Báo cáo khoa học: Methylene analogues of adenosine 5¢-tetraphosphate Their chemical synthesis and recognition by human and plant mononucleoside tetraphosphatases and dinucleoside tetraphosphatases

Methylene analogues of adenosine 5¢-tetraphosphate Their chemical synthesis and recognition by human and plant mononucleoside tetraphosphatases and dinucleoside tetraphosphatases Andrzej Guranowski1, El_zbieta Starzyn´ ska1, Małgorzata Pietrowska-Borek1, Jacek Jemielity2, Joanna Kowalska2, Edward Darzynkiewicz2, Mark J. Thompson3 and G. Michael Blackburn3

1 Department of Biochemistry and Biotechnology, Agricultural University, Poznan´ , Poland 2 Department of Biophysics, Institute of Experimental Physics, Warsaw University, Poland 3 Department of Chemistry, Krebs Institute, University of Sheffield, UK

Keywords adenosine 5¢-tetraphosphate; p4A; methylene analogues of p4A; nucleoside tetraphosphatase; dinucleoside tetraphosphatase

c,d-methylene-adenosine

5¢-tetraphosphate

5¢-tetraphosphate

(pCH2ppCH2pA),

Correspondence A. Guranowski, Katedra Biochemii i Biotechnologii, Akademia Rolnicza ul. Wołyn´ ska 35, 60–637 Poznan´ , Poland Fax: +48 61 8487146 Tel: +48 61 8487201 E-mail: guranow@au.poznan.pl Website: http://www.au.poznan.pl

Note This study is dedicated to Professor Wojciech J. Stec on the occasion of his 65th birthday.

(Received 9 November 2005, revised 15 December 2005, accepted 21 December 2005)

doi:10.1111/j.1742-4658.2006.05115.x

Adenosine 5¢-polyphosphates have been identified in vitro, as products of certain enzymatic reactions, and in vivo. Although the biological role of these compounds is not known, there exist highly specific hydrolases that degrade nucleoside 5¢-polyphosphates into the corresponding nucleoside 5¢-triphos- phates. One approach to understanding the mechanism and function of these enzymes is through the use of specifically designed phosphonate analogues. We synthesized novel nucleotides: a,b-methylene-adenosine 5¢-tetraphosphate (pppCH2pA), b,c-methylene-adenosine 5¢-tetraphosphate (pCH2pppA), (ppCH2ppA), ab, ab,cd-bismethylene-adenosine bc-bismethylene-adenosine 5¢-tetraphosphate (ppCH2pCH2pA) and bc, cd-bis(dichloro)methylene-adenosine 5¢-tetraphosphate (pCCl2pCCl2ppA), and tested them as potential substrates and ⁄ or inhibitors of three specific nu- cleoside tetraphosphatases. In addition, we employed these p4A analogues with two asymmetrically and one symmetrically acting dinucleoside tetra- phosphatases. Of the six analogues, only pppCH2pA is a substrate of the two nucleoside tetraphosphatases (EC 3.6.1.14), from yellow lupin seeds and human placenta, and also of the yeast exopolyphosphatase (EC 3.6.1.11). Surprisingly, none of the six analogues inhibited these p4A-hydrolysing enzymes. By contrast, the analogues strongly inhibit the (asymmetrical) dinu- cleoside tetraphosphatases (EC 3.6.1.17) from human and the narrow-leafed lupin. ppCH2ppA and pCH2pppA, inhibited the human enzyme with Ki val- ues of 1.6 and 2.3 nm, respectively, and the lupin enzyme with Ki values of 30 and 34 nm, respectively. They are thereby identified as being the strongest inhibitors ever reported for the (asymmetrical) dinucleoside tetraphospha- tases. The three analogues having two halo ⁄ methylene bridges are much less potent inhibitors for these enzymes. These novel nucleotides should prove valuable tools for further studies on the cellular functions of mono- and di- nucleoside polyphosphates and on the enzymes involved in their metabolism.

Abbreviations Ap3A, diadenosine 5¢,5¢¢¢-P1,P3-triphosphate; Ap4A, diadenosine 5¢,5¢¢¢-P1,P4-tetraphosphate; NpnN¢, dinucleoside 5¢,5¢¢¢-P1,P n-polyphosphate; p4A, adenosine 5¢-tetraphosphate; p5A, adenosine 5¢-pentaphosphate; pCCl2pCCl2ppA, bc,cd-bis(dichloro)methylene-adenosine 5¢- tetraphosphate; pCH2ppCH2pA, ab,cd-bismethylene-adenosine 5¢-tetraphosphate; pCH2pppA, c,d-methylene-adenosine 5¢-tetraphosphate; pnN, nucleoside 5¢-polyphosphate; ppCH2pCH2pA, ab,bc-bismethylene-adenosine 5¢-tetraphosphate; pppCH2pA, a,b-methylene-adenosine 5¢-tetraphosphate; pppCH2ppA, b,c-methylene-adenosine 5¢-tetraphosphate.

FEBS Journal 273 (2006) 829–838 ª 2006 The Authors Journal compilation ª 2006 FEBS

829

A. Guranowski et al.

Methylene analogues of adenosine 5¢-tetraphosphate

(pnNs, where n ¼ 4),

are diadenosine

Fig. 1. Structures of p4A analogues.

In addition to the canonical nucleoside mono-, di-, and triphosphates, cells contain various minor nucleo- tides. Among these are the nucleoside 5¢-polyphos- such as adenosine phates 5¢-tetraphosphate (p4A or ppppA) [1–5] and adenosine 5¢-pentaphosphate (p5A or pppppA) [2], and the dinu- cleoside 5¢,5¢¢¢-P1,Pn-polyphosphates (NpnN¢s, where N and N¢ are 5¢-O-nucleosides and n represents the num- ber of phosphate residues in the polyphosphate chain that links N and N¢ through their 5¢-positions). Typical 5¢,5¢¢¢-P1,P3-triphosphate examples (Ap3A) and diadenosine 5¢,5¢¢¢-P1,P4-tetraphosphate (Ap4A) [6–12]. The biological roles of these NpnN¢s are partially understood. In particular, ApnA has been implicated in various intracellular processes [13,14] and also in extracellular signalling [15,16]. By contrast, the role of pnNs is inadequately recognized. Almost 20 years ago, the accumulation of p4A and p5A in yeast was correlated with sporulation [2] and only recently, p4A was identified in human myocardial tissue and shown to modulate coronary vascular tone [4]. This compound has also been found as a constituent of the nucleotide pool present in the aqueous humour of New Zealand rabbits where it is proposed to act as a physiological regulator of intraocular pressure in the normotensive rabbit eye [5].

produced already and tested with numerous enzymes [28,29], p4A analogues have been synthesized only recently. ab,bc-bismethylene-p4A and bc,cd-bis(dichlo- ro)methylene-p4A were tested as agonists or antago- nists of the P2X2 ⁄ 3 receptor [30] and a short report has appeared on the synthesis of pCH2pppA, pCH2pppG and pCH2pppm7G [31].

