Portal de Periódicos da CAPES

An Extremely Potent Inhibitor of Xanthine Oxidoreductase

2003; Elsevier BV; Volume: 278; Issue: 3 Linguagem: Inglês

10.1074/jbc.m208307200

1083-351X

Ken Okamoto, B.T. Eger, Tomoko Nishino, Shiro Kondo, E.F. Pai, Takeshi Nishino,

Metalloenzymes and iron-sulfur proteins

TEI-6720 (2-(3-cyano-4-isobutoxyphenyl)-4-methyl-5-thiazolecarboxylic acid) is an extremely potent inhibitor of xanthine oxidoreductase. Steady state kinetics measurements exhibit mixed type inhibition withK i and K i′ values of 1.2 ± 0.05 × 10−10m and 9 ± 0.05 × 10−10m, respectively. Fluorescence-monitored titration experiments showed that TEI-6720 bound very tightly to both the active and the inactive desulfo-form of the enzyme. The dissociation constant determined for the desulfo-form was 2 ± 0.03 × 10−9m; for the active form, the corresponding number was too low to allow accurate measurements. The crystal structure of the active sulfo-form of milk xanthine dehydrogenase complexed with TEI-6720 and determined at 2.8-Å resolution revealed the inhibitor molecule bound in a long, narrow channel leading to the molybdenum-pterin active site of the enzyme. It filled up most of the channel and the immediate environment of the cofactor, very effectively inhibiting the activity of the enzyme through the prevention of substrate binding. Although the inhibitor did not directly coordinate to the molybdenum ion, numerous hydrogen bonds as well as hydrophobic interactions with the protein matrix were observed, most of which are also used in substrate recognition. TEI-6720 (2-(3-cyano-4-isobutoxyphenyl)-4-methyl-5-thiazolecarboxylic acid) is an extremely potent inhibitor of xanthine oxidoreductase. Steady state kinetics measurements exhibit mixed type inhibition withK i and K i′ values of 1.2 ± 0.05 × 10−10m and 9 ± 0.05 × 10−10m, respectively. Fluorescence-monitored titration experiments showed that TEI-6720 bound very tightly to both the active and the inactive desulfo-form of the enzyme. The dissociation constant determined for the desulfo-form was 2 ± 0.03 × 10−9m; for the active form, the corresponding number was too low to allow accurate measurements. The crystal structure of the active sulfo-form of milk xanthine dehydrogenase complexed with TEI-6720 and determined at 2.8-Å resolution revealed the inhibitor molecule bound in a long, narrow channel leading to the molybdenum-pterin active site of the enzyme. It filled up most of the channel and the immediate environment of the cofactor, very effectively inhibiting the activity of the enzyme through the prevention of substrate binding. Although the inhibitor did not directly coordinate to the molybdenum ion, numerous hydrogen bonds as well as hydrophobic interactions with the protein matrix were observed, most of which are also used in substrate recognition. xanthine oxidoreductase xanthine dehydrogenase xanthine oxidase 2-(3-cyano-4-isobutoxyphenyl)-4-methyl-5-thiazolecarboxylic acid sodium-8-(3-methoxy-4-phenylsulfonylphenyl)pyrazolo[1,5-a]-1,3,5-triazine-4-olate monohydrate activity to flavin ratio phenazine methosulfate molybdenum-pterin Xanthine oxidoreductase (XOR)1 enzymes have been isolated from a wide range of organisms, from bacteria to man, and they accelerate the hydroxylation of a wide variety of purine, pyrimidine, pterin, and aldehyde substrates. All of these proteins have similar molecular weights and composition of redox centers (1Nishino T. J. Biochem. (Tokyo). 1994; 116: 1-6Crossref PubMed Scopus (189) Google Scholar, 2Hille R. Nishino T. FASEB J. 1995; 9: 995-1003Crossref PubMed Scopus (378) Google Scholar). In humans, the enzyme catalyzes the last two steps of purine catabolism, the oxidation of hypoxanthine to xanthine and of xanthine to uric acid. This reaction occurs at a molybdenum-pterin center and from there the electrons are transferred via two Fe2S2clusters to the isoalloxazine ring of FAD, which then passes them on to the second substrate NAD+ (1Nishino T. J. Biochem. (Tokyo). 