619 of β1), and contains the catalytic site and the pseudosymmetric substrate binding site. This result localizes the allosteric nucleotide binding site to the C-terminus of sGC and establishes that the catalytic domains can be further used as a model to study the allosteric nucleotide binding site.
Different binding sites have been proposed for YC-1, including the N-terminus of the α1 subunit (13, 22) and the pseudosymmetric substrate site (11, 23, 24). Others have proposed that both ATP and YC-1 bind to the pseudosymmetric site (10, 11). To examine this possibility, full-length sGC and α1catβ1catactivity were examined in the presence of both ATP and YC-1. If YC-1 and ATP compete for the same site, then ATP should inhibit YC-1 activation of sGC. Figure 2A shows that YC-1 activates full-length sGC 10– 15-fold regardless of the presence of ATP (2.5 mM). Furthermore, there was no decrease in the fold activation of sGC by YC-1 at any concentration of ATP examined (0–5 mM). The sGC catalytic domain constructs, which contain the allosteric ATP binding site, were insensitive to YC-1 activation and YC-1 had no effect on the ATP-induced inhibition (Figure 2B). Figure 2 shows full-length sGC and α1catβ1cat activity at a single ATP concentration, but we observed consistent results at several nucleotide concentrations (0.1, 0.5, 1.0, 2.5, and 5.0 mM ATP). Additionally, the more soluble YC-1 derivative, BAY 41–2272, had no effect on the guanylate cyclase activity of α1catβ1cat (data not shown). The overall conclusion is that ATP and the allosteric activators YC-1 and BAY 41–2272 bind to distinct sites on sGC.
The mixed-type inhibition observed for both full-length sGC and α1catβ1cat (Figure 1A and B) indicates that ATP binds to the sGC catalytic site in addition to an allosteric nucleotide binding site, and indicates the potential for ATP turnover by the enzyme. sGC-catalyzed adenylate cyclase (AC) activity has been reported (25, 26); however, a thorough kinetic characterization of this activity had not been previously performed. Therefore, cAMP formation at various ATP concentrations for both NO-stimulated sGC and α1catβ1cat was measured.
Adenylate cyclase activity was observed for both NO-stimulated full-length sGC and α1catβ1cat (Figure 3A and B). The amount of turnover observed for sGC in the absence of NO was very low (0.1 nmol cAMP/min/mg at 1 mM ATP). Due to the low specific activity of the unligated enzyme and limitations in acquiring purified protein, the sGC KM for ATP in the absence of NO was not determined; however, it is clear that the adenylate cyclase activity of purified sGC is significantly activated by the presence of NO. This observation is in agreement with an earlier study with protein isolated from various rat tissues (26). The presence of YC-1 also stimulated the AC activity of full-length sGC ~10-fold (data not shown). Substrate inhibition of adenylate cyclase activity was observed at high concentrations of ATP (0.5–10 mM) in both NO-stimulated full-length sGC and α1catβ1cat. Fitting the plots in Figure 3 to a non-competitive substrate inhibition equation yields a KM ATP of 52 µM and KI ATP of 1.4 mM for full-length sGC. α1catβ1catexhibits a decreased affinity for ATP relative to sGC as the KM ATP and KI ATP were determined to be 459 µM and 5.4 mM, respectively (Table 1). This suggests that ATP binding to the allosteric site leads to a conformational change at the catalytic site to inhibit turnover of both ATP and GTP. Additionally, the observation of this kinetic phenomena in both sGC and α1catβ1catconfirms that the allosteric nucleotide binding site is contained on the C-terminus of α1 and β1.
GTP inhibition of adenylate cyclase activity was also examined. Figure 4 shows that the presence of GTP (0.5 mM) decreased the Vmax for both NO-stimulated full-length sGC and α1catβ1catAC activity. This indicates that GTP is not strictly a competitive inhibitor of ATP, and suggests that GTP also binds to two sites on full-length sGC and α1catβ1cat (Figure 6).
