Irreversible Inhibition of Lysyl Oxidase by Homocysteine Thiolactone and Its Selenium and Oxygen Analogues

Homocysteine thiolactone, selenohomocysteine lactone, and homoserine lactone were found to be competitive, irreversible inhibitors of lysyl oxidase, with K I values of 21 ± 3 μm, 8.3 ± 2.2 μm, and 420 ± 56 μm, respectively. The first order rate constants for inactivation (k 2) of the enzyme varied over a much smaller range, ranging from 0.12 to 0.18 to 0.28 min−1 for the Se-, thio-, and O-lactones, respectively. Mutually exclusive labeling of the enzyme by [1-14C]β-aminopropionitrile, [U-14C]phenylhydrazine, or [35S]homocysteine thiolactone was observed. These labeling results, together with the closely similar perturbations of the near UV-visible spectra of lysyl oxidase and of a model of its lysine tyrosylquinone cofactor by the thiolactone, indicate that the lactones likely derivatize and reduce the active site carbonyl cofactor. Substitution with deuterium at the α-carbon of the thiolactone caused a deuterium kinetic isotope effect on k 2 of 3.2 ± 0.2, consistent with the involvement of rate-limiting α-proton abstraction during lactone-induced inactivation of the enzyme. The activities of plasma amine oxidase and diamine oxidase were only minimally reduced at concentrations of the sulfur or selenium lactones that fully inhibited lysyl oxidase. Thus, these lactones constitute a new category of mechanism-based inactivators selective for lysyl oxidase. Further, these results may relate to the development of connective tissue defects seen in homocystinuria. Homocysteine thiolactone, selenohomocysteine lactone, and homoserine lactone were found to be competitive, irreversible inhibitors of lysyl oxidase, with K I values of 21 ± 3 μm, 8.3 ± 2.2 μm, and 420 ± 56 μm, respectively. The first order rate constants for inactivation (k 2) of the enzyme varied over a much smaller range, ranging from 0.12 to 0.18 to 0.28 min−1 for the Se-, thio-, and O-lactones, respectively. Mutually exclusive labeling of the enzyme by [1-14C]β-aminopropionitrile, [U-14C]phenylhydrazine, or [35S]homocysteine thiolactone was observed. These labeling results, together with the closely similar perturbations of the near UV-visible spectra of lysyl oxidase and of a model of its lysine tyrosylquinone cofactor by the thiolactone, indicate that the lactones likely derivatize and reduce the active site carbonyl cofactor. Substitution with deuterium at the α-carbon of the thiolactone caused a deuterium kinetic isotope effect on k 2 of 3.2 ± 0.2, consistent with the involvement of rate-limiting α-proton abstraction during lactone-induced inactivation of the enzyme. The activities of plasma amine oxidase and diamine oxidase were only minimally reduced at concentrations of the sulfur or selenium lactones that fully inhibited lysyl oxidase. Thus, these lactones constitute a new category of mechanism-based inactivators selective for lysyl oxidase. Further, these results may relate to the development of connective tissue defects seen in homocystinuria. Lysyl oxidase (EC 1.4.3.13) is unique among the mammalian copper amine oxidases by catalyzing a critical post-translational modification essential to the biogenesis of connective tissue matrices. This enzyme initiates covalent cross-linking between and within the molecular units of elastin and of collagen by oxidizing peptidyl lysine in these proteins to peptidyl α-aminoadipic-δ-semialdehyde (1Pinnell S.R. Martin G.R. Proc. Natl. Acad. Sci. U. S. A. 1968; 61: 708-716Crossref PubMed Scopus (426) Google Scholar, 2Kagan H.M. Mecham R.P. Biology of Extracellular Matrix. 