Portal de Periódicos da CAPES

Structural Basis and Mechanism of the Inhibition of the Type-3 Copper Protein Tyrosinase from Streptomyces antibioticusby Halide Ions

2002; Elsevier BV; Volume: 277; Issue: 34 Linguagem: Inglês

10.1074/jbc.m202461200

1083-351X

Armand W.J.W. Tepper, Luigi Bubacco, Gerard W. Canters,

Bioactive Compounds and Antitumor Agents

The inhibition of the type-3 copper enzyme tyrosinase by halide ions was studied by kinetic and paramagnetic1H NMR methods. All halides are inhibitors in the conversion of l-3,4-dihydroxyphenylalanine (l-DOPA) with apparent inhibition constants that follow the order I− < F− ≪ Cl− < Br− at pH 6.80. The results show that the inhibition arises from the interaction of halide with both the oxidized (affinity F− > Cl− > Br− ≫ I−) and reduced (affinity I− > Br− > Cl− ≫ F−) enzyme. The paramagnetic 1H NMR of the oxidized enzyme complexed with the halides is consistent with a direct interaction of halide with the type-3 site and shows that the (Cu-His3)2coordination occurs in all halide-bound species. It is surmised that halides bridge both of the copper ions in the active site. Fluoride and chloride are shown to bind only to the low pH form of oxidized tyrosinase, explaining the strong pH dependence of the inhibition by these ions. We further show that p-toluic acid and the bidentate transition state analogue, Kojic acid, displace chloride from the oxidized active site, whereas the monodentate substrate analogue, p-nitrophenol, forms a ternary complex with the enzyme and the chloride ion. On the basis of the experimental results, a model is formulated for the inhibitor action and for the reaction of diphenols with the oxidized enzyme. The inhibition of the type-3 copper enzyme tyrosinase by halide ions was studied by kinetic and paramagnetic1H NMR methods. All halides are inhibitors in the conversion of l-3,4-dihydroxyphenylalanine (l-DOPA) with apparent inhibition constants that follow the order I− < F− ≪ Cl− < Br− at pH 6.80. The results show that the inhibition arises from the interaction of halide with both the oxidized (affinity F− > Cl− > Br− ≫ I−) and reduced (affinity I− > Br− > Cl− ≫ F−) enzyme. The paramagnetic 1H NMR of the oxidized enzyme complexed with the halides is consistent with a direct interaction of halide with the type-3 site and shows that the (Cu-His3)2coordination occurs in all halide-bound species. It is surmised that halides bridge both of the copper ions in the active site. Fluoride and chloride are shown to bind only to the low pH form of oxidized tyrosinase, explaining the strong pH dependence of the inhibition by these ions. We further show that p-toluic acid and the bidentate transition state analogue, Kojic acid, displace chloride from the oxidized active site, whereas the monodentate substrate analogue, p-nitrophenol, forms a ternary complex with the enzyme and the chloride ion. On the basis of the experimental results, a model is formulated for the inhibitor action and for the reaction of diphenols with the oxidized enzyme. tyrosinase l-3,4-dihydroxyphenylalanine nuclear Overhauser effect WEFT, water-suppressed equilibriumFourier transform freeinduction decay One of the unresolved questions in the enzymology of the type-3 copper-containing tyrosinases (EC 1.14.18.1) is the detailed molecular mechanism of both inhibitor action and substrate conversion. This report focuses on the mechanism of their inhibition by halides. Tyrosinases are monooxygenating enzymes catalyzing theortho-hydroxylation of monophenols and the subsequent oxidation of the diphenolic products to the corresponding quinones. The reactions take place under concomitant reduction of molecular oxygen to water. The formed quinones are reactive precursors in the synthesis of melanin pigments. In fruits, vegetables, and mushrooms, Ty1 is a key enzyme in the browning that occurs upon bruising or long-term storage. In mammals, Ty is responsible for skin pigmentation. Defects in the enzyme may lead to some forms of oculocutaneous albinism or vitiligo (1Oetting W.S. Pigment Cell Res. 2000; 13: 320-325Crossref PubMed Scopus (181) Google Scholar). Furthermore, the enzyme has been linked to Parkinson's and other neurodegenerative diseases (2Tief K. Schmidt A. Beermann F. Brain Res. Mol. Brain Res. 1998; 53: 307-310Crossref PubMed Scopus (88) Google Scholar, 3Xu Y. Stokes A.H. Freeman W.M. Kumer S.C. Vogt B.A. Vrana K.E. Brain Res. Mol. Brain Res. 1997; 45: 159-162Crossref PubMed Scopus (201) Google Scholar, 4Berman S.B. Hastings T.G. J. Neurochem. 