Identification and Partial Purification of an Endogenous Inhibitor of Soluble Guanylyl Cyclase from Bovine Lung*

An endogenous inhibitor of soluble guanylyl cyclase from bovine lung has been partially purified by use of anion exchange, hydrophobic interaction, and gel filtra- tion chromatography. This inhibitor is a protein with a molecular weight of about 149,000 which was estimated from its elution behavior, versus that of a series of standards, on a Sephacryl S-300-HR column. Its activity was measured by comparison of the level of cGMP production from soluble guanylyl cyclase in the presence and absence of the inhibitor protein. Soluble guanylyl cyclase is inhibited by this protein when either Mg2' or M n 2 + is used as a cofactor. The insensitivity of its inhibi- tory activity to both isobutylmethylxanthine and the presence of cGMP demonstrates that this new protein is not a phosphodiesterase. This new protein inhibits both activated and unactivated soluble guanylyl cyclase. Be- cause the inhibition was found to be noncompetitive with respect to the substrate, GTP, it appears that this inhibitor may be an allosteric regulator of soluble guanylyl cyclase.


Identification and Partial Purification of an Endogenous
An endogenous inhibitor of soluble guanylyl cyclase from bovine lung has been partially purified by use of anion exchange, hydrophobic interaction, and gel filtration chromatography. This inhibitor is a protein with a molecular weight of about 149,000 which was estimated from its elution behavior, versus that of a series of standards, on a Sephacryl S-300-HR column. Its activity was measured by comparison of the level of cGMP production from soluble guanylyl cyclase in the presence and absence of the inhibitor protein. Soluble guanylyl cyclase is inhibited by this protein when either Mg2' or M n 2 + is used as a cofactor. The insensitivity of its inhibitory activity to both isobutylmethylxanthine and the presence of cGMP demonstrates that this new protein is not a phosphodiesterase. This new protein inhibits both activated and unactivated soluble guanylyl cyclase. Because the inhibition was found to be noncompetitive with respect to the substrate, GTP, it appears that this inhibitor may be an allosteric regulator of soluble guanylyl cyclase.
cGMP is now known to act as an intracellular second messenger in numerous physiological responses including retinal phototransduction, smooth muscle relaxation, and intestinal secretion, among others. The cGMP produced by guanylyl cyclase appears to directly activate cGMP-dependent protein kinases (l), and the ultimate consequence of cGMP elevation is a reduction of intracellular Ca" concentration (2). The intermediate steps in this pathway are not yet fully understood, but it is known that cGMP plays a role in regulating Ca2+-activated ATPases (3-51, controlling protein phosphorylation (6, 7) and forming inositol triphosphate and diacylglycerol (8). Any or all of these activities may be involved in the cGMP signaling pathway.
Within cells, cGMP formation is mediated by two distinct subtypes of guanylyl cyclase, a membrane-bound form, which is responsive to certain peptide hormones, and a cytosolic form, which is the target for nitric oxide (9,10). The importance of cGMP in mediating cellular response has lead many researchers to study the various factors which regulate the activity of these enzymes. The soluble guanylyl cyclase is at present the only known receptor for nitric oxide and many of the physiological effects of NO are mediated through the activation of this enzyme (9). Other factors which are known to affect the activity of the soluble cyclase include the activators arachidonic acid, protoporphyrin IX, and certain free radicals and the in-* This work was supported by a grant-in-aid from the American Heart Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 (9). Our efforts to understand the regulation of soluble guanylyl cyclase have led to the discovery of an endogenous protein in bovine lung which is an inhibitor of this enzyme. In this study, we report the partial purification and preliminary characterization of this new inhibitor protein.

EXPERIMENTAL PROCEDURES
Materials-Trypsin-TPCK' was obtained from Worthington. All other reagents, unless otherwise noted, were obtained from Sigma and were of analytical grade.
Preparation of Bovine Lung Soluble Fraction-All steps were performed at 4 "C unless otherwise noted. Fresh bovine lung (1 kg) was finely minced in a food processor. The mince was thoroughly washed in 3 volumes of 50 m M triethanolamine-HC1 buffer (TEA), pH 7.8, containing 2 m M EDTA followed by 3 volumes of TEA-DTT buffer (50 m M TEA at pH 7.8 containing 10 m~ dithiothreitol). The washed mince was homogenized in TEA-DTT buffer (40% (w/v)) in a Waring blender. The homogenate was centrifuged at 13,000 x g for 30 min aRer which the supernatant was centrifuged at 100,000 x g for 1 h. The resulting soluble fraction was used for the isolation of soluble guanylyl cyclase and the endogenous inhibitor.
