Cellular Signalling 12 (2000) 233–237

http://www.elsevier.com/locate/cellsig

Nitric oxide selectively inhibits adenylyl cyclase isoforms 5 and 6Jennifer Hilla, Allyn Howlettb, Claudette Kleina,*

aE.A. Doisy Department of Biochemistry and Molecular Biology, Saint Louis University School of Medicine, 1402 S. Grand Blvd., St. Louis,MO 63104, USA

bDepartment of Pharmacological and Physiological Science, Saint Louis University School of Medicine, 1402 S. Grand Blvd., St. Louis, MO63104, USA

Received 18 August 1999; revised 29 November 1999; accepted 8 December 1999

Abstract

We have previously shown that N18TG2 neuroblastoma cells express the type 6 adenylyl cyclase and that preincubation withnitric oxide (NO) attenuates Gs- and forskolin-stimulated activity. Here we show that this inhibition reflects a direct action ofNO on the adenylyl cyclase. Preincubation of N18TG2 cell membranes and insect cell membranes expressing recombinant type5 and type 6 isoforms with NO donors leads to an inhibition of forskolin-stimulated adenylyl cyclase activity. NO donors do notalter the type 1 (representative of the type 1,3,8 family) or type 2 (representative of the type 2,4,7 family) isoforms expressedin insect cells, even under conditions of compromised assay conditions or a range of temperatures. Thus, the ability of NO toinhibit adenylyl cyclase stimulation is dependent upon the nature of the isoform present, and appears to represent a uniqueregulation of the type 5,6 isoform family. 2000 Elsevier Science Inc. All rights reserved.

Keywords: Adenylyl cyclase; Isoform specific regulation; Nitric oxide

1. Introduction targets of NO. The regulatory properties of the predom-inant isoform of adenylyl cyclase expressed in N18TG2

Nitric oxide (NO) regulates a variety of biologicalindicate that it is a type 5 or 6 isoform [5–8] while

phenomena including platelet aggregation and adher-our recent Northern analyses indicate that the type 6

ence, vascular smooth muscle relaxation, and neuro-enzyme is the one expressed in N18TG2 cells [9].transmitter release (reviewed in [1–3]). When present

Experiments using intact cells eliminated hormonein relatively low levels, NO is believed to function ho-receptors, Gs and Gi/o as targets for NO. Upon duplica-meostatically but when present in chronic or elevatedtion of the effects of NO in purified membrane prepara-levels, it is believed to be a cytostatic/cytotoxic agent. Totions, we also eliminated indirect actions of NO (e.g.,understand how NO elicits its numerous effects, manyvia changes in cellular ATP levels or on ion channels)researchers have attempted to identify the target mole-[9]. Thus, either the adenylyl cyclase itself, or an as yetcules that are altered by NO. Although NO-inducedunknown regulator of the type 5,6 isoform family is thevasodilation appears to reflect changes in soluble guany-target of NO. To demonstrate that the adenylyl cyclaselyl cyclase activity, the functional significance of NO-protein is the target for NO, and to evaluate whichmediated changes in cGMP in other systems is not soisoform families are sensitive to NO, we have studiedevident.the effects of NO on recombinant isoforms. We demon-We have demonstrated that the addition of NO tostrate that the types 5 and 6 isoforms are specifically,N18TG2 neuroblastoma cells, either in the form of NOand directly, regulated by NO.donor compounds, or NO gas, attenuates the ability of

cells to accumulate cAMP in response to both hormoneand forskolin stimulation [4]. Experiments ruled out any

2. Materials and methodseffect of NO on cAMP degradation and thus focused ourattention to regulators of cAMP synthesis as possible 2.1. Membrane pretreatment with NO donors

Sucrose gradient purified N18TG2 plasma mem-* Corresponding author. Tel.: 314-577-8155; fax: 314-577-8156.E-mail address: Kleinc@slu.edu (C. Klein) branes [7] were resuspended in buffer A (50 mM Na-