1997; 69: 1185-1195Crossref PubMed Scopus (100) Google Scholar, 5Berman S.B. Hastings T.G. J. Neurochem. 1999; 73: 1127-1137Crossref PubMed Scopus (572) Google Scholar, 6Higashi Y. Asanuma M. Miyazaki I. Ogawa N. J. Neurochem. 2000; 75: 1771-1774Crossref PubMed Scopus (33) Google Scholar). Consequently, the enzyme poses considerable interest from medical, agricultural, and industrial points of view. The current knowledge of Ty at the biological, mechanistic, and structural levels has recently been reviewed (7Decker H. Tuczek F. Trends Biochem. Sci. 2000; 25: 392-397Abstract Full Text Full Text PDF PubMed Scopus (306) Google Scholar, 8Sánchez-Ferrer A. Rodrı́guez-López J.N. Garcı́a-Cánovas F. Garcı́a-Cánovas F. Biochim. Biophys. Acta. 1995; 1247: 1Crossref PubMed Scopus (1113) Google Scholar, 9Solomon E.I. Sundaram U.M. Machonkin T.E. Chem. Rev. 1996; 96: 2563-2605Crossref PubMed Scopus (3192) Google Scholar, 10van Gelder C.W. Flurkey W.H. Wichers H.J. Phytochemistry. 1997; 45: 1309-1323Crossref PubMed Scopus (406) Google Scholar). Ty harbors a dinuclear so-called type-3 copper center, the occurrence of which has also been established in hemocyanins, which act as oxygen carriers in arthropods and mollusks, and the catechol oxidases, which oxidize o-diphenols to the corresponding quinones. The two closely spaced copper ions in the type-3 active site are coordinated each by 3 histidine residues through the Nε nitrogen atoms (9Solomon E.I. Sundaram U.M. Machonkin T.E. Chem. Rev. 1996; 96: 2563-2605Crossref PubMed Scopus (3192) Google Scholar). Although the known type-3 centers are found to be similar both in structure and in their ability to bind molecular oxygen, they perform different functions. These differences are believed to result from variations in the substrate binding pocket or the accessibility of substrates to the active site, although the exact reasons remain to be defined (7Decker H. Tuczek F. Trends Biochem. Sci. 2000; 25: 392-397Abstract Full Text Full Text PDF PubMed Scopus (306) Google Scholar). The reduced species (Tyred; [Cu(I) (CuI)], binds oxygen to render (b) the oxygenated form (Tyoxy; [Cu(II)-O22−-Cu(II)]. In Tyoxy, molecular oxygen is bound as peroxide in a μ-η2:η2 side-on bridging mode, which destabilizes the O-O bond and activates it for reaction with mono- or diphenols (9Solomon E.I. Sundaram U.M. Machonkin T.E. Chem. Rev. 1996; 96: 2563-2605Crossref PubMed Scopus (3192) Google Scholar). The Tyoxy species shows a strong LMCT (ligand to metal chargetransfer transition) at ∼345 nm (ε ≈ 18.5 mm−1cm−1) and is EPR silent. The latter property also holds for (c) the resting form of the enzyme, i.e. the oxidized derivative (Tymet; [Cu(II)-Cu(II)] where antiferromagnetic coupling between the unpaired spins of the Cu2+ ions occurs through spin super-exchange mediated by a Cu2 bridging ligand (11Himmelwright R.S. Eickman N.C. Solomon E.I. Biochem. Biophys. Res. Commun. 1979; 86: 628-634Crossref PubMed Scopus (27) Google Scholar). Because of the diamagnetic nature of both Tyoxy and Tymet in the ground state, magnetic resonance studies on Ty and its inhibitor bound species until now mainly dealt with (d) the EPR active half-reduced species (Tyhalf-met; [Cu(I) Cu(II)]), which can be prepared by partial reduction of Tymet (12Wilcox D.E. Porras A.G. Hwang Y.T. Lerch K. Winkler M.E. Solomon E.I. J. Am. Chem. Soc. 1985; 107: 4015Crossref Scopus (332) Google Scholar). Albeit a nonphysiological derivative, EPR studies on Tyhalf-met have yielded a considerable amount of information on the structure of the active site and its ligand-bound derivatives (12Wilcox D.E. Porras A.G. Hwang Y.T. Lerch K. Winkler M.E. Solomon E.I. J. Am. Chem. Soc. 