[18–22], or

uridylyltransferase

cerevisiae)

Enzymatic reactions that can lead to the accumula- tion of p4A and other p4Ns in cells fall into three cat- egories. The first comprises enzymes that catalyse transfer of a phosphate residue from a phosphate donor to ATP (e.g. the muscle adenylate kinase) [17]. The second category of enzymes includes those able to transfer adenylate or nucleotide residue onto tri- polyphosphates. The pA residue comes either from a mixed acyl–pA anhydride, as in the case of some li- gases and firefly luciferase from an enzyme–pA complex, as in the case of the DNA- and RNA-ligases [23,24]. Recently, the yeast UTP ⁄ glucose- 1-phosphate (EC 2.7.7.9) was shown to function according to the same pattern and to synthesize p4U by transferring the uridylyl moiety from UDP-glucose onto tripolyphosphate [25]. The third category includes several enzymes that degrade Ap5A or Ap6A yielding p4A as one of the reaction products [26]. Degradation of p4A can be controlled by various nonspecific and specific pnN-degrading enzymes [26,27].

that

[36]

Here, we describe details on the synthesis of and the results of enzymatic studies on a series of novel p4A analogues that have a single methylene bridge substitu- ting one of the three bridging oxygens in the tetraphos- phate chain, or have two methylene bridges, or contain two dichloromethylene groups. The structures of these compounds are shown in Fig. 1. We prepared these nucleotides for evaluation first, as potential substrates and ⁄ or inhibitors of three enzymes that hydrolyse the pyrophosphate bond between the c- and d-phosphates of p4A and second, as inhibitors of two types of for which p4A itself acts as a Ap4A hydrolase, strong inhibitor. The p4A-hydrolysing enzymes are the two highly specific mononucleoside tetraphosphatases (EC 3.6.1.14), from yellow lupin (Lupinus luteus) seeds [32] and from human placenta [33], and the yeast (Saccharomyces exopolyphosphatase that can hydrolyse p4A to ATP and (EC 3.6.1.11) phosphate [34]. The Ap4A hydrolases investigated are the two asymmetrically acting ones (EC 3.6.1.17), from human [35] and from narrow-leaved lupin (Lupinus angustifolius) split Ap4A into ATP and AMP, and the Co2+-dependent symmetrically acting

Among studies that shed light on the mechanism of the action of these phosphohydrolases are investiga- tions of the interaction of a given enzyme with its sub- strate analogues. Whereas many analogues of Ap3A and Ap4A, modified in the polyphosphate chain, in adenine(s) or in the ribose moiety(-ies), have been

FEBS Journal 273 (2006) 829–838 ª 2006 The Authors Journal compilation ª 2006 FEBS

830

A. Guranowski et al.

Methylene analogues of adenosine 5¢-tetraphosphate

dinucleoside tetraphosphatase (EC 3.6.1.41) that con- verts Ap4A into two ADPs [37].

characteristics readily identify the nature and location of the oxygen and methylene groups bridging the four phosphorus atoms (see Supplementary material).

Results and Discussion

Comments on the synthesis of p4A analogues

Recognition of p4A analogues as substrates by the p4A hydrolysing enzymes

reaction mixture

reverse-phase

The preparation of intermediate ADP and ATP ana- logues followed standard methods. Their conversion into p4A analogues called for condensation with phos- phate (for ATP analogues), or with pyrophosphate or a methylenebisphosphonate (for ADP and ADP ana- logues). Although a variety of options were explored initially, the use of phosphoroimidazolates [31] proved to be the most reliable method and gave satisfactory yields without detailed optimization (Fig. 2). The prod- ucts were first, purified by ion-exchange chromato- graphy on DEAE-Sephadex 25A, which separates nucleotides according to net charge at pH 7.9, and readily resolved the desired products as tetra-to-penta anions from the corresponding reactants (di-to-tetra chromatography anions). Additional provided the product p4A analogues in high purity. The MS and 1H NMR spectra of these nucleotides are unexceptional. The 31P NMR spectra, however, pro- vide examples of ABCD spectra, whose chemical shift

Each compound was checked as a potential substrate for two highly specific nucleoside tetraphosphatases (EC 3.6.1.14), from yellow lupin seeds and from human, and for the soluble exopolyphosphatase (EC 3.6.1.11) that has an inherent capacity to hydrolyse the distal pyrophosphate bond in p4Ns thus acting as a nucleoside tetraphosphatase. A typical (see Experimental procedures) contained 1 mm analogue and excess of enzyme, i.e. an amount that, under the same conditions, completely hydrolysed 1 mm p4A to ATP and Pi in < 15 min. Incubation was for up to 16 h and the progress of potential hydrolysis was analysed by TLC System A. Of six p4A analogues, only pppCH2pA was susceptible to hydrolysis and the relative velocities the reactions were estimated only for the pair of p4A ⁄ pppCH2pA. Figure 3 shows typical elution pat- terns of the substrate ⁄ product pairs on the reverse-phase HPLC column. Satisfactory separation of p4A from ATP was obtained by isocratic elution with potassium

Fig. 2. Chemical synthesis of pppCH2pA (A), ppCH2ppA (B) and pCH2pppA (C). ‘A’ represents adenosine, DMF dimethylformamide, PPh3 tri- phenylphosphine, TEA triethylammonium, and TEAB triethylammonium bicarbonate.