0898-6568/00/$ – see front matter 2000 Elsevier Science Inc. All rights reserved.PII: S0898-6568(99)00082-0

234 J. Hill et al. / Cellular Signalling 12 (2000) 233–237

Hepes pH 8, 3 mM MgCl2, 1 mM EDTA) at 0.5 mg/ml.Sf9 and Hi-5 membranes were resuspended in buffer B(20 mM NaHepes pH 8, 2 mM MgCl2, 1mM EDTAplus 3.2 mg/ml leupeptin and trypsin inhibitor, 2 mg/mlaproptinin, 22 mg/ml TLCK, and 1 mM PMSF). SNP(sodium nitroprusside) and SNAP (S-nitroso-N-acetyl-D,L-penicillamine) were prepared as stock solutions inMillipore-filtered water and methanol, respectively. So-lutions were kept on ice and used within 30 minutes ofpreparation. Membranes were preincubated with NOdonors for 30 min on ice prior to assay. Controls (0donor) were given either vehicle alone or the largestaliquot of donor solution that had been made 16 hoursprior to addition. Release of NO, in the form of nitrite,was measured using the Greiss reagent as described [10].

2.2. Adenylyl cyclase activity

Enzyme activity monitors the conversion of[a32P]ATP to [32P]-cAMP at 308C for 10 min as pre-viously described [6]. Where indicated, forskolin waspresent at 10 mM. The phosphodiesterase inhibitors forN18TG2 and Sf9 or Hi-5 insect membranes were 0.1 mMRo20-1724 and isobutylmethylxanthine, respectively. Theassays were initiated by the addition of 15 mg N18TG2or 5 mg Sf9 or Hi-5 membrane protein. Determinationswere made in triplicate. The data are presented in pmol/min/mg protein in order to facilitate comparison of spe-cific activities between experiments. Each figure is rep-resentative of experiments performed a minimum ofthree times, with the percent inhibition of enzyme activ- Fig. 1. NO donors inhibit adenylyl cyclase activity in N18TG2 cellity between experiments agreeing within 10%. membranes. Membranes were preincubated with the indicated con-

centrations of either SNP (A) or SNAP (B) and assayed for basal (b)2.3. Materials or forskolin- (fsk) stimulated adenylyl cyclase activity. The data are

means 6 S.E.M. of triplicates. 0 donor represents the activity ofSNP and SNAP were purchased from Alexis Bio- membranes pretreated with vehicle alone, which was equal to that in

which donor solution had been made 16 h prior to addition.chemicals (San Digeo, CA, USA). All other reagentswere purchased from Sigma (St. Louis, MO, USA).

the time period of our experiments [10]. Indeed, assays3. Results of the amount of nitrite (the oxidation product of NO)

released during a 30-min incubation indicated that the3.1. NO donors inhibit N18TG2 cell adenylylmaximum levels of NO that could accumulate from 10cyclase activitymM SNP and SNAP was only 3.6 and 20 mM, respec-tively. The relative difference in NO release is consistentWe have previously observed that when membranes

of N18TG2 cells are preincubated with SNP and then with the relative effectiveness of these two donors. Ifmembranes were washed free of donor prior to theassayed for adenylyl cyclase activity, hormone and for-

skolin-stimulated activity is attenuated [9]. In Fig. 1, assay, the inhibition in activity persisted, eliminatingthe possibility that the donors functioned by altering athe concentration-dependent effects of two structurally

distinct NO donors, SNP (Fig. 1a) or SNAP (Fig. 1b) component of the assay (data not shown). We used twochemically distinct NO donors to assure that the effectswere examined. In either case, NO donors had no effect

on basal activity. In contrast, forskolin-stimulated activ- were mediated by NO and not an alternative decomposi-tion product. This was confirmed by allowing the solu-ity was inhibited in a concentration-dependent manner.