1994; 116: 1-6Crossref PubMed Scopus (189) Google Scholar, 2Hille R. Nishino T. FASEB J. 1995; 9: 995-1003Crossref PubMed Scopus (378) Google Scholar, 3Bray R.C. Boyer P.D. The Enzymes XII. Academic Press, New York1975: 300-419Google Scholar, 4Hille R. Massey V. Spiro T.G. Molybdenum Enzymes. Wiley Interscience, New York1985: 443-518Google Scholar, 5Hille R. Chem. Rev. 1996; 96: 2757-2816Crossref PubMed Scopus (1475) Google Scholar). XOR is synthesized as xanthine dehydrogenase (XDH; EC 1.1.1.204) with very low reactivity toward molecular oxygen but high reactivity toward NAD+ (6Stirpe F. Della Corte E. J. Biol. Chem. 1969; 244: 3855-3863Abstract Full Text PDF PubMed Google Scholar, 7Della Corte E. Stirpe F. Biochem. J. 1972; 126: 739-745Crossref PubMed Scopus (349) Google Scholar). In mammals, however, XDH can easily be converted to xanthine oxidase (XO; EC 1.1.3.22), which does not interact with NAD+ but is very efficient in producing superoxide anion (O2⨪) and H2O2instead. The conversion is initiated either by formation of intramolecular disulfide bonds or by proteolytic cleavage of a loop region connecting the FAD-binding domain and the molybdenum-binding domain (8Nishino T. Nishino T. J. Biol. Chem. 1997; 272: 29859-29864Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar). This conversion has been implicated in O2⨪-mediated stress phenomena such as postischemic reperfusion injury (9McCord J.M. N. Engl. J. Med. 1985; 312: 159-163Crossref PubMed Scopus (4980) Google Scholar). In an effort to elucidate the structural basis for these effects, we recently determined the crystal structures of both the XDH and XO forms of the bovine milk enzyme, a very close homologue of the human enzyme, at 2.1- and 2.5-Å resolutions, respectively. These analyses showed that structural rearrangements next to the FAD cause this remarkable change of reactivity (10Enroth C. Eger B.T. Okamoto K. Nishino T. Pai E.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 10723-10728Crossref PubMed Scopus (583) Google Scholar). Allopurinol (Fig. 1 A), an analogue of hypoxanthine, was developed by Elion et al. (11Elion G.B. Callahan S. Nathan H. Bieber S. Rundles R.W. Hitching G.H. Biochem. Pharmacol. 1963; 12: 85-93Crossref Scopus (193) Google Scholar) as an XDH inhibitor 30 years ago and has been widely prescribed as a treatment of hyperuricemia and gout since (12Reginato A.J. Braunwald E. Harrison's Principles of Internal Medicine. 15th Ed. McGraw-Hill, Columbus, OH2001: 1994-1998Google Scholar). In addition, administration of allopurinol has been reported to prevent postischemic tissue damage by inhibiting XO activity (9McCord J.M. N. Engl. J. Med. 1985; 312: 159-163Crossref PubMed Scopus (4980) Google Scholar). In some cases, however, severe life-threatening side effects have been reported, such as a toxicity syndrome dramatized by eosinophila, vasculitits, rash hepatitis, and progressive renal failure (12Reginato A.J. Braunwald E. Harrison's Principles of Internal Medicine. 15th Ed. McGraw-Hill, Columbus, OH2001: 1994-1998Google Scholar). The intrinsic radical-scavenging features of allopurinol (13Moorhouse P.C. Grootveld M. Halliwell B. Quinlan J.G. Gutteridge J.M.C. FEBS Lett. 1987; 213: 23-28Crossref PubMed Scopus (352) Google Scholar) make it difficult to distinguish between the effects it causes directly and the effects produced by XOR inhibition, e.g. in the production of radical species in reperfusion injury. As oxypurinol, the oxidation product of allopurinol, is the actual inhibiting species (it coordinates to the reduced molybdenum center of XOR), a lag phase for complete inhibition is to be expected. Once inhibited, the enzyme can also be reactivated by spontaneous reoxidation of the metal cofactor (t 12 = 300 min at 25 °C) (14Massey V. Komai H. Palmer G. Elion G.B. J. Biol. Chem. 1970; 245: 2837-2844Abstract Full Text PDF PubMed Google Scholar). This feature requires the administration of at least three relatively high doses of the drug per day to keep the plasma level of the drug at an effective concentration. Because of these shortcomings, new potent inhibitors, preferentially those with well defined inhibition mechanisms, are still most useful both under clinical and scientific experimental aspects. A potential single-dose, low-concentration regiment would be highly welcomed by patients and physicians but the varied features of inhibition could also provide the biochemically inclined experimentalist with a more refined knowledge of the character of the active site of the enzyme and its exact chemical mechanism (15Massey V. Edmondson D. J. Biol. Chem. 1970; 245: 6595-6598Abstract Full Text PDF PubMed Google Scholar). We have undertaken a series of investigations aimed at providing more information about the kinetic and structural properties of recently developed XOR inhibitors. As a first step into this direction, we described the inhibition mechanism of a newly introduced inhibitor, BOF-4272 (sodium-8-(3-methoxy-4-phenylsulfinyl-phenyl) pyrazolo[1,5-a]-1,3,5-triazine-4-olate monohydrate) (16Okamoto K. Nishino T. J. Biol. Chem. 1995; 270: 7816-7821Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar), which has been tested in animal studies and in in vitroexperiments as an agent to specifically inhibit XOR-based superoxide generation (17Sanders S.A. Eisenthal R. Harrison R. Eur. J. Biochem. 