1985; 107: 4015Crossref Scopus (332) Google Scholar, 13van Gastel M. Bubacco L. Groenen E.J. Vijgenboom E. Canters G.W. FEBS Lett. 2000; 474: 228-232Crossref PubMed Scopus (43) Google Scholar). The various species discussed above, with the exception of Tyhalf-met, fit into a reaction cycle that is represented by SchemeFS1. Recently, we have shown that the oxidized enzyme, Tymet, is amenable to paramagnetic NMR spectroscopy (14Bubacco L. Salgado J. Tepper A.W. Vijgenboom E. Canters G.W. FEBS Lett. 1999; 442: 215-220Crossref PubMed Scopus (65) Google Scholar,15Bubacco L. Vijgenboom E. Gobin C. Tepper A.W.J.W. Salgado J. Canters G.W. J. Mol. Cat. B. 2000; 8: 27-35Crossref Scopus (47) Google Scholar). Although the Cu(II)-Cu(II) ground state is a diamagneticS = 0 singlet, the paramagnetic S = 1 triplet state appears to be populated at room temperature, providing sufficient paramagnetism for the signals originating from nuclei of the coordinated His residues and from exogenous ligands to shift outside the diamagnetic envelope, thereby providing a detailed fingerprint of the active site. Especially in the presence of chloride, 1H NMR resonances appear remarkably sharp for a Cu(II) system, by virtue of which it was possible to show that Ty contains a classical type-3 copper site as found in hemocyanins and catechol oxidases (14Bubacco L. Salgado J. Tepper A.W. Vijgenboom E. Canters G.W. FEBS Lett. 1999; 442: 215-220Crossref PubMed Scopus (65) Google Scholar). The1H paramagnetic NMR spectrum furthermore drastically changes when competitive inhibitors that mimic the transition state are bound to the active site (15Bubacco L. Vijgenboom E. Gobin C. Tepper A.W.J.W. Salgado J. Canters G.W. J. Mol. Cat. B. 2000; 8: 27-35Crossref Scopus (47) Google Scholar). The inhibition of Ty by halides has already been reported a long time ago (16Martı́nez J.H. Solano F. Garcia-Borrón J.C. Iborra J.L. Lozano J.A. Biochem. Int. 1985; 11: 729-738PubMed Google Scholar, 17Martı́nez J.H. Solano F. Peñafiel R. Galindo J.D. Iborra J.L. Lozano J.A. Comp. Biochem. Physiol. B. 1986; 83: 633-636Crossref PubMed Scopus (24) Google Scholar, 18Peñafiel R. Galindo J.D. Solano F. Pedreño E. Iborra J.L. Lozano J.A. Biochim. Biophys. Acta. 1984; 788: 327-332Crossref PubMed Scopus (13) Google Scholar) but, despite the universal presence of chloride in biological systems ([Cl−] 5–200 mm), it has never been addressed in kinetic and/or structural detail. The current knowledge concerning Ty halide inhibition is mainly limited to the observations that the inhibition is strongly pH-dependent and that the order of inhibition strength of F−, Cl−, Br−, and I− appears to be dependent on the source of the enzyme studied (16Martı́nez J.H. Solano F. Garcia-Borrón J.C. Iborra J.L. Lozano J.A. Biochem. Int. 1985; 11: 729-738PubMed Google Scholar, 17Martı́nez J.H. Solano F. Peñafiel R. Galindo J.D. Iborra J.L. Lozano J.A. Comp. Biochem. Physiol. B. 1986; 83: 633-636Crossref PubMed Scopus (24) Google Scholar, 18Peñafiel R. Galindo J.D. Solano F. Pedreño E. Iborra J.L. Lozano J.A. Biochim. Biophys. Acta. 1984; 788: 327-332Crossref PubMed Scopus (13) Google Scholar). The latter differences have been explained mainly in terms of the accessibility of the halide ion to the Ty active site (17Martı́nez J.H. Solano F. Peñafiel R. Galindo J.D. Iborra J.L. Lozano J.A. Comp. Biochem. Physiol. B. 1986; 83: 633-636Crossref PubMed Scopus (24) Google Scholar). Furthermore, EPR studies on halide-bound half-met hemocyanin derivatives have been reported (11Himmelwright R.S. Eickman N.C. Solomon E.I. Biochem. Biophys. Res. Commun. 1979; 86: 628-634Crossref PubMed Scopus (27) Google Scholar, 19Himmelwright R.S. Eickman N.C. Solomon E.I. Biochem. Biophys. Res. Commun. 