SNAP appeared to be the more potent of the inhibitors, tions of SNP and SNAP to decompose overnight priorto addition to membranes. Under such conditions, theresulting in approximately a 50% inhibition with 0.2

mM SNAP whereas 1 mM SNP was necessary for a NO released is converted to alternative oxidation prod-ucts, nitrite/nitrate [11] while the other decompositionsimilar inhibition. These concentrations of donor would

result in low micromolar concentrations of NO within products are unaltered. Under such conditions, no ef-

J. Hill et al. / Cellular Signalling 12 (2000) 233–237 235

Fig. 2. The forskolin-stimulated activity of recombinant types 5 and6 isoforms are attenuated by NO donors. Insect cell membranes ex-pressing either the type 5 (A) or 6 (B) recombinant enzymes werepreincubated with the indicated concentrations of NO donors. Mem- Fig. 3. NO donors do not inhibit forskolin-stimulated activity of thebranes were then assayed for basal (b) or forskolin- (fsk) stimulated types 1 and 2 isoforms. Insect cell membranes expressing the recombi-activity. The data are means 6 S.E.M. of triplicate determinations. nant type 1 (A) or type 2 (B) were preincubated with the indicatedControl samples represent the activity of membranes pretreated with concentrations of NO donors. Membranes were then assayed for basalvehicle alone, which was equal to that in which donor solution had (b) or forskolin- (fsk) stimulated activity. The data are means 6been made 16 h prior to addition. S.E.M. of triplicates.

fects were observed on either basal or forskolin-stimu- crease in forskolin-stimulated activity, with SNAP beinglated activity. the more effective donor. As with N18TG2 membranes,

basal adenylyl cyclase activity was not altered. The other3.2. NO donors inhibit selective isoforms of

member of this isoform family, the type 6 enzyme, wasadenylyl cyclase

similarly regulated by NO (Fig. 2b). Because the activityof the type 6 isoform was relatively low, we primarilyTo determine if the adenylyl cyclase protein is theused membranes expressing the type 5 isoform for ourdirect target of NO, we examined the sensitivity of re-studies of this family of enzymes.combinant type 5 and 6 isoforms (Fig. 2). Forskolin

To see if other adenylyl cyclase isoforms could serveproduced a 10-fold increase in recombinant type 5 activ-as NO targets, we examined the type 1 (representativeity expressed in Sf9 membranes preincubated withoutof the type 1, 3, 8 family) and type 2 (representative ofadditions (Fig. 2a, control). The same results were seenthe type 2, 4, and 7 family) isoforms (reviewed in [12–with membranes preincubated with NO donor solutions14]). As shown in Fig. 3, neither SNP nor SNAP alteredthat had been allowed to decompose overnight beforethe ability of forskolin to stimulate the recombinantaddition. Membrane preincubation with freshly pre-

pared solutions of either SNP or SNAP caused a de- types 1 and 2 isoforms. To eliminate the possibility that

236 J. Hill et al. / Cellular Signalling 12 (2000) 233–237

a component present in Buffer B might interfere withthe action of the NO donors, we pretreated N18TG2membranes with NO donors in Buffer B and found thatthe inhibition of the enzyme was unaltered. Addition-ally, we determined that NO donors remained effectiveinhibitors of the activity in N18TG2 membranes whenthose membranes were mixed with Sf9 membranes con-taining recombinant AC2. We also considered the possi-bility that the enzyme must attain an NO-sensitive con-formation. If so, membrane fluidity would be expectedto influence the freedom with which the enzyme attainsthat conformation. Therefore, we used a range of tem-peratures during the preincubation of Sf9 membranescontaining the types 1 and 2 isoforms with NO donors,but no effect of NO was observed (data not shown).

