Báo cáo khoa học: Xanthosine and xanthine Substrate properties with purine nucleoside phosphorylases, and relevance to other enzyme systems

doi:10.1046/j.1432-1033.2002.03097.x

Eur. J. Biochem. 269, 4048–4057 (2002) (cid:2) FEBS 2002

Xanthosine and xanthine Substrate properties with purine nucleoside phosphorylases, and relevance to other enzyme systems

Gerasim Stoychev1, Borys Kierdaszuk1 and David Shugar1,2 1Department of Biophysics, Institute of Experimental Physics, University of Warsaw, Poland; 2Institute of Biochemistry and Biophysics, Polish Academy of Sciences, Warsaw, Poland

product inhibition. Replacement of Pi by arsenate led to complete arsenolysis of Xao. Kinetic parameters are reported for the phosphorolytic and reverse synthetic path- ways at several selected pH values. Phosphorolysis of 200 lM Xao by the human enzyme at pH 5.7 is inhibited by Guo (IC50 ¼ 10 ± 2 lM), Hx (IC50 ¼ 7 ± 1 lM) and Gua (IC50 ¼ 4.0 ± 0.2 lM). With Gua, inhibition was shown to be competitive, with Ki ¼ 2.0 ± 0.3 lM. By contrast, Xao and its products of phosphorolysis (Xan and R1P), were poor inhibitors of phosphorolysis of Guo, and Xan did not inhibit the reverse reaction with Gua. Possible modes of binding of the neutral and anionic forms of Xan and Xao by mammalian PNPs are proposed. Attention is directed to the fact that the structural properties of the neutral and ionic forms of XMP, Xao and Xan are also of key importance in many other enzyme systems, such as IMP dehydrogenase, some nucleic acid polymerases, biosynthesis of caffeine and phosphoribosyltransferases.

Keywords: Purine nucleoside phosphorylases; xanthine/ xanthosine; enzyme kinetics; enzyme–ligand interactions; pH-dependence.

Substrate properties of xanthine (Xan) and xanthosine (Xao) for purine nucleoside phosphorylases (PNP) of mammalian origin have been reported previously, but only at a single arbitrarily selected pH and with no kinetic con- stants. Additionally, studies have not taken into account the fact that, at physiological pH, Xao (pKa ¼ 5.7) is a mono- anion, while Xan (pKa ¼ 7.7) is an equilibrium mixture of the neutral and monoanionic forms. Furthermore the monoanionic forms, unlike those of guanosine (Guo) and inosine (Ino), and guanine (Gua) and hypoxanthine (Hx), are still 6-oxopurines. The optimum pH for PNP from human erythrocytes and calf spleen with both Xao and Xan is in the range 5–6, whereas those with Guo and Gua, and Ino and Hx, are in the range 7–8. The pH-dependence of substrate properties of Xao and Xan points to both neutral and anionic forms as substrates, with a marked preference for the neutral species. Both neutral and anionic forms of 6-thioxanthine (pKa ¼ 6.5 ± 0.1), but not of 2-thioxan- thine (pKa ¼ 5.9 ± 0.1), are weaker substrates. Phosphor- olysis of Xao to Xan by calf spleen PNP at pH 5.7 levels off at 83% conversion, due to equilibrium with the reverse synthetic pathway (equilibrium constant 0.05), and not by

inorganic phosphate (Pi), a reaction reversible with natural substrates, as follows:

The ubiquitous purine nucleoside phosphorylases (PNP, purine nucleoside phosphorylase ribosyl transferases), cat- alyse the cleavage (phosphorolysis) of the glycosidic bond of ribo- and 2¢-deoxyribo- nucleosides in the presence of

purine base

PNP b-nucleoside+Pi (cid:2)(cid:2)*)(cid:2)(cid:2) þ a-D-ribose-1-phosphate

In mammalian cells, phosphorolysis is the predominant reaction, due to coupling with guanase and xanthine oxidase, leading to stepwise formation of xanthine (Xan) and, finally, urate. PNP functions in the so-called purine salvage pathway, wherein the purines liberated by phos- phorolysis are converted by hypoxanthine-guanine phos- phoribosyltransferase (HGPRTase) to the monophosphates of inosine (Ino) and guanosine (Guo).

The natural substrates of the mammalian enzymes are the 6-oxopurine nucleosides, Ino and Guo, and their 2¢-deoxy counterparts, but not the 6-aminopurine nucleosides, ade- nosine (Ado) and dAdo. By contrast, all the aforementioned are substrates for the enzyme from E. coli, a product of the deoD gene [1], as well as for the enzyme from many prokaryotes, e.g. S. typhimurium. It was first shown by Hammer-Jespersen et al. [2] that cultivation of E. coli K-12 cells in the presence of xanthosine (Xao), but no other led to the expression of a second nucleoside or base,

Correspondence to B. Kierdaszuk, Department of Biophysics, Institute of Experimental Physics, University of Warsaw, 93 _ZZwirki i Wigury Street, 02-089 Warsaw, Poland. Fax: + 48 22 554 0001, Tel.: + 48 22 554 0715, E-mail: borys@biogeo.uw.edu.pl Abbreviations: PNP, purine nucleoside phosphorylase; HGPRTase, hypoxanthine-guanine phosphoribosyltransferase; Xao, xanthosine; dXao, 2¢-deoxyxanthosine; Xan, xanthine; 6-thio-Xan, 6-thioxan- thine; 2-thio-Xan, 2-thioxanthine; m7Guo, N(7)-methylguanosine; m7-6-thioGuo, N(7)-methyl-6-thioguanosine; R1P, a-D-ribose- 1-phosphate; dR1P, 2¢-deoxy-a-D-ribose-1-phosphate; NR+, nicotinamide 1-b-D-ribose. Enzymes: PNP (EC 2.4.2.1); E. coli PNP-I (EC 2.4.2.1); E. coli PNP-II (EC 2.4.2.1); HGPRTase (EC 2.4.2.8), guanase (EC 3.5.4.3); xanthine oxidase (EC 1.1.3.22); IMP dehydrogenase (EC 1.1.1.205); nucleoside triphosphate pyrophosphatase (EC 3.6.1.19). (Received 30 April 2002, revised 27 June 2002, accepted 5 July 2002)

Xanth(os)ine and purine nucleoside phosphorylases (Eur. J. Biochem. 269) 4049

(cid:2) FEBS 2002

contrast (see Fig. 1), it was shown long ago that Xao (pKa ¼ 5.7) is predominantly a monoanion at physiological pH, at which Xan (pKa ¼ 7.7) is an equilibrium mixture of the neutral and monoanionic species. Moreover, unlike Ino and Guo, and hypoxanthine (Hx) and guanine (Gua), where monoanion formation at pH > 8 is due to dissoci- ation of the N(1)-H [11] so that they are no longer 6-oxo purines, it is the N(3)-H which dissociates in Xao and Xan [11–13], so that their monoanions, like the neutral forms, are still 6-oxopurines (see Fig. 1). This is further supported by the finding that it is the N(3)-H which is dissociated in the crystal structure of the monoanion of Xan [14], and that the pKa of 1-methyl-Xao, where only the N(3)-H can dissociate, is 5.85 [12], close to 5.7 for Xao.