1978; 84: 300-305Crossref PubMed Scopus (18) Google Scholar), showing that halides interact directly with copper in the type-3 hemocyanin active site. Here we report on a detailed study of the inhibition of the 31-kDa Ty from Streptomyces antibioticus by halide ions, using paramagnetic 1H NMR as a complementary technique to the more conventional kinetic and optical spectroscopic methods. The pH dependence of halide inhibition in the conversion of l-DOPA and of halide binding to Tymet were studied, providing insight into the halide inhibition at a structural and at a mechanistic level and resulting in the proposal of a halide binding mode. Our results address for the first time the Ty halide inhibition by considering the interaction of halide with the physiologically relevant Ty species that participate in the enzymatic reaction pathway. The results show that the halide inhibition derives from the interaction of halide with the Cu2 center of both the oxidized and the reduced Ty species, where the halide binding affinity is found to be dependent on the nature of the halide ion as well as on the Ty oxidation state. Furthermore, we show that halides can be used as probes in paramagnetic 1H NMR investigations of the coordination mode of exogenously added ligands such as substrate and transition-state analogues as well as aromatic carboxylic acid inhibitors. The relevance of the findings toward understanding the reaction of diphenolic substrates with Tymet is discussed, forming the basis of a proposed structural scheme of Ty inhibition and of the reaction of Tymet with diphenolic substrates. The enzyme was obtained from the growth medium of liquid cultures of S. antibioticusharboring the pIJ703 Ty expression plasmid (15Bubacco L. Vijgenboom E. Gobin C. Tepper A.W.J.W. Salgado J. Canters G.W. J. Mol. Cat. B. 2000; 8: 27-35Crossref Scopus (47) Google Scholar). The protein was purified according to published procedures (15Bubacco L. Vijgenboom E. Gobin C. Tepper A.W.J.W. Salgado J. Canters G.W. J. Mol. Cat. B. 2000; 8: 27-35Crossref Scopus (47) Google Scholar). Purity was checked by SDS-PAGE and exceeded 95% in all preparations. Protein concentrations in pure samples were routinely determined optically using a value of 82 mm−1cm−1 for the extinction coefficient at 280 nm (20Jackman M.P. Hajnal A. Lerch K. Biochem. J. 1991; 274: 707-713Crossref PubMed Scopus (52) Google Scholar). Enzyme activity assays usingl-DOPA as a substrate were performed at 21 °C by optically following the formation of the DOPAchrome reaction product at 475 nm according to the method described previously (21Lerch K. Ettlinger L. Eur. J. Biochem. 1972; 31: 427-437Crossref PubMed Scopus (169) Google Scholar). Over the time course of the experiment, linear product formation was observed in all cases. Inhibition constants were obtained by measuring values ofVmax/Km, corresponding to the slopes of the Michaelis-Menten plots, in the presence of at least 5 different concentrations of halide inhibitor, [I], and by using five different substrate concentrations at each [I]. The plots ofVmax/Kmversus[I] appeared linear within the error of the experiment (estimated at ±10%) in all cases, allowing for an estimation of Kiapp from the slope and the intercept. The halide concentrations in the assay mixture were chosen both well below and above the value of the apparent inhibition constant for the halide under consideration. For the pH dependence of halide inhibition, a 75 mm phosphate, 25 mm borate instead of a 100 mm Pi buffer was used to ensure efficient buffering at all pH values used. The enzyme was stored in the form of a 5 mm Pi stock solution at pH 7.2 to prevent protein degradation during storage prior to the experiments. Enzyme was added to the assay medium directly from this stock. The reported pH values were measured on the final mixed solutions. All steps were performed in a cold chamber (4 °C) using air-saturated buffers obtained by vigorous shaking under air for at least 1 h. The pH of the buffers used was 6.80 (measured at 21 °C) throughout the procedure. A mixture containing Tyoxy and Tyred was prepared by the incubation of a 5 mmPi solution containing the protein (∼5 μm) with 0.5 mm hydroxylamine resulting in complete reduction of Tymet to Tyred. The binding of oxygen from the air-saturated reaction buffer (∼ 0.27 mmO2) to Tyred results in the formation of a mixture of ∼95% Tyoxy and ∼5% Tyred(Kd = 17 μm). 