1. Academic Press, Orlando, FL1986: 321-328Google Scholar). The peptidyl aldehyde can then condense with neighboring ε-amino groups or peptidyl aldehydes to form the covalent cross-linkages found in fibrillar collagen and elastin. Lysyl oxidase contains a tightly bound copper cofactor as well as a covalently bound carbonyl prosthetic group recently identified as lysine tyrosylquinone (3Wang S.X. Mure M. Medzihradszky K.F. Burlingame A.L. Brown D.E. Dooley D.M. Smith A.J. Kagan H.M. Klinman J.P. Science. 1996; 273: 1078-1084Crossref PubMed Scopus (308) Google Scholar). Lysyl oxidase catalyzes primary amine oxidation through a ping pong bi ter kinetic mechanism (4Williamson P.R. Kagan H.M. J. Biol. Chem. 1987; 262: 8196-8201Abstract Full Text PDF PubMed Google Scholar, 5Williamson P.R. Kagan H.M. J. Biol. Chem. 1986; 261: 9477-9482Abstract Full Text PDF PubMed Google Scholar). Following initial Schiff base formation with the LTQ 1The abbreviations used are: LTQ, lysine tyrosylquinone; BAPN, β-aminopropionitrile fumarate; HCTL, homocysteine thiolactone; HSL, homoserine lactone; LO, lysyl oxidase; PAGE, polyacrylamide gel electrophoresis; SeHCL, selenohomocysteine lactone; HPLC, high performance liquid chromatography. cofactor, the bound substrate undergoes rate-limiting, general base-facilitated α-proton abstraction (6Gacheru S.N. Trackman P.C. Kagan H.M. J. Biol. Chem. 1988; 263: 16704-16708Abstract Full Text PDF PubMed Google Scholar). Electrons migrating from the resulting carbanion reduce the carbonyl cofactor, followed by hydrolysis of the product imine intermediate to release the aldehyde product. The reduced enzyme, retaining the amino function of the substrate, is reoxidized by molecular oxygen to produce hydrogen peroxide and ammonia, regenerating the oxidized enzyme and completing the catalytic cycle. The role of lysyl oxidase in the growth and repair of connective tissues has been well documented. Markedly increased levels of LO activity are observed in a variety of fibrotic diseases in which excess collagen is deposited in the affected tissues, as in models of atherosclerosis, hypertension, and liver and pulmonary fibrosis (2Kagan H.M. Mecham R.P. Biology of Extracellular Matrix. 1. Academic Press, Orlando, FL1986: 321-328Google Scholar). The possibility that the development of fibrosis may be restricted by the specific suppression of lysyl oxidase activity has stimulated the search for selective and potent inhibitors of this enzyme. These efforts have identified mechanism-based and ground-state inhibitors, including β-substituted haloethylamines (7Tang S. Simpson D.E. Kagan H.M. J. Biol. Chem. 1984; 259: 975-979Abstract Full Text PDF PubMed Google Scholar), benzylamines substituted with electronegative para-substituents (8Williamson P.R. Kagan H.M. J. Biol. Chem. 1987; 262: 14520-14524Abstract Full Text PDF PubMed Google Scholar), and 1,2-diamines (9Gacheru S.N. Trackman P.C. Calaman S.D. Greenaway F.T. Kagan H.M. J. Biol. Chem. 1989; 264: 12963-12969Abstract Full Text PDF PubMed Google Scholar), each of which appear to inhibit as adducts of the carbonyl cofactor. In the present report, we describe our observations that homocysteine thiolactone and the oxygen and selenium lactone analogues of this compound are active site-directed, irreversible inhibitors of lysyl oxidase. The selenium and sulfur lactones are the most potent of these and are selective for lysyl oxidase, among mechanistically similar copper-dependent mammalian amine oxidases tested in the present study. Notably, HCTL occurs in mammalian systems as a metabolic by-product of methyl transfer from S-adenosylhomocysteine. Moreover, the accumulation of HCTL has been suggested to be related to mechanisms of carcinogenesis and atherogenesis (10McCully K.S. Ann. Clin. Lab. Sci. 1994; 24: 27-59PubMed Google Scholar, 11Tsai J.