INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Among the distinct regulatory properties of adenylyl cyclases, one remains most intriguing- the inhibition of the enzyme by 3'-nucleotides via a domain referred to as the P-site (1). Inhibition via this domain is a property of all known isoforms of mammalian adenylyl cyclases, possibly save the enzyme from sperm (2, 3), and was originally so designated because of the increased inhibitory potency of ligands containing an intact purine. We now know this to be a simplification (4-6). Considerable pharmacological evidence (e.g. Refs. 1, 4, and 7-10) and, specifically, recent observations with a newly synthesized family of 3'-nucleotides from this laboratory (5, 6, 10) demonstrate three key characteristics of inhibitory ligands. These are (a) the nearly absolute requirement for an intact adenine, (b) the enhanced inhibitory potency of 2'-deoxy- and especially 2',5'-dideoxy derivatives, and (c) the markedly enhanced potency elicited by the addition of 3'-phosphates. The family of ligands exhibiting the greatest potency follow the order: 3'-mono- < 3'-di < 3'-triphosphate and Ado¹ < 2'-deoxyadenosine < 2',5'-dd-Ado. Given the dependence of potency on 3'-phosphate group content and because the effectiveness of nucleotides to induce conformational transitions in some proteins has been proportional to the number of phosphates (11), the possibility is suggested that 3'-tetra-phosphate derivatives would be yet more potent inhibitors of adenylyl cyclases. Consequently, 2'-deoxy- and 2',5'-dideoxyadenosine 3'-tetraphosphates (2'-d-3'-A₄P and 2',5'-dd-3'-A₄P) were synthesized and tested for inhibition of crude and purified forms of adenylyl cyclase.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Preparation and Assay of Adenylyl Cyclase-- Detergent-solubilized preparations of adenylyl cyclase from rat and bovine brains were prepared and assayed as described previously (4, 12, 13). Bovine brain adenylyl cyclase was purified as described by Pfeuffer et al. (14). Inhibition kinetics were determined on enzyme assayed in duplicate or triplicate, with concentrations of divalent cation fixed in excess of the 5'ATP concentration as described previously (12, 15). IC₅₀ values were derived from inhibition curves and K_i values from replots of slopes from Lineweaver-Burk plots comprising six to eight concentrations of inhibitor in at least two experiments for each condition. cAMP formation was linear with respect to reaction time and enzyme concentration under these reaction conditions.

Na⁺/K⁺-ATPase-- Activity of Na⁺/K⁺-ATPase was determined in a reaction mixture containing 50 mM triethanolamine·HCl, pH 7.5, 100 mM NaCl, 20 mM KCl, 5 mM MgCl₂, 100 µM 5'ATP, and ~ 10⁵ cpm [-³²P]5'ATP in a volume of 100 µl. Reactions were for 10 min at 30 ^oC and were terminated by the addition of a 1-ml slurry of ice-cold 0.1 M H₃PO₄ containing 25 mg of charcoal/ml. These suspensions were centrifuged, and a measured portion of the supernatant fraction containing the released ³²P_i was quantified by herenkov radiation in a liquid scintillation counter. The rate of ³²P_i release was linear with reaction time and enzyme concentrations under the conditions of these experiments.

Ca²⁺/Calmodulin-dependent Cyclic Nucleotide Phosphodiesterase-- Activity was determined as described previously (16) in a reaction mixture containing 50 mM triethanolamine·HCl, pH 7.5, 2 mM MgCl₂, 40 µM cAMP, 1 mg of bovine serum albumin/ml, 1 mM dithiothreitol, 50 µM CaCl₂, and ~10⁵ cpm [³²P]cAMP in a volume of 100 µl. Reactions were for 10 min at 30 ^oC and were stopped by the addition of 20 µl of a solution containing 100 mM triethanolamine·HCl, 25 mM EDTA, 2.5 mM 3-isobutyl-1-methylxanthine, and 4.53 mM cAMP. 5'-Nucleotidase from Crotalus atrox snake venom was then added to a final concentration of 0.8 mg/ml, and samples were incubated an additional 15 min at 30 ^oC. This second reaction was terminated by the addition of a 1-ml slurry of ice-cold 0.1 M H₃PO₄ containing 25 mg of charcoal/ml. These suspensions were centrifuged, and a measured portion of the supernatant fraction containing the released ³²P_i was quantified by herenkov radiation in a liquid scintillation counter. Calmodulin (~0.7 µM) caused a 3.1 ± 0.1-fold (n = 8) increase in the rate of cAMP hydrolysis by this phosphodiesterase preparation. This rate was linear with reaction time and enzyme concentrations under the conditions of these experiments.