2A. W. J. W. Tepper, L. Bubacco, and G. W. Canters, unpublished data. The reaction mixture was applied immediately to a small column containing ∼0.5 ml of CM-Sepharose column material, which had previously been equilibrated with 5 mm Pi buffer. To remove excess hydroxylamine, the column was washed extensively with the equilibration buffer after which the protein was eluted with 100 mm Pi buffer in a volume varying between 2 and 5 ml. Because the Tyoxy protein is unstable,2the protein solution was made immediately prior to the experiments. A solution of ∼95% Tyoxy and ∼5% Tyred (1 ml in a sealed cuvette; typically 3 μm total concentration) in 100 mm air-saturated Pi buffer at pH 6.80 and 4 °C was allowed to equilibrate to 21 °C after which halide was added in 5–20 μl volume steps from concentrated stock solutions of 2.00 m F−, 5.00 mCl−, 1.00 m Br−, or 100 mm I− made up in the assay buffer. Absorption changes were measured at 345 nm. NMR Tymet samples (∼0.6 mm in 100 mm NaPi at pH 6.80) were prepared as described previously (14Bubacco L. Salgado J. Tepper A.W. Vijgenboom E. Canters G.W. FEBS Lett. 1999; 442: 215-220Crossref PubMed Scopus (65) Google Scholar). Exogenous ligands were added to the samples from concentrated stock solutions prepared by using the same buffer as for the measurement. The pH of the NMR samples was varied by adding small aliquots of either dilute NaOH or 100 mm H3PO4 under continuous mixing and monitoring of the pH. 1H spectra were recorded at 600 MHz using a Bruker DMX-600 spectrometer with the super-WEFT pulse sequence (22Inubushi T. Becker E.D. J. Magn. Reson. 1983; 51: 128-133Google Scholar). Depending on the required signal-to-noise ratio, 4,000–128,000 FIDs were recorded and Fourier-transformed using a 60 Hz exponential window function (LB); the base line was corrected using software provided by Bruker. One-dimensional NOEs on TymetF were measured as described previously (14Bubacco L. Salgado J. Tepper A.W. Vijgenboom E. Canters G.W. FEBS Lett. 1999; 442: 215-220Crossref PubMed Scopus (65) Google Scholar,23Calzolai L. Gorst C.M. Bren K.L. Zhou Z.H. Adams M.W.W. LaMar G.N. J. Am. Chem. Soc. 1997; 119: 9341-9350Crossref Scopus (41) Google Scholar). To characterize the effect of halides on the enzymatic activity, we performed steady-state kinetic measurements using diphenolic l-DOPA as the substrate at pH 6.80 and room temperature. In all cases, product formation linear in time was observed, and the dependence of the reaction rate versus[l-DOPA] obeyed Michaelis-Menten type kinetics. For all halides studied, the plots ofKm/Vmax values obtained from the slope of the Michaelis-Menten plots versus[X−] showed the linear dependence (see Supplemental Material), indicating that the binding of a single halide ion is responsible for the inhibition (24Boyer P.D. The Enzymes: Kinetics and Mechanism 2. 3rd Ed. Academic Press, New York1970: 18-25Google Scholar). The fluoride and chloride inhibitions appeared competitive, whereas iodide and bromide inhibit through an apparent noncompetitive mechanism (supporting information). The order of strength of inhibition is I− > F− ≫ Cl− > Br− with apparent inhibition constants of 3.8 mm, 11 mm, 0.16 m, and 0.23 m, respectively. We studied the pH dependence of fluoride inhibition. The fluoride inhibition appeared very sensitive to the pH, as depicted in Fig.1; in the plot of p KI,F−appversus pH, an approximately linear dependence is observed. A linear least-squares fit to the data yields the relationship p KI,F−app = 8.9 − 1.06 × pH. These data can be explained by adopting Scheme 2,Ty⇋KaH+Ty*⇋Ki,F−F−TyFSCHEME 2 where Ty* denotes the acidic form of the enzyme, which is capable of binding fluoride. Assuming that TymetF is catalytically inactive, the equation relating the pH to the observed inhibition constant becomesKi,F−app=Ki,F−Ka[H+]+1(Eq. 1) In the region where [H+] ≪Ka, Equation 1 may be simplified to becomepKi,F−app=−log(KaKi,F−)−pH=(pKi,F−+pKa)−pH(Eq. 2) and p KI,F−appbecomes directly proportional to the pH. The data in Fig. 1 exhibit no curvature down to pH 5.5, indicating, according to Equation 1, that the pKa value is 55 ppm down-field region in all cases. We