-C. Perrela M.A. Yoshizumi M. Hsieh C.-M. Haber E. Schlegel R. Lee M.-E. Proc. Natl. Acad. Sci. U. S. A. 1994; 91 (6369): 63-73Crossref PubMed Scopus (49) Google Scholar, 12McCully K.S. Ann. Clin. Lab. Sci. 1993; 23: 477-493PubMed Google Scholar), whereas it has also been shown to thiolate proteins, including low density lipoproteins in vitro (13Vidal M. Sainte-Marie J. Philippot J. Bienvenue A. Biochimie. 1986; 68: 723-730Crossref PubMed Scopus (22) Google Scholar, 14McCully K.S. Cancer Res. 1976; 36: 3198-3202PubMed Google Scholar). Elucidation of interactions of these compounds with lysyl oxidase should increase options for the design of antifibrotic agents and aid in the understanding of the biological effects of HCTL. Homocysteine thiolactone hydrochloride, homocysteine, methionine, seleno-dl-methionine, d-homoserine, homoserine lactone hydrochloride, homovanillic acid, horseradish peroxidase, diaminopentane dihydrochloride, BAPN, pyridoxal hydrochloride, [14C]NaCN, deuterated water, porcine kidney diamine oxidase, and bovine plasma amine oxidase were obtained from Sigma. Phenylhydrazine hydrochloride and Darco G-60 activated carbon were obtained from Fisher. Pyrroloquinoline quinone was a product of Fluka Chemical Corp. Hydrogen iodide, deuterium iodide, and β-bromoethylamine hydrobromide were products of Aldrich. [U-14C]Phenylhydrazine was purchased from ICN Pharmaceuticals, Irvine, CA. [35S]Methionine and [1-14C]iodoacetamide were purchased from Amersham Life Science. 4-Ethyl-5-(butylamino)-1,2-benzoquinone (Fig. 1) served as a model of the LTQ carbonyl cofactor of lysyl oxidase and was generously provided by Dr. Judith P. Klinman of the Departments of Chemistry and Molecular and Cell Biology, University of California, Berkeley. Lysyl oxidase was isolated from calf aorta as a co-purified mixture of four individual variants (15Williams M.A. Kagan H.M. Anal. Biochem. 1985; 149: 430-437Crossref PubMed Scopus (29) Google Scholar) resolving as a single band at 32 kDa by SDS-PAGE (16Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (207521) Google Scholar). The individual ionic forms exhibit common structural and apparently identical mechanistic features (17Kagan H.M. Sullivan K.A. Olsson T.A. Cronlund A.L. Biochem. J. 1979; 177: 203-214Crossref PubMed Scopus (78) Google Scholar, 18Sullivan K.A. Kagan H.M. J. Biol. Chem. 1982; 257: 13520-13526Abstract Full Text PDF PubMed Google Scholar), further noted by the presence of a single dipeptide sequence containing the carbonyl cofactor, which was isolated from a proteolytic digest of the mixture of variants of bovine aorta lysyl oxidase (3Wang S.X. Mure M. Medzihradszky K.F. Burlingame A.L. Brown D.E. Dooley D.M. Smith A.J. Kagan H.M. Klinman J.P. Science. 1996; 273: 1078-1084Crossref PubMed Scopus (308) Google Scholar). Lysyl oxidase activity was assayed either against 125,000 cpm of a recombinant tropoelastin substrate expressed and labeled withl-[4,5-3H]lysine in a bacterial expression system (19Bedell-Hogan D. Trackman P. Abrams W. Rosenbloom J. Kagan H. J. Biol. Chem. 1993; 268: 10345-10350Abstract Full Text PDF PubMed Google Scholar) or against 500,000 cpm ofl-[4,5-3H]lysine-labeled chick calvarial collagen substrate (20Siegel R.C. Proc. Natl. Acad. Sci. U. S. A. 1974; 71: 4826-4830Crossref PubMed Scopus (136) Google Scholar). Tritiated water formed by lysyl oxidase action was isolated by vacuum distillation and quantified by liquid scintillation spectrometry. Enzyme activity was also assayed against nonpeptidyl amine substrates by a horseradish peroxidase-coupled fluorescence method (21Trackman P.C. Zoski C.G. Kagan H.M. Anal. Biochem. 