cAMP-dependent Protein Kinase A-- Holoenzyme was isolated from bovine muscle and enriched by ammonium sulfate precipitation and chromatography on DEAE-Sephadex by established procedures (17). Activity was measured in a reaction mixture containing 10 mM Na-phosphate, pH 7.5, 1 mg of bovine serum albumin/ml, 5 mM MgCl₂, 100 µM ATP, and 10 µg of histone, in a volume of 70 µl. Reactions were started with the addition of protein kinase A and were for 15 min at 30 ^oC. They were terminated by spotting 50-µl portions of each sample on Whatman P81 phospho-cellulose paper filters (25 mm diameter) and placing these in ice-cold 75 mM phosphoric acid. Filters were then washed with two exchanges of the phosphoric acid and one wash with ethanol, each after perhaps 10 min. Filters were air dried, and adsorbed ³²P-histone was quantified by herenkov radiation in a liquid scintillation counter. Histone phosphorylation was linear with enzyme concentration and reaction time under these assay conditions, with ATP utilization always less than 2%. The addition of 1 µM cAMP caused a 9.5 ± 0.4-fold (n = 9) increase in the rate of histone phosphorylation by this preparation of protein kinase A.

Quantification of Nucleotides by High Performance Liquid Chromatography-- Nucleotides were quantified after high performance liquid chromatography (HPLC) as areas under peaks determined with a Waters 996 photo-diode array detector and the accompanying Millennium software (Version 2.10). Ion exchange chromatography was on an Altex Spherogel TSK DEAE-5PW column (5 µm, 7.5 × 75 mm) developed with sequential step gradients of triethylammonium bicarbonate, pH 8.5, to separate nucleoside 3'-mono-, 3'-di-, 3'-tri-, and 3'-tetraphosphates.

Tributylammonium Salt of Triphosphoric Acid-- Pentasodium triphosphate (21 g, 56.7 mmol) was dissolved in 200 ml of cold water, and the resulting solution was added to a Büchner funnel loaded with 1.3 liter of regenerated and washed Dowex 50 (H⁺ form). Fractions of 100 ml were collected, and the free acid form of triphosphoric acid was eluted with cold water under aspiration. Pooled fractions were adjusted to pH 5.5 with tributylamine and then were lyophilized to give the tributylammonium salt of triphosphoric acid in the form of a gum.

2',5'-dd-3'-A₄P-- 2',5'-Dideoxyadenosine 3'-O-[(2,2,2-tribromoethyl)morpholinophosphonate] (0.89 g, 1.5 mmol) was added to a solution of pyridine (80 ml) containing activated zinc (0.15 g) (18) and tributylammonium salt of triphosphoric acid (15 g, 15 mmol), under the exclusion of moisture. The mixture was stirred at room temperature for 2 days. The reaction was then concentrated in vacuo, diluted with cold water (300 ml), filtered, and then purified by chromatography on QAE-Sephadex (HCO₃ form) with a linear gradient of triethylammonium bicarbonate (0.1-1 M). The appropriate fractions were lyophilized and then coevaporated several times with methanol, yielding 0.18 mmol of 2',5'-dd-3'-A₄P. This nucleotide was isolated as its sodium salt by addition of 1 M sodium iodide in acetone to a methanol solution of the triethylammonium nucleotide. The precipitate was collected by centrifugation and washed three times with cold acetone and dried in vacuo giving the sodium salt of 2',5'-dd-3'-A₄P: ¹H NMR (D₂O)  1.41 (d, 3H, J = 6.5 Hz, 3H-5'), 2.80-3.00 (m, 2H, H-2', and H-2"), 6.52 (t, 1H, J = 6.9 Hz, H-1'), 8.28 (s, 1H, H-2), 8.47(s, 1H, H-8). ³¹P NMR (D₂O)  18.55(t, J = 17.1 Hz, P-3), 17.80 (t, J = 19.2 Hz, P-2), 8.09 (dd, J_P-H = 9.1 Hz, J_P-P = 18.1 Hz, P-1), 1.55 (d, J = 13.8 Hz, P-4).