Based on the surprising ability of sGC to cyclize ATP in vitro we sought to determine if NO-induced adenylate cyclase activity is observed in vivo. Using PC12 cells, an immortalized cell line derived from a rat pheochromocytoma which is known to express sGC, we found that sGC does not produce cAMP in response to NO in vivo. In both intact cells and PC12 lysate there was no increase in cAMP in response to NO, whereas cGMP levels increased several hundred fold in response to NO (Table 2). This suggested that a small molecule or protein present in PC12 cells could be inhibiting sGC adenylate cyclase activity. To test this possibility the activity of purified sGC in the presence and absence of PC12 lysate was measured. Figure 5 shows that the NO-induced guanylate cyclase activity of purified sGC (~370-fold) is not inhibited by the presence of PC12 lysate, but the NO-induced adenylate cyclase activity of purified sGC is completely inhibited by PC12 lysate, suggesting that ATP turnover by sGC is regulated by an intracellular protein or small molecule.
sGC is a critical cellular receptor for the gaseous signaling agent NO. In vivo sGC is regulated by a complex interplay between NO, GTP and ATP, and this regulation is essential for the physiological functions controlled by cGMP. Understanding sGC regulation by both nucleotides and synthetic allosteric activators, such as YC-1 and BAY 41–2272, is necessary to develop therapeutic agents to treat diseases related to dysfunction in sGC. To this end, we examined the effects of ATP, GTP, and YC-1 on the guanylate and adenylate cyclase activity of full-length sGC and α1catβ1cat.
Several groups have studied the interaction of ATP with sGC. These detailed studies provided kinetic and thermodynamic data to support the existence of two distinct nucleotide binding sites on sGC; a catalytic site and an allosteric nucleotide binding site (5, 11). Based on primary sequence analysis the C-terminus of sGC is predicted to contain the allosteric nucleotide binding site, but limitations in acquiring purified full-length protein, and the lack of sensitive tools for studying nucleotide binding have hindered the identification of this site. By using the sGC catalytic domains, α1catβ1cat, we were able to localize the allosteric nucleotide binding site to the C-terminus of sGC. Specifically, we found that ATP is a mixed-type inhibitor for α1catβ1cat (Figure 1) which indicates that ATP inhibits cGMP production by binding to two different sites on the C-terminus of sGC. We then used the catalytic domains as a probe to investigate the YC-1 binding site. We found that α1catβ1cat was unresponsive to YC-1, and that ATP did not alter the fold activation of full-length sGC induced by YC-1 (Figure 2). These observations suggest that YC-1 does not bind to the allosteric nucleotide binding site that is contained on the C-terminus of sGC. The N-terminus of the α1 subunit has been shown to bind YC-1 (13). YC-1 binding to a hydrophobic pocket within this domain (22) could induce a conformational change that effects both the sGC heme environment and the catalytic domain. This model of YC-1 activation is consistent with our current results and our previously reported spectroscopic studies (17, 19).
While researchers have studied the inhibition of sGC by ATP for over thirty years, the potential for adenylate cyclase activity with purified protein had not been thoroughly investigated. Recently the catalytic domain from a bacterial guanylate cyclase, Cya2, was shown to cyclize both GTP and ATP (27). Cya2 shows specificity for GTP verses ATP, and sequence analysis predicts that this mechanism of nucleotide specificity varies from that of eukaryotic sGCs (27). We have now found that both NO-stimulated sGC and α1catβ1cat cyclize ATP to cAMP (Figure 3). The AC activity of full-length sGC was significantly lower in the absence of NO (Table 2), indicating that NO stimulates cAMP synthesis several hundred fold. YC-1 also stimulates cAMP synthesis ~10-fold (data not shown) suggesting that NO and YC-1 are not selective for GTP-bound sGC. Interestingly, AC activity is inhibited at high ATP concentrations due to substrate inhibition. This phenomenon was not observed in the Cya2 homodimer and represents a unique effect of ATP binding to the allosteric nucleotide binding site on the eukaryotic heterodimeric enzyme.