Isolation of Soluble Guanylyl Cyclase-Soluble guanylyl cyclase was batch adsorbed onto Q-Sepharose Fast Flow by stirring the 100,000 x g supernatant slowly for 1 h with 500 ml of resin which had been preequilibrated with TEA-DTT buffer. The supernatant was decanted and the resin was poured into a column (5.0-cm diameter) and packed by washing with TEA-DTT buffer. Once the column reached a constant bed height, the resin was washed for an additional 90 min with TEA-DTT buffer. Bound proteins were eluted with a 1500-ml linear NaCl gradient (0 to 0.46 M) in TEA-DTT buffer. Guanylyl cyclase activity was detected via the assay described below. The soluble guanylyl cyclase containing fractions were pooled, concentrated in a stirred-cell concentrator with a PM-30 membrane (Amicon) and stored at -80 "C. This partially purified soluble guanylyl cyclase preparation was used to assay the activity of the inhibitor.
Partial Purification of Inhibitor-The fractions from the aforementioned ion exchange column were assayed for inhibitor by the use of partially purified guanylyl cyclase to monitor the activity. Those fractions containing significant inhibitory activity were pooled and concentrated as above. The inhibitor protein was further purified on a phenyl-Sepharose CL-4B column (2.6 x 18 cm). The concentrated inhibitor sample, in 0.2 M NaCl, was loaded onto the phenyl-Sepharose column, which was equilibrated with TEA-DTT buffer. After the column was washed with 5 volumes of TEA-DTT buffer, the inhibitor was eluted with a linear gradient of Triton X-100 (0 to 2% (v/v)) in TEA-DTT buffer. Fractions containing inhibitory activity were pooled and concentrated in a stirred cell concentrator with a PM-30 membrane. The inhibitor was further purified by gel filtration on a Sephacryl S-300-HR (Pharmacia Biotech Inc.) column (2.6 cm x 80 cm). Partially purified inhibitor was then concentrated and stored at -80 "C.
Guanylyl Cyclase Activity Assay-Guanylyl cyclase activity was measured by the production of [32PlcGMP from [n-32PlGTP by a modification of the procedure described previously (11).  Both guanylyl cyclase and inhibitor were isolated from a 100,000 x g supernatant fraction by ion exchange chromatography on Q-Sepharose Fast Flow eluted with a linear NaCl gradient. Guanylyl cyclase eluted in a single peak of activity centered about fraction 74. Inhibitor eluted from the column in a single peak of activity centered about fraction 63.
was terminated by the addition of 10 p1 of 0.5 M EDTA after 10 min of incubation at 37 "C. [32PlcGMP was isolated by thin-layer chromatography on polyethyleneimine-cellulose (12) and counted in 10 ml of Biosafe I1 mixture in a Beckman LS-6000 liquid scintillation counter.
Inhibitor Activity Assuy--To measure inhibitor activity, aliquots (2-40 p1) of inhibitor-containing fraction were included in the assay samples of soluble guanylyl cyclase described above. The percent inhibition was determined by the formula (A -B)/A x loo%, where A and B represent guanylyl cyclase activity in the absence and presence of inhibitor, respectively. One unit of inhibitor activity denotes the amount of inhibitor that results in 50% inhibition of activity for 0.5 mg of guanylyl cyclase sample in the assay.
Test for Cyclic Phosphodiesterase Actiuity-Unlabeled cGMP (final concentrations, 5-50 p~) was included in assays of the inhibitor at varying concentrations to determine if the loss of guanylyl cyclase activity was due to removal of the product [32P]cGMP by phosphodiesterase activity. Guanylyl cyclase and inhibitor protein were added to the assay tube last to initiate the reaction and to prevent premature consumption of either cGMP or GTP. Assays were otherwise as described above. At the concentrations tested, cGMP itself had no inhibitory effect of the soluble guanylyl cyclase activity.
Total Protein Assay-Protein concentrations were measured by the Biuret method (13) using bovine y-globulin as a standard.

RESULTS
The presence of an endogenous inhibitor was revealed in the early stages of the purification of soluble guanylyl cyclase. The first chromatographic step in the purification of soluble guanylyl cyclase regularly resulted in the recovery of total activity yields in excess of 100% and an examination of previously reported purification schemes for soluble guanylyl cyclase reveals this inconsistency is characteristic of this enzyme system (14-17). It was inferred from this observation that an endogenous inhibitor was being separated from the cyclase at this step of the purification. Its presence was confirmed by the addition of all the ion exchange column fractions to the isolated guanylyl cyclase and measurement of the effect of the added fraction on the activity of the cyclase. Reconstitution of the active guanylyl cyclase fraction with inactive fractions showed the presence of an inhibiting factor which elutes from the ion exchange column just prior to guanylyl cyclase (Fig. 1). The position of the inhibitory fractions on the elution profile relative to the gradient and to the position of the cyclase was found to be reproducible in all preparations.