phosphorylase, a product of the xapA gene, which, in addition to Ino and Guo, accepts Xao as a substrate but not Ado. This enzyme was initially referred to as Xao phosphorylase, subsequently as Ino–Guo phosphorylase; we refer to it as E. coli PNP-II. It was subsequently partially purified and some of its properties further characterized [3]. This directed our attention to the specificity of the mammalian enzymes, earlier reviewed [4,5], but with limited reference to Xan and Xao as substrates. A more recent review [6] has redirected attention to this and, amongst others, to a much earlier report by Friedkin [7]. In this report, phosphorolysis of dGuo by a partially purified PNP from rat liver, known to contain guanase, led to the appearance of the then unknown 2¢-deoxyxanthosine (dXao), which was isolated in crystalline form. This can be interpreted by the following reaction sequence:

PNP dGuo + Pi (cid:2)(cid:2)*)(cid:2)(cid:2)

Gua+dR1P # Guanase Xan+dRIP (cid:2)(cid:2)*)(cid:2)(cid:2) dXao+Pi

Bearing in mind the physiological significance of Xan and Xao in the purine salvage pathway and the differences in structure between the neutral and monoanionic forms of these relative to those of Hx and Ino, and Gua and Guo, it is clearly desirable to determine the substrate properties of the neutral and monoanionic forms of Xan and Xao. This is also relevant to the properties of the E. coli PNP-II, referred to above, which exhibits a marked preference for Xao, as well as to a number of other enzyme systems, discussed below.

M A T E R I A L S A N D M E T H O D S

Materials

Phosphorolysis of Guo by the same enzyme preparation led to the appearance of Xao, also shown more recently by Giorgelli et al. [8], who do not refer to the paper by Friedkin [7]. It is assumed that the rat liver preparation was devoid of xanthine oxidase activity. In the presence of xanthine oxidase and an excess of Pi, the isolated dXao was a substrate for phosphorolysis, at about 2% of the rate for dGuo at pH 7.4 [7].

Purine nucleoside phosphorylase from human erythrocytes and calf spleen (Sigma, St Louis, MO, USA) was further purified by size-exclusion chromatography, followed by concentration, as described previously [15]. Specific activi- ties of the enzymes are given in the footnote to Table 2.

Guo, Ino, formycin B, disodium arsenate, mono- and disodium phosphate and a-D-ribose-1-phosphate (R1P) were also obtained from Sigma, and Xao and Xan from

It is rather surprising that, in light of this and other reports on the substrate properties of Xao and Xan with mammalian PNPs [9,10], no account has been taken of their structures and the pH-dependence of these structures. Both Ino and Guo, with pKa values of 8.8 and 9.2, respectively, due to dissociation of the N(1)-H, exist predominantly as the neutral 6-oxo forms at physiological pH. In striking

Fig. 1. Structures of the neutral and monoanionic forms of hypoxanthine (Hx) and inosine (Ino), guanine (Gua) and guanosine (Guo), and xanthine (Xan) and xanthosine (Xao). Note that the monoanions of the latter are still 6-oxopurines, like the neutral forms of Gua and Guo, and Hx and Ino.

4050 G. Stoychev et al. (Eur. J. Biochem. 269)

(cid:2) FEBS 2002

NMR spectroscopy. This is relevant to earlier reports on the substrate properties of Xao.

Serva (Heidelberg, Germany). N-methylated xanthines, and thioxanthines, were prepared as described previously [16,17]. The purity of compounds was confirmed by chromatography and pH-dependent UV absorption spec- tra. All solutions were prepared with Milli-Q water (Millipore), using reagents of the highest quality commer- cially available.

Buffering media, acetate (pH 3.6, 4.5, 5.0 and 5.5), Mes (pH 6.0 and 6.5), Hepes (pH 7.0, 7.5 and 8.2), Ches (pH 8.5 and 9.0) and Caps (pH 10.0) (Sigma) were selected to avoid buffer effects on enzyme activity previously noted with Tris and other buffers [4,15,18,19]. Enzyme activities with these buffers were unaffected, within experimental error, when acetate was replaced by Mes at pH 5.0, Mes by Hepes at pH 6.8, and Hepes by Ches at pH 8.4.

From amongst four commercially available preparations of Xao, only that from Serva was chromatographically homogeneous, with pH-dependent UV spectra (Table 1) consistent with those reported previously [11,12], and the absence of contaminants further confirmed by 1H and 13C

–1Æcm)1)

Table 1. Spectral properties of compounds.

Measurements of pH (± 0.05) were carried out with a CP315m pH meter (Elmetron, Poland) equipped with a combination semimicro electrode (Orion, UK) and temper- ature sensor. UV absorption was monitored with a Kontron Uvikon 922 recording instrument, fitted with a thermostat- ically controlled cell compartment, using 1-, 2-, 5- or 10-mm pathlength cuvettes.

pH Compound kmax (nm) emax (M pKa

Enzyme kinetics

Xao 5.7a

Xan 7.7a

a pKa values for Xao and Xan were taken from references [11,12,60], independently confirmed in this study by spectropho- tometric titration at 25 (cid:3)C. b Determined by spectrophotometric titration.

Phosphorolysis was monitored spectrophotometrically at 25 (cid:3)C in 50 mM buffers in the presence of 8 mM (substrate saturation) Pi, by following the maximal changes in absorption of the substrate Xao at 242 nm and Guo at 257 nm (Fig. 2) due to formation of Xan and Gua, respectively. The concentration of Pi (8 mM) at each pH was well above its Km (< 1 mM). The absorption spectra of each reaction showed isosbestic points, at each pH, e.g. 223, 260 and 279 nm (with Xao), and 240 and 287.5 nm (with Guo) at pH 5.7, which permitted the monitoring of product formation in the reaction mixture. The reverse synthetic reaction was monitored in the presence of 1 mM (substrate saturation) R1P and no Pi. For pH effects on enzyme activity, enzyme samples were preincubated at each pH and

2-thio-Xan 6-thio-Xan Guo Gua 5.9b 6.5b 9.2 9.3 3.6 5.7 6.0 6.5 9.0 3.6 6.0 7.5 10.0 8.0 7.0 7.0 7.0 235, 264 253 251, 271 249, 276 248, 278 268 267 271 241, 278 278 342 253 246 8100, 9200 8700 9100, 8100 9700, 8700 10 100, 9000 10 400 10 300 9200 9000, 9300 15 000 24 600 13 700 10 700

Fig. 2. pH-Dependence of relative activities, expressed as initial rates, of PNP from (A, B) human erythrocytes and (C, D) calf spleen, for phos- phorolysis of 1.2 mM Xao (d), and 200 lM (pH < 8), 500 lM (pH = 8) and 780 lM (pH = 8.5) Guo (m), and for the reverse synthetic reaction with 1.2 mM (pH < 7) and 1.8 mM (pH ‡ 7) Xan (O) and 100 lM Gua (m). Activities of both enzymes vs. 200 lM Guo and 100 lM Gua at pH 7 were taken as 100%, for the phosphorolytic and synthetic reactions, respectively. Measurements were in 50 mM buffers containing 8 mM Pi for phosphorolysis, and 1 mM R1P for the reverse reaction at 25 (cid:3)C. Reactions for Xao and Xan were monitored spectrophotometrically at 242 nm, for which values of De were: 4030 (pH 3.6), 4040 (pH 4.5), 4320 (pH 5.0), 4640 (pH 5.7), 4740 (pH 6.0), 4820 (pH 6.5), 4130 (pH 7.0), 3240 (pH 7.6), 1720 (pH 8.1) and 440 (pH 8.5); and for Guo and Gua at 257 nm, with De of 4600 in the pH range 3.6–8.5.