1997; 245: 541-548Crossref PubMed Scopus (145) Google Scholar, 18Matsumura F. Yamaguchi Y. Goto M. Ichiguchi O. Akizuki E. Matsuda T. Okabe K. Liang J. Ohshiro H. Iwamoto T. Yamada S. Mori K. Ogawa M. Hepatology. 1998; 28: 1578-1587Crossref PubMed Scopus (36) Google Scholar, 19Millar T.M. Stevens C.R. Benjamin N. Eisenthal R. Harrison R. Blake D.R. FEBS Lett. 1998; 427: 225-228Crossref PubMed Scopus (410) Google Scholar, 20Suzuki H. DeLano F.A. Parks D.A. Jamshidi N. Granger D.N. Ishii H. Suematsu M. Zweifach B.W. Schmid-Schonbein G.W. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 4754-4759Crossref PubMed Scopus (223) Google Scholar). TEI-6720 (2-(3-cyano-4-isobutoxyphenyl)-4-methyl-5-thiazolecarboxylic acid) (Fig. 1 B) is another recently developed inhibitor of XOR. When tested in rats and chimpanzees, its inhibition of uric acid production in vivo was stronger and lasted longer than that of allopurinol, without causing any noticeable side effects (21Osada Y. Tsuchimoto M. Fukushima H. Takahashi K. Kondo S. Hasegawa M. Komoriya K. Eur. J. Pharmacol. 1993; 241: 183-188Crossref PubMed Scopus (157) Google Scholar, 22Komoriya K. Osada Y. Hasegawa M. Horiuchi H. Kondo S. Couch R.C. Griffin T.B. Eur. J. Pharmacol. 1993; 250: 455-460Crossref PubMed Scopus (63) Google Scholar). The molecular structure of this inhibitor is quite dissimilar to that of substrates like xanthine or hypoxanthine, strongly suggesting that its mode of action will be different from that of allopurinol. In an effort to characterize in detail the way TEI-6720 interacts with its target, we investigated its inhibitory mechanism based on kinetic measurements and determined the crystal structure of the complex formed by TEI-6720 with XDH at 2.8-Å resolution. Bovine milk XO was purified using the method of Ball (23Ball E.G. J. Biol. Chem. 1939; 128: 51-67Abstract Full Text PDF Google Scholar) and the fully active form was obtained by further purification following the procedure described by Nishino et al. (24Nishino T. Nishino T. Tsushima K. FEBS Lett. 1981; 131: 369-372Crossref PubMed Scopus (70) Google Scholar). XO prepared in this way routinely exhibits an activity:flavin ratio (AFR) 2AFR is enzyme activity defined as the absorbance change/min at 295 nm (monitoring conversion of xanthine to uric acid) divided by the enzyme absorbance at 450 nm under standard assay conditions.of 200 at 25 °C indicating that more than 95% of the protein sample is in the active form (24Nishino T. Nishino T. Tsushima K. FEBS Lett. 1981; 131: 369-372Crossref PubMed Scopus (70) Google Scholar, 25Massey V. Komai H. Palmer G. Elion G.B. Vitam. Horm. 1970; 28: 505-531Crossref PubMed Scopus (24) Google Scholar). The XDH form of the enzyme was prepared according to Eger et al. (26Eger B.T. Okamoto K. Enroth C. Sato M. Nishino T. Pai E.F. Nishino T. Acta Crystallogr. Sect. D Biol. Crystallogr. 2000; 56: 1656-1658Crossref PubMed Scopus (29) Google Scholar). Purified enzyme was stored on ice without freezing in a solution containing 20 mmpyrophosphate buffer (pH 8.5), 40 mm Tris-HCl buffer (pH 7.8), 1 mm salicylate, and 0.2 mm EDTA. The desulfo-form of XO was prepared by incubating the enzyme in storage buffer containing 10 mm KCN for 2 h at 25 °C immediately followed by gel filtration to remove excess cyanide (15Massey V. Edmondson D. J. Biol. Chem. 1970; 245: 6595-6598Abstract Full Text PDF PubMed Google Scholar). The concentration of XO and XDH was determined by spectrophotometry using a molar extinction coefficient of 37,800 at 450 nm (27Massey V. Brumby P.E. Komai H. J. Biol. Chem. 1969; 244: 1682-1691Abstract Full Text PDF PubMed Google Scholar). The inhibitor TEI-6720 was provided by Teijin Co., Tokyo, Japan. All other reagents were of the highest purity commercially available. Before crystallization, XDH was passed through a Sephadex G-25 column (Amersham Biosciences AB) to remove salicylate. The eluate was then brought to a concentration of about 75 mg/ml using YM-100 concentrators (Amicon). Crystals could be grown under conditions very similar to the ones described by Eger et al. (26Eger B.T. Okamoto K. Enroth C. Sato M. Nishino T. Pai E.F. Nishino T. Acta Crystallogr. Sect. D Biol. Crystallogr. 2000; 56: 1656-1658Crossref PubMed Scopus (29) Google Scholar),i.e. crystallization was carried out at 20 °C employing an enzyme concentration of 7.5 mg/ml in a solution containing 50 mm potassium phosphate buffer (pH 6.5), 5 mmdithiothreitol, 1 mm salicylate, 0.2 mm EDTA, 30% glycerol, 0.5 mm TEI-6720 as well as 6–10% (w/v) PEG 4,000 as precipitant. Crystals of the enzyme-inhibitor complex were flash-frozen with their mother liquor as cryoprotectant and mounted in cryoloops. Diffraction data were collected at beamline BL40B2, SPring8, Harima Garden City, Japan; a temperature of 100 K, radiation of 1.00 Å wavelength, and a Q4 area detector (ACSD) were used. Data were reduced with the help of the program package DENZO and scaled using SCALEPACK (28Otwinoski Z. Minor W. Methods Enzymol. 