1981; 113: 336-342Crossref PubMed Scopus (56) Google Scholar). Error ranges for all assay data, including those used to express kinetic isotope effects, were determined as the 90% confidence intervals from unweighted linear regression analyses according to Mendenhall (22Mendenhall W. Introduction to Probability Statistics. Duxbury Press, North Scituate, MA1975: 250-276Google Scholar). Porcine kidney diamine oxidase and bovine plasma amine oxidase activities were determined at 37 °C at pH 7.2 or 7.4, respectively, by a horseradish peroxidase-coupled fluorescence assay modified from that used for lysyl oxidase (21Trackman P.C. Zoski C.G. Kagan H.M. Anal. Biochem. 1981; 113: 336-342Crossref PubMed Scopus (56) Google Scholar). Reaction mixtures contained enzyme, 2.5 mm putrescine or benzylamine for assay of diamine oxidase or plasma amine oxidase, respectively, 40 μg of horseradish peroxidase, and 0.7 mm sodium homovanillate, in 16 mm potassium phosphate in a total volume of 2 ml. Background rates for each enzyme were determined in the presence of enzyme-saturating concentrations of inhibitors, i.e. 100 μm aminoguanidine for diamine oxidase (23Tang S.-S. Chichester C.O. Kagan H.M. Connect. Tiss. Res. 1989; 19: 93-103Crossref PubMed Scopus (35) Google Scholar) or 100 μm β-bromoethylamine for plasma amine oxidase (24Neumann R. Hevey R. Abeles R.H. J. Biol. Chem. 1975; 250: 6362-6367Abstract Full Text PDF PubMed Google Scholar). [1-14C]BAPN (55 mCi mmol−1) was synthesized from [14C]NaCN (55 mCi mmol−1) and β-bromoethylamine hydrobromide. [14C]Sodium cyanide (1 mCi; 0.02 mmol) was completely dissolved in dimethyl sulfoxide, and then 1.3 μl (0.01 mmol) of triethylamine was added, followed by addition of 0.01 mmol of β-bromoethylamine in dimethyl sulfoxide. The reaction mixture was stirred gently at 37 °C for 4 days to assure complete reaction. An identical reaction mixture containing [12C]NaCN was incubated in parallel as a control to monitor the status of the reaction by thin layer chromatography on silica gel plates using 1-butanol:acetic acid:water (4:1:1) as the mobile phase. At the completion of the reaction, concentrated HCl was added to bring the pH to 4–5 to quench traces of [14C]NaCN remaining in the solution. A Dynamax-C8 preparative scale HPLC column (Rainin Instruments, Woburn, MA) was employed for the purification of the synthesized compound. Elution gradients were generated between Buffer A (0.05% trifluoroacetic acid in water) and Buffer B (0.03% trifluoroacetic acid in acetonitrile) at 25 °C at a flow rate of 2 ml/min. Peak elution was monitored at 232 nm. Optimal separation was achieved by the addition of Buffer B to 85% from 0% within 1 min at 25 min after application of the sample to the HPLC column as indicated by TLC and UV absorption. Radioactive BAPN eluted between 45 and 48 min and was detected by monitoring fractions for radioactivity and for the appearance of BAPN by TLC (R F value, 0.36), visualizing by reaction with ninhydrin. The final product was >95% pure, and the overall yield was 36%. [35S]HCTL was prepared from [35S]methionine (82 mCi mmol−1) according to Baernstein (25Baernstein H.D. J. Biol. Chem. 1934; 106: 451-456Abstract Full Text PDF Google Scholar) with a few modifications. Thus, 0.5 mCi ofl-[35S]methionine was added to 1.2 ml of 57% hydrogen iodide, and the mixture was refluxed at 137 °C for 3.5 h. The resulting solution was extracted several times with ether, the aqueous phase was diluted with water and then dried by lyophilization. The product was dissolved in hot absolute alcohol and then precipitated by addition of three volumes of pure ether. The precipitate was recovered by centrifugation and was further washed with ether. The purity of the product was confirmed by its absorption spectrum, by TLC, and by NMR spectroscopy. The