Adenine Nucleoside 3′-Tetraphosphates Are Novel and Potent Inhibitors of Adenylyl Cyclases*

2′-Deoxyadenosine 3′-tetraphosphate (2′-deoxy-3′-A4P) and 2′,5′-dideoxyadenosine 3′-tetraphosphate (2′,5′-dideoxy-3′-A4P) were synthesized, and their effects were tested on crude and purified forms of native adenylyl cyclases isolated from brain. Syntheses combined the method of alkoxide activation with the use of tribromoethyl phosphoromorpholino-chloridate as an initial phosphorylating agent. Inhibition of adenylyl cyclase was rapid in onset. With 2′-d-3′-A4P or 2′,5′-dd-3′-A4P inhibition of a purified native enzyme conformed to a linear noncompetitive behavior with respect to substrate, metal-5′ATP. Order of potency was 2′,5′-dideoxy- > 2′-deoxyadenosine and 3′-tetraphosphate > 3′-triphosphate. Both mechanism of inhibition and rank order of potency were consistent with inhibition via the 3′-nucleotide-(P)-site on adenylyl cyclase. Neither 2′,5′-dd-3′-ATP nor 2′,5′-dd-3′-A4P had any effect on the activities of other adenosine nucleotide binding proteins such as Ca2+/calmodulin-sensitive cyclic nucleotide phosphodiesterase, Na+/K+-ATPase, or cAMP-dependent protein kinase. With purified adenylyl cyclase from bovine brain 2′,5′-dd-3′-A4P and 2′-d-3′-A4P gave, respectively, IC50 values of 9.3 and 15 nm and Ki values of 23 and 53 nm. These 3′-nucleotides are the most potent regulators described for adenylyl cyclases.

diation in a liquid scintillation counter. Calmodulin (ϳ0.7 M) caused a 3.1 Ϯ 0.1-fold (n ϭ 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 Naphosphate, pH 7.5, 1 mg of bovine serum albumin/ml, 5 mM MgCl 2 , 100 M ATP, and 10 g of histone, in a volume of 70 l. Reactions were started with the addition of protein kinase A and were for 15 min at 30 o C. They were terminated by spotting 50-l 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 32 P-histone was quantified by Ĉ 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 M cAMP caused a 9.5 Ϯ 0.4-fold (n ϭ 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 m, 7.5 ϫ 75 mm) developed with sequential step gradients of triethylammonium bicarbonate, pH 8.5, to separate nucleoside 3Ј-mono-, 3Ј-di-, 3Ј-tri-, and 3Ј-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 Bü chner funnel loaded with 1.3 liter of regenerated and washed Dowex 50 (H ϩ 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.
2Ј,5Ј-dd-3Ј-A 4 P-2Ј,5Ј-Dideoxyadenosine 3Ј-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 (HCO 3 Ϫ 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 2Ј,5Јdd-3Ј-A 4 P. 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 2Ј,5Ј-dd-3Ј-A 4 P: 1 (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 NaHCO 3 , diluted to two liters, filtered, and then purified by chromatography on QAE-Sephadex as above, yielding 0.21 mmol of 2Ј-d-3Ј-A 4 P. This nucleotide was also isolated as its sodium salt as above. No impurities were noted on anion exchange HPLC: 1  Materials-[␣-32 P]5ЈATP and [␥-32 P]5ЈATP 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. [ 32 P]cAMP was prepared from [␣-32 P]ATP by reaction with a purified preparation of adenylyl cyclase with 5 mM MnCl 2 as divalent cation and in the absence of added unlabeled 5ЈATP, 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 ϩ /K ϩ -ATPase was from Sigma, A7305. Pyridine was redistilled over CaH 2 . 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 31 P spectra, with an 85% solution of H 3 PO 4 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 m; 4.6 ϫ 250 mm), eluted with a gradient from 10 mM tetrabutylammonium hydroxide, 10 mM KH 2 PO 4 , 1% methanol, pH 5.5 to 2.8 mM tetrabutylammonium hydroxide, 100 mM KH 2 PO 4 , 30% methanol, pH 7.0. 2Ј,5Ј-dd-3Ј-ATP and 2Ј-d-3Ј-ATP were synthesized as described previously (5,6).