1997; 276: 307-326Crossref Scopus (38526) Google Scholar). The program package EPMR (29Kissinger C.R. Gehlhaar D.K. Fogel D.B. Acta Crystallogr. Sect. D Biol. Crystallogr. 1999; 55: 484-491Crossref PubMed Scopus (690) Google Scholar) established the correct solutions of the respective molecular replacement function (20.0 to 4.0-Å resolution range). One subunit of bovine milk XDH (Protein Data Bank code 1FO4) without its cofactors was employed as the search model. The molecular models were built with the help of the program package O (30Jones T.A. Zou J.Y. Cowan S.W. Kjeldgaard M. Acta Crystallogr. Sect. A. 1991; 47: 110-119Crossref PubMed Scopus (13009) Google Scholar). Subsequent refinement, including rigid body, simulated annealing, grouped B factors, and least square minimization were carried out with CNS, version 1.0 (31Brunger A.T. Adams P.D. Clore G.M. DeLano W.L. Gros P. Grosse-Kunstleve R.W. Jiang J.S. Kuszewski J. Nilges M. Pannu N.S. Read R.J. Rice L.M. Simonson T. Warren G.L. Acta Crystallogr. D Biol. Crystallogr. 1998; 54: 905-921Crossref PubMed Scopus (16957) Google Scholar). TheR-free reflection set was chosen in random resolution shells using the DATAMAN program from CCP4 (32Kleywegt G.J. Jones T.A. Acta Crystallogr. D Biol. Crystallogr. 1996; 52: 826-828Crossref PubMed Scopus (334) Google Scholar). No NCS constraints were used in the final round of refinement (TableI). Figures were generated with MOLSCRIPT (33Kraulis P. J. Appl. Crystallogr. 1991; 24: 946-950Crossref Google Scholar) and RASTER3D (34Merritt E.A. Bacon D.J. Methods Enzymol. 1997; 277: 505-524Crossref PubMed Scopus (3875) Google Scholar).Table IData collection and refinement statisticsSpace groupC2Resolution range (Å)20–2.8Number of unique reflections70,003(used for R-free calculation)(2827)R sym11.6 (43.8)I/sigma (I)9.2 (2.2)Completeness (%)99.0 (98.3)R cryst1-aR cryst = Σhkl‖F obs −F calc‖/F obs, whereF obs and F calc are the observed and the calculated structure factors, respectively, and the summation is over the reflections used for model refinement. Values in parentheses refer to the highest resolution shell (2.85 − 2.8 Å).(R free)0.244 (0.275)Rmsd bond length (Å)0.010Rmsd bind angles (°)1.4Number of nonhydrogen atoms10,1401-a R cryst = Σhkl‖F obs −F calc‖/F obs, whereF obs and F calc are the observed and the calculated structure factors, respectively, and the summation is over the reflections used for model refinement. Values in parentheses refer to the highest resolution shell (2.85 − 2.8 Å). Open table in a new tab Xanthine oxidase activity was determined by following the rate of uric acid formation at 295 nm. Assays were performed in solutions of 0.15 mm xanthine and 0.2 mm EDTA in 0.1 m pyrophosphate buffer (pH 8.5) under air-saturated conditions at 25 °C (26Eger B.T. Okamoto K. Enroth C. Sato M. Nishino T. Pai E.F. Nishino T. Acta Crystallogr. Sect. D Biol. Crystallogr. 2000; 56: 1656-1658Crossref PubMed Scopus (29) Google Scholar). Xanthine-phenazine methosulfate (PMS) activity was measured by following the PMS-linked reduction of horse heart cytochromec (Roche Molecular Biochemicals) (35Nishino T. Tsushima K. Hille R. Massey V. J. Biol. Chem. 1982; 257: 7348-7353Abstract Full Text PDF PubMed Google Scholar). The reduction of cytochrome c was determined by monitoring the absorbance change at 550 nm in 0.1 m pyrophosphate buffer (pH 8.5) containing 0.2 mm EDTA, 0.15 mm xanthine, 16.7 μm PMS, and 16.7 μm cytochrome cat 25 °C. The PMS solution was kept on ice and in the dark before use. As TEI-6720 exhibits time-dependent inhibition, enzyme activities were determined not from initial rates but from the steady state rates after absorbance changes had sufficiently stabilized. Dissociation constants (K d value) for enzyme-inhibitor complexes were determined by titrating the enzyme with TEI-6720. The mixture was incubated for 10 min in the dark at 25 °C to ensure equilibrium before the fluorescence (excitation at 314 nm and emission at 390 nm) of the solution was measured. As previously described (16Okamoto K. Nishino T. J. Biol. Chem. 1995; 270: 7816-7821Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar),K d values were calculated from the plots of fluorescence versus the total concentrations of added inhibitor. As is commonly observed in tight-binding inhibitors (36Cha S. Biochem. Pharmacol. 