overall yield was >90%. SeHCL was prepared by the same method, substituting selenomethionine for methionine. [α-2H]HCTL was synthesized in two steps designed to minimize the opening of the thiolactone ring. Racemic [α-2H]methionine was prepared froml-[α-1H]methionine according to Fujihara and Schowen (26Fujihara H. Schowen R.L. J. Org. Chem. 1984; 49: 2819-2820Crossref Scopus (23) Google Scholar). l-Methionine (10.7 mmol) was dissolved at room temperature in 10 ml of deuterium oxide containing 20 mmol of sodium hydroxide followed by the addition of 1.0 mmol of pyridoxal hydrochloride. The mixture was then autoclaved at 15 p.s.i. steam pressure for 30 min. The mixture was then cooled on ice and adjusted to pH 4 with HCl. Continued incubation at 0 °C resulted in the precipitation of crystalline material. The crystalline precipitate was washed with cold water and methanol by filtration, the product dissolved in water containing 50 mg of Darco G-60 activated carbon and filtered. Methanol was added, and the product was precipitated by incubation at 4 °C. The crystalline α-deuterated methionine was isolated by filtration and dried under vacuum. Deuterium incorporation was determined by NMR spectroscopy to be >90% of theory. Deuterated HCTL was then synthesized from the deuterated methionine and deuterium iodide (57%) according to the procedure for synthesizing [35S]HCTL (25Baernstein H.D. J. Biol. Chem. 1934; 106: 451-456Abstract Full Text PDF Google Scholar). Deuterium incorporation into the final product was estimated as >83% by NMR, and overall yield was >80%. When hydrogen iodide was employed in the second step, the deuterium incorporation decreased for unknown reasons. d-HSL was prepared by refluxing 250 mg ofd-homoserine in 2 ml of 2 n HCl for 2.5 h at 120 °C. At the completion of the reaction, the mixture was cooled to room temperature, diluted with water, and freeze-dried. The resulting solid was redissolved in a minimum amount of hot absolute alcohol, reprecipitated, and washed with pure ether. The stereochemical purity of the product was confirmed by circular dichroism, and the overall yield was >90%. SDS-PAGE was carried out in 0.1% SDS on slab gels (14 × 10 × 0.15 cm) of 12.5% acrylamide using the Tris/glycine buffer system of Laemmli (16Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (207521) Google Scholar). Gels were stained for 30 min in 0.1% Coomassie Blue (R-250) and then destained with 40% methanol, 10% acetic acid. Gels were also treated with EN3HANCE™ (NEN Life Science Products) according to the protocol of the manufacturer before autoradiography. Lysyl oxidase was reacted with [1-14C]iodoacetamide by a modification of a published procedure (27Kagan H.M. Cai P. Methods Enzymol. 1995; 258: 122-132Crossref PubMed Scopus (28) Google Scholar). Purified bovine aorta lysyl oxidase (15 μg) was incubated at 37 °C in the presence or absence (control) of 80 μm HCTL in 16 mm potassium phosphate, 1.2m urea, pH 7.8, in a total volume of 35 μl for 2 h, fully inactivating the enzyme incubated in the presence of the inhibitor. Solid urea as well as 10 μl of 8 m urea, 0.22m NH4HCO3, pH 8.5, was then added, bringing the urea concentration to 8 m. The mixture was further incubated at 50 °C for 0.5 h to denature the protein. The mixture was cooled to room temperature, [1-14C]iodoacetamide was added at a 100-fold molar excess relative to disulfide bond content of the native protein (27Kagan H.M. Cai P. Methods Enzymol. 