2'-d-3'-A₄P-- 5'-O-(Dimethoxytrityl)-2'-deoxyadenosine 3'-O-[(2, 2,2-tribromoethyl)morpholinophosphonate] (1.42 g, 2.5 mmol) was added to a solution of pyridine (50 ml) containing activated zinc (0.5 g) and tributylammonium salt of triphosphoric acid (15 g, 15 mmol), under the exclusion of moisture. The mixture was stirred at room temperature for 2 days. The reaction mixture was then concentrated in vacuo and treated with 80% acetic acid at room temperature for 30 min. The medium was then neutralized with a cold solution of 0.5 M NaHCO₃, diluted to two liters, filtered, and then purified by chromatography on QAE-Sephadex as above, yielding 0.21 mmol of 2'-d-3'-A₄P. This nucleotide was also isolated as its sodium salt as above. No impurities were noted on anion exchange HPLC: ¹H NMR (D₂O)  2.81-2.99 (m, 2H, H-2' and H-2''), 3.88 (d, 2H, 2H-5'), 4.40-4.45 (m, 1H, H-4'), 5.06-5.19 (m, 1H, H-3'), 6.55 (t, 1H, J = 6.7 Hz, H-1'), 8.26 (s, 1H, H-2), 8.40 (s, 1H, H-8); ³¹P NMR (D₂O)  18.28 (t, J = 17.1 Hz, P-3), 17.24 (t, J = 17.9 Hz, P-2), 8.07 (dd, J_P-H = 7.7 Hz, J_P-P = 19.2 Hz, P-1), 1.44 (d, J = 18.4 Hz, P-4).

Materials-- [-³²P]5'ATP and [-³²P]5'ATP were purchased from ICN Pharmaceuticals. Lubrol-PX (from Sigma, L-3753), used for solubilizing the enzyme, was filtered through alumina (Neutral, AG7, from Bio-Rad) to remove peroxides. [³²P]cAMP was prepared from [-³²P]ATP by reaction with a purified preparation of adenylyl cyclase with 5 mM MnCl₂ as divalent cation and in the absence of added unlabeled 5'ATP, as described previously (16). Calmodulin-sensitive cAMP-phosphodiesterase, isolated from chicken gizzard and chromatographically enriched on DEAE-cellulose, was a gift from Dr. Jack N. Wells, Department of Pharmacology, Vanderbilt University, Nashville, TN. Calmodulin was purified to homogeneity from porcine testes by established procedures (19). Dog kidney Na⁺/K⁺-ATPase was from Sigma, A7305. Pyridine was redistilled over CaH₂. Charcoal used was carbon decolorizing alkaline Norit A (C-176) from Fisher Scientific that was then activated. Tetrahydrofuran was distilled from benzophenone ketyl before use. NMR spectra were recorded with a Bruker AC250 at 250 MHz for proton spectra and at 101 MHz for ³¹P spectra, with an 85% solution of H₃PO₄ as external standard. The purity of the nucleotides was checked by HPLC with a DEAE column (above), eluted with gradient of triethylammonium bicarbonate and with ion pair chromatography on an Ultrasphere C18 column (5 µm; 4.6 × 250 mm), eluted with a gradient from 10 mM tetrabutylammonium hydroxide, 10 mM KH₂PO₄, 1% methanol, pH 5.5 to 2.8 mM tetrabutylammonium hydroxide, 100 mM KH₂PO₄, 30% methanol, pH 7.0. 2',5'-dd-3'-ATP and 2'-d-3'-ATP were synthesized as described previously (5, 6).

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

2'-d-3'-A₄P and 2',5'-dd-3'-A₄P Synthesis-- The previously described synthesis of a family of adenine nucleoside 3'-polyphosphates combined the method of alkoxide activation with the use of tribromoethyl phosphoromorpholino-chloridate as an initial phosphorylating agent (5, 6, 20). This method was extended for the synthesis of 2'-d-3'-A₄P and 2',5'-dd-3'-A₄P by the second stage use of the tributylammonium salt of inorganic triphosphate (Scheme 1). Although yields for these syntheses were low (8.4 and 12%, respectively), they were sufficient to permit the preparation of research quantities of these 3'-nucleotides.

View larger version (12K): [in this window] [in a new window]   Scheme 1.   Reagents and conditions. i, tert-BuMgCl, tetrahydrofuran, room temperature; ii, phosphorylating agent, room temperature; iii, zinc, (Bu3N)3 H5P3O10, pyridine, room temperature, followed by AcOH 80%, room temperature, 30 min for 2.

Inhibition of Adenylyl Cyclases-- The 3'-tetraphosphates, 2'-d-3'-A₄P and 2',5'-dd-3'-A₄P, exhibited notably more potent inhibition of adenylyl cyclase than did the respective 3'-triphosphates (Fig. 1). A comparison of potencies of these nucleotides, determined from several experiments, are given in Table I. The addition of the fourth phosphate increased potency for the 2'-deoxy derivative more than for the 2',5'-dideoxy derivative, whereas in both cases, the increase in potency was not quite an order of magnitude. With IC₅₀ values of 32 and 106 nM (Table I), respectively, for 2',5'-dd-3'-ATP and 2'-d-3'-ATP, the 3'-triphosphates exhibited potencies with this rat brain extract similar to those previously reported (6). The 3'-tetraphosphates exhibited IC₅₀ values of ~10.5 nM for 2'-d-3'-A₄P and ~7.4 nM for 2',5'-dd-3'-A₄P and K_i values of 53 and 23 nM, respectively, making them the most potent known regulators of adenylyl cyclase activity. This inhibition occurred at an estimated enzyme concentration of approximately 0.9 nM, with the assumptions of a mass of 116 kDa and a specific activity of 7 µmol/(min·mg of protein) for the purified type I adenylyl cyclase (14, 21). This is not a large excess of inhibitor relative to enzyme, and slightly lower IC₅₀ values might be observed with lower enzyme concentrations.

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Inhibition of rat brain adenylyl cyclase by adenine nucleoside 3'-tri- and 3'-tetraphosphates. Activities were determined with a detergent-dispersed preparation of adenylyl cyclase from rat brain as described under "Experimental Procedures." Velocities are relative to an uninhibited initial velocity of 5.6 nmol cAMP formed/(min·mg of protein). Protein was 10.8 µg/100 µl of reaction volume.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Time course for inhibition of rat brain adenylyl cyclase by 2'-d-3'-A4P, 2',5'-dd-3'-A4P, and 2',5'-dd-3'-ATP. Enzyme and determinations were as for Fig. 1. Uninhibited control velocity was 7.1 nmol of cAMP formed/(min·mg of protein). Inhibitor additions and concentrations were as indicated.

View larger version (34K): [in this window] [in a new window]   Fig. 3.   Double reciprocal plot for inhibition of rat brain adenylyl cyclase by 2',5'-dd-3'-A4P. Enzyme and determinations were as for Fig. 1. Units for velocity are nmol of cAMP formed/(min·mg of protein). Concentrations of 2',5'-dd-3'-A4P were as indicated.

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Double reciprocal plot for inhibition of adenylyl cyclase purified from bovine brain by 2',5'-dd-3'-A4P. Enzyme was prepared and assayed as described under "Experimental Procedures." Units for velocity are pmol of cAMP formed/(min·100 µl of assay volume). Concentrations of 2',5'-dd-3'-A4P were as indicated.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Eadie-Hofstee plot for inhibition of adenylyl cyclase purified from bovine brain by 2',5'-dd-3'-A4P. Data are from the experiment represented in Fig. 4. Units for velocity are pmol of cAMP formed/(min·100 µl of assay volume). Concentrations of 2',5'-dd-3'-A4P were as indicated.