We then considered that use of substrate and divalentcation concentrations that provide for maximum activitymight limit the possibility of observing an effect of NOon isoforms other than the types 5 and 6. We thereforeused assay conditions that were suboptimal. One ap-proach was to reduce the Mg21 concentration to limitthe amount of divalent cation present as free Mg21. Ithas been shown that one Mg21 ion associates with theadenylyl cyclase as a complex with ATP while anotherfunctions kinetically as free Mg21 to stimulate activity[15]. As shown in Fig. 4a, although enzyme activity wasreduced by approximately 50% when measured in thepresence of 3 mM MgCl2, as opposed to the normal 10mM, we were still unable to evidence an effect of NOdonors on the type 2 isoform. SNP and SNAP were stilleffective inhibitors of the forskolin-stimulated activityof the type 5 isoform (Fig. 4b), and of the activity presentin N18TG2 membranes (Fig. 4c). In a similar set ofexperiments in which the ATP concentration was re-duced 5-fold to 0.1 mM, we were unable to observe aneffect of NO donors. Additionally, reducing the concen-

Fig. 4. Modification of the assay conditions in order not to materializetration of forskolin by 10-fold also did not materialize an effect of NO donors on the types 1 and 2 isoforms. Insect cellan inhibition of type 1 or type 2 adenylyl cyclase activity membranes expressing the recombinant type 2 (A) or type 5 (B)by NO donors (data not shown). isoforms, or N18TG2 cell membranes (C) were preincubated with the

indicated concentrations of NO donors. Membranes were then assayedfor basal (b) or forskolin- (fsk) stimulated activity in the presence of3 mM MgCl2. The data are means 6 S.E.M. of triplicates.

4. Discussion

The experiments presented here indicate that NO CaM-regulation of the type 1 isoform is inhibited bycan attenuate forskolin-stimulated adenylyl cyclase ac- NO treatment of intact cells.tivity of both type 5 and 6 isoforms by directly targeting That forskolin-stimulated activity of select isoformsthe enzyme. We also show that this regulation by NO of adenylyl cyclase is inhibited by NO argues that theis specific to this isoform family and does not occur catalytic site of the enzyme is unlikely to be the targetwith either the type 1 or type 2 isoforms, enzymes that of NO. This premise is supported by our finding thatrepresent the regulatory features of the type 1,3,8 and basal enzyme activity of the types 5 and 6 isoforms istype 2,4,7 isoform families [12–14]. That NO does not not altered by NO. That forskolin stimulation of thealter forskolin-stimulated activity of recombinant type other isoforms of adenylyl cyclase is not similarly attenu-1 enzyme has been previously demonstrated by Duhe ated would indicate that either the amino acid residue(s)et al. [16]. In those experiments, the authors showed targeted by NO in the types 5 and 6 isoforms are notthat the CaM stimulation of the type 1 isoform could present in the other isoforms, or that the conformation

of the latter enzymes precludes accessibility to NO.be inhibited by NO. It remains to be determined if the

J. Hill et al. / Cellular Signalling 12 (2000) 233–237 237

We observed that, in progressing from intact cells to Sustained production of NO, generally associated withcytostatic/cytotoxic effects, would be expected to resultplasma membranes, higher concentrations of NO do-in a more chronic inhibition of adenylyl cyclase activity.nors were required to attenuate forskolin-stimulated ad-Continued investigations into the mechanisms by whichenylyl cyclase activity. For example, 0.1–0.2 mM SNPNO specifically modulates adenylyl cyclase activitywas maximally effective in intact N18TG2 cells whileshould not only provide valuable information with re-3–10 mM SNP was needed when using membranes, ei-gard to the structure/function relationship of the variedther from N18TG2 or insect cells. Similar concentrationisoforms but also provide insights into how NO func-ranges of SNP have generally been used to attenuatetions to modulate signal transduction.the functions of a variety of proteins [12,16,17]. In partic-

ular, such concentrations of SNP were necessary forDuhe et al. [16] to observe an inhibition of CaM regula- Acknowledgmentstion of the type 1 isoform in Sf9 membranes. Our previ-