As evaluated by the kcat for each nucleotide, sGC exhibits high specificity for GTP as a substrate, but the observed AC activity enabled us to further investigate the allosteric nucleotide binding site. By fitting the plots in Figure 3 to the non-competitive substrate inhibition equation, we obtained both the KM ATP and KI ATP. These values provide a means to estimate the binding affinity of ATP at the catalytic site (KM) and the allosteric nucleotide binding site (KI). From this we observe that the allosteric nucleotide binding site has a lower affinity for ATP (Table 1), which is in agreement with Kd estimations based on equilibrium dialysis experiments with ATP analogues (11). Since substrate inhibition of GC activity is not observed with GTP for either full-length sGC or α1catβ1cat, it suggests that either GTP has a significantly lower affinity to the allosteric site or that GTP binding does not inhibit GTP turnover. To evaluate these possibilities we tested the mechanism of GTP inhibition of AC activity. Figure 4 shows that GTP decreases Vmax of cAMP formation for both NO-stimulated full-length sGC and α1catβ1cat. This indicates that GTP is a mixed-type inhibitor of AC activity and suggests that both nucleotides bind to both the catalytic site and the allosteric nucleotide binding site. Interestingly, GTP binding to the allosteric nucleotide binding site inhibits ATP turnover but not GTP turnover. This differential effect indicates that distinct conformational changes occur within the catalytic site depending on the occupation of the allosteric nucleotide binding site. Furthermore, this differential nucleotide regulation reflects a broken symmetry that would not be possible with a homodimeric enzyme. Understanding these conformational changes is essential to elucidate the mechanism of nucleotide regulation of sGC.
To evaluate the role of ATP in regulating sGC activity in vivo we examined the effect of PC12 lysate on both GC and AC sGC activity. sGC produced endogenously in PC12 cells synthesizes cGMP in response to NO but not cAMP (Table 2). This discrepancy between AC activity in purified protein versus PC12 cells led us to question if an endogenously produced small molecule or protein could be influencing sGCs ability to discriminate between ATP and GTP in vivo. To test this proposal we assayed AC and GC activity of purified sGC in the presence and absence of PC12 lysate. Interestingly we observed the selective inhibition of NO-stimulated AC activity in the presence of PC12 lysate. It is unlikely that endogenous ATP or GTP in PC12 lysate could account for this inhibition as sGC maintains partial NO-stimulated activity even with saturating nucleotide (Figure 3 and Figure 4). Additionally, we observed that salt (50 mM NaCl) selectively inhibits sGC AC activity (data not shown). While the degree of sGC inhibition induced by the presence of salt is not enough to account for the effect observed in PC12 lysate, it lends support to the proposal that other small molecules may also influence nucleotide specificity. This intriguing data resolves a discrepancy in the literature between the absence of adenylate cyclase activity of sGC observed in COS-7 cells overexpressing sGC (28) and the activity observed with purified protein (25).
Here, we report the effects of GTP and ATP binding to both the catalytic site and the allosteric site in vitro. These nucleotides function as substrates as well as allosteric modulators that regulate the activity of purified protein (Figure 6). However, it is apparent that additional factors are involved in regulating sGC activity in vivo. Additionally, this work localizes the allosteric nucleotide binding site to the C-terminus of sGC and strongly supports the existence of the proposed pseudosymmetric site. Clearly mapping the sites that GTP and ATP bind to sGC or α1catβ1cat is essential to confirm if this proposed psuedosymmetirc site is important for regulating cGMP synthesis.
The authors thank Dr. Jacquin Niles, Dr. Jonathan Winger, and Joshua Woodward for helpful discussions. Additionally, we thank Dr. Jonathan Winger for providing α1catβ1cat for preliminary experiments.
Funding was provided by NIH grant GM077365 to M.A.M.
2sGC amino acid numbering is that of the rat enzyme unless otherwise noted.