1975; 24: 2177-2185Crossref PubMed Scopus (451) Google Scholar), both allopurinol and TEI-6720 show time-dependent inhibition (Fig. 2). The underlying reasons for their time dependence, however, are quite different. A relatively large excess of 3.3 μm allopurinol reduced the rate gradually until complete inactivation was achieved; in contrast, a slight excess of 33 nm TEI-6720 caused a progressive rate decline and finally reached a steady state level of catalytic activity. Massey et al. (15Massey V. Edmondson D. J. Biol. Chem. 1970; 245: 6595-6598Abstract Full Text PDF PubMed Google Scholar) have studied the mechanism of allopurinol inhibition of XOR in detail. They found that the enzyme oxidizes allopurinol to oxypurinol, which then in turn coordinates tightly to the pterin-bound Mo(IV) ion, preventing further catalysis. As a result, the enzyme activity decreases in proportion to the accumulation of the oxypurinol-Mo(IV) complex, a classical example of suicide inhibition. Crystal structures of the oxypurinol complexes of bacterial (37Truglio J.J. Theis K. Leimkuhler S. Rappa R. Rajagopalan K.V. Kisker C. Structure (Lond.). 2002; 1: 115-125Abstract Full Text Full Text PDF Scopus (176) Google Scholar) and bovine XDH 3B. T. Eger, K. Okamoto, T. Nishino, T. Nishino, and E. F. Pai, unpublished results. show this inhibitor replaces the hydroxyl ligand of the molybdenum ion. TEI-6720, as recovered from its XOR complex, was chemically unchanged, no hydroxylation had occurred (data not shown). Therefore, we interpret the time dependence of the inhibition as caused by multiple steps for the finally settled enzyme-inhibitor complex after initial binding of the inhibitor, a process rather commonly observed in the formation of such tight binding. No hydroxylation during binding is consistent with the fact that no coordination to the metal cofactor was observed in the crystal structure of the XDH·TEI-6720 complex discussed below. The coexistence of inhibitory effects caused by TEI-6720 and product inhibition by NADH (38Della Corte E. Stirpe F. Biochem. J. 1970; 117: 97-100Crossref PubMed Google Scholar) conspired to prevent a meaningful steady state analysis of XDH activity. As the Mo-pterin sites of both XDH and XO are structurally equivalent (10Enroth C. Eger B.T. Okamoto K. Nishino T. Pai E.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 10723-10728Crossref PubMed Scopus (583) Google Scholar), steady state kinetics of the product inhibition-free oxidase reaction were measured instead. These analyses were performed varying the concentrations of xanthine and TEI-6720 under air-saturated conditions. A representative Lineweaver-Burk plot is given in Fig. 3. As described for BOF-4272 (16Okamoto K. Nishino T. J. Biol. Chem. 1995; 270: 7816-7821Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar), TEI-6720 exhibits mixed-type inhibition. Binding of TEI-6720 to the active enzyme was too tight to allow concentrations of free inhibitor (Ifree) to be set equal to the initial concentrations of TEI-6720 (I0). In the inset, concentrations of free TEI-6720 were corrected according to Equations 1 and 2 (16Okamoto K. Nishino T. J. Biol. Chem. 1995; 270: 7816-7821Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar,36Cha S. Biochem. Pharmacol. 1975; 24: 2177-2185Crossref PubMed Scopus (451) Google Scholar). [Ifree]=[I0]−[EI]Equation 1 [EI]=1/2Ki+I0+E0−[(Ki+I0+E0)2−4I0E0]Equation 2 Analysis of the kinetic results indicates a K ivalue of 1.2 ± 0.05 × 10−10m for TEI-6720, 1 order of magnitude smaller than that of BOF-4272 (16Okamoto K. Nishino T. J. Biol. Chem. 1995; 270: 7816-7821Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar). Plotting the apparent V max versus the inhibitor concentrations gives a K i′ value of 9 ± 0.05 × 10−10m. In contrast, when PMS was used as an electron donor, TEI-6720 showed a competitive inhibition pattern (Fig. 4), with aK i value of 1.2 ± 0.03 × 10−10m. Again, Ifree was calculated as described above. When catalyzing the transfer of electrons from xanthine to oxygen as the terminal acceptor, XO applies a ping-pong mechanism alternating the positive charge of molybdenum between Mo(VI) and Mo(IV) (27Massey V. Brumby P.E. Komai H. J. Biol. Chem. 1969; 244: 1682-1691Abstract Full Text PDF PubMed Google Scholar, 39Olson J.S. Ballow D.P. Palmer G. Massey V. J. Biol. Chem. 1974; 249: 4350-4362Abstract Full Text PDF PubMed Google Scholar). However, when PMS is used as an electron acceptor, it very rapidly oxidizes Mo(IV), not allowing the collection of meaningful information about this state during turnover. The same is true in the case of allopurinol and BOF-4272 inhibition (15Massey V. Edmondson D. J. Biol. Chem. 1970; 245: 6595-6598Abstract Full Text PDF PubMed Google Scholar, 16Okamoto K. Nishino T. J. Biol. Chem. 1995; 270: 7816-7821Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar, 40Spector T. Johns D.G. Biochem. Biophys. Res. Commun. 