Xanth(os)ine and purine nucleoside phosphorylases (Eur. J. Biochem. 269) 4051

(cid:2) FEBS 2002

Table 2. Kinetic parameters for phosphorolysis of Xao and Guo (in presence of 8 mM Pi), and for the reverse synthetic reaction with Xan and Gua (in presence of 1 mM R1P), for human and calf PNP at various pH values.

a

a

human PNP calf PNP

a Vmax %

a Vmax %

Km lM Vmax/Km % Km lM Vmax/Km % Compound pH

Xao

Guob 0.9 ± 0.2 0.14 ± 0.07 7 ± 5 68 ± 5

Xan

a Values for Xao and Xan are relative to Guo and Gua, respectively, at pH 7. b Values of Vmax and Vmax/Km at pH 5.7 and pH 6.5 are relative to those at pH ¼ 7. c 12 ± 1 lM in [31]. d Values of kcat and kcat/Km are 33 ± 4 s)1 and 2.8 ± 0.5 s)1ÆlM )1, and 43 ± 5 s)1 and )1 for Guo and Gua, respectively (cf. Stoeckler et al. [31]). e 11 lM in [61]. f Values of kcat and kcat/Km are 31 ± 4 s)1 and 3.1 ± 0.6 s)1lM 2.8 ± 0.5 s)1lM )1 for Guo and Gua, respectively (cf. Porter [62]). g Values of Vmax and Vmax/Km at )1, and 23 ± 3 s)1 and 3.8 ± 0.7 s)1lM pH 6.0 and 7.5 are relative to those at pH 7. h 6 ± 1 lM in [63].

their activities were measured, at concentrations as close to saturation as possible, with Xao ((cid:5) 1.5 mM), Guo ((cid:5) 0.5 mM) and Pi (8 mM) for phosphorolysis, and Xan ((cid:5) 1.2 mM), Gua ((cid:5) 100 lM) and R1P (1 mM) for the reverse reaction. Concentrations of substrates may be considered as saturated only in the pH range of 5–6 (for phosphorolysis of Xao) and 5.0–7.5 (for the reverse reaction with Xan), where they are at least threefold higher than the appropriate Km values (Table 2). Due to low solubility of Gua and Xan, they were initially dissolved in slightly alkaline medium and then diluted with buffer to the appropriate pH. Concentrations of nucleosides and bases were determined from absorbance measurements, using molar extinction coefficients (Table 1).

Guag 100d 1.0 ± 0.2 0.8 ± 0.2 39 ± 5 100d 5.7 6.5 5.7 6.5 7.0 6.0 7.5 6.0 7.0 7.5 580 ± 150 1600 ± 300 91 ± 13 14 ± 2 12 ± 2c 380 ± 15 425 ± 85 29 ± 2 13.8 ± 0.6 22 ± 1 42 ± 10 18 ± 4 55 ± 5 79 ± 5 100d 28 ± 5 26 ± 5 81 ± 6 100d 99 ± 2 62 ± 10 400 ± 100 1100 ± 100 46 ± 3 15.9 ± 1.6 11 ± 2e 280 ± 20 306 ± 40 11 ± 1 6.0 ± 1.2h 9 ± 1 126 ± 23 54 ± 12 50 ± 5 79 ± 5 100f 95 ± 15 13 ± 4 60 ± 6 100f 89 ± 5 3.5 ± 0.8 0.5 ± 0.3 12 ± 5 55 ± 5 100f 2.0 ± 0.5 0.3 ± 0.1 33 ± 4 100f 59 ± 8

Kinetic constants were determined using the initial rate method. Initial rates (v) were determined from linear regression fitting to at least 10 experimental points for the linear course of the reaction (1–2 min), with an accuracy of £ 5%. The values of the Michaelis constant (Km) and maximal velocity (Vmax) were determined from nonlinear regression fitting of the Michaelis–Menten eqn (1) to initial rates (v) measured for the whole concentration range of substrate ([S]):

ð1Þ

v ¼ Vmax=ð1 þ Km=½S(cid:8)Þ

groups [7–10,20], but in each case only at a single arbitrarily selected pH. These experiments did not take into account the existence of a mixture of neutral and monoanionic forms, and with no measurements of the kinetic constants.

Inhibition constants (Ki) were calculated using the Dixon equation for competitive inhibition:

ð2Þ

1=v ¼ ½ðKm=[S])ð1 þ [I]=KiÞ þ 1(cid:8)=Vmax

Equation (2) was fitted to initial rates measured at four concentrations of inhibitor [I] for each substrate concentra- tion (Fig. 3), and apparent values of Ki calculated.

Fig. 3. Dixon plot for the inhibition of phosphorolysis of Xao by Gua with human PNP at pH 5.7 and 25 (cid:2)C: (j) 290 lM Xao (d) 580 lM Xao (m) 1160 lM Xao. The solid lines represent linear equations fitted independently using the linear regression method.

R E S U L T S

Reaction equilibrium for phosphorolysis of Xao

The phosphorolytic conversion of 100 lM Xao to Xan by calf spleen PNP was followed in the presence of 8 mM Pi at pH 5.7, where the population of the neutral form of Xao is (cid:5) 50%, and that of the neutral form of Xan is (cid:5) 100%. The reaction levels off at about 83% conversion, corresponding to an equilibrium constant of 0.05. This is not due to enzyme inactivation, as addition of fresh enzyme at this point was without effect. Nor is it due to product inhibition, because the initial rate of the reaction was unaffected in the presence of 1 mM Xan, and the IC50 of R1P was 1 mM. The levelling off of the reaction must therefore be due to establishment of equilibrium with the reverse synthetic reaction, confirmed by addition of 0.25 mM R1P, which led to reduction of the

Substrate properties of Xao and Xan for PNPs from mammalian sources have been reported previously by several

4052 G. Stoychev et al. (Eur. J. Biochem. 269)

(cid:2) FEBS 2002

plateau level to (cid:5) 53%, and lack of an effect of addition of 30 lM Xan. When Pi was replaced by arsenate, arsenolysis proceeded at a slower rate, but on prolonged incubation went virtually to completion. This is because arsenolysis is not reversible, due to very rapid hydrolysis of a-D-ribose-1- arsenate [21].

pH-dependence of substrate properties

Xan, both of which would be expected to be more acidic than the parent Xan, and hence with higher populations of the monoanions at physiological pH. We have confirmed this by spectrophotometric titration, which gave pKa values of 5.9 ± 0.1 for 2-thio-Xan and 6.5 ± 0.1 for 6-thio-Xan. For 6-thio-Xan the predominant tautomer of the neutral form was identified by means of UV spectroscopy in aqueous medium and by NMR spectroscopy in dimethyl- sulfoxide-water [16,17] as the 6-thione-2-oxo-N(7)-H. To our knowledge there are no experimental data on the structure of the anionic species, but a recent theoretical study [25] suggests that the neutral form of 6-thio-Xan is 6-thione-2-oxo-N(7)-H, and that monoanion formation involves dissociation of the N(3)-H, as for the parent Xan (Fig. 1). Furthermore, it points to 6-oxo-2-thione-N(9)-H as being more stable than 6-oxo-2-thione-N(7)-H in aqueous medium [26].