did not attempt to detect paramagnetically affected signals under the diamagnetic envelope. The addition of iodide to a final concentration of 0.2 m to a sample of Tymet did not lead to changes in the 1H NMR spectrum apart from a significant loss of signal intensity, possibly indicating that iodide ion reduces the copper ions under the conditions of the experiment or that it destabilizes the Tymetprotein. The observed changes upon the addition of halide cannot be assigned to the increase in ionic strength, as the addition of 0.25m Na2SO4 to a sample of native Tymet in 100 mm Pi at pH 6.8 did not affect the paramagnetic part of the spectrum. The spectra of native Tymet and TymetCl have been discussed previously (14Bubacco L. Salgado J. Tepper A.W. Vijgenboom E. Canters G.W. FEBS Lett. 1999; 442: 215-220Crossref PubMed Scopus (65) Google Scholar). The spectra of the three halide-bound derivatives each displayed several well resolved paramagnetically shifted signals. The shift pattern is rather similar for the chloride- and bromide-bound species, whereas the signal distribution of the TymetF species appears to be quite different. Yet, in each halide-bound derivative, six sharp signals together with several broader, partially overlapping signals can be distinguished. For TymetCl, the hyperfine shifted resonances could be assigned (14Bubacco L. Salgado J. Tepper A.W. Vijgenboom E. Canters G.W. FEBS Lett. 1999; 442: 215-220Crossref PubMed Scopus (65) Google Scholar) based on H2O/D2O exchange experiments, intra-residue NOE patterns, and T1 relaxation data. The total signal intensity in the 10–50 ppm range appeared compatible with what can be expected for the combined signals of six histidines (14Bubacco L. Salgado J. Tepper A.W. Vijgenboom E. Canters G.W. FEBS Lett. 1999; 442: 215-220Crossref PubMed Scopus (65) Google Scholar). The sharp solvent-exchangeable signals, marked withasterisks in Fig. 2, were assigned to the histidine Nδ protons, whereas the broader signals could be assigned to the His-Cε and His-Cδ protons of the coordinating histidines. The latter are closer to the copper and therefore experience stronger paramagnetic relaxation and hence more broadening compared with the Nδ protons. Furthermore, the observed NOE patterns allowed us to couple each of the six sharp Nδ proton signals with a broader Cε signal, thereby identifying each histidine residue in the six-coordinate ligand sphere of the Cu2 site. Because the observed signal shift pattern is quite different for the TymetCl than for the TymetF species, we repeated the assignment procedure for TymetF. Fig. 2B shows the spectrum recorded in D2O. It can be observed that five of the six sharp signals are readily solvent-exchangeable similar to what is observed for the chloride-bound species (14Bubacco L. Salgado J. Tepper A.W. Vijgenboom E. Canters G.W. FEBS Lett. 1999; 442: 215-220Crossref PubMed Scopus (65) Google Scholar). The detection of NOE couplings between the paramagnetically shifted 1H signals in TymetF is complicated due to both the relatively fast TymetFT1 proton relaxation2 and the presence of considerable signal overlap, for example for the signal at 31 ppm that is composed of both a sharp and a broad signal. Yet, we were able to establish four NOE connectivities as indicated in Fig.2. The expected NOE couplings from the sharp signals at 31.0 and 14.8 ppm to broader signals remain undetectable thus far. Further assignments must await the outcome of additional experimentation. The NMR spectra of samples of ∼0.5 mm Tymet containing 0.2 m fluoride or 0.5 m chloride are clearly pH-dependent, as shown in Fig. 3 for fluoride ion. The observed effects are fully reversible. For both of the halide-bound species similar titration behavior is observed. The spectra of Tymet recorded in the presence of fluoride or chloride at pH 9.6 superimpose well on the spectrum of native Tymet at pH 6.8 (Fig. 2A), showing that the protein reverts to native Tymet when the pH is increased. No signal broadening effects or shifts are observed over the whole titration range for all observable signals for both species, showing