1968; 32: 1039-1044Crossref PubMed Scopus (16) Google Scholar), suggesting that the inhibitor-Mo(VI) complex is the main molecular species formed and represented in a competitive inhibition pattern in Fig. 4. As theK i value determined for the xanthine oxidase activity (estimated by plotting the slopes in the secondary plot) is almost identical to the one found for xanthine-PMS activity, these values would be representative for the Mo(VI) state of the enzyme, whereas the K i′ value for the xanthine oxidase activity, the same as the one established for BOF-4272 (16Okamoto K. Nishino T. J. Biol. Chem. 1995; 270: 7816-7821Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar), would refer to the Mo(IV) state. When fully active XO was titrated with TEI-6720 (Fig. 5 A), significant spectral perturbations were observed in the UV and visible regions of the spectrum. The difference spectrum exhibited two negative peaks; the larger one at 400 nm (Δε = 1.3 × 103m−1 cm−1) and a rather shallow, ill-defined one centered at 550 nm. Their amplitudes increased in proportion to the amount of inhibitor added until the inhibitor concentration reached that of the enzyme, which indicated that the formation of the inhibitor-enzyme complex was causing the spectral perturbation. However, both the spectral difference and the dissociation constant were too small to allow the direct determination of the K d value from a spectral titration experiment. If, however, desulfo-XO was used in the titration experiment, no absorption changes were recorded (Fig. 5 B) despite the facile formation of a tight-inhibitor complex. The absorbance changes described above are different from those Ryan et al. (41Ryan M.G. Ratnam K. Hille R. J. Biol. Chem. 1995; 270: 19209-19212Abstract Full Text Full Text PDF PubMed Scopus (25) Google Scholar) assigned to the change of the redox state of the molybdenum ion. Upon the formation of the enzyme complex, the fluorescence of TEI-6720 with a peak at 390 nm (excitation at 319 nm) is quenched. We used this property to measure the dissociation constants of the TEI-6720·XO complexes of both desulfo-XO (AFR = 3) and fully active XO (AFR = 190). A plot of fluorescence intensity against the TEI-6720 concentration is given in Fig. 6. The fluorescence signal increased markedly after equimolar amounts of TEI-6720 had been added to the enzyme solution, implying a 1:1 ratio of TEI-6720 to enzyme in the complex and making a single, specific binding site of the inhibitor on the enzyme very plausible. TheK d value for TEI-6720 binding to desulfo-XO was calculated as 2 ± 0.03 × 10−9m. Again, the corresponding value for the fully active enzyme could not be determined because the combination of very tight binding by the inhibitor and rather low fluorescence intensity prevented the sufficiently accurate determination of the concentration of free inhibitor (Fig. 6, inset). Given the very small structural difference between the sulfo- and desulfo-forms of the enzyme (replacement of a sulfur atom by an oxygen), one might expect the interactions between the inhibitor molecule and the protein matrix to stay the same in the two forms of the enzyme. The distinctive spectral changes together with the difference in K d values between the respective TEI-6720 complexes, however, seem to reflect subtle differences in the local electrostatic fields and/or minor structural rearrangements that occur around the molybdenum ion when the active sulfo-form of the enzyme is transformed into the inactive desulfo-form. Freshly purified XDH produced crystals in space group C2 with unit cell axesa = 168.3 Å, b = 124.6 Å,c = 147.3 Å, and β = 91.0°. These parameters correspond closely to those of the free or salicylate-bound crystals of XDH (26Eger B.T. Okamoto K. Enroth C. Sato M. Nishino T. Pai E.F. Nishino T. Acta Crystallogr. Sect. D Biol. Crystallogr. 2000; 56: 1656-1658Crossref PubMed Scopus (29) Google Scholar). The crystals diffracted to 2.8-Å resolution and contained two subunits in the asymmetric unit. The overall structure of the protein chain in the inhibitor complex was identical to the one found in the salicylate-bound enzyme (10Enroth C. Eger B.T. Okamoto K. Nishino T. Pai E.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 10723-10728Crossref PubMed Scopus (583) Google Scholar). Clear electron density representing the bound TEI-6720 molecule was identified and easily interpreted (Fig. 7). TEI-6720 bound in the channel leading from bulk solvent to the buried Mo-pterin cofactor, closing it off like a plug. With a distance of 4.9 Å, the methyl carbon in the thiazole ring was the atom closest to the molybdenum ion. However, no electron density representing a potential covalent bond between TEI-6720 and molybdenum was observed. The torsion angle between the planes of the thiazole and benzonitrile rings was 30°. Six hydrogen bonds and one charge-charge interaction were observed between TEI-6720 and the protein. The most tightly bound part of the inhibitor molecule was its carboxylate group. It was located at almost the same position as the carboxylate group of the salicylate molecule in the original XDH crystal structure (10Enroth C. Eger B.T. Okamoto K. Nishino T. Pai E.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 10723-10728Crossref PubMed Scopus (583) Google Scholar) and displayed an identical binding pattern: its oxygen atoms interacted with the side chain guanidinium group of Arg880 and, in addition, one of them formed hydrogen bonds to the side chain hydroxyl and the backbone amide of Thr1010. Although the interactions between protein and inhibitor that involve the carboxylate group of the latter strongly contributed to the formation of the tight complex, there had to exist other interactions to explain the 106-fold difference between the K i values of TEI-6720 and salicylate. Indeed, several such interactions could be identified. The side chain amide of Asn768 and the nitrile group of TEI-6720 were 2.9 Å apart. In addition to providing binding energy, this interaction could be essential in stabilizing the position of the benzene ring. When the 3-cyano moiety of TEI-6720 was replaced by hydrogen, its binding affinity was significantly decreased. In contrast, when the CN-group of the inhibitor was substituted with a nitro ligand, which is also capable of accepting a hydrogen bond, the resulting derivative showed a binding affinity very similar to that of TEI-6720. 4S. Kondo, unpublished results. The carboxylate group of Glu802 was located 2.8 Å from N-3 of the thiazole ring. Given this distance and pK avalues of 2.5 and 4.6 for thiazole and the side chain of glutamate, respectively, the most probable scenario has the binding of the inhibitor causing the protonation of Glu802 and the formation of a hydrogen bond between the thiazole nitrogen and the carboxylate side chain. The hydrophobic character of large parts of the TEI-6270-binding channel would favor a shift of the pK a value of glutamate in the required direction. The thiazole ring as a whole was sandwiched between two phenylalanine residues, Phe914 and Phe1009 (Fig. 8 A). The aromatic ring of Phe914 lay parallel to the plane of the thiazole ring at a distance of 3.4 Å, whereas the side chain of Phe1009pointed perpendicularly to the center of the thiazole ring, approaching it to 4.0 Å. This arrangement of energetically favorable aromatic/aromatic interactions (42Burley S.K. Petsko G.A. Science. 1985; 229: 23-28Crossref PubMed Scopus (2235) Google Scholar) had also been seen in the crystal structure of the salicylate complex and its conservation argues for an important role in stabilizing the binding positions of aromatic substrates; it might well represent one of the key features of substrate recognition. Hydrophobic interactions also contributed to the binding of TEI-6720 to XOR. The benzonitrile portion of TEI was inserted between Leu873 and Leu1014, keeping a distance of 3.4 and 3.7 Å from each side chains. Together with the Asn768-nitrile bond, this arrangement guides the orientation of the benzonitrile part of the inhibitor. The hydrophobic 4-isobutoxy tail of TEI-6720 was surrounded by amino acids Leu648, Phe649, Val1011, and Leu1013, with distances ranging from 3.7 to 4.2 Å (Fig. 8 A). Although some of these crystallographically determined values are too large to argue for direct van der Waals binding, they establish a pocket well suited to accommodate bulky hydrophobic moieties, which are often found as part of the molecular structures of good substrates or inhibitors of XOR (16Okamoto K. Nishino T. J. Biol. Chem. 1995; 270: 7816-7821Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar, 43Beedham C. Ellis G.P. West G.B. Progress in Medical Chemistry. Elsevier, Amsterdam1987: 85-127Google Scholar). The numerous interactions, which the TEI-6720 molecule displayed when assuming its position in the elongated access channel leading to the Mo-pterin group, are illustrated in Fig. 8 B. The space filling representation conveys the generally very tight fit between the inhibitor molecule and surrounding residues. Almost all amino acids, whose interactions with the TEI-6720 molecule are discussed above, are conserved among bovine (44Berglund L. Rasmussen J.T. Andersen M.D. Rasmussen M.S. Petersen T.E. J. Dairy Sci. 1996; 79: 198-204Abstract Full Text PDF PubMed Scopus (59) Google Scholar) and human XOR (45Ichida K. Amaya Y. Noda K. Minoshima S. Hosoya T. Sakai O. Shimizu N. Nishino T. Gene (Amst.). 