Figure 2 exhibits the substrate properties of Xao and Xan with the human and calf enzymes over the pH range 3.6–8.7, and, for comparison, those of Guo and Gua. Note that, because of their poorer substrate properties relative to Guo and Gua, the concentrations of Xao and Xan employed were necessarily several-fold higher than those of Guo and Gua (Fig. 2). Despite this, the concentrations of Xao and Xan may be considered as saturating at pH values where these concentrations are several-fold higher than the appropriate Km values (Table 2), i.e. for Xao at pH 5–6, and for Xan at pH 5–7.5. The use of higher concentrations of Xan was limited by the low solubility, and by the decrease in accuracy of reaction rates monitored by small changes of very high absorbancy of substrates, even with a 1-mm optical path length.

Hitherto, thioxanthines and their nucleosides have not been examined as potential substrates of PNP, and studies of their substrates’ properties with other enzymes have not considered their physico-chemical properties, notwithstand- ing that 6-thio-Xan is an intermediate in the metabolism of thiopurines [27]. These compounds are considered as potential antitumour agents [28], and 6-thio-Xan is a prodrug in gene therapy [29].

To compare the substrate properties of 6-thio-Xan and Xan with both enzymes, we estimated De for the following conversion

6-thio-Xan (cid:2)(cid:2)*)(cid:2)(cid:2) 6-thio-Xao

It is clear from Fig. 2 that, whereas the optimum pH for both Guo and Gua is in the range 7–8, that for Xao and Xan is in the range pH 5–6, particularly pronounced for the calf spleen enzyme. It is of interest, in this context, that with E. coli PNP-II, the pH profile for Xao (optimum 6.7) has been shown to overlap those for dGuo (optimum 6.7) and Guo (optimum 6.9) [22]. With both calf and human enzymes, phosphorolysis of Xao is optimal at about pH 5 (where the population of the neutral species is (cid:5) 70%), and decreases with increase in pH, as compared to an increase for Guo. This points to the neutral form of Xao as the preferred substrate, further supported by its high activity with both enzymes at pH 3.5 (Fig. 2B,D), where it exists exclusively as the neutral form. The marked decrease in the rate of phosphorolysis above pH 6, where phosphorolysis of Guo increases, further suggests that the neutral form of Xao may be the exclusive substrate.

The same applies to Xan in the reverse synthetic reaction with both enzymes, the rate of which decreases sharply above pH 6, at which the monoanionic form appears (pKa ¼ 7.7). The pH profiles suggest that the monoanionic form of Xan is two orders of magnitude weaker as a substrate than the neutral form.

from the differences between the absorption spectra of 6-thio-Xan, and those reported for 6-thio-Xao [30], at pH 2, 7 and 11. At pH 5, where the population of the neutral form is > 90%, conversion of 6-thio-Xan (400 lM) to 6-thio-Xao (De (cid:5) 4000 at kobs ¼ 355 nm) in the presence of 1 mM R1P was 10-fold slower than for the parent Xan. Raising the pH to 8.2, where the population of the monoanion of 6-thio- Xan is (cid:5) 98% (as compared to (cid:5) 75% for Xan) reduced its reaction rate, which at this pH was similar to that for Xan, further pointing to substrate activity of the monoanion. This is consistent with the proposed existence of the neutral form as 6-thio-2-oxo, and dissociation of the N(3)-H to form the monoanion [25], as for the parent Xan monoanion. With 2-thio-Xan, quantitative measurements of enzyme activity were not possible, because the UV absorption spectra of its nucleoside are unknown. However, spectral changes at pH ¼ 5 ((cid:5) 90% neutral form), and at pH 8 ((cid:5) 100% anion), in the presence of the enzyme and R1P, were barely detectable, pointing to its being a very feeble substrate, if at all. The poor, if any, substrate properties of 2-thio-Xan clearly call for an investigation of the structures of its neutral and monoanionic species.

Kinetic constants

We then compared the pH-dependence of enzyme activity for Xao and Xan with a substrate which does not undergo ionization in the pH range 6–9. Such a substrate is the cationic nicotinamide riboside (NR+), which, like the cation of m7Guo [23], undergoes nonreversible phosphorolysis by the enzymes from mammalian sources and E. coli [24]. Vmax and Vmax/Km for NR+, which is exclusively in the cationic form (pKa (cid:5) 11.9), is unchanged over the pH range 7–10 [24]. It follows that the pH-dependence of reaction rates for Xao and Xan in the pH range 6–9 should reflect changes in substrate properties due to ionization of the base moiety.

Substrate properties of thioxanthines

Table 2 presents several kinetic constants for Xao and Xan at selected pH values, and the corresponding constants for Guo and Gua, with the human and calf spleen enzymes. Surprisingly, the Vmax for Xao at pH 5.7 with the calf, but not human, enzyme is 2.5-fold higher than for Guo. However, it should be noted that the Km values for both Xao and Xan are very high relative to those for Guo and

The apparent substrate properties of the monoanion of xanthine directed our attention to 2-thio-Xan and 6-thio-

Xanth(os)ine and purine nucleoside phosphorylases (Eur. J. Biochem. 269) 4053

(cid:2) FEBS 2002

Gua, accounting in large part for the lower rate constants, Vmax/Km, of the former in both the phosphorolytic and reverse reaction pathways.

1 mM R1P. This is consistent with the finding of Krenitsky et al. [32] that the reverse reaction for Hx with human PNP is inhibited by Xan with Ki ¼ 40 lM.

Formycin B, a structural analogue of Ino, is a weak inhibitor of phosphorolysis of Ino by the human and calf enzymes at pH 7, with Ki (cid:5) 100 lM, and an even weaker inhibitor of both Ino and Xao phosphorolysis by E. coli PNP-II [3], with Ki (cid:5) 300 lM. We have examined the effect of formycin B on phosphorolysis of Xao and Guo by the human enzyme at pH 5.7, at substrate concentrations comparable to their Km values, 500 lM and 90 lM, respec- tively. This led to IC50 values of 160 ± 30 lM versus Xao and 500 ± 100 lM versus Guo, and shows that more effective inhibition of Xao correlates with its lower activity as substrate.

The Vmax/Km for phosphorolysis of Xao at pH 5.7 is, with the human enzyme, 13% of that for Guo, and with the calf enzyme, 29% that for Guo. These values decrease dramatically in going from pH 5.7 to 6.5, i.e. with a large increase in population of the monoanionic species of Xao, indicating that the enzymes highly prefer the neutral form. Similarly, the Vmax for Xan in the reverse synthetic reaction with the calf enzyme decreases sevenfold in going from pH 6 to 7.5, i.e. with an increase in population of the monoanionic form, accounting for the sevenfold lower rate constant at pH 7.5. By contrast, for Xan in the reverse synthetic reaction with the human enzyme, both Vmax and Vmax/Km are barely affected by an increase of pH from 6 to 7.5, in line with the smaller effect of pH, in this pH range, on the Vmax and Vmax/Km for Gua in the reverse reaction with the human, than with the calf, enzyme (Table 2).