Effects on Other Enzymes-- Because few adenine nucleotides interact solely with a single protein, the possibility was considered that adenosine 3'-polyphosphates might affect enzymes other than adenylyl cyclase. As an initial investigation in this direction, we tested effects on Na⁺/K⁺-ATPase and on two enzymes participating in the cAMP signaling cascade, cAMP phosphodiesterase and cAMP-dependent protein kinase.

View larger version (21K): [in this window] [in a new window]   Fig. 6.   Lack of effect of 2',5'-dd-3'-ATP or 2',5'-dd-3'-A4P on cAMP-dependent protein kinase. Activities were determined as described under "Experimental Procedures" and are expressed as pmol of 32P transferred from [-32P]5'-ATP to histone per min. Additions of 3'-nucleotides and cAMP (1 µM) were as indicated.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

2'-Deoxy- and 2',5'-dideoxyadenosine 3'-tetraphosphates represent important additions to the family of 3'-polyphosphates constituting the most potent known regulators of adenylyl cyclases. The rapid onset and linear noncompetitive nature of inhibition are characteristics that conform to those expected for P-site ligands. Inhibition by either 2'-d-3'-A₄P (IC₅₀ ~10.5 nM) or 2',5'-dd-3'-A₄P (IC₅₀ ~7.4 nM) was not because of the formation of inorganic polyphosphates, whether by enzymatic or nonenzymatic means. Significant hydrolysis of the 3'-tetraphosphates did not occur, notwithstanding the crude nature of some of the adenylyl cyclase preparations with which they were tested. This stability was consistent with that previously reported for the homologous 3'-triphosphates (6) and implies a lack of significant 3'-nucleotidase activity in these preparations. Moreover, the possibly resulting inorganic polyphosphates, i.e. PPPP_i, PPP_i, PP_i, and P_i, are not potent inhibitors (6). PPPP_i and PPP_i were three orders of magnitude less potent (IC₅₀ ~55 µM) than were the title compounds, and inhibition was competitive with respect to substrate (6). PP_i was less potent (IC₅₀ ~1.4 mM) and inhibition was mixed (12). These observations argue convincingly that it is 2'-d-3'-A₄P and 2',5'-dd-3'-A₄P per se that cause inhibition in the present study.

The striking inhibitory potency of 2'-d-3'-A₄P and 2',5'-dd-3'-A₄P is due both to the specificity of interaction of the adenine moiety, for which there is nearly an absolute requirement, and to the binding energy contributed by the addition of phosphates at the 3'-position, which for some ligand-protein interactions has been as high as ~10 kcal/phosphate (26). Both moieties are obviously important for interaction with adenylyl cyclases, and the similarities in structure of inhibitor and substrate imply similarity of catalytic and inhibitory configurations of the enzyme. Available data argue that both cytosolic domains (C₁ and C₂) of adenylyl cyclases contain potential binding domains for adenine nucleotide, that both are required for catalysis and inhibition, and that catalysis occurs along a cleft located at the interface of these two domains (27-30). The simplest inhibitory mechanism is that of dead-end inhibition, suggested by Wolin (31) for the enzyme from Brevibacterium liquifaciens and by Dessauer and Gilman (32) for a truncated chimeric construct of the mammalian enzyme. Although not supported by all the evidence,² this model allows for the common observations that P-site-mediated inhibition is more potent with activated forms of the enzyme (4, 8-10, 12, 13, 23, 24, 33) and that inhibition of adenylyl cyclase by adenine nucleosides or by adenine nucleoside monophosphates is enhanced by PP_i (12, 31, 32, 34). Recently solved structures of truncated constructs of adenylyl cyclase C₁ and C₂ domains are consistent with 2'-d-3'-AMP binding in what appears to be a catalytic site (29, 30). P-site ligands such as 2',5'-dd-Ado or 2'-d-3'-AMP would bind at the adenosine-3':5'-cyclic monophosphate leaving site in the presence of PP_i (12, 31-34), the other product of the reaction. Structures with either bound substrate or bound nucleoside 3'-polyphosphate have not been solved (29, 30).³ Presumably, though, 2',5'-dd-3'-ATP would occupy both adenine and PP_i binding domains and 2',5'-dd-3'-A₄P would take advantage of a locus within the cleft capable of accepting the fourth phosphate group. Whether 3'-mono- or 3'-polyphosphate, inhibitor would bind to that configuration of the enzyme from which product leaving occurs and would prevent a subsequent conformational shift to a form capable of interacting again with substrate metal-5'ATP.