2Ј-d-3Ј-A 4 P and 2Ј
,5Ј-dd-3Ј-A 4 P Synthesis-The previously described synthesis of a family of adenine nucleoside 3Јpolyphosphates 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 2Ј-d-3Ј-A 4 P and 2Ј,5Ј-dd-3Ј-A 4 P 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 3Ј-nucleotides.
Inhibition of Adenylyl Cyclases-The 3Ј-tetraphosphates, 2Јd-3Ј-A 4 P and 2Ј,5Ј-dd-3Ј-A 4 P, exhibited notably more potent inhibition of adenylyl cyclase than did the respective 3Јtriphosphates ( 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 2Ј-deoxy derivative more than for the 2Ј,5Ј-dideoxy derivative, whereas in both cases, the increase in potency was not quite an order of magnitude. With IC 50 values of 32 and 106 nM (Table I), respectively, for 2Ј,5Ј-dd-3Ј-ATP and 2Ј-d-3Ј-ATP, the 3Ј-triphosphates exhibited potencies with this rat brain extract similar to those previously reported (6). The 3Ј-tetraphosphates exhibited IC 50 values of ϳ10.5 nM for 2Ј-d-3Ј-A 4 P and ϳ7.4 nM for 2Ј,5Ј-dd-3Ј-A 4 P 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 mol/(min⅐mg of protein) for the purified type I adenylyl cyclase (14,21). This is not a large excess of inhibitor relative to SCHEME 1. Reagents and conditions. i, tert-BuMgCl, tetrahydrofuran, room temperature; ii, phosphorylating agent, room temperature; iii, zinc, (Bu 3 N) 3 H 5 P 3 O 10 , pyridine, room temperature, followed by AcOH 80%, room temperature, 30 min for 2. enzyme, and slightly lower IC 50 values might be observed with lower enzyme concentrations.
Inhibition was rapid in onset for inhibition by 2Ј,5Ј-dd-3Ј-ATP or by either of two concentrations of 2Ј,5Ј-dd-3Ј-A 4 P (Fig.  2). As expected from their IC 50 values, inhibition by 10 nM 2Ј,5Ј-dd-3Ј-A 4 P elicited an inhibited rate similar to that achieved with 50 nM 2Ј,5Ј-dd-3Ј-ATP; 100 nM 2Ј,5Ј-dd-3Ј-ATP almost completely suppressed activity. The lack of lag phase argues for a rapid equilibrium mechanism of interaction of adenosine 3Ј-polyphosphates with the enzyme.
In our previously reported kinetic determinations for inhibition conducted either with adenine nucleosides or with adenine nucleoside 3Ј-phosphates, inhibition typically conformed to a linear noncompetitive mechanism (5,6,12). Although some exceptions to this observation have been noted (7,12,(22)(23)(24), this behavior has become a characteristic of P-site inhibition. Both 2Ј-d-3Ј-A 4 P and 2Ј,5Ј-dd-3Ј-A 4 P (Fig. 3) elicited a nonlinear noncompetitive inhibition of the detergent-extracted adenylyl cyclase from rat brain, the preparation we have used for many comparisons of P-site ligands (cf. Refs. 2, 4 -6, 10, 12, and 13). By comparison, a linear noncompetitive inhibition was observed with either 3Ј-tetraphosphate with the enzyme puri-fied from bovine brain, shown in Fig. 4 for 2Ј,5Ј-dd-3Ј-A 4 P. The noncompetitive character of this inhibition is perhaps more clearly shown with a Hofstee plot (V versus V/S) (Fig. 5). For this plot, a series of parallel lines is expected for straightforward noncompetitive inhibition, whereas intersecting lines would be consistent with competitive inhibition. The reason for the apparent discrepancy in kinetic behavior between adenylyl cyclases in the rat brain and purified bovine brain preparations is not known. It does not lie in nonlinear behavior of the reaction per se (cf. Fig. 2) but likely lies in the complex character of the crude detergent extract or differences in the conformations of the enzymes being tested. Nor was the nonlinear behavior because of an effect of breakdown products of the 3Ј-tetraphosphates. No significant hydrolysis of either 2Ј-d-3Ј-A 4 P or 2Ј,5Ј-dd-3Ј-A 4 P was noted during the course of typical 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/ (min⅐mg of protein). Protein was 10.8 g/100 l of reaction volume.  adenylyl cyclase incubations (Table II). This lack of hydrolysis, whether enzymatic or nonenzymatic, is consistent with the established stability of both 2Ј-d-3Ј-ATP and 2Ј,5Ј-dd-3Ј-ATP that we reported earlier (6).