The behavior of the inhibitory fraction offers strong support that it is a protein. Boiling the inhibitor fraction resulted in the elimination of its inhibitory activity, demonstrating that the inhibitor is heat-labile. When the inhibitor was passed through a Sephadex G-50 column, the inhibitory activity eluted in the void volume, as predicted for a large molecule. Exposure of the inhibitor to trypsin substantially decreased its activity, as illustrated in Fig. 2. Based on these observations it was concluded that inhibition of soluble guanylyl cyclase was due to another protein.
The activity of the inhibitor was determined by its effect on the activity of partially purified soluble guanylyl cyclase. The dose-response curve for inhibition of guanylyl cyclase by the inhibitor fraction is sigmoid, as shown in Fig. 3. The inflection point of this curve, at the concentration of inhibitor causing 50% inhibition (IC6,), was used to define the activity of the inhibitor. To simplify the determination of IC,, values, the doseresponse curve was transformed by use of a probit percent table (18) to a straight line with an IC,, equivalent to the original curve (Fig. 3). Both the purity of the inhibitor and the recovery of activity units were monitored by this method.
The inhibitor was partially purified by hydrophobic affinity chromatography and gel filtration, as summarized in Table I. The purification of the inhibitor resulted in a 1226-fold reduction in total mass with a 23% recovery of total activity. Detection of inhibitory activity in the 100,000 x g supernatant was impaired by the presence of guanylyl cyclase activity in the supernatant. Isolation of the inhibitor by elution on a Sephacryl S-300-HR column allowed an approximation of the molecular weight for the inhibitor. Its mass was estimated to be about 149,000 Da, by comparison to a series of known molecular weight standards (Fig. 4). The degree of purification was insufficient for the conclusive identification of bands corresponding to the inhibitor on a SDS-polyacrylamide gel (Fig. 5).
The inhibition of soluble guanylyl cyclase was examined to rule out the possibility that the inhibitor was simply a cGMP phosphodiesterase or a GTP specific phosphatase. The presence of a phosphodiesterase or phosphatase would result in a lower apparent guanylyl cyclase activity due to the destruction of the product and starting material of the cyclization reaction, respectively. The activity of cyclic nucleotide phosphodiesterases is typically controlled by the inclusion of IBMX in the assay  NA, activity of inhibitor in 100,000 x g , could not be determined due to the presence of guanylyl cyclase. mixture and IBMX was routinely included in all assays to eliminate phosphodiesterase activity. Because a new class of IBMX-insensitive cyclic nucleotide phosphodiesterases has recently been discovered (19), it was necessary to rule out the presence of such enzymes. Excess unlabeled cGMP had no effect on inhibitor activity (Fig. 6). This result demonstrates that the reduction of cGMP concentration in the activity assays is not due to the action of a IBMX-insensitive phosphodiesterase. The insensitivity of its inhibition to both IBMX and the presence of cGMP indicates that the inhibitor is not a cGMP phosphodiesterase. Inhibitor assays were also performed in the presence of a GTP-regenerating system. When creatine kinase and creatine phosphate were included in the assay, no effect on  the inhibition of guanylyl cyclase was observed (data not shown). The inability of this GTP-regenerating system to reduce inhibition demonstrates that inhibition of guanylyl cyclase is not the result of GTP degradation by phosphatase enzymes.
Inhibition of guanylyl cyclase occurs with either Mg2' or Mn2+ present as cofactor in the activity assay (Table 11). Neither the levels of inhibition nor the slopes of the dose-response curves differ significantly when inhibition of guanylyl cyclase is measured in the presence of either metal cofactor.