Inhibition by adenosine 3'-polyphosphates is a characteristic conserved in all mammalian adenylyl cyclases sequenced to date. Although examples from bacteria and possibly sperm indicate that catalysis can occur without susceptibility to P-site ligands (2, 31), it is nonetheless a characteristic that might well have been lost by natural selection did it not serve an important function. It provides an exquisite means for inhibition of this crucial signal transduction pathway. Both 2'-d-3'-AMP and 3'-AMP occur naturally, their levels may be chronically regulated (39), and 3'-polyphosphorylation of nucleoside 3'-monophosphates is known to occur in mammalian systems (40). There are, however, very few reports related to nucleoside-3'-polyphosphates in animal cells. If nucleoside 3'-tri- or 3'-tetraphosphates were to exist naturally in animal cells and to exert a physiological effect in the regulation of adenylyl cyclases, they might be expected to exist in minute quantities, in a range consistent with their effects on adenylyl cyclases. For this reason alone, their presence may not have been detected. By comparison, adenosine 5'-tetraphosphate occurs naturally (e.g. Refs. 35 and 36), and there is considerable literature on diadenosine tetraphosphates (A(5')p4(5')A; e.g. Ref. 37).⁴ However, the former has not been tested as a P-site ligand, and the latter is not an inhibitor (4) although a number of 3'  5'-dinucleotides do inhibit the enzyme (cf. Table II of Ref. 4). That adenosine 3'-polyphosphates can inhibit adenylyl cyclase in intact cells was confirmed by Hempel et al. (38), who observed complete obliteration of serotonin-induced elevations in cAMP and the accompanying hyperpolarization response in lobster stomatogastric ganglion cells. Because of the nature of microinjection experiments, though, it was not possible to determine precisely the actual intracellular concentration of 2',5'-dd-3'-ATP that was effective.

Given that few if any adenine nucleotides interact solely with one protein, it is also likely that, as part of their overall action in cells, adenosine 3'-polyphosphates will affect proteins other than adenylyl cyclases. Neither 2',5'-dd-3'-ATP nor 2',5'-dd-3'-A₄P affected activities of the three other adenine nucleotide binding enzymes we tested. This does not preclude effects on other enzymes. DNA polymerase is an interesting example and comparisons of it and adenylyl cyclase may suggest structural motifs required for regulation by this class of nucleotide. Prokaryotic DNA polymerase is well known to bind 2'-deoxynucleoside 3'-triphosphates, and even 3'-tetraphosphates, via the triphosphate domain (41). The affinity for nucleotide (2'-d-3'-ATP: K_D ~38-68 µM (41)), though, is almost 3 orders of magnitude less than that exhibited by adenylyl cyclase, implying at once that the 3'-nucleotide binding domains of these two enzyme families exhibit similarities and important differences. That adenylyl cyclases share structural as well as catalytic characteristics with "palm" domains of the polymerase I family of prokaryotic DNA polymerases is also evident from comparisons of three-dimensional topologies of both proteins, as recently pointed out by Artymiuk et al. (42), though qualified by Bryant et al. (43). These observations would suggest that other enzymes sharing this topology and/or yielding nucleoside monophosphate and inorganic pyrophosphate as products would be potential targets for additional interactions with nucleoside 3'-polyphosphates.