The authors thank Dr. A.G. Gilman and Dr. R. Iy-ous experiments, in which substrate occupancy of theengar for providing us with Sf9 and Hi-5 cell membranesadenylyl cyclase in N18TG2 membranes protected theexpressing the recombinant isoforms of adenylyl cy-activity from the effects of NO, suggested that the sensi-clase. We also thank Dr. R. Sunahara and Dr. R. Iyengartivity of the enzyme is related to its conformation [9].for insights into the assays of these enzymes, and Dr.In that case, it is reasonable to propose that dynamicJ. Corbett for performing the nitrite determinations forconformational changes occur more readily in intactus. We thank Jill Pigg and Mirya Kim for preparationcells than in mechanically disrupted membranes. Weof N18TG2 cell membranes. This work was supportedalso note that the release of NO from donors is relativelyby research funds from the NIH (A.H.) and AHA (C.K.).complex, and influenced by a number of factors includ-

ing the incubation medium, light, temperature and pH[10]. Thus differences in the media in which the experi- Referencesments were performed could also account for the altered

[1] Bredt DS, Snyder SH. Annu Rev Biochem 1994;63:175–95.effectiveness of donor compounds in cells versus mem-[2] Garthwaite J. Trends Neurosci 1991;14:60–7.branes. Indeed, the levels of NO released by these do- [3] Stamler JS. Cell 1994;78:931–6.

nors when incubated with membranes was found to be [4] Tao Y-P, Najafi L, Shipley S, Howlett A, Klein C. J Pharm Exptlin the low micromolar range, a level of NO which we Ther 1998;286:298–304.

[5] Matsuzawa H, Nienberg M. Proc Natl Acad Sci USAhave previously shown to attenuate maximally cAMP1978;75:1788–91.production in intact cells [4]. In either case, we note

[6] Howlett AC. Mol Pharmacol 1982;21:664–70.that the effects of both donors were quite selective, [7] Howlett AC. Mol Pharmacol 1985;27:429–36.modulating the forskolin-stimulated, but not basal, ac- [8] Howlett AC, Qualy JM, Khachatrian LL. Mol Pharmacol

1986;29:307–13.tivity of the types 5 and 6 isoforms, with no effect on[9] McVey M, Hill J, Howlette AC, Klein C. J Biol Chemeither basal or forskolin-stimulated activity of the types

1999;274:18887–92.1 and 2 isoforms.[10] Green LC, Wagner DA, Glogowski J, Skipper PL, Wishnok JS,

The Gs- and forskolin-stimulated cAMP production Tannenbaum SR. Anal Biochem 1982;126:131–8.by N18TG2 cells proved to be very responsive to low [11] Feelisch M, Stamler JS. In: Feelisch M, Stamler JS, editors. Meth-

ods in Nitric Oxide Research. J. Wiley, New York, NY, USAmicromolar concentrations of NO [4]. With intact cells,1997. pp. 71–98.the inhibitory effects of such low concentrations were

[12] Iyengar R. FASEB J 1993;7:768–75.also found to be readily reversible [4], suggesting that [13] Mons N, Cooper DMF. Trends Neurosci 1995;18:536–42.this regulation of adenylyl cyclase could reflect a homeo- [14] Sunahara RK, Dessauer CW, Gilman AG. Annu Rev Pharmacol

Toxicol 1996;36:461–80.static mechanism. In situ, however, the sources of NO[15] Hurley JJ. J Biol Chem 1999;274:7599–602.that may regulate cAMP production can be quite varied.[16] Duhe RJ, Nielsen MD, Dittman AH, Villacres EC, Choi EJ,Transient, low levels of NO would be provided by cells Storm, DR. J Biol Chem 1994;269:7290–6.

expressing a cNOS whereas cells expressing iNOS would [17] Lei SZ, Pan ZH, Aggarwal SK, Chen HS, Hartman J, SucherNJ, Lipton SA. Neuron 1992;8:1087–99.produce relatively high and chronic levels of NO [1–3].