that the halide exchange process is slow on the NMR time scale in both cases. As a consequence, the signals of native Tymet and TymetF or TymetCl can be followed independently. The midpoint of the titration occurs at a pH of ∼8.2 for TymetF ([F−] = 0.2 m) and at pH ∼7.8 For TymetCl ([Cl−] = 0.5m). We did not quantify signal intensities because partial protein degradation at the extremes of the pH values used prevented accurate comparison between individual spectra. A quantitative analysis of the pH dependence of fluoride binding is presented below. A pH titration of native Tymet toward the lower pH values was unsuccessful because of irreversible protein degradation leading to a rapid loss of signal intensity. The observed pH dependence of fluoride binding and inhibition prompted us to investigate fluoride binding at fixed pH. More specifically, we performed a titration of native Tymet with fluoride at two pH values (pH 7.06, [F−] 0–38 mm, and pH 8.03, [F−] 0–375 mm). These pH values were chosen in the region of maximal stability of the protein to prevent protein degradation during the experiment. The observed titration behavior is the reverse of that observed in the pH titration described above (Fig.3), in agreement with a two-state model where both species are in slow exchange on the NMR time scale. There is no indication of the occurrence of intermediates during the titration, in agreement with the binding of a single fluoride ion. The relative amounts of native Tymet and TymetF were determined by measuring peak heights. For native Tymet, this was carried out on the isolated signals at 47.2 and 22.0 ppm where there is no overlap with signals originating from the fluoride-bound species. The relative amount of TymetF was determined by measuring the intensity of the sharp Nδ proton signals 28.8, 33.3, 35.9, and 41.4 ppm. The intensity of the signals were normalized and then fitted to the general equation for two-state binding under the conditions that the ligand is in large excess over the enzyme, Iobs,AImax,A=1−[F−]Kdapp+[F−](Eq. 3) Iobs,FImax,F=[F−]Kdapp+[F−](Eq. 4) where Iobs,F and Iobs,A represent the observed signal intensities of TymetF and native Tymet, respectively. Kdapp represents the apparent value for the dissociation constant of the fluoride-bound complex, Imax, A, the signal intensity of the native species at zero [F−] and Imax, Fthe TymetF signal intensity at [F−] ≫ Kdapp. For each of the two titration experiments, Kdapp was set as a shared parameter between the Iobs,F and Iobs,Aversus [F−] data sets, resulting in a single value for Kdapp, Imax, A, and Imax, Fat each pH. At both pH values, good fits were obtained as depicted in Fig. 4, left panel (pH 7.06) and right panel (pH 8.03). The values obtained for Kdapp amounted to 5.8 and 51 mm at pH 7.06 and 8.03 with corresponding p Kdapp values of 2.24 and 1.29, respectively. The difference in the p Kdapp values is 0.94, comparing well with the difference in the pH values of the experiments (0.97). This shows that the fluoride dissociation constant, like the fluoride inhibition constant (Fig. 1), is inversely proportional to the proton concentration in the pH range of 7 to 8, and it explains why the halide is displaced from TymetX at fixed [X−] upon increasing the pH (see Fig. 3). To obtain a better insight into the mode of halide and inhibitor binding in the oxidized [Cu(II)-Cu(II)] species, we recorded 1H NMR spectra of ∼0.5 mm native Tymet and TymetCl ([Cl−] = 0.5 m) in the presence of the bidentate inhibitor Kojic acid or the monodentate ligand p-nitrophenol. The resulting spectra are represented in Fig.5. It appears that the spectra of Tymet + Kojic acid recorded in the absence and presence of chloride are nearly identical (Fig. 5, C and D), indicating that Kojic acid displaces chloride from the active site. The small differences between the spectra can be explained by assuming that there is still a small fraction of TymetCl present in solution (compare Fig. 2C with Fig. 5D). Analogous behavior is observed with the bidentate carboxylic acid inhibitor p-toluic acid (not shown). The situat