Competition and product inhibition

Possible competition between Xao and Guo was investi- gated by monitoring phosphorolysis of each by human PNP in a medium containing 200 lM Xao and 10 lM Guo, in the presence of 8 mM Pi at pH 5.7. Phosphorolysis of Guo was followed at 260 nm, where there is an isosbestic point for the interconversion

All four N-monomethyl xanthines were found to be very poor, or barely detectable, inhibitors of phosphorolysis by both the calf and human enzymes at pH 5.7, where the 1-methyl-, 3-methyl- and 7-methyl- xanthines are predomi- nantly in the neutral forms, and 9-methylxanthine (pKa (cid:5) 6.3) is a mixture of neutral and monoanionic species [11]. Krenitsky et al. [32] had previously reported that all of these were very poor inhibitors of the reverse synthetic pathway by the human enzyme at pH 7.2, where they are all mixtures of neutral and monoanionic forms. It follows that both the neutral and monoanionic species of all four monomethyl xanthines are very poor inhibitors of both the phosphorolytic and synthetic pathways, and that dimethyl xanthines should also be poor inhibitors, as found.

Xao (cid:2)(cid:2)*)(cid:2)(cid:2) Xan

ðDe ¼ 0Þ

Guo (cid:2)(cid:2)*)(cid:2)(cid:2) Gua

ðDe ¼ 4600Þ

D I S C U S S I O N

Comparison with earlier data

Phosphorolysis of Xao was monitored at 287.5 nm, the isosbestic point for

Guo (cid:2)(cid:2)*)(cid:2)(cid:2) Gua

ðDe ¼ 0Þ

Xao (cid:2)(cid:2)*)(cid:2)(cid:2) Xan

ðDe ¼ 2200Þ

Bearing in mind differences between enzymes from different sources, as shown here between the human and calf enzymes, it is instructive to note that our results are in general accord with data reported earlier, but only at single pH values, e.g. for (a) phosphorolysis of Xao at pH ¼ 6 by human erythrocytic PNP [10], (b) synthesis of Xao from Xan by bovine liver PNP at pH ¼ 8 [9], (c) synthesis of Xao from Xan by the calf spleen enzyme at pH ¼ 7 [20], and (d) phosphorolysis of dXao and Xao by rat liver PNP at pH ¼ 7.4 [7].

Possible modes of binding of Xao and Xan by PNP

Information now available, largely from crystallographic studies, of the modes of binding of Hx and Gua, and their nucleosides, as well as nucleoside analogue inhibitors, by the PNPs from various sources [33–37], permits inferences of modes of binding of Xao and Xan, for which no experi- mental data are available.

The rate of phosphorolysis of 10 lM Guo at pH 5.7 was only minimally affected in the presence of 200 lM Xao. As the latter is a 1 : 1 mixture of the neutral and monoanionic forms at this pH, it follows that both are poor inhibitors. By contrast, the initial rate of phosphorolysis of 200 lM Xao at this pH was inhibited by about 50% in the presence of 10 lM Guo, and this inhibition was markedly accentuated as the reaction proceeded, pointing to the involvement of some product of phosphorolysis. Both Xan and R1P were very poor inhibitors, with IC50 > 1 mM. However, Gua proved to be a good inhibitor of phosphorolysis of Xao by human PNP (IC50 (cid:5) 4 lM), as was Hx (IC50 (cid:5) 7 lM). In the case of Gua, inhibition was shown to be competitive (Fig. 3), with Ki ¼ 2.0 ± 0.3 lM. It is to be expected that, at pH > 6.5, inhibition will be more pronounced because of the threefold higher Km for Xao, whereas Km values of Guo and Gua are unchanged (Table 2). However, the high Km at pH ¼ 6.5 proved to be an obstacle to measurement of Ki for Gua at this pH.

The active center of E. coli PNP-I [37], which does not accept Xao and Xan, differs from those of the mammalian enzymes in that it does not contain the Glu201 of the latter. This residue is proposed to play a key role in the catalytic process via (cid:2)two-hydrogen bond binding(cid:3) of the Glu201Oe1 and Glu201Oe2 to the C(2)-NH2 and the N(1)-H of Gua [or the N(1)-H of Hx] thus stabilizing the intermediate state of the base [33–36]. This is in line with the preference of the mammalian enzymes for the neutral 6-oxo forms of Gua and Guo [38–40], Hx and Ino, and the cationic 6-oxo forms

The reverse reaction for human PNP with 10 lM Gua and 1 mM (saturated) R1P was not affected in the presence of 100 lM Xan at pH 7, where the latter is a 6 : 1 mixture of the neutral and monoanionic forms. By contrast, Xan was a good inhibitor (IC50 (cid:5) 20 lM) of the reverse reaction, with human PNP and 10 lM Hx (i.e. at its Km value [31]) and

4054 G. Stoychev et al. (Eur. J. Biochem. 269)

(cid:2) FEBS 2002

of m7Guo [23] and m7-6-thioGuo [41]. With the mono- anions of Gua and Guo, and the zwitterions of m7Guo and its m7-6-thioGuo, there will be electrostatic repulsion between the negative charge on N(1) and the anionic form of the Glu201 carboxyl in the active site of the mammalian enzymes (and E. coli PNP-II), which is absent in the active site of PNP-I.

irrespective of

monas, the properties of which are similar to those of the mammalian enzymes, was also proposed by Tebbe et al. [42], based on the structure of its complex with 8-iodogua- nine and sulfate (or phosphate). This is more feasible with the N(7)-H tautomer, further supported by data on the ternary complex bovine PNP/9-deazaIno/Pi [35], and some 9-deazaGuo inhibitors, where N(7) is protonated, complexed with human erythrocyte PNP [43]. A similar pattern was observed for the ternary complex of human PNP with the transition-state analogue inhibitor immucillin-H and Pi [34], again pointing to possible involvement of the N(7)-H form of purine bases in the reverse reaction. We propose that the N(7)-H tautomers of the neutral and ionic forms of Xan are preferentially bound by the active sites of the human and calf enzymes, and the N(7)-H donates a hydrogen to Asn243Od, while the Asn243Nd1 donates a hydrogen to the exocyclic O6 of the neutral and ionic forms of Xan (Fig. 4). Additionally, a bridging water molecule between O6 and Glu201Oe2 (not shown) could also be present here, as observed for purines and purine nucleosides in the active site of human erythrocyte PNP [40], for Hx in the binary complex with the calf spleen enzyme [36], and for the ternary complex bovine-PNP/immucillin-H/Pi [34]. By contrast, in E. coli PNP-I, Asn is replaced by Asp204, so that binding of the ligand is additionally dependent on ionization of the Asp side chain, which would then electrostatically repel the the site of anions of Xan and Gua, dissociation in these purines. Furthermore a bridging water molecule is not observed, e.g. in the formycin B complex with E. coli PNP-I [37], which in the case of mammalian enzymes may be involved in enzyme-ligand binding and/or the enzymatic reaction.