Effects on Other Enzymes-Because few adenine nucleotides interact solely with a single protein, the possibility was considered that adenosine 3Ј-polyphosphates might affect enzymes other than adenylyl cyclase. As an initial investigation in this direction, we tested effects on Na ϩ /K ϩ -ATPase and on two enzymes participating in the cAMP signaling cascade, cAMP phosphodiesterase and cAMP-dependent protein kinase.
Under conditions with which strophanthadin exhibited an IC 50 of ϳ3 M for inhibition of Na ϩ /K ϩ -ATPase, neither 2Ј-d-3Ј-A 4 P at concentrations ranging from 10 nM to 3.3 M nor 2Ј,5Ј-dd-3Ј-A 4 P at concentrations from 10 nM to 11 M exhibited any effect whatsoever on enzyme activity.
Ca 2ϩ /calmodulin-sensitive cAMP phosphodiesterase was tested in the absence or presence of 0.7 M calmodulin, which elicited Ͼ3-fold activation of this enzyme preparation. Neither 2Ј,5Ј-dd-3Ј-A 4 P at concentrations from 0.1 to 10 M nor 2Ј,5Ј-dd-3Ј-ATP at concentrations from 3 nM to 10 M had any effect on phosphodiesterase activity, without or with Ca 2ϩ /calmodulin. Although Flockhart et al. (25) evaluated effects of numerous nucleosides and nucleotides on both cAMP-and cGMP-dependent protein kinases, none was a nucleoside 3Ј-phosphate. The rabbit muscle cAMP-dependent protein kinase we tested was unaffected by either 2Ј,5Ј-dd-3Ј-A 4 P or 2Ј,5Ј-dd-3Ј-ATP, at concentrations from 10 nM to 10 M, whether in the absence or presence of 1 M cAMP, which elicited Ͼ9-fold activation with this enzyme preparation (Fig. 6). Even though these experiments were conducted with a concentration of substrate tenfold greater than that of the 3Ј-nucleotides, it was clear that these 3Ј-nucleotides had no effect on either catalytic (5ЈATP) or regulatory (cAMP) domains of this protein kinase. DISCUSSION 2Ј-Deoxy-and 2Ј,5Ј-dideoxyadenosine 3Ј-tetraphosphates represent important additions to the family of 3Ј-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 2Ј-d-3Ј-A 4 P (IC 50 ϳ10.5 nM) or 2Ј,5Ј-dd-3Ј-A 4 P (IC 50 ϳ7.4 nM) was not because of the formation of inorganic polyphosphates, whether by enzymatic or nonenzymatic means. Significant hydrolysis of the 3Ј-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 3Ј-triphosphates (6) and implies a lack of significant 3Ј-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 (IC 50 ϳ55 M) than were the title compounds, and inhibition was competitive with respect to substrate (6). PP i was less potent (IC 50 ϳ1.4 mM) and inhibition was mixed (12). These observations argue convincingly that it is 2Ј-d-3Ј-A 4 P and 2Ј,5Ј-dd-3Ј-A 4 P per se that cause inhibition in the present study.
The striking inhibitory potency of 2Ј-d-3Ј-A 4 P and 2Ј,5Ј-dd-3Ј-A 4 P 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 FIG. 4. Double reciprocal plot for inhibition of adenylyl cyclase purified from bovine brain by 2,5-dd-3-A 4 P. Enzyme was prepared and assayed as described under "Experimental Procedures." Units for velocity are pmol of cAMP formed/(min⅐100 l of assay volume). Concentrations of 2Ј,5Ј-dd-3Ј-A 4 P were as indicated.
FIG. 5. Eadie-Hofstee plot for inhibition of adenylyl cyclase purified from bovine brain by 2,5-dd-3-A 4 P. Data are from the experiment represented in Fig. 4. Units for velocity are pmol of cAMP formed/(min⅐100 l of assay volume). Concentrations of 2Ј,5Ј-dd-3Ј-A4P were as indicated.