Inhibition is also observed for the activated form of guanylyl cyclase. A significant reduction in guanylyl cyclase activity is seen when inhibitor is added to guanylyl cyclase samples activated by either nitric oxide or sodium nitroprusside (SNF'). Both SNP-activated samples and unactivated samples of gua- nylyl cyclase are inhibited to a similar degree (Fig. 7).2 Inhibition of SNP-activated guanylyl cyclase is independent of the order of addition of the activator and inhibitor. Samples of guanylyl cyclase to which SNP was added prior to the inhibitor were inhibited to the same extent as samples in which the inhibitor was added before the SNP. The consistent inhibition, regardless of the order of addition, clearly demonstrates that the inhibition of guanylyl cyclase is independent of the mechanism for NO activation of soluble guanylyl cyclase. Experiments were performed to probe the nature of the interaction between the inhibitor protein and soluble guanylyl cyclase. To determine whether the inhibition was reversible, pre-mixed guanylyl cyclase and inhibitor were separated and tested for activity in independent assays. A sample containing a fured amount of guanylyl cyclase and inhibitor was assayed for guanylyl cyclase activity and then chromatographed on a Q-Sepharose column. The fractions containing the inhibitor and the guanylyl cyclase were re-isolated and assayed independently for activity. Neither guanylyl cyclase nor the inhibitor showed any loss of activity, demonstrating that the inhibition was completely reversible. Inhibition was studied at several substrate and inhibitor concentrations to determine the mechanism of inhibition. Construction of a Lineweaver-Burk plot from the resulting data shows classic noncompetitive behavior, as seen in Fig. 8. The inhibitor altered the V, , for guanylyl cyclase but had no effect on the apparent K, for the GTP substrate. These data cumulatively suggest that the new protein may be an allosteric regulator of soluble guanylyl cyclase. DISCUSSION Cyclic GMP is a second messenger of considerable physiological importance; however, less is known about the function of cGMP than the more widely studied CAMP (3, 10). These two second messengers frequently play antagonistic roles in the regulation of cellular metabolism. Great advances have been made in understanding the regulation of adenylyl cyclases via G-proteins; however, it is only recently that exploration of the regulation of the analogous guanylyl cyclases has begun (10). Guanylyl cyclases are responsible for the formation of intracellular cGMP from GTP and play a significant role in intestinal secretion, phototransduction, vascular relaxation, and neurotransmission (20-22). Two forms of guanylyl cyclase have been identified, a particulate membrane-bound form and a soluble cytosolic form. These proteins are genetically distinct and are regulated by different mechanisms (9).
Regulation of the soluble form of guanylyl cyclase by NO is physiologically important in the maintenance of vascular tone (23, 24) in the establishment of penile erection (25), and it is hypothesized to play a role in long-term potentiation in nerve somewhat heme depleted as indicated by the low (9-fold) activation by a The guanylyl cyclase samples used to assay inhibitor activity were SNP. Despite the deficiency of endogenous heme in guanylyl cyclase, the effect of the inhibitor can be clearly seen for both SNP-activated and unactivated guanylyl cyclase.
0.60 I I.
(-)SNP (+) SNP   FIG. 7. Effect of partially purified inhibitor on SNP-activated and unactivated soluble guanylyl cyclase activity. Soluble guanylyl cyclase was activated by the inclusion of sodium nitroprusside (100 p~) in the assay mixture. Both activated and unactivated guanylyl cyclase samples were assayed for activity in the presence and absence of partially purified inhibitor. Inhibition of soluble guanylyl cyclase has been less well studied and the known inhibitors are unlikely to be physiologically important in the in vivo regulation of activity. Because this enzyme is known to contain one or more essential thiols, thiolmodifying agents eliminate enzymatic activity (32). The precise role of these thiol groups in the enzymatic mechanism is unknown. A number of agents are known which inhibit activation of soluble guanylyl cyclase by NO, but do not influence the basal (unactivated) enzyme activity. Redox active agents including methylene blue and the superoxide anion may exert their inhibitory effect by direct reaction with NO, thereby preventing the NO activator from reaching the heme site (14, 33). While the reaction of NO with superoxide anion is undoubtedly physiologically important (34), the significance of this reaction in the regulation of soluble guanylyl cyclase activity is not known. Metalloporphyrins can inhibit both the activation of soluble guanylyl cyclase and its basal activity, apparently by different mechanisms. Certain metalloporphyrins bind avidly to NO, thereby preventing it from reaching the heme in soluble guanylyl cyclase and preventing activation in a manner analogous to that described for the redox-active agents above. Inhibition of the basal activity of soluble guanylyl cyclase by metalloporphyrins appears to be a result of competition for the heme-binding site in the protein. The presence of a metalloporphyrin in the heme site, including the native heme (Fe-protoporphyrin E), suppresses the basal activity of soluble guanylyl cyclase; heme-deficient soluble guanylyl cyclase has consistently higher basal activity than the heme-containing form of the protein (17). The endogenous heme is weakly bound, and certain metalloporphyrins, such as Zn-protoporphyrin M, apparently displace the native heme (17). Consistent with this hypothesis is the observation that metalloporphyrins can prevent activation of heme-deficient guanylyl cyclase by protoporphyrin M (35).
Despite the obvious importance of the regulation of this enzyme, no physiological inhibitors have been identified. As part of our ongoing investigations into the regulation of soluble guanylyl cyclase, we have sought to understand the suppression as well as the activation of this essential enzyme. Systematic variations of its activity occur during the purification of soluble guanylyl cyclase. Observations, in our laboratory and in others (15-17), that activity yields in initial stages of purification are consistently greater than loo%, have led us to postulate the existence of an endogenous regulator of guanylyl cyclase. The discrepancy in activity yields has been postulated previously to result from the presence of heme or hemoproteins in crude preparations (16,17); however, no data to support this hypothesis were presented. A more logical explanation of the excess activity is the presence of a reversible non-heme inhibitor in the crude extracts, since inhibition by heme requires that the soluble guanylyl cyclase be present in the heme-deficient form. Isolation and characterization of such an endogenous inhibitor could reveal new information about the overall mechanism of soluble guanylyl cyclase regulation in vivo.
The data in this study support the claim that the new protein, described herein, is a previously undiscovered inhibitor of soluble guanylyl cyclase. This inhibitor protein is present in the cytosolic fraction of bovine lung tissue and is separated from soluble guanylyl cyclase in the early stages of purification. This new protein inhibits both the activated and unactivated states of soluble guanylyl cyclase to the same extent. The inhibition of the activated soluble guanylyl cyclase occurs regardless of the order of addition of the NO source and inhibitor protein, demonstrating that the inhibitor is not simply preventing NO from reaching soluble guanylyl cyclase. These results suggest that the activation and inhibition are independent processes and that the site for interaction of the inhibitor is unique and remote from the heme binding site. The enzyme-inhibitor complex is kinetically competent to carry out cyclization of GTP, as evidenced by the noncompetitive nature of the inhibition. The binding of the inhibitor influences the maximal velocity of the enzyme but not the substrate binding affinity, supporting the hypothesis that the site of interaction is remote from the substrate binding site. Together these observations suggest that the new protein is a n allosteric regulator of soluble guanylyl cyclase.
An endogenous inhibitor of soluble guanylyl cyclase has been detected in human platelets (36), but the characteristics of the platelet-derived inhibitor are different from those of the protein described in this study. In the report of the human platelet inhibitor, the authors observed two discrete protein fractions after ion exchange chromatography, with the latter fraction containing cyclase activity. Recombining the separated fractions caused the cyclase activity to be diminished. The relative positions of the inhibitor and cyclase-containing fractions with respect to the ion exchange column profile were similar to those observed during our purification. The platelet-derived inhibitor inhibited the cyclase only with Mg2' present as cofactor; no effect of the inhibitor on cyclase activity was observed in the presence of Mn". In addition, the inhibitor fraction restored the activation of soluble guanylyl cyclase by SNP, a fact which was attributed to the presence of a heme in the inhibitor fraction. Visible spectra of the inhibitor fraction showed the distinct Soret absorption characteristic of heme, and the authors postulate that the inhibition of the cyclase resulted from the presence of the heme.
Although the bovine lung-and platelet-derived inhibitor fractions were isolated in a similar fashion on a n ion exchange column, there are marked differences between the two. The bovine lung protein inhibits equally well in the presence of either Mg2' or Mn2+ as cofactor for the cyclase. Both the SNPactivated and the unactivated forms of soluble guanylyl cyclase are inhibited, and the extent of inhibition of the two forms of the cyclase is equal. The partially purified bovine lung inhibitor protein shows no ability to potentiate the activation of the cyclase nor is there evidence for a heme in the visible spectrum. The contrasting characteristics of the two inhibitor fractions suggest that these are two different substances. Given the different roles of platelet and endothelial cells, it is not surprising to find more than one mechanism for down-regulation of an enzyme as widely distributed as soluble guanylyl cyclase. Since the platelet-derived inhibitory fraction was not purified or characterized in any detail, it is impossible to speculate on the role of such an inhibitor for the platelet system. CONCLUSION The isolation of a protein inhibitor from bovine lung reveals the existence of a new mechanism for the regulation of soluble guanylyl cyclase. A great deal of attention has been given to the activation of soluble guanylyl cyclase by nitric oxide; however, little is known about how this activation might be reversed or limited. This new inhibitor protein may play a role in the suppression of cGMP levels under activated or unactivated conditions. Relatively little is known about the regulation of cGMP levels in vivo, and this protein may play a role in the control of the cGMP cascade. The existence of an inhibitor of soluble guanylyl cyclase, as characterized in the present study, provides evidence of a new pathway for the regulation of cGMP levels. Further purification and characterization of this new protein will provide a greater understanding of this complex regulatory pathway.