FEBS Journal 273 (2006) 829–838 ª 2006 The Authors Journal compilation ª 2006 FEBS

831

A. Guranowski et al.

Methylene analogues of adenosine 5¢-tetraphosphate

Fig. 3. Time course of p4A (A) and pppCH2pA (B) hydrolysis catalysed by yeast exopolyphosphatase. Reaction mixtures (0.25 mL) were pre- pared and incubated as described in the Experimental procedures. Aliquots (0.05 mL) were withdrawn after the indicated time of incubation, the reaction was stopped by heating (96 (cid:1)C, 3 min) and 2-lL sample subjected to HPLC on the Supelcosil LC-18-T reverse-phase column (25 cm · 4.6 mm). Satisfactory separation of ATP from p4A (A) was obtained by eluting the column with an isocratic system using 0.1 M KH2PO4 buffer (pH 6.0), and separation of ppCH2pA from pppCH2pA (B) when the eluting system was a linear gradient (0–100%) of buffer A–buffer B, applied within 20 min at the flow rate 1.3 mLÆmin)1 [buffer A was 0.1 M KH2PO4 + 0.008 M (CH3CH2CH2CH2)4N+HSO4 –, pH 6.0 and buffer B was buffer A: ⁄ methanol (70 : 30 v ⁄ v)].

Table 1. Comparison of the hydrolysis of ppppA and pppCH2pA by specific p4A-hydrolysing enzymes The velocities of conversion of the nucleoside tetraphosphates (0.5 mM) to corresponding nucleo- side triphosphates were calculated based on the HPLC profiles the (exemplified in Fig. 3). For each enzyme the velocity of pppCH2pA hydrolysis was related to that of the ppppA degradation.

Enzyme

Relative velocity of the pppCH2pA hydrolysis

in pppCH2pA. Thus

0.8

ppppA hydrolase from human placenta ppppA hydrolase from yellow lupin seeds Exopolyphosphatase from the yeast

45 42

not tolerate methylene modification of their substrates at the scissile P–O–P bond; (b) none of the enzymes residue hydrolysed the from terminal phosphate in which the ppCH2ppA, or from ppCH2pCH2pA, P–O–P bond between c and d phosphate remains unchanged; (c) the human hydrolase is sensitive to the –CH2– insert even in the most distant position from the i.e. between the a- and b-phosphorus reaction site, the p4N hydrolysing atoms enzymes are more stringent with respect to recognition of their substrates than the (asymmetrical) Ap4A hydro- lases, which cleave the Pa–O–Pb bridge not only in Ap4A [27], but also in AppCH2ppA, AppCF2ppA, and AppCCl2ppA [28].

There are obvious

phosphate buffer (Fig. 3A), and of pppCH2pA from ppCH2pA by the use of a more complex solvent system and methanol gradient (Fig. 3B). Integrated peaks of the products were used for calculating the reaction velo- the yellow lupin p4A cities. As shown in Table 1, hydrolase and the yeast exopolyphosphatase hydrolysed pppCH2pA slightly more than twofold slower than p4A, and the human p4A hydrolase 125-fold slower. This result shows that (a) the p4N hydrolysing enzymes do

reasons why analogues with proximate methylene bridges should resist cleavage. For pCH2pppA, removal of the terminal phosphate is frustrated by the (largely a dissociative process) stability of the Pc–C–Pd bridge. For ppCH2ppA, its stability can be attributed, in at least part, to the impaired leaving group ability of b,c-methyleneATP. Neither of these explanations accounts for the much reduced activity of pppCH2pA for the human p4A

FEBS Journal 273 (2006) 829–838 ª 2006 The Authors Journal compilation ª 2006 FEBS

832

A. Guranowski et al.

Methylene analogues of adenosine 5¢-tetraphosphate

Table 2. Analogues of p4A as inhibitors of (asymmetrical) Ap4A hydrolases. The Km values for Ap4A estimated for the human and lupin enzyme were 2 lm (this study) and 1 lm [35], respectively. The Ki values are means of three independent estimations; stand- ard errors did not exceed 20%. For details of assays see Experi- mental procedures.

Human

Narrow-leafed lupin (Lupinus angustifolius)

hydrolase. It does not seem likely that such an iso- steric and isopolar analogue [47] of p4A could have a conformational bias that impairs access to the cata- lytic site of the enzyme by over 100-fold because it has the strongest affinity for the symmetrically clea- ving bacterial Ap4A hydrolase. This brings into focus the possibility of direct recognition of the Pa–O–Pb bridge by the protein, a possibility that might be explored by the synthesis and use of the imino ana- logue, pppNHpA.

Do the analogues inhibit the p4A hydrolysing enzymes?

0.05 0.18 0.0016 0.0023 1.3 0.25

0.40 0.36 0.030 0.034 0.07 0.62

53

40 10

0.16

16 2

n.d. n.d. n.d. n.d.

ppppA pppCH2pA ppCH2ppA pCH2pppA ppCH2pCH2pA pCH2ppCH2pA pCCl2pCCl2ppA ppppRib ppCH2ppRib pppA pCH2ppA

recently solved structures

All three p4A hydrolysing enzymes were tested with each of the six p4A analogues to see whether they inhibit normal hydrolysis of p4A. None of the ana- logues used at concentrations up to 0.5 mm retarded the conversion of p4A (1 mm) into ATP. This unex- pected result suggests that the active sites of these three enzymes recognize and bind only nucleotides with tetraphosphate chains having intact P–O–P brid- ges, even though all of the analogues are formally it isopolar and isosteric to p4A [47]. In this regard, is noteworthy that for dUMPNPP in complex with dUTP hydrolases from Escherichia coli [48] and Mycobacterium tuberculosis [49] show a key hydrogen bond from a conserved serine hydroxyl to the imino bridge in the catalyti- cally active complex, whereas the complex between the methylene analogue dUMPCPP and the human enzyme, which cannot form such a hydrogen bond, is folded into an inactive conformation [J A Tainer, personal communication].