The above suggests that, for binding of Xan and Xao in the active sites of the mammalian enzymes (and E. coli PNP-II), the absence of dissociation of the N(1)-H is of key importance for the substrate properties of their mono- anions, inasmuch as dissociation of the N(3)-H still permits interaction of the N(1)-H with either the neutral or anionic forms of Glu201 carboxyl (Fig. 4), and hence their substrate properties, as observed. Dissociation of these protons reduced substrate activity at slightly alkaline pH, similarly to that observed at pH < 5, where protonation of His64 also led to a significantly reduced enzyme efficiency (Fig. 2). The proposed modes of binding of Xan by calf and human enzymes should also incorporate data showing that, in aqueous medium, Xan exists as a mixture of the N(7)-H and N(9)-H tautomeric forms [13]. This is often overlooked in analysis of binding and reverse reactions with Gua and Hx, with only the N(9)-H tautomer taken into account, because of its structural similarity to natural purine nucleosides. One possible mode of binding of the N(9)-H form of Xan is similar to that shown for binding of Xao (Fig. 4), originally proposed by Mao et al. [33] for Ino and sulfate in the active site of bovine spleen PNP, although in their PDB entry (1A9S), Asn243 is rotated in such a way that Asn243Nd donates a hydrogen to O6 of the base. Involvement of O6 in binding to the enzyme from Cellulo-

Fig. 4. Proposed models of binding by mam- malian PNPs of the neutral (A, B, D, E) and anionic (C, F) forms of xanthine (A–C) and xanthosine (D–F), based on the enzyme–ligand interactions observed in the crystal structures of immucillin-H [34] and hypoxanthine [33,36] with calf spleen PNP, and 5¢-iodo-9-deazaino- sine with human PNP [35]. Note proposed binding with the neutral (A, D) and anionic (B, C, E, F) forms of the Glu201 carboxylate. See text for further details.

Xanth(os)ine and purine nucleoside phosphorylases (Eur. J. Biochem. 269) 4055

(cid:2) FEBS 2002

In line with the above, we suggest that both the neutral and anionic forms of the Glu201 carboxyl hydrogen bonds the N(1)-H of Xan and Xao, irrespective of the ionization of N(3)-H (Fig. 4), and together with the interactions main- tained by Asn243, play a key role in transition state formation, as well as in phosphorolysis of Xao and the reverse reaction with Xan.

Relevance to other enzyme systems

The foregoing would be incomplete without at least passing reference to current efforts to develop noncanon- ical base pairs for replication [54] and transcription [55,56]. The 5¢-triphosphates of Xao (XTP and dXTP) have been widely employed for this purpose, and are complementarily incorporated into RNA and DNA by some polymerases with moderate selectivity. However, the Xan moiety has been assumed to be in the neutral 2,6- diketo form, notwithstanding that it was shown long ago that poly(xanthylate) forms multistranded helices with different structures in acid and alkaline media, related to dissociation of the N(3)-H of Xao [57,58]. This is further confirmed by X-ray diffraction of fibers, which form one helix at acid pH with the Xan residues in the neutral form, and another at pH 8 with dissociation of the N(3)- H of the Xan residues [59].

A C K N O W L E D G E M E N T S

The pKa values, and unique structures of the monoanions of Xao and Xan, as well as of XMP [12], are of equal relevance in other enzyme systems, for which they are substrates or intermediates. One case in point is IMP dehydrogenase, the rate-limiting enzyme in the de novo synthesis of guanine nucleotides, which catalyses the NAD- dependent oxidation of IMP to XMP, which is then converted to GMP [44]. A novel nucleoside triphosphate pyrophosphatase from the thermophilic Methanococcus jannaschii has been reported as highly specific for the noncanonical nucleotides ITP and XTP, even at high alkaline pH [45], where both exist exclusively as mono- anions, albeit with different structures.

The biosynthesis of caffeine, recently extensively reviewed by Ashihara & Crozier [46], proceeds through a number of enzymatic steps involving as intermediates Xan and Xao, XMP and m7XMP, followed by cleavage of the latter and stepwise methylation of the liberated N(7)-methylxanthine to give N(1),N(3),N(7)-trimethylxanthine (caffeine).

We are indebted to Prof Wolfgang Pfleiderer (University of Konstanz, Germany) for several authentic samples of N-methyl xanthines. This investigation was supported by the State Committee for Scientific Research (KBN, Grant no. 6P04A03812, and partially BW 1482/BF and BST 661/BF); and by an International Research Scholar’s award of the Howard Hughes Medical Institute (Grant No. HHMI 75195– is also indebted for support to KBN (Grant no. 543401). G. S. 6PO4A03813), and to the Fellowship Program of the Institute of Biochemistry and Biophysics.

R E F E R E N C E S

the

purine

relevant

Particularly

1. Buxton, R.S., Hammer-Jespersen, K. & Valentin-Hansen, P. (1980) A second purine nucleoside phosphorylase in Escherichia coli K-12. I. Xanthosine phosphorylase regulatory mutants iso- lated as secondary-site revertants of a deoD mutant. Mol. Gen. Genet. 179, 331–340.

2. Hammer-Jespersen, K., Buxton, R.S. & Hansen, T.D. (1980) A second purine nucleoside phosphorylase in Escherichia coli K-12. II. Properties of xanthosine phosphorylase and its induction by xanthosine. Mol. Gen. Genet. 179, 341–348.

phospho- are ribosyltransferases, which function in the salvage pathway by addition of a preformed purine base to the a–carbon of a–D-phosphoribosylpyrophosphate to generate purine nucleotides. These include a family of enzymes specific for 6-oxopurines. The human enzyme accepts Hx and Gua, but only very minimally Xan [47,48]. E. coli contains two such enzymes, one with a preference for Hx, the other for Gua and Xan [49]. Parasitic protozoa, which are incapable of de novo synthesis of purine nucleotides, express a unique complement of purine salvage enzymes; for example Leishmania donovani possesses one such enzyme with a marked preference for Xan [50,51].

3. Koszalka, G.W., Vanhooke, J., Short, S.A. & Hall, W.W. (1988) Purification and properties of inosine-guanosine phosphorylase from Escherichia coli K-12. J. Bacteriol. 170, 3493–3498.

4. Parks, E. Jr & Agrawal, R.B. (1972) Purine nucleoside phos- phorylase. In The Enzymes, 7 (Boyer, P.D., ed.), pp. 483–514. Academic Press, New York.

5. Stoeckler, J.D. (1984) Purine nucleoside phosphorylase: a target for chemotherapy. In Developments in Cancer Chemotherapy (Glazer, R.I., ed.), pp. 35–60. CRC Press, Boca Raton.

6. Bzowska, A., Kulikowska, E. & Shugar, D. (2000) Purine nucleoside phosphorylases: properties, functions, and clinical aspects. Pharmacol. Ther. 88, 349–425.

The X-ray structures of many of these enzymes in complexes with 6-oxopurines or their nucleotides, including Xan and XMP, have been reported, in all instances with the explicit assumption that the xanthine moiety is uniquely 2,6- dioxo. However, whereas the modes of binding of Hx and Gua in the forward reaction, and IMP and GMP in the reverse reaction, have been reasonably well assigned, the situation is less clear for the modes of binding of Xan and XMP [49]. It is conceivable that, in the crystal structures, dissociation of the xanthine N(3)-H is blocked. However, the resolution of the crystal structures is insufficient to distinguish between a C¼O and C-O–.

7. Friedkin, M. (1952) Enzymatic synthesis of desoxyxanthosine by the action xanthosine phosphorylase in mammalian tissue. J. Am. Chem. Soc. 74, 112–115.

8. Giorgelli, F., Bottai, C., Mascia, L., Scolozzi, C., Camici, M. & Ipata, P.L. (1997) Recycling of alpha-D-ribose 1-phosphate for nucleoside interconversion. Biochim. Biophys. Acta 1335, 6–22. 9. Korn, E.D. & Buchanan, J.M. (1955) Biosynthesis of the purines VI. J. Biol. Chem. 217, 183–191.