1995; 258: 122-132Crossref PubMed Scopus (28) Google Scholar), and the mixture was incubated at room temperature in the dark for 1 h. The samples were initially and intermittently flushed with nitrogen to prevent oxidation of free sulfhydryls that may have formed during the reaction. At the end of the reaction, loading buffer containing SDS in the absence of a disulfide reductant was added and the samples were subjected to SDS-PAGE immediately. LO (15 μg) was preincubated in the presence or absence of 100 μm BAPN, 80 μm phenylhydrazine, or 60 μm HCTL in 16 mm potassium phosphate buffer, 1.2 m urea, pH 7.8, at 37 °C for 1 h and then incubated for an additional 1 h with 60 μm [35S]HCTL (82 mCi mmol−1), 100 μm [1-14C]BAPN (55 mCi mmol−1), or 80 μm[U-14C]phenylhydrazine (7.5 mCi mmol−1), in the presence or absence of nonisotopic inhibitors as specified. Unbound ligands were removed by exhaustive dialysis against 2 murea, 16 mm potassium phosphate, pH 7.8, followed by dialysis against distilled water. The dialyzed samples were concentrated in vacuo and resolved by SDS-PAGE and analyzed by autoradiography. UV absorption spectra were recorded at 25 °C in a Hewlett Packard model 8452A diode array spectrophotometer. HCTL (95 μm) was incubated with 44 μm LO in 16 mm potassium phosphate, 6m urea, 30 mm NaCl, pH 7.8, in a 1-cm cuvette. Scanning was initiated within 1 min of mixing of the enzyme and inhibitor and repeated every 90 s for 30 min. Difference spectra were obtained by subtracting the spectrum of the initial scan (equivalent to that of free enzyme) from those of the subsequent scans. The structures of the aminolactones and of other compounds used in these studies are shown in Fig. 1. Each of the aminolactones proved to inhibit the activity of purified lysyl oxidase. As shown (Fig.2), Lineweaver-Burk plots of initial rate assay data obtained in the presence and absence of varied concentrations of HCTL, SeHCL, or HSL yielded plots that intersect at the 1/v axis, indicative of competitive modes of inhibition in each case. The stereospecificity of the inhibition varied. Thus, thed- and l-isomers of HCTL were equally inhibitory, as indexed by the K I values determined from Lineweaver-Burk plots (28Lineweaver H. Burk D. J. Am. Chem. Soc. 1934; 56: 658-666Crossref Scopus (8066) Google Scholar), whereas only the l-isomer of HSL inhibited the enzyme. The individual isomers of SeHCL were not available, and all data obtained with this compound were determined with the racemic mixture. The reversible or irreversible nature of the inhibition was determined by preincubating lysyl oxidase at 37 °C with varied concentrations of each lactone followed by assay for residual activity by the peroxidase-coupled assay method. Aliquots of the preincubated enzyme-inhibitor mixtures were sufficiently diluted to reduce lactone levels to noninhibitory concentrations in the assay. Each of the lactones caused the time-dependent loss of lysyl oxidase activity, which was not reversed by dilution. Graphing the data as the logarithm of the percent remaining activity against time of preincubation resulted in a series of linear plots with increasing negative slope at increasing inhibitor concentrations (data not shown), indicating that the development of irreversibility followed first order kinetics. Inhibition was also seen to be fully irreversible by exhaustive dialysis of enzyme preincubated with sufficient concentrations of these inhibitors to fully inhibit activity. Secondary plots of the t½ values derived from the first order plots against 1/[lactone] were also linear (Fig.3), consistent with site-saturation interactions between the inhibitors and the enzyme. This kinetic behavior is consistent with the mechanism summarized in Scheme I (29Walsh C. Enzymatic Reaction Mechanisms. W. H. Freeman and Co., San Francisco1979: 86-89Google Scholar), according to which irreversible inhibition develops by conversion of the lactone, initially bound in a reversible EI complex, to a