1993; 133: 279-284Crossref PubMed Scopus (132) Google Scholar). The only exceptions are Leu648 and Phe649(bovine XOR), which are Ile and Cys in human XOR, respectively. This change still preserves the hydrophobic character of the side chains, the property important in their interaction with the inhibitor molecule. Therefore, the mechanistic conclusions drawn in our discussion are fully applicable to the human enzyme as well. TEI-6720 bound more tightly to the sulfo-form than to the desulfo-form of XOR (16Okamoto K. Nishino T. J. Biol. Chem. 1995; 270: 7816-7821Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar) and only in the case of the catalytically competent sulfo-form were spectral perturbations observed. The display of biphasic inhibition kinetics was probably caused by two binding modes of the inhibitor (e.g. 1, attached to the channel entrance and 2, fully inserted) and not by structural rearrangements of the accommodating binding site, as there were no overall changes observed in the location and orientation of the protein matrix when TEI-6720 was bound to XOR. In the crystal structure, the methyl substituent of the thiazole ring represented the part of the inhibitor closest to the molybdenum complex approaching its hydroxy ligand to a distance of 3.5 Å. The oxygen atom of the latter has been proposed as the one incorporated into substrate molecules (46Xia M. Dempski R. Hille R. J. Biol. Chem. 1999; 274: 3323-3330Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar). However, an analogous compound, in which the methyl moiety of TEI-6720 has been removed, will also perturb the spectrum in a way very similar to TEI-6720, 5K. Okamoto, unpublished observations. effectively ruling out this interaction as the source of spectral variation. As the spectral changes are only observed in the sulfo-form, it is interesting to note that the shortest distance between the sulfur ligand of Mo-pterin and the TEI-6720 molecule was 5.0 Å, again making it very improbable that the reason for the modified absorption behavior would be a direct influence by the inhibitor on the sulfur ligand of the cofactor. Although a crystal structure of the desulfo-form of XOR is not yet available, the structure of the salicylate-bound form of bovine XDH (10Enroth C. Eger B.T. Okamoto K. Nishino T. Pai E.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 10723-10728Crossref PubMed Scopus (583) Google Scholar) supported by the recent results of a 1.7-Å resolution refinement3 can serve as comparison because its absorption spectrum is identical to the one displayed by the desulfo-form, not showing the conspicuous changes in the 400–500 nm region. The only significant difference found was the movement of the side chain of Glu802. Above, we have discussed the potential of this side chain to undergo protonation and to engage in a hydrogen bond with the N-3 atom of the thiazole ring. On the other hand, in the salicylate-bound form, Glu802 is 3.3 Å apart from the sulfur ligand and likely to have a hydrogen bond to the sulfur (10Enroth C. Eger B.T. Okamoto K. Nishino T. Pai E.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 10723-10728Crossref PubMed Scopus (583) Google Scholar). We believe that such a change in charge close to the sulfur ligand combined with an increase in distance between these groups, at the moment represents the best explanation of how inhibitor binding can influence the electronic structure of the Mo-pterin cofactor. The crystal structure of the TEI-6720 complex of bovine XOR showed the inhibitor bound in a narrow channel leading to the molybdenum center of the enzyme. The potential drug molecule fills the entire pocket thereby inhibiting the activity of the enzyme simply by obstructing substrate binding. Although no direct coordination was observed between the molybdenum ion and the inhibitor, numerous hydrogen bonds and hydrophobic interactions are evident, some of them conserved in their contribution to substrate recognition. A slight reorientation of a glutamate side chain, probably accompanied by its protonation, is postulated to be the cause of the spectral changes observed upon binding of TEI-6720 to the sulfo-form of XOR.