10. Tsuboi, K.K. & Hudson, P.B. (1957) Enzymes of the human specific II. Purine nucleoside phosphorylase, erythrocytes. properties. J. Biol. Chem. 224, 889–897.

We are aware of only one enzyme system, xanthine oxidase, where attention was directed to possible substrate properties of the monoanion. On the basis of kinetic and pH-titration studies, it was proposed that the neutral forms of Xan [52] and 1-Me-Xan [53] are the preferred substrates, but with the erroneous assumption that monoanion formation involves dissociation of an imidazole proton. These findings do not unequivocally exclude weak substrate properties of the monoanions.

11. Cavalieri, L., Fox, J., Stone, A. & Chang, N. (1954) On the nature of xanthine and substituted xanthines in solution. J. Am. Chem. Soc. 76, 1119–1122.

4056 G. Stoychev et al. (Eur. J. Biochem. 269)

(cid:2) FEBS 2002

12. Roy, K.B. & Miles, H.T. (1983) Tautomerism and ionization of studies on some sulfur-containing ribonucleosides. J. Phys. Chem. 73, 4218–4227. Xanthosine. Nucleosides Nucleotides 2, 231–242.

13. Shugar, D. & Psoda, A. (1990) Numerical data and functional relationships in science and technology. In Biophysics of Nucleic Acids, Landoldt-Bo¨rnstein New Series VII/Id. (Saenger, W., ed.), pp. 308–348. Springer-Verlag, Berlin. 31. Stoeckler, J.D., Poirot, A.F., Smith, R.M., Parks Jr, R.E., Ealick, S.E., Takabayashi, K. & Erion, M.D. (1997) Purine nucleoside phosphorylase. 3. Reversal of purine base specificity by site- directed mutagenesis. Biochemistry 36, 11749–11756.

14. Mizuno, H., Fujiwara, T. & Tomita, K. (1969) The crystal and molecular structure of the sodium salt of xanthine. Bull. Chem. Soc. Japan 42, 3099–3105. 32. Krenitsky, T.A., Elion, G.B., Henderson, A.M. & Hitchings, G.H. (1968) Inhibition of human purine nucleoside phosphorylase. Studies with intact erythrocytes and the purified enzyme. J. Biol. Chem. 243, 4779–4784.

15. Kierdaszuk, B., Modrak-Wo´ jcik, A. & Shugar, D. (1997) Binding of phosphate and sulfate anions by purine nucleoside phospho- rylase from E. coli: ligand-dependent quenching of enzyme in- trinsic fluorescence. Biophys. Chem. 63, 107–118. 33. Mao, C., Cook, W.J., Zhou, M., Federov, A.A., Almo, S.C. & Ealick, S.E. (1998) Calf spleen purine nucleoside phosphorylase complexed with substrates and substrate analogues. Biochemistry 37, 7135–7146.

16. Lichtenberg, D., Bergman, F. & Neimann, Z. (1971) Tautomeric forms and ionisation processes in xanthine and its N-methyl derivatives. J. Chem. Soc. C 1971, 1676–1682.

34. Fedorov, A., Shi, W., Kicska, G., Fedorov, E., Tyler, P.C., Furneaux, R.H., Hanson, J.C., Gainsford, G.J., Larese, J.Z., Schramm, V.L. & Almo, S.C. (2001) Transition state structure of purine nucleoside phosphorylase and principles of atomic motion in enzymatic catalysis. Biochemistry 40, 853–860. 17. Twanmoh, L.M., Wood, H.B. Jr & Driscoll, J.S. (1973) NMR spectral characteristics of N-H protons in purine derivatives. J. Heterocycl. Chem. 10, 187–190.

18. Ropp, P.A. & Traut, T.W. (1991) Purine nucleoside phospho- rylase. Allosteric regulation of a dissociating enzyme. J. Biol. Chem. 266, 7682–7687.

35. Ealick, S.E., Babu, Y.S., Bugg, C.E., Erion, M.D., Guida, W.C., Montgomery, J.A. & Secrist 3rd, J.A. (1991) Application of crystallographic and modeling methods in the design of purine nucleoside phosphorylase inhibitors. Proc. Natl Acad. Sci. USA 88, 11540–11544.

to enzyme,

36. Koellner, G., Luic, M., Shugar, D., Saenger, W. & Bzowska, A. (1997) Crystal structure of calf spleen purine nucleoside phos- phorylase in a complex with hypoxanthine at 2.15 A˚ resolution. J. Mol. Biol. 265, 202–216. 19. Kierdaszuk, B., Modrak-Wo´ jcik, A., Wierzchowski, J. & Shugar, D. (2000) Formycin A and its N-methyl analogues, specific inhibitors of E. coli purine nucleoside phosphorylase (PNP): induced tautomeric and shifts on binding enzyme fi ligand fluorescence resonance energy transfer. Biochim. Biophys. Acta 1476, 109–128. 20. Krenitsky, T.A. (1967) Purine nucleoside phosphorylase: kinetics, mechanism, and specificity. Mol. Pharm. 3, 526–536.

37. Koellner, G., Luic, M., Shugar, D., Saenger, W. & Bzowska, A. (1998) Crystal structure of the ternary complex of E. coli purine nucleoside phosphorylase with formycin B, a structural analogue of the substrate inosine, and phosphate (Sulphate) at 2.1 A˚ resolution. J. Mol. Biol. 280, 153–166. 21. Kline, P.C. & Schramm, V.L. (1993) Purine nucleoside phos- phorylase. Catalytic mechanism and transition-state analysis of the arsenolysis reaction. Biochemistry 32, 13212–13219.

38. Erion, M.D., Takabayashi, K., Smith, H.B., Kessi, J., Wagner, S., Honger, S., Shames, S.L. & Ealick, S.E. (1997) Purine nucleoside phosphorylase. 1. Structure–function studies. Biochemistry 36, 11725–11734. 22. Bezirjian, KhO., Kocharian, Sh.M. & Akopyan, Zh.I. (1986) Isolation of a hexameric form of purine nucleoside phosphorylase II from E. coli, Comparative study of trimeric and hexameric forms of the enzyme. Biokhimia 51, 1085–1092.

39. Erion, M.D., Stoeckler, J.D., Guida, W.C., Walter, R.L. & Ealick, (1997) Purine nucleoside phosphorylase. 2. Catalytic S.E. mechanism. Biochemistry 36, 11735–11748. 23. Stoychev, G., Kiedaszuk, B. & Shugar, D. (2001) Interaction of Escherichia coli purine nucleoside phosphorylase (PNP) with the cationic and zwitterionic forms of the fluorescent substrate N(7)- methylguanosine. Biochim. Biophys. Acta 1544, 74–88.

24. Wielgus-Kutrowska, B., Kulikowska, E., Wierzchowski, J., Bzowska, A. & Shugar, D. (1997) Nicotinamide riboside, an unusual, non-typical, substrate of purified purine-nucleoside phosphorylases. Eur. J. Biochem. 243, 408–414. 40. Montgomery, J.A. (1993) Structure-based drug design: inhibitors of purine nucleoside phosphorylase. Farmaco. 48, 297–308. 41. Farutin, V., Masterson, L., Andrcopulo, A.D., Cheng, J., Riley, B., Hakimi, R., Frazer, J.W. & Cordes, E.H. (1999) Structure– activity relationship for a class of inhibitors of purine nucleoside phosphorylase. J. Med. Chem. 42, 2422–2431.

