Purification and properties of a protein required for sodium azide activation of guanylate cyclase.

Sodium azide activation of guanylate cyclase requires a protein factor that has been partially purified from rat liver. While the activator factor and guanylate cyclase in rat liver supernatant fractions were not separable with gel filtration, they were separable with DEAE-cellulose chromatography. The activator factor was heat-labile and inactivated with bacterial protease and cu-chymotrypsin. Its activity was not altered by a variety of other hydrolytic enzymes including trypsin. With gel filtration the protein factor was estimated to be about 200,000 daltons. Polyacrylamide gel electrophoresis resulted in a single peak of activator that in the presence of NaN, increased guanylate cyclase activity assayed with either Mn*+ or Mg2+. The activator in the absence of NaN, also increased guanylate cyclase activity partially purified from liver or cerebral cortex. However, the activation was greater in the presence of NaN, and the degree of activation was a function of the concentration of both NaN, and activator. Guanyl-5’-yl imidodiphosphate was 5 to 10% as effective as the substrate GTP for rat liver guanylate cyclase and the relative stimulatory effect of NaN, was similar with either substrate. ATP inhibited basal and azide-activated guanylate cyclase activity similarly. The effect of NaN, was not altered by NaF. Thus, the activation by NaN, does not seem to require a phosphorylation or dephosphorylation process. The effects of sulfhydryl compounds, heavy metal ions, H,O,, CN-, and N, suggest that oxidation-reduction is required for the NaN, effect. Either catalase, peroxidase, or cytochrome b, could replace the requirement for the activator protein in NaN, activation. The activator factor and catalase were obtained in similar fractions from DEAEcellulose chromatography and gel electrophoresis. The conversion of azide to nitric oxide or another material by the activator protein, catalase, or an analogous factor may be the mechanism of guanylate cyclase activation by azide and other agents.

In most instances the mechanism by which agents increase cyclic GMP accumulation is unknown. Although several preliminary reports have described activation of guanylate cyclase in cell-free systems by some hormones and neurohormones (3-V, these observations have not been confirmed (8,9). Possible mechanisms to regulate guanylate cyclase activity have been reviewed recently (10, 11). We have described the activation of guanylate cyclase from several tissues with sodium azide, hydroxylamine, phenylhydrazine, sodium nitrite, nitroglycerin, and sodium nitroprusside (12,13). Some of these agents also increase cyclic GMP levels in incubations of slices from liver, cerebral cortex, cerebellum (14), tracheal smooth muscle (15), taenia coli, atria (16), and kidney (17). Activation of guanylate cyclase with sodium azide is dependent upon sodium azide concentration, incubation time, and temperature (12,18  and ammonium sulfate treatment and applied (82 mg of protein) to a DEAE-cellulose column as described under "Materials and Methods." Fractions (4.5 ml) were assayed for activator without (0) and with (0) 1 rnM NaN, using a DEAE-cellulose preparation of soluble rat cerebral cortex guanylate cyclase.
The molecular size of the activating factor was estimated to be about 200,000 with gel filtration on a calibrated Sepharose 6B column (Fig. 3). The molecular size was similar with activator that was prepared with or without prior trypsin treatment (Fig. 3). The elution of activator activity from the Sepharose 6B column was also similar when either Mn2+ or Mg2+ was used as the sole cation in guanylate cyclase incubations (Fig. 3).
Although the activator was resistant to trypsin and other hydrolases, its activity was markedly decreased with treatment with bacterial protease or a-chymotrypsin (Table II). Phospholipase C treatment decreased the activation by activator and NaN, slightly. Similar effects were observed when Mg2+ rather than MnZ+ was used as sole cation for guanylate cyclase. None of the hydrolytic enzymes at the concentrations tested and introduced into the cyclase assay had significant effects on basal guanylate cyclase activity. The activator prepared through the DEAE-cellulose step was inactive after heating at 90" for 2 min. At 50" for 5 min about one-third of the activity was lost (not shown).  Electrophoresis of the activating factor from the DEAEcellulose step on polyacrylamide gels yielded a single peak of activity upon elution of gel slices (Fig. 4). The peak of activity was similar when assayed with enzyme, 1 mM NaN,, and either Mn*+ or Mg2+ as sole cation. We previously reported that guanylate cyclase from several tissues activated with NaN, can effectively use Mn2+ or Mg2+ as the sole cation cofactor (18).
With partially purified guanylate cyclase and the activator a time lag was not observed with activation by the protein factor alone or the activator plus NaN, (Fig. 5). We have previously observed that a time lag of several minutes occurred with NaN3 activation of crude preparations of guanylate cyclase (12). As summarized in Fig. 1 4. Electrophoresis of the activator on polyacrylamide gels. The 50 /*g of activator prepared through DEAE-cellulose chromatography was applied to a 5% polyacrylamide gel and run for 5 h at 3 mA/tube using 50 mM Trislglycine buffer (pH 8.0) containing 1 rnM dithiothreitol and 1 mM EDTA. The 3-mm segments of the gel were eluted with 50 mM Tris/HCl (pH 7.6) buffer. Eluates were assayed for activator using partially purified rat cerebral cortex guanylate cyclase (11.4 pg of protein) with and without 1 mM NaN, and either Mg2+ (B) or Mnz+ (Cl as sole cation. Samples were also assayed for catalase activity (A).  dependent manner when 1 mM NaN, was present. Higher concentrations of activator (1 to 3 pg/assay tube) also increased guanylate cyclase activity in liver or cerebral cortex preparations in the absence of NaNR when Mn*+ or Mg*+ was used as sole cation cofactor (Fig. 6). However, the stimulator-y effects were much greater when NaN, was present. The degree of activation of guanylate cyclase was a function of the concentration of activator and the concentration of NaN, (Fig. 7). As discussed earlier the effects of NaN, and activator were independent of guanylate cyclase concentration (Fig. 1). With increasing amounts of the activator the concentration of NaN,, required for half-maximal activation decreased. With 2.1 pg of activator protein/assay tube the half-maximal concentration for NaN, was 0.2 PM compared to 1.5 p,~ with 0.021 pg of activator.
With partially purified soluble rat cerebral cortex guanylate cyclase Mn2+ or Mg*+ were maximally effective as sole cation cofactors at about 4 mM with basal or azide-activated enzyme. Activation with NaN, did not alter the stimulatory effect of free Mn2+ in excess of GTP (not shown). This is in contrast to the effect with liver preparations where the requirement for free Mn*+ is markedly diminished with azide activation (12,18). Similar to studies with liver guanylate cyclase concentrations of MnZ+ or Mg2+ greater than 4 mM were less effective after azide activation of cerebral cortex preparations.   arations from various tissues (9, 28-30). ATP inhibited partially purified soluble cerebral cortex enzyme from rat (Fig. 8). NaN,, activation is not due to a phosphorylation or dephospho-The degree of inhibition was similar when activity was in-rylation mechanism. However, additional studies are recreased with the protein factor or the protein factor plus NaN,,.
quired. The concentration of ATP that resulted in 50% inhibition was As summarized in Table III a variety of agents and condiabout 0.25 mM under all of these conditions. ATP at 0.5 mM tions altered NaN,, activation of guanylate cyclase in crude rat also inhibited basal and NaN:]-activated crude soluble gua-liver supernatant fractions. Several compounds with sulfhynylate cyclase from rat liver 60 to 80%. The inhibition of ATP dry1 groups increased the effect of NaN,, as did FeCl,. The was similar when 4 mM Mg2+ and 50 PM cyclic AMP were effect of NaN,, was decreased with H,Oy and KCN without included in incubations. Gpp(NH)p is about 5 to 10% as effec-effects on basal activity. CuCl, decreased basal activity withtive as GTP as a substrate for soluble rat liver guanylate out altering the relative stimulatory effect of NaN,, (4.3-fold). cyclase. The relative stimulatory effect of 1 mM NaN,, with this The effect of azide was markedly decreased when incubations substrate was not altered. Similarly the effect of NaN, was not were carried out in a nitrogen atmosphere (Table III). The altered by 10 to 100 mM NaF. These observations suggest that effects of cysteine and CN-were reported previously (12).

Guanylate
Cyclase Activation with Sodium Azide 4389 (A) Catalase Activity 6 4. _ (B) Mg*+ ble of increasing cyclic GMP levels in intact tissues (1,2). Since most of these agents do not increase cyclic GMP accumulation in cell-free systems, the mechanism(s) for these effects are not known. Possible mechanisms to regulate guanylate cyclase activity and cyclic GMP accumulation have recently been reviewed (10, 11). Sodium azide, hydroxylamine, sodium nitrite, nitroglycerin, and sodium nitroprusside can increase both cyclic GMP accumulation in intact tissues (14-17) and guanylate cyclase in cell-free preparations (12,13,18,19). Recently, nitrosamines have also been reported to produce effects similar to these agents in liver preparations (31). The mechanism(s) in which these agents increase guanylate cyclase activity could provide some clues about hormonal regulation of the enzyme. However, these agents, unlike most hormonal effects on cyclic GMP accumulation in intact tissue do not require Ca2+ in the external medium (14,15,17,32). Thus, the ultimate mechanisms may prove to be quite different. The 200 pg of commercially obtained beef liver catalase were applied to gel electrophoresis under the same conditions of Fig. 4. The 3-mm segments of gel were eluted and assayed for catalase activity.
Eluates were also incubated with partially purified cerebral cortex guanylate cyclase (11.4 pg of protein) without and with 1 rnM NaN,. Mg*+ (B) or Mn*+ (Cl was used in cyclase assays. These studies suggested that an oxidative-reductive process was required for the NaN, effect. The addition of either beef liver catalase, horseradish peroxidase, or cytochrome 6, to partially purified guanylate cyclase could replace the requirement for the activator factor for NaN:, activation (Table IV). Xanthine oxidase or superoxide dismutase did not alter guanylate cyclase activity in the absence or presence of 1 mM NaN:, (not shown). Gel electrophoresis of beef liver catalase preparations revealed several protein bands. Two peaks of catalase activity were obtained and fractions from both of these regions could activate guanylate cyclase in the presence of NaNR (Fig. 9). DEAE-cellulose fractions of the activator also contained catalase activity. With gel electrophoresis of these preparations catalase and activator activities were obtained in similar regions from gels (Fig. 4).
NaN, can increase guanylate cyclase activity in soluble and particulate preparations from rat liver and kidney and particulate preparations of cerebral cortex and cerebellum (12, 18). The absence of an effect in heart, lung, and other tissue preparations may be due to the absence of the activator factor, the presence of a protein inhibitor, or both (11,12). Addition of the macromolecular activator factor to azide-nonresponsive guanylate cyclase preparations permits NaN,, activation to occur (18,19). The activator factor from rat liver was partially purified and characterized.
It has a molecular weight of about 200,000, is heat-labile, and is inactivated with bacterial protease and a-chymotrypsin.
However, it was resistant to trypsin treatment and other hydrolases under the conditions used. In our earlier report the azide effect was decreased by trypsin treatment of lower concentrations of crude rat liver preparations (12). The activator protein alone increased liver and cerebral cortex guanylate cyclase activity when either Mn2+ or Mg*+ was used as sole cation. In the presence of NaN:, greater stimulator-y effects were observed. The effects of both the activator and NaN,, were concentration-dependent and were not altered by the concentration of guanylate cyclase in incubations. The concentration of NaN, required for half-maximal activation of partially purified rat cerebral cortex granulate cyclase was 0.2 to 1.5 pM depending upon the concentration of activator. This value is similar to that observed with activator and partially purified liver guanylate cyclase of 1 PM (18) and much less than the K,, of 40 PM in crude rat liver supernatant fractions (12). Unlike effects in crude liver guanylate cyclase preparations no time lag was observed with NaN, activation of partially purified preparations. DISCUSSION A variety of agents such as choline esters, histamine, serotonin, catecholamines, prostaglandin F,, and others are capa-ATP inhibited both basal and NaN,-activated cerebral cortex guanylate cyclase to a similar degree. Although Gpp(NH)p is a poor substrate for liver guanylate cyclase, the relative stimulatory effect of NaN, was not altered. NaF did not modify the NaN,, effect. These studies suggest that the mechanism of NaN, activation does not involve a phosphate transfer. The effects of several agents indicated that an oxidative-reductive process was required for NaN, activation of guanylate cyclase. With crude liver supernatant fractions sodium azide activation was not observed in a nitrogen atmosphere, enhanced with sulfhydryl agents, enhanced with Fez+ and inhibited by cyanide and H,O, (Table III). Since azide and cyanide can interact with catalase (331, this enzyme was added to incubations and found to replace the requirement for the activator factor. Both catalase activity and the activator activity migrated to similar areas when catalase or activator preparations were applied to polyacrylamide gel electrophoresis. Cata-