TABLE II
Stability of 2Ј-d-3Ј-A 4 P and 2Ј,5Ј-dd-3Ј-A 4 P Either 2Ј-d-3Ј-A 4 P (100 M) or 2Ј,5Ј-dd-3Ј-A 4 P (100 M) was incubated with a preparation of adenylyl cyclase extracted from rat brain (108 g of total protein per ml) at 30°C for the indicated times. The medium contained 5 mM MnCl 2 , 100 M 5Ј-ATP, 1 mM isobutylmethylxanthine, 1 mM dithiothreitol, 100 M forskolin, 2 mM creatine phosphate, 100 g of creatine kinase per ml, 1 mg of bovine serum albumin per ml, 0.1% (w/v) Lubrol-PX, and 50 mM triethanolamine hydrochloride, pH 7.5. Following incubations, samples were placed on ice. Nucleotides were quantified after purification on anion exchange HPLC as described under "Experimental Procedures." Values are averages from duplicate determinations. at the 3Ј-position, which for some ligand-protein interactions has been as high as ϳ10 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 (C 1 and C 2 ) 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)(28)(29)(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 C 1 and C 2 domains are consistent with 2Ј-d-3Ј-AMP binding in what appears to be a catalytic site (29,30). P-site ligands such as 2Ј,5Ј-dd-Ado or 2Ј-d-3Ј-AMP would bind at the adenosine-3Ј:5Ј-cyclic monophosphate leaving site in the presence of PP i (12,(31)(32)(33)(34), the other product of the reaction. Structures with either bound substrate or bound nucleoside 3Ј-polyphosphate have not been solved (29,30). 3 Presumably, though, 2Ј,5Ј-dd-3Ј-ATP would occupy both adenine and PP i binding domains and 2Ј,5Ј-dd-3Ј-A 4 P would take advantage of a locus within the cleft capable of accepting the fourth phosphate group. Whether 3Ј-mono-or 3Ј-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-5ЈATP. Inhibition by adenosine 3Ј-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 2Ј-d-3Ј-AMP and 3Ј-AMP occur naturally, their levels may be chronically regulated (39), and 3Ј-polyphosphorylation of nucleoside 3Јmonophosphates is known to occur in mammalian systems (40). There are, however, very few reports related to nucleoside-3Ј-polyphosphates in animal cells. If nucleoside 3Ј-tri-or 3Ј-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 5Ј-tetraphosphate occurs naturally (e.g. Refs. 35 and 36), and there is considerable literature on diadenosine tetraphosphates (A(5Ј)p4(5Ј)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 3Ј 3 5Јdinucleotides do inhibit the enzyme (cf.  (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 2Ј,5Ј-dd-3Ј-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 3Ј-polyphosphates will affect proteins other than adenylyl cyclases. Neither 2Ј,5Ј-dd-3Ј-ATP nor 2Ј,5Ј-dd-3Ј-A 4 P 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 2Ј-deoxynucleoside 3Ј-triphosphates, and even 3Ј-tetraphosphates, via the triphosphate domain (41). The affinity for nucleotide (2Ј-d-3Ј-ATP: K D ϳ38 -68 M (41)), though, is almost 3 orders of magnitude less than that exhibited by adenylyl cyclase, implying at once that the 3Ј-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 2 Although we reported kinetic evidence consistent with this mechanism, a characteristic uncompetitive inhibition for inhibition in the presence of Gs␣ (12) also supported by studies of Dessauer and Gilman (24) with recombinant soluble enzyme, we observed no effect of inorganic pyrophosphate on the affinity of the enzyme for the P-site ligand 2Ј-d-3Ј-AMP (12). 3 These solved crystal structures also do not show the two metalbinding sites known to participate in catalysis with native enzyme (12,15). 4 The nomenclature for adenosine 3Ј,5Ј-bis-polyphosphates implies 5Ј-substitution to the left of the base and 3Ј-substitution to the right of the base, e.g. ppp(5Ј)A(3Ј)ppp, pppApp, ppApp, or alternatively p n Ap n or p n dAp n . By this norm, 3Ј 3 5Ј-linked dinucleotide polyphosphates are correctly denoted as Ap nϾ1 A or dAp nϾ1 dA. Unfortunately, this latter abbreviation (e.g. AppppA or Ap 4 A) is used inappropriately also for 5Ј 3 5Ј-linked diadenosine polyphosphates by many authors and can lead to considerable confusion, causing some data base searching protocols to cite incorrect references. monophosphate and inorganic pyrophosphate as products would be potential targets for additional interactions with nucleoside 3Ј-polyphosphates.