25. Civcir, P.U. (2001) Tautomerism of 6-thioxanthine in gas and aqueous phases using AM1 and PM3 methods. J. Chem. Phys. 114, 1582–1588.

42. Tebbe, J., Bzowska, A., Wielgus-Kutrowska, B., Schro¨ der, W., Kazimierczuk, Z., Shugar, D., Saenger, W. & Koellner, G. (1999) Crystal structure of the purine nucleoside phosphorylase (PNP) from Cellulomonas sp. and its implication for the mechanism of trimeric PNPs. J. Mol. Biol. 294, 1239–1255. 26. Civcir, P.U. (2001) Tautomerism of 2-thioxanthine in the gas and aqueous phases using AM1 and PM3 methods. J. Mol. Struct. (Teochem.) 546, 163–173.

43. Montgomery, J.A., Niwas, S., Rose, J.D., Secrist, J.A., 3rd, Babu, Y.S., Bugg, C.E., Erion, M.D., Guida, W.C. & Ealick, S.E. (1993) Structure-based design of inhibitors of purine nucleoside phos- phorylase. 1. 9-(arylmethyl) derivatives of 9-deazaguanine. J. Med. Chem. 36, 55–69. 27. Keuzenkamp-Jansen, C.W., De Abreu, R.A., Bokkerink, J.P., Lambooy, M.A. & Trijbels, J.M. (1996) Metabolism of intravenously administered high-dose 6-mercaptopurine with and without allopurinol treatment in patients with non-Hodgkin lymphoma. J. Pediatric Hematol./Oncol. 18, 145–150.

44. Franchetti, P. & Grifantini, M. (1999) Nucleoside and non- nucleoside IMP dehydrogenase inhibitors as antitumor and antiviral agents. Curr. Med. Chem. 6, 599–614. 28. Tamiya, T., Ono, Y., Wei, M.X., Mroz, P.J., Moolten, F.L. & Chiocca, E.A. (1996) Escherichia coli gpt gene sensitizes rat glioma cells to killing by 6-thioxanthine or 6-thioguanine. Cancer Gene Ther. 3, 155–162.

45. Chung, J.H., Back, J.H., Park, Y.I. & Han, Y.S. (2001) Bio- chemical characterization of a novel hypoxanthine/xanthine dNTP pyrophosphatase from Methanococcus jannaschii. Nucleic Acids Res. 29, 3099–3107. 29. Ono, Y., Ikeda, K., Wei, M.X., Harsh 4th, G.R., Tamiya, T. & Chiocca, E.A. (1997) Regression of experimental brain tumors with 6-thioxanthine and Escherichia coli gpt gene therapy. Human Gene Ther. 8, 2043–2055.

30. Cheong, K.K., Fu, Y.Ch, Robins, R.K. & Eyring, H. (1969) Absorption, optical rotatory dispersion, and circular dichroism 46. Ashihara, H. & Crozier, A. (1999) Biosynthesis and catabolism of caffeine in low-caffeine-containing species of Coffea. J. Agric. Food Chem. 47, 3425–3431.

Xanth(os)ine and purine nucleoside phosphorylases (Eur. J. Biochem. 269) 4057

(cid:2) FEBS 2002

analog uridine triphosphates as candidates for in vitro selection of nucleic acids. Nucleic Acids Res. 28, 3316–3322.

47. Vos, S., Parry, R.J., Burns, M.R., de Jersey, J. & Martin, J.L. (1998) Structures of free and complexed forms of Escherichia coli xanthine-guanine phosphoribosyltransferase. J. Mol. Biol. 282, 875–889. 56. Ohtsuki, T., Kimoto, M., Ishikawa, M., Mitsui, T., Hirao, I. & Yokoyama, S. (2001) Unnatural base pairs for specific transcrip- tion. Proc. Natl Acad. Sci. USA 98, 9422–9425.

57. Fikus, M. & Shugar, D. (1969) Properties of poly-xanthylic acid reactions with potentially complementary homo- 48. Shivashankar, K., Subbayya, I.N. & Balaram, H. (2001) Devel- opment of a bacterial screen for novel Hx-Gua PRTase substrates. J. Mol. Microbiol. Biotechnol. 3, 557–562. and its polynucleotides. Acta Biochim. Polon. 16, 55–84.

49. Vos, S., de Jersey, J. & Martin, J.L. (1997) Crystal structure of Escherichia coli xanthine phosphoribosyltransferase. Biochemistry 36, 4125–4134. 50. Craig 3rd, S.P. & Eakin, A.E. (2000) Purine phosphoribosyl- transferases. J. Biol. Chem. 275, 20231–20234. 58. Roy, K.B., Frazier, J. & Miles, H.T. (1979) Polixanthylic acid: Structures of the ordered forms. Biopolymers 18, 3077–3087. 59. Arnott, S., Chandrasekaran, R., Day, A.W., Puigjaner, L.C. & Watts, L. (1981) Double-helical structures for polyxanthylic acid. J. Mol. Biol. 149, 489–505.

51. Jardim, A., Bergeson, S.E., Shih, S., Carter, N., Lucas, R.W., Merlin, G., Myler, P.J., Stuart, K. & Ullman, B. (1999) Xanthine phosphoribosyltransferase from Leishmania donovani. Molecular cloning, biochemical characterization, and genetic analysis. J. Biol. Chem. 274, 34403–34410.

52. Kim, J.H., Ryan, M.G., Knaut, H. & Hille, R. (1996) The reductive half-reaction of xanthine oxidase. The involvement of prototropic equilibria in the course of the catalytic sequence. J. Biol. Chem. 271, 6771–6780. 60. Pfleiderer, W. & Nu¨ bel, G. (1961) Zur struktur des xanthins und seiner N-methyl-xanthin-derivate. Ann. Chem. 647, 155–160. 61. Bzowska, A., Ananiev, A.V., Ramzaeva, N., Alksins, E., Maurins, J.A., Kulikowska, E. & Shugar, D. (1994) Purine nucleoside phosphorylase: inhibition by purine N(7)- and N(9)-acyclonu- cleosides; and substrate properties of 7-beta-D-ribofurano- sylguanine and 7-beta-D-ribofuranosylhypoxanthine. Biochem. Pharmacol. 48, 937–947.

62. Porter, D.J. (1992) Purine nucleoside phosphorylase. Kinetic mechanism of the enzyme from calf spleen. J. Biol. Chem. 26, 7342–7351.

53. Sau, A.K., Mondal, M.S. & Mitra, S. (2000) Effects of pH and temperature on the reaction of milk Xanthine oxidase with 1-methylxanthine. J. Chem. Soc. Dalton Trans. 2000, 3688–3692. 54. Horlacher, J., Hottiger, M., Podust, V.M., Hubscher, U. & Benner, S.A. (1995) Recognition by viral and cellular DNA polymerases of nucleosides bearing bases with nonstandard hydrogen bonding patterns. Proc. Natl Acad. Sci. USA 92, 6329–6333. 63. Wierzchowski, J., Wielgus-Kutrowska, B. & Shugar, D. (1996) Fluorescence emission properties of 8-azapurines and their nucleosides, and application to the kinetics of the reverse synthetic reaction of purine nucleoside phosphorylase. Biochim. Biophys. Acta 1290, 9–17.

55. Vaish, N.K., Fraley, A.W., Szostak, J.W. & McLaughlin, L.W. (2000) Expanding the structural and functional diversity of RNA: