Activation of a Cytosolic ADP-ribosyltransferase by Nitric Oxide-generating Agents*

Sodium nitroprusside is a vasodilator and an inhibi- tor of platelet activation. It is thought that these effects are mediated by the spontaneous release of nitric oxide and stimulation of cytosolic guanylate cyclase. We have found that sodium nitroprusside (5-200 pM) greatly increased a cytosolic ADP-ribosyltransferae that ADP-ribosylates a soluble 39-kDa protein. This activ- ity causes the mono-ADP-ribosylation of the 39-kDa protein, since digestion with snake venom phosphodiesterase releases 5’-AMP. This enzyme is present in platelets, brain, heart, intestine, liver, and lung. The effect of sodium nitroprusside is not related to stimu- lation of soluble guanylate cyclase and the production of cyclic GMP because cyclic GMP, dibutyryl cyclic GMP, and 8-bromo-cyclic GMP are ineffective. 3-Mor-pholinosydnonimine formation of oxide as sodium ADP-ribosylation the 39-kDa protein. activation the cytosolic ADP-ribosyltransferase. studies demon- a novel action of nitric oxide related to the activation of an endogenous ADP-ribosyltransferase.

Recent evidence suggests the existence of cytosolic ADPribosyltransferase in human platelets (17,18). We have now found that compounds that liberate nitric oxide such as sodium nitroprusside and 3-morpholinosydnonimine can greatly enhance this activity in the cytosolic fraction of different tissues. It is widely known that nitric oxide, which has been identified as endothelium-derived relaxing factor (N), can (14-16).
* 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.
$ Recipient of financial support from the Deutsche Forschungsgemeinschaft. stimulate soluble guanylate cyclase (20). However, here we show that the effect of nitric oxide on the activation of ADPribosyltransferase is totally independent of the activation of guanylate cylase.

EXPERIMENTAL PROCEDURES
Isolation of Human Platelets-Human blood (100-200 ml) was drawn from healthy volunteers, using trisodium citrate (0.38%, w/v) as an anticoagulant, and centrifuged at 180 x g for 25 min to yield platelet-rich plasma. Platelets were obtained by centrifuging the platelet-rich plasma at 1200 X g for 15 min in the presence of prostacyclin (0.1 pglml) and were washed once by centrifugation at 800 X g for 10 min in 50 ml of a Hepesl-buffered Tyrode's solution (138 mM NaCl, 0.36 mM NaH2P04, 2.9 mM KC1, 1 mM MgC12, 1 mM EGTA, 10 mM glucose, 20 mM Hepes, pH 7.4) containing prostacyclin (0.3 pg/ml). At this stage the platelets were either resuspended in cold hypotonic buffer for fractionation or in Tyrode's solution for incubation with different agonists. In the latter case, incubations were performed at 37 "C for the times indicated, and platelets were centrifuged at 800 X g for 10 min and resuspended in hypotonic buffer for fractionation.
Fractionation of Platelets-A modification of a published method (21) was used. The platelet pellet was quickly resuspended at 4 "C in 1 ml of hypotonic buffer (5 mM Tris. HC1, 5 mM EDTA, pH 7.5) and frozen in liquid nitrogen. Four cycles of rapid freezing and thawing lysed the cells, and after centrifugation at 800 X g for 10 min to remove any unbroken cells, the homogenate was centrifuged at 160,000 X g for 15 min in an air-driven microcentrifuge. The resulting supernatant, representing the platelet cytosol fraction, was removed and stored in liquid nitrogen.
Preparation of Cytosolic Fraction from Different Rat Tissues-A female rat was decapitated, and the organs were immediately withdrawn, extensively washed in cold hypotonic buffer, homogenized, and centrifuged at 800 X g for 10 min to remove unbroken cells. The homogenate was centrifuged at 160,000 X g for 20 min in an airdriven microcentrifuge. The supernatant was used as the cytosolic fraction.
f2P/ADP-ribosylation-ADP-ribosylation was carried out by generally following the procedure of Ribeiro-Net0 et al. (22). The reaction was performed in a final volume of 65 pl containing 10 mM thymidine, 1 mM ATP, 2 mM dithiothreitol, 0.1 mM GTP, 138 mM NaC1, 0.36 mM NaHZPO4, 2.9 mM KCl, 12 mM NaHCO3, 5 mM HEPES, 1 p M NAD, 3 mM MgCI2, 200 mM potassium phosphate (pH 7), 1.5 pCi of ["P)NAD/assay, and 30-38 pg of platelet protein. After a 30-min incubation (or otherwise as indicated) at 37 "C, samples were precipitated with 1 ml of cold 10% (w/v) trichloroacetic acid and centrifuged for 10 min at 2000 X g. The pellets were finally washed twice with 2 ml of cold water-saturated ether, and proteins were resolved in 11% sodium dodecyl sulfate-polyacrylamide gels. Gels were subjected to autoradiography, and the bands were cut and counted using a liquid scintillation counter. Representative results of at least three similar experiments are shown.
Snake Venom Phosphodiesterase Digestion-The ADP-ribosylation of the 39-kDa protein was carried out as described above, but 200 pg of protein/assay was used. The radioactive protein in the sodium dodecyl sulfate gels was located by autoradiography, and bands corresponding to two assays were removed from the dried gels and subjected to electroelution according to the method of Hunkapiller et al. (23). The eluted protein was dialyzed against 160 mM Tris (pH 8) containing 10 mM MgClz and lyophilized. Snake venom phosphodiesterase cleavage of the [3ZP]ADP 39-kDa protein was done as previously reported (24)(25)(26).
Richard Gryglewski, Cracow, Poland. The G, , , , . antiserum was provided by Dr. Susanne Mumby and Dr. Alfred Gilman, University of Texas Health Center, Dallas, TX.

RESULTS
We have studied the effects of different platelet agonists and known platelet inhibitors on cytosolic ADP-ribosyltransferase activity. We found that sodium nitroprusside stimulates the ADP-ribosylation of a cytosolic 39-kDa protein. Fig. 1, A and R, shows the time course of this reaction, which is linear up to 120 min. Activation by sodium nitroprusside is observable at 10 p~ and increases up to 200 p~ (Fig. 2), although it was not observed in the particulate fraction (not shown).
ADP-ribosylation of the 39 kDa is not affected when a 200fold excess of nonradioactive NAD is added at the end of the incubation, which suggests a covalent modification of the A.

SODIUM NITROPRUSSIDE, 200 r~
Mol.WI.  Release of "P-labeled 5'-AMP after snake venom phosphodiesterase digestion of the "*P-labeled 39-kDa protein. ADP-ribosylation of the 39-kDa protein was carried out in the presence of 200 PM sodium nitroprusside. The :'*P-labeled 39-kDa protein was electroeluted from the gels and digested as detailed under "Experimental Procedures." An aliquot corresponding to one-tenth of the total sample was chromatographed on polyethyleneimine plates as described by Lehmann et al. (25). All the radioactivity comigrated with a standard of "C-labeled 5'-AMP. ADP-ribose, adenosine, and NAD were distinctly separated from AMP as indicated in the figure. protein. Moreover, the increase of the concentration of nonradioactive NAD (1-100 p~) resulted in a gradual decrease of the labeling of the 39-kDa protein. A concentration of 100 p M NAD produced 80% inhibition of the labeling. The covalent modification of the 39-kDa protein was also confirmed by snake venom phosphodiesteratic cleavage of the [32P]ADP 39-kDa electroeluted protein, which caused only the release of "P-labeled 5'-AMP as shown in Fig. 3. The release of "Plabeled 5'-AMP paralleled the loss of radioactivity associated with the 32P-labeled ADP 39-kDa protein. Sodium nitroprusside did not produce degradation of [32P]NAD in our assay system.
It is known that sodium nitroprusside liberates nitric oxide, which in turn activates soluble guanylate cyclase. Therefore, we studied whether cyclic GMP, 8-bromo-cyclic GMP, dibutyryl cyclic GMP, and other agents that increase cellular cyclic GMP levels such as hydroxylamine and aniline influence the ADP-ribosylation of the 39-kDa protein. Fig. 4 illustrates that the activation of ADP-ribosyltransferase is not related to cyclic GMP. This activity is lost when the cytosolic fraction is boiled (not shown). We have also treated intact platelets for 2 min with sodium nitroprusside and then prepared the cytosolic fraction to assess endogenous ADPribosyltransferase activity. Under this condition, sodium nitroprusside stimulated ADP-ribosylation of the 39-kDa protein, but hydroxylamine and aniline were much less active while 8-bromo-cyclic GMP was not active (not shown).
3-Morpholinosydnonimine (29) is another compound that releases nitric oxide. 3-Morpholinosydnonimine stimulates the ADP-ribosylation of the 39-kDa protein (Fig. 5A). Hemoglobin binds nitric oxide (30) and inhibits, in this way, its effects. Fig. 5B shows that hemoglobin inhibits the effect of sodium nitroprusside on the ADP-ribosylation of the 39-kDa protein in a concentration-dependent manner.
The effect of sodium nitroprusside on the activation of cytosolic ADP-ribosyltransferase was also shown in different rat tissues. Fig. 6A shows this activation in rat brain, liver, intestine, heart, and lung. The basal enzymatic activity varies in the different tissues but in all cases is activated by sodium nitroprusside. In brain and heart (Fig. 6B), the effect of sodium nitroprusside is not mimicked by agents that increase the levels of cyclic GMP.
It has previously been shown that sodium cyanide can stimulate a mitochondrial ADP-ribosyltransferase (31). Therefore, we have also determined the effect of thiocyanide (100-500 PM) and sodium cyanide (100-500 PM) on the ADPribosylation of the 39-kDa protein; however, activity was not affected (not shown).

DISCUSSION
These studies describe the stimulation of an endogenous ADP-ribosyltransferase in various tissues by nitric oxidereleasing compounds. The enzyme mono-ADP-ribosylates a specific 39-kDa protein in all tissues analyzed. This cytosolic protein has not been characterized, but we determined that it is not recognized by GcOmmOn (A-569) antiserum (17) in platelets, brain, or heart. Therefore, this protein is not one of the known G, subunits of the GTP-binding proteins that have similar molecular masses.

Mol WI kDa
3 0 - 1, radioactive protein standards; 2, control; 3, 500 p~ sodium nitroprusside; 4, 500 pM hydroxylamine; 5, 500 p M aniline; 6, 500 p M 8bromo-cyclic GMP; 7, 500 p~ dibutyryl cyclic GMP; 8,500 p~ cyclic GMP. This figure also shows that sodium nitroprusside increases the labeling of a high molecular weight protein. This only occurred when assays were carried out with a high concentration of sodium nitroprusside and a long incubation time (compare to Fig. 1). Sodium nitroprusside and 3-morpholinosydnonimine are known nitrovasodilators and inhibitors of platelet aggregation (32,33). This action is mediated through the spontaneous release of nitric oxide that stimulates soluble guanylate cyclase (34). Nitric oxide is also produced by stimulation of cells, and it is recognized as an endothelium-derived relaxing factor (19,35). Our results now indicate that the release of nitric oxide can affect another enzymatic activity, an ADPribosyltransferase. This effect of nitric oxide is independent of guanylate cyclase activation, because cyclic GMP or agents that increase cellular levels of cyclic GMP do not cause stimulation of this ADP-ribosyltransferase. Sodium nitroprusside and 3-morpholinosydnonimine produce a spontaneous nonenzymatic production of nitric oxide, while formation of nitric oxide by hydroxylamine and aniline is not spontaneous. This could explain why hydroxylamine and aniline are not stimulators of the cytosolic ADP-ribosyltransferase.
Sodium nitroprusside and 3-morpholinosydnonimine can produce vasodilation and the inhibition of the physiological responses of platelets. It has been assumed that this was the result of the activation of guanylate cyclase and the subsequent protein phosphorylation induced by cyclic GMP-dependent protein kinase. It should also be considered that nitric oxide release by sodium nitroprusside and 3-morpholinosydnonimine also affects the ADP-ribosylation of a cellular protein and that this covalent modification might explain, at least in part, the physiological action of nitric oxide. Prostacyclin, which increases cellular cyclic AMP levels and consequent protein phosphorylation, also causes vasodilation and inhibition of platelet responses. Contrary to sodium nitroprusside and 3-morpholinosydnonimine, prostacyclin does not activate the ADP-ribosylation of the 39-kDa protein (not shown). Therefore, the specific contributions of the ADPribosylation and phosphorylation of proteins to physiological responses induced by nitric oxide remain to be elucidated.
Pertussis toxin, cholera toxin, diphtheria toxin, and botulinum toxin are known to have intrinsic ADP-ribosyltransferase activity that can ADP-ribosylate specific cellular proteins. This protein modification has been associated with the changes of specific biochemical reactions and physiological responses (36,37). Endogenous ADP-ribosyltransferase has not as yet been associated with specific cellular changes. It is now imperative to study the possible influences of physiological agonists on this activity and the functional consequences related to this ADP-ribosylation.