The methylene analogues of p4A as inhibitors of the (asymmetrical) Ap4A hydrolases

and pCH2pppA appear to be the strongest inhibitors of both the human and plant enzymes. The Ki values esti- mated for the human enzyme, 1.6 and 2.3 nm, respect- ively, were over 30- and 20-fold lower than the Ki estimated for the same enzyme for p4A (50 nm). More- over, these values are five and three times smaller than the lowest Ki estimated yet reported (7.5 nm) for the reaction of Ap4A hydrolysis catalysed by the firefly enzyme [52]. Significantly, the analogue, pppCH2pA, with its methylene bridge in the position most distant from the reaction site, is (cid:1) 100-fold less potent an inhibitor than ppCH2ppA. Two analogues having two methylene bridges are generally poorer inhibitors than those that possess a single methylene group. In every cases, however, the Ki values were below the Km values for the Ap4A substrate (1 lm for the lupin and 2 lm for the human hydrolase). Finally, the analogue with the bulkiest modification, the dichloromethylene groups, was a rather poor inhibitor with Ki values exceeding the Km values for Ap4A by some 20–50-fold. Both p4A and its two strongly binding methylene ana- logues inhibited the lupin enzyme 8–20-fold less effect- ively than they inhibit the human enzyme. The differential recognition of the ligands by these two hydrolases may relate to structural differences within the substrate-binding sites seen in the recently estab- lished three-dimensional structures of the lupin Ap4A hydrolase [55] and the human enzyme [56]. The the human enzyme by p4A stronger inhibition of and its analogues may be explained by the more restric- ted space in the substrate-binding cleft in the lupin enzyme.

Adenosine tetraphosphate itself has been known for a long time as an effective competitive inhibitor of the (asymmetrical) Ap4A hydrolases. Examples of the reported inhibition constants are 48 nm for the rat liver enzyme [50], 30 nm for the enzyme from Ehrlich ascites tumour cells [51] and 7.5 nm, the lowest value reported to date, for the enzyme from firefly lanterns [52]. Owing to such low Ki values, this nucleotide has been used for the (asymmetrical) Ap4A hydrolases the elution of adsorbed to dye–ligand affinity columns as homogen- eous proteins [53,54]. We tested all six methylene and chloromethylene p4A analogues as potential inhibitors of two (asymmetrical) Ap4A hydrolases, from human and from narrow-leafed lupin, and the results are sum- marized in Table 2. Of all the analogues, ppCH2ppA

FEBS Journal 273 (2006) 829–838 ª 2006 The Authors Journal compilation ª 2006 FEBS

833

A. Guranowski et al.

Methylene analogues of adenosine 5¢-tetraphosphate

into account their inhibition of the asymmetrically act- ing Ap4A hydrolases. Adenosine tetraphosphate itself inhibited the enzyme with Ki threefold lower than the Km for Ap4A (27 lm). The lowest Ki value was estima- ted for pppCH2pA (6.7 lm) and the highest values were for ppCH2pCH2pA and pCH2ppCH2pA, 20 and 34 lm, respectively.

Conclusion

The data presented here show the potential usefulness of certain p4A analogues for the further study of the metabolism of mononucleoside polyphosphates and dinucleoside polyphosphates as well as of the function- ing of different purine ⁄ nucleotide receptors. In partic- ular, they have shown a remarkable selectivity in their behaviour as inhibitors for enzymes having super- ficially related functions as nucleoside polyphosphate hydrolases as well as showing nanomolar activity against selected enzymes.

The newly discovered ATP N-glycosidase [38] allowed us to generate the depurinated derivatives of p4A and of the best inhibitor analogue, ppCH2ppA, and evaluate the two polyphosphorylated riboses obtained as inhibi- tors of the human (asymmetrical) Ap4A hydrolase. It emerged that p4Ribose is 200 times weaker an inhibitor than p4A, whereas ppCH2ppRibose is 100 times weaker than ppCH2ppA. Finally, therefore, we compared the inhibition of the human recombinant Ap4A hydrolase by p4A and ppCH2ppA with that by ATP (p3A) and pCH2ppA, both compounds truncated by one phos- phate (and one negative charge) relative to the nucleo- side tetraphosphates. Both ATP and its b,c-methylene analogue were definitively weaker inhibitors than their d-phosphate homologues. Altogether, it is evident that both the adenine ring and the length of the polyphos- phate chain contribute to the strength of binding of the mononucleoside polyphosphates by the (asymmetrical) Ap4A hydrolases, whereas a single methylene bridge, preferably at or adjacent to the P–O–P reaction site, the binding. Because ppCH2ppA and potentiates pCH2pppA are the strongest inhibitors of the asymmet- rically acting Ap4A hydrolases ever reported and they are not degraded, in marked contrast to p4A which is both an inhibitor and a slow substrate for these enzymes [57,58], they clearly have excellent potential to serve as ‘true inhibitors’ and be valuable tools in biochemical and physiological studies, e.g. on nucleotide receptors.

The methylene analogues of p4A as inhibitors of the (symmetrical) Ap4A hydrolase from Escherichia coli

This new group of nucleotide analogues complements a different set of synthetic nucleotides, the adenosine- phosphorothioylated and adenosine-phosphorylated polyols, which has recently been proved to inhibit sym- metrically acting bacterial Ap4A hydrolases particularly strongly, with Ki values as low as 40 nm [59]. These new, nonhydrolysable p4A nucleotide analogues are promis- ing tools for those who would like specifically to inhibit the asymmetrically acting Ap4A hydrolases. In partic- ular, they should help in structural studies of these enzymes [55,56,60]. The apparent lack of inhibition of the p4A hydrolysing enzymes by the methylene and chloromethylene analogues of p4A further challenges chemists to create other types of p4A analogues that may need to reach beyond the isopolar–isosteric princi- ples that have governed their design for 25 years [47].

Experimental procedures

This Co2+-dependent enzyme was shown to hydrolyse p4A slowly, within a range of substrates from which it always liberates ADP as one of the reaction products [37]. The p4A analogues studied here are not substrates for this enzyme. However, as shown in Table 3, all act as inhibitors, albeit relatively moderate ones taking

Enzymes

phosphohydrolase 5¢-tetraphosphate

Table 3. Analogues of p4A as inhibitors of (symmetrical) Ap4A hydrolase from Escherichia coli. The Km value for Ap4A estimated for the bacterial enzyme was 25 lM [36]. Ki values are means of three independent determinations; standard errors do not exceed 15%. For details of assays see Experimental procedures.

Compound

Ki (lM)

10.5 6.7 16.2 21.8 20.0 34.0 8.3

ppppA pppCH2pA ppCH2ppA pCH2pppA ppCH2pCH2pA pCH2ppCH2pA pCCl2pCCl2ppA

(S. cerevisiae)

FEBS Journal 273 (2006) 829–838 ª 2006 The Authors Journal compilation ª 2006 FEBS

834

Adenosine was obtained from yellow lupin seeds [32] and the recombinant exopolyphosphatase from yeast [34] was kindly donated by Dr Sh. Liu (Stanford University, CA). Adenosine 5¢-tetraphosphate phosphohydrolase from human placenta [33] was partially purified according to the following procedure. The placenta extract was fractionated with ammonium sulfate and the protein precipitated between 30 and 50% of saturation was subjected to ion-exchange chro- matography on a DEAE-Sephacel column. The enzyme was eluted with a 0–0.5 m KCl gradient, concentrated and chro- matographed on a Sephadex G-100 column from which it

A. Guranowski et al.

Methylene analogues of adenosine 5¢-tetraphosphate

Structures of the six p4A analogues are presented in Fig. 1.

Other chemicals

Unlabelled mono- and dinucleotides were from Sigma (St. Louis, MO), and [3H]Ap4A (740 TBqÆmol)1) was purchased from Moravek, Biochemicals (Brea, CA).

2 mm nucleotides eluted as a protein with molecular mass around 84 kDa. This preparation was free of any competing activity and was used for the studies of the p4A analogues. The recombinant human (asymmetrical) Ap4A hydrolase was kindly donated by Professor A. G. McLennan (University of Liverpool, UK) and the enzyme from narrow-leafed lupin from Drs D. Maksel and K. Gayler (University of Melbourne, Australia). We also used an extract from the marine sponge Axinella polypoides that contained an ATP N-glycosidase [38]. This unusual hydrolase is able to depurinate p4A and ppCH2ppA, giving d-ribose 5-O-tetraphosphate and its corresponding b,c-methylene analogue, respectively. (The sponge extract was kindly donated by Dr T. Reintamm, Tallinn, Estonia.)

the reactions,

Chemical synthesis of p4A analogues

Analogues with one methylene bridge

ppppRibose and ppCH2ppRibose were obtained enzy- matically by incubating p4A and ppCH2ppA with the sponge ATP N-glycosidase. The progress of depurination in 50 mm Hepes ⁄ KOH buffer of (pH 8.0) was monitored by TLC (System A) in which the liberated adenine migrated with the solvent front. After completion of the glycosidase was heat inactivated (3 min at 96 (cid:1)C) and the mixtures used directly as a source of the depurinated compounds. The highest concentration of these compounds in the inhibi- tion assays with the (asymmetrical) Ap4A hydrolases was 0.05 mm. converted into the

Enzyme assays

coupled with pyrophosphate to Adenosine 5¢-methylenebisphosphonate was obtained by regioselective 5¢-phosphonylation of adenosine with methyl- enebis(dichlorophosphonate) using recent methodology [39]. The product was imidazolidate, ImpCH2pA, using imidazole with triphrenylphosphone ⁄ 2,2¢-dithio-dipyridine as the condensing agent, and this intermediate give pppCH2pA in 80% yield. Activation of

exopolyphosphatase yeast

Each methylene analogue of p4A was tested as a potential substrate for the three p4A-hydrolysing enzymes under the conditions established earlier as optimal for p4A hydrolysis. The reaction mixtures (0.05 mL final volume) contained 50 mm buffer, chloride of a divalent cation, 1 mm substrate (p4A or its analogue) and the investigated enzyme. For the yellow lupin p4A hydrolase the mixture contained Hepes ⁄ KOH buffer (pH 8.2) and 5 mm MgCl2, for the human enzyme Hepes ⁄ KOH (pH 7.0) and 1 mm CoCl2, and for the sodium acetate buffer (pH 4.7) and 1 mm CoCl2. Incubations were carried out at 30 (cid:1)C. The results were analysed either by TLC or HPLC (see below).

the b-phosphate group of ADP was achieved by conversion into imidazolidate, ImppA. The activated compound was reacted with a fourfold excess of the triethylammonium salt of methylenebisphosphonate in DMF to give pCH2pppA [31]. The rate of pyrophosphate bond formation was greatly accelerated when carried out in the presence of an eightfold excess of ZnCl2 [40]. Similarly, AMP was converted into adenosine 5¢-phosphoroimidazoli- date, ImpA, which was efficiently coupled with the triethyl- ammonium salt of methylenebisphosphonic acid. The resulting pCH2ppA was again activated with imidazole to give imidazolidate ImpCH2ppA, and this intermediate cou- pled with triethylammonium phosphate in a ZnCl2-mediated reaction to give ppCH2ppA in 25% yield. Schemes of these syntheses are shown in Fig. 2.

Reaction mixtures were separated using DEAE-Sepha- dex 25A (triethylammonium bicarbonate gradient, pH 7.9) and ⁄ or reverse-phase HPLC (C18 column, water ⁄ methanol gradient). HPLC analyses showed that all p4A analogues were at least 96% pure. Structures of all compounds syn- thesized were fully confirmed using 1H and 31P NMR spectroscopy and MS (see Supplementary material).

Analogues with two methylene or dichloromethylene bridges

the procedures that

FEBS Journal 273 (2006) 829–838 ª 2006 The Authors Journal compilation ª 2006 FEBS

835

Asymmetrically acting Ap4A hydrolases were assayed in a reaction mixture (0.05 mL total volume) containing 50 mm Hepes ⁄ KOH (pH 7.6), 0.02 mm dithiothreitol, 5 mm MgCl2, 0.05 mm [3H]Ap4A (300 000 c.p.m.), various con- centrations of p4A or its analogue and a rate-limiting quan- tity of enzyme ((cid:1) 0.3 mU). For assaying the symmetrically acting Ap4A hydrolase from E. coli, 5 mm MgCl2 was replaced with 0.1 mm CoCl2. Incubations were carried out at 30 (cid:1)C. To estimate reaction rates, 0.005 mL aliquots were spotted on to TLC plates (aluminium plates precoated with silica gel containing fluorescent indicator; Merck cat. no. 5554), usually after 6, 12, 18 and 24 min of incubation. Unlabelled standards of the product [ATP for (asymmetri- cal) Ap4A hydrolases and ADP for (symmetrical) Ap4A hydrolase] were applied at the origin, and plates were devel- oped for 90 min in dioxane ⁄ ammonia ⁄ water (6 : 1 : 4 v ⁄ v ⁄ v). Spots of the products, visualized under short-wave UV light, were excised, immersed in scintillation cocktail, Details of led to pCCl2pCCl2ppA, ppCH2pCH2pA and pCH2ppCH2pA are given in the Sup- plementary material [41–45].

A. Guranowski et al.

Methylene analogues of adenosine 5¢-tetraphosphate

cleoside Polyphosphates (McLennan AG, ed.), pp. 29– 61. CRC Press, Boca Raton, FL. 7 Ogilvie A & Jacob P (1983) Diadenosine 5¢,5¢¢¢-P1,P3-

triphosphate in eucaryotic cells: identification and quan- titation. Anal Biochem 134, 382–392. 8 Bochner BR, Lee PC, Wilson SW, Cutler CW & and the radioactivity measured. Ki values were calculated according to the method of Dixon and Webb [46] from the slopes of plots v ⁄ vi against [I] (where v and vi are velocities in the absence and presence of inhibitor, respectively, the inhibitor concentration), where slope ¼ is and [I] Km ⁄ Ki(1 ⁄ Km + S).

Chromatographic systems

Ames BN (1984) Ap4A and related nucleotides are synthesized as consequence of oxidation stress. Cell 37, 225–232.

9 Miller D & McLennan AG (1986) Changes in intra- cellular levels of Ap3A and Ap4A in cysts and larvae of Artemia do not correlate with changes in protein synth- esis after heat-shock. Nucleic Acids Res 14, 6031–6040. 10 Pa´ lfi Z, Sura´ nyi G & Borbe´ ly G (1991) Alterations in

the accumulation of adenylylated nucleotides in heavy- metal-ion-stressed and heat-stressed Synechococcus sp. Strain PCC 6301, cyanobacterium in light and dark. Biochem J 276, 487–491. Analyses of the hydrolysis of p4A or its analogues to their corresponding NTPs were performed on silica gel TLC plates developed in dioxane ⁄ ammonia ⁄ water (6 : 1 : 6 v ⁄ v ⁄ v) (System A). Inhibitory effects of the analogues exer- ted on the Ap4A hydrolysing enzymes were analysed by developing the same TLC plates in dioxane ⁄ ammonia ⁄ water mixed at the 6 : 1 : 4 ratio (System B). The velocities of p4A and pppCH2pA hydrolysis were estimated by the use of HPLC on the reverse-phase column (for details see legend to Fig. 3).

Acknowledgements

11 Brevet A, Plateau P, Best-Belpomme M & Blanquet S (1985) Variation of Ap4A and other dinucleoside poly- phosphates in stressed Drosophila cells. J Biol Chem 260, 15566–15570.

and

PBZ-KBN-059 ⁄ T09 ⁄ 10,

12 Coste H, Brevet A, Plateau P & Blanquet S (1987) Non- adenylated bis (5¢-nucleosidyl) tetraphosphates occur in Saccharomyces cerevisiae and in Escherichia coli and accumulate upon temperature shift or exposure to cad- mium. J Biol Chem 262, 12096–12103. 13 McLennan AG (2000) Dinucleoside polyphosphates –

Financial support from the State Committee for Scien- tific Research (KBN, Poland), within grants PBZ- KBN-059 ⁄ T09 ⁄ 04 is gratefully acknowledged. We thank the Wellcome Trust for generous financial support (to MJT) Grant no. 057599 ⁄ Z ⁄ 99.

friend or foe? Pharmacol Ther 87, 73–89.

References

14 McLennan AG, Barnes LD, Blackburn GM, Brenner Ch, Guranowski A, Miller AD, Rotlla´ n P, Soria B, Tanner JA & Sillero A (2001) Recent progress in the study of the intracellular function of diadenosine poly- phosphates. Drug Dev Res 52, 249–259. 1 Small GD & Cooper C (1966) Studies on the occurrence and biosynthesis of adenosine tetraphosphate. Biochem- istry 5, 26–33.

15 Hoyle CHV, Hilderman RH, Pintor JJ, Schlu¨ ter H & King BF (2001) Diadenosine polyphosphates as extra- cellular signal molecules. Drug Dev Res 52, 260–273. 16 Jankowski V, Tolle M, Vanholder R, Scho¨ nfelder G, 2 Jakubowski H (1986) Sporulation of the yeast Saccharo- myces cerevisiae is accompanied by synthesis of adeno- sine 5¢-tetraphosphate and adenosine 5¢-pentaphosphate. Proc Natl Acad Sci USA 83, 2378–2382.

van der Giet M, Henning L, Schlu¨ ter H, Paul M, Zidek W & Jankowski J (2005) Uridine adenosine tetrapho- sphate: a novel endothelium-derived vasoconstrictive factor. Nat Med 11, 223–227.

3 Gu¨ nther Sillero MA, Del Valle M, Zaera E, Michelena P, Garcı´ a AG & Sillero A (1994) Diadenosine 5¢,5¢¢¢- P1,P4-tetraphosphate (Ap4A), ATP and catecholamine content in bovine adrenal medulla, chromaffin granules and chromaffin cells. Biochimie 76, 404–409.

17 Kuprianov VV, Ferretti JA & Balaban RS (1986) Mus- cle adenylate kinase catalyzes adenosine 5¢-tetraphos- phate synthesis from ATP and ADP. Biochim Biophys Acta 869, 107–111.

4 Westhoff T, Jankowski H, Schmidt S, Luo J, Giebing G, Schlu¨ ter H, Tepel M, Zidek W & van der Giet M (2003) Identification and characterization of adenosine 5¢-tetraphosphate in human myocardial tissue. J Biol Chem 278, 17735–17740. 5 Pintor J, Pelaez T & Peral A (2004) Adenosine tetra-

18 Randerath K, Janeway CM, Stephenson ML & Zamec- nik PC (1966) Isolation and characterization of dinu- cleoside tetra- and tri- phosphates formed in the presence of lysyl-sRNA synthetase. Biochem Biophys Res Commun 24, 98–105.

FEBS Journal 273 (2006) 829–838 ª 2006 The Authors Journal compilation ª 2006 FEBS

836

phosphate, Ap4, a physiological regulator of intraocular pressure in normotensive rabbit eyes. J Pharmacol Exp Ther 308, 468–473. 6 Garrison PN & Barnes LD (1992) Determination of 19 Ortiz B, Sillero A & Gu¨ nther Sillero MA (1993) Specific synthesis of adenosine (5¢)tetraphospho (5¢)nucleoside and adenosine (5¢)oligophospho (5¢)adenosine (n>4) dinucleoside polyphosphates. In Ap4A and Other Dinu-

A. Guranowski et al.

Methylene analogues of adenosine 5¢-tetraphosphate

33 Pietrowska-Borek M, Szalata M & Guranowski A catalyzed by firefly luciferase. Eur J Biochem 212, 263–270. (1999) Nucleoside 5¢-tetraphosphate hydrolase from human placenta. Cell Mol Biol Lett 4, 439. 34 Guranowski A, Starzyn´ ska E, Barnes LD, Robinson

20 Guranowski A, Gu¨ nther Sillero MA & Sillero A (1994) Adenosine 5¢-tetraphosphate and adenosine 5¢-penta- phosphate are synthesized by yeast acetyl coenzyme A synthetase. J Bacteriol 1243, 78–84.

AK & Liu S (1998) Adenosine 5¢-tetraphosphate phos- phohydrolase activity is an inherent property of soluble exopolyphosphatase from yeast Saccharomyces cerevisiae. Biochim Biophys Acta 1380, 232–238. 35 Thorne NMH, Hankin S, Wilkinson MC, Nu´ n˜ ez C, 21 Dieckman R, Pavela-Vrancic M & von Do¨ hrn H (2001) Synthesis of (di)adenosine polyphosphates by non-ribo- somal peptide synthetases (NRPS). Biochim Biophys Acta 1546, 234–241.

22 Pietrowska-Borek M, Stuible H-P, Kombrink E & Gur- anowski A (2003) 4-Coumarate: coenzyme A ligase has the catalytic capacity to synthesize and reuse various (di)adenosine polyphosphates. Plant Physiol 131, 1401– 1410.

Barraclaugh R & McLennan AG (1995) Human diade- nosine 5¢,5¢¢¢-P1,P4-tetraphosphate pyrophosphohydro- lase is a member of the MutT family of nucleotide pyrophosphohydrolases. Biochem J 311, 717–721. 36 Maksel D, Gooley PR, Swarbrick JD, Guranowski A, Gange Ch, Blackburn GM & Gayler KR (2001) Char- acterization of active residues in diadenosine tetraphos- phate hydrolase from Lupinus angustifolius. Biochem J 357, 399–405. 23 Madrid O, Martı´ n D, Atencia EA, Sillero A & Gu¨ nther Sillero MA (1998) T4 DNA ligase synthesizes dinucleo- side polyphosphates. FEBS Lett 433, 283–286.

24 Atencia EA, Madrid O, Gu¨ nther Sillero MA & Sillero A (1999) T4 RNA ligase catalyzes the synthesis of dinucleoside polyphosphates. Eur J Biochem 261, 802–811.

37 Guranowski A, Jakubowski H & Holler E (1983) Cata- bolism of diadenosine 5¢,5¢¢¢-P1,P4-tetraphosphate in prokaryotes; purification and properties of diadenosine 5¢,5¢¢¢-P1,P4-tetraphosphate (symmetrical) pyrophos- phohydrolase from Escherichia coli K12. J Biol Chem 258, 14784–14789.

25 Guranowski A, de Diego A, Sillero A & Gu¨ nther Sillero MA (1998) Uridine 5¢-polyphosphates (p4U and p5U) and uridine (5¢)polyphospho (5¢)nucleosides (UpnNs) can be synthesized by UTP: glucose-1-phosphate uri- dylyltransferase from Saccharomyces cerevisiae. FEBS Lett 561, 83–88. 38 Reintamm T, Lopp A, Kuusksalu A, Pehk T & Kelve M (2003) ATP N-glycosidase a novel ATP-converting activity from marine sponge Axinella polypoides. Eur J Biochem 270, 4122–4132. 39 Kalek M, Jemielity J, Ste˛pin´ ski J, Stolarski R & Dar-

26 Guranowski A (2004) Metabolism of diadenosine tetra- phosphate (Ap4A) and related nucleotides in plants; review with historical and general perspective. Front Biosci 9, 1398–1411. zynkiewicz E (2005) A direct metod for the synthesis of nucleoside 5¢-methylene bis(phosphonate)s from nucleo- sides. Tetrahedron Lett 46, 2417–2421. 27 Guranowski A (2000) Specific and nonspecific enzymes 40 Jemielity J, Fowler T, Zuberem J, Ste˛pin´ ski J, Lewd-

orowicz M, Nedzwiecka A, Stolarski R, Darzynkiewicz E & Rhoads RE (2003) Novel ‘anti-reverse’ cap analogs with superior translational properties. RNA 9, 1108– 1122. 41 Maier L & Gredig R (1969) Organophosphorus

involved in the catabolism of mononucleoside and dinu- cleoside polyphosphates. Pharmacol Ther 87, 117–139. 28 Blackburn MG, Guo M-J & McLennan AG (1992) Syn- thetic structural analogues of dinucleoside polyphos- phates. In Ap4A and Other Dinucleoside Polyphosphates (McLennan AG, ed.), pp. 305–342. CRC Press, Boca Raton, FL. 29 Guranowski A (2003) Analogs of diadenosine tetraphos-

phate (Ap4A). Acta Biochim Polon 50, 947–972. 30 Spelta V, Mekhalfia A, Rejman D, Thompson M, compounds. XXXVI. Preparation and properties of bis(dialkoxyphosphonylmethyl)-, bis(alkoxyphosphinyl- methyl)-, and bis(oxophosphonarylmethyl) phosphinic acid esters and the corresponding acids. Helv Chim Acta 52, 827–845. 42 Thompson MJ, Mekhalfia A, Hornby DP & Blackburn

GM (1999) Synthesis of two stable nitrogen analogues of S-adenosyl-l-methionine. J Org Chem 64, 7467–7473. Blackburn GM & North RA (2003) ATP analogues with modified phosphate chains and their selectivity for rat P2X2 and P2X2 ⁄ 3. Br J Pharmacol 140, 1027–1034. 31 Jemielity J, Pietrowska-Borek M, Starzyn´ ska E, Kow- 43 Davisson VJ, Davis DR, Dixit VM & Poulter CD

alska J, Stolarski R, Guranowski A & Darzynkiewicz E (2005) Synthesis and enzymatic characterization of methylene analogs of adenosine 5¢-tetraphosphate (p4A). Nucleosides Nucleotides Nucleic Acids 24, 589–593. 32 Guranowski A, Starzyn´ ska E, Brown P & Blackburn

FEBS Journal 273 (2006) 829–838 ª 2006 The Authors Journal compilation ª 2006 FEBS

837

(1987) Synthesis of nucleoside 5¢-diphosphates from 5¢- O-tosyl nucleosides. J Org Chem 52, 1794–1801. 44 Moffatt JG & Khorana HG (1961) Nucleoside poly- phosphates. X. The synthesis and some reactions of nucleoside-5¢-phosphoromorpholidates and related com- pounds. Improved methods for the preparation of nucleoside-5¢-polyphosphates. J Am Chem Soc 83, 649–658. GM (1997) Adenosine 5¢-tetraphosphate phosphohydro- lase from yellow lupin seeds; purification to homogen- eity and properties. Biochem J 328, 257–262.

A. Guranowski et al.

Methylene analogues of adenosine 5¢-tetraphosphate

45 Trowbridge DB & Kenyon GL (1970) Adenosine strate-binding mechanism of human Ap4A hydrolase. J Biol Chem 280, 8471–8481.

5¢-bis(dihydroxyphosphinylmethyl)-phosphinate, the a,b: b,c-bis(methylene) analog of adenosine 5’-triphosphate. J Am Chem Soc 92, 2181–2182. 46 Dixon M & Webb EC (1964) Enzymes, 2nd edn. Aca- demic Press, New York.

57 Jakubowski H & Guranowski A (1983) Enzymes hydro- lyzing ApppA and ⁄ or Ap4A in higher plants: purifica- tion and some properties of diadenosine triphosphatase, diadenosine tetraphosphatase, and phosphodiesterase from yellow lupin (Lupinus luteus) seeds. J Biol Chem 258, 9982–9989. 47 Blackburn GM (1981) Phosphonates as analogues of biological phosphates. Chem Industry (London) 134– 138. 58 Abdelghany HM, Gasmi L, Cartwright JL, Bailey S, 48 Barabas O, Pongraez V, Kovari J, Wilmanns M &

Vertessy BG (2004) Structural insights into the catalytic mechanisms of phosphate ester hydrolysis by dUTPase. J Biol Chem 279, 42907–42915.

Raffarty JB & McLennan AG (2001) Cloning, characteri- zation and crystallization of a diadenosine 5¢,5¢¢¢-P1,P4- tetraphosphate pyrophosphohydrolase from Caenor- habditis elegans. Biochim Biophys Acta 1550, 27–36. 59 Guranowski A, Starzyn´ ska E, McLennan AG, Baraniak J & Stec WJ (2003) Adenosine-5¢-O-phosphorylated and adenosine-5¢-O-phosphorothioylated polyols as strong inhibitors of (symmetrical) and (asymmetrical) dinucleo- side tetraphosphatases. Biochem J 373, 635–640. 49 Chan S, Segelke B, Lekin T, Krupka H, Cho US, Kim M-Y, So M, Kim C-Y, Naranjo CM, Rogers YC et al. (2004) Crystal structure of the Mycobacterium tuberculo- sis dUTPase: insights into the catalytic mechanism. J Mol Biol 341, 503–517. 50 Lobato´ n CD, Vallejo CG, Sillero A & Gu¨ nther Sillero

60 Bailey S, Sedelnikova SE, Blackburn GM, Abdelghany HM, Baker PJ, McLennan AG & Raffarty JR (2002) The crystal structure of diadenosine tetraphosphate hydrolase from Caenorhabditis elegans in free and bin- ary complex forms. Structure 10, 589–600.

Supplementary material

MA (1975) Diguanosine tetraphosphatase from rat liver: activity on diadenosine tetraphosphate and inhibition by adenosine tetraphosphate. Eur J Biochem 50, 495–501. 51 Moreno A, Lobato´ n CD, Gu¨ nther Sillero MA & Sillero A (1982) Dinucleosidetetraphosphatase from Ehrlich ascites tumor cells: inhibition by adenosine, guanosine and uridine 5¢-tetraphosphates. Int J Biochem 14, 629–634.

is available

The following supplementary material online:

Characterization of

the p4A analogues with one

methylene group by MS and NMR spectroscopy.

Syntheses of

the p4A analogues with two halo ⁄ methylene bridges: General remarks on preparation of the precursors of p4A analogues.

Synthesis of isopropyl bis(diethyl phosphonodichloro-

methyl)phosphinate, pCCl2pCCl2p pentaester.

52 McLennan AG, Mayers E, Walker-Smith I & Chen H (1995) Lanterns of the firefly Photinus pyralis contain abundant diadenosine 5¢,5¢¢¢-P1,P4-tetraphosphate pyro- phosphohydrolase activity. J Biol Chem 270, 3706–3709. 53 Costas MJ, Pinto RM, Ferna´ ndez A, Canales J, Garcı´ a- Agu´ ndez JA & Cameselle JC (1990) Purification to homogeneity of rat liver dinucleoside tetraphosphatase by affinity elution with adenosine 5¢-tetraphosphate. J Biochem Biophys Methods 21, 25–33.

Synthesis of bis(phosphonodichloromethyl)phosphi-

nic acid, pCCl2pCCl2p free acid.

54 Lazewska D, Starzyn´ ska E & Guranowski A (1993) Human placenta (asymmetrical) diadenosine 5¢,5¢¢¢- P1,P4-tetraphosphate hydrolase: purification to homo- geneity and some properties. Protein Expr Purif 4, 45–51.

Synthesis of adenosine-5’-[b,c,c,d-bis(dichlorometh- ylene)]tetraphosphate, pCCl2pCCl2ppA. Synthesis of a,b;b,c-bis(methylene)-ATP, tris(triethylammonium) salt, pCH2pCH2pA. Synthesis of

adenosine-5’-[a,b,b,c-bis(methylene)]-

tetraphosphate, ppCH2pCH2pA.

Synthesis of

adenosine-5’-[a,b;c,d-bis(methylene)]-

55 Swarbrick JD, Bashtannyk T, Maksel D, Zheng X-R, Blackburn GM, Gayler KR & Gooley PR (2000) The three-dimensional structure of the nudix enzyme diade- nosine tetraphosphate hydrolase from Lupinus angustifo- lius L. J Mol Biol 302, 1165–1177.

tetraphosphate, pCH2ppCH2pA.

FEBS Journal 273 (2006) 829–838 ª 2006 The Authors Journal compilation ª 2006 FEBS

838

56 Swarbrick JD, Buyya S, Gunawardana D, Gayler KR, McLennan AG & Gooley PR (2005) Structure and sub-