covalently linked, inactivated enzyme-inhibitor complex ([EI]*, Scheme I). E+I↔k−1k+1[EI]→k+2[EI]*Scheme I The inhibitor dissociation constant, K I, is equal to k −1/k 1, whereas k 2 is the limiting first order rate constant for inactivation (29Walsh C. Enzymatic Reaction Mechanisms. W. H. Freeman and Co., San Francisco1979: 86-89Google Scholar). The first order rate of inactivation caused by 25 μm HCTL decreased 10-fold from 0.1 min−1 to 0.01 min−1 when 5 mm n-hexylamine, a productive substrate for lysyl oxidase, was present in the enzyme-inhibitor preincubation mixture, consistent with the competitive kinetics obtained (see Fig. 2) and with interaction of lactone inhibitors at the active site. Table I summarizes the constants derived from these kinetic studies with the three lactones. As seen, SeHCL has the greatest affinity for the enzyme, with a K I value 2.5- and 50-fold less than those of HCTL and HSL, respectively. In contrast, the k 2 values of the three lactones are similar, although there is a tendency for the value of k 2 to increase in the order SeHCL < HCTL < HSL. The second order rate constants (k 2/K I) are predominantly influenced by the K I values (Table I).Table ILactone constantsConstantHSL (O)HCTL (S)Se-HCL (Se)K I(L) 420 ± 56 μm21 ± 3 μm8.3 ± 2.2 μmk 2(L) 0.28 ± 0.01 min−10.18 ± 0.01 min−10.12 ± 0.03 min−1k 2/K I11 s−1m−1143 s−1m−1241 s−1m−1 Open table in a new tab As shown in the top panelof Fig. 3, substitution of the α-proton of HCTL with a deuteron has a marked effect on the t½ for inactivation, determined at 37 °C, and, thus, on the k 2value for this thiolactone. The k 2 at 37 °C for [α-1H]HCTL is 0.18 min−1, whereas that for [α-2H]HCTL, prepared and tested as the racemate, is 0.055 min−1, giving a kinetic isotope effect of 3.3-fold. The magnitude of this deuterium kinetic isotope effect is similar to the primary kinetic isotope effects found for k cat with productive substrates undergoing enzyme-catalyzed α-proton abstraction, including n-butylamine (D kcat, 4.3 at 37 °C; Ref. 9Gacheru S.N. Trackman P.C. Calaman S.D. Greenaway F.T. Kagan H.M. J. Biol. Chem. 1989; 264: 12963-12969Abstract Full Text PDF PubMed Google Scholar) and tyramine (D kcat, 2.6 at 55 °C; Ref. 30Shah M.A. Scaman C.H. Palcic M.M. Kagan H.M. J. Biol. Chem. 1993; 268: 11573-11579Abstract Full Text PDF PubMed Google Scholar). While similar (31Callery P.S. Subramanyam B. Yuan Z. Pou S. Geelhaar L.A. Reynolds K.A. Chem. Biol. Interact. 1992; 85: 15-26Crossref PubMed Scopus (2) Google Scholar), smaller (32Davidson V.L. Biochem. J. 1989; 261: 107-111Crossref PubMed Scopus (43) Google Scholar) and significantly larger (33Hartmann C. Klinman J.P. Biochemistry. 1991; 30: 44611-46056Google Scholar) D kcat values have been found reflecting α-proton abstraction steps in other copper-dependent amine oxidases, the values for lysyl oxidase suggest that additional rate contributing steps may exist that partially suppress the isotope effect of the isotope-sensitive proton abstraction step. In toto, these results point to the conclusion that the bound lactone undergoes α-proton abstraction as a rate-limiting step in the development of irreversible inhibition. The substitution of the α-proton with a deuteron did not change the apparent K I, as determined from Lineweaver-Burk plots of assays performed at 37 °C, resulting in aD K I of 1 and aD k 2/K I equal to that for D k 2. Thus, the K I for the protonated lactone was 21.6 ± 0.5 μm, whereas that for the deuterated lactone was 22 ± 0.7 μm. These K I values are presumed to reflect the true dissociation constants predominantly while largely neglecting the contribution of k 2. Thus,k 2 for HCTL is relatively slow (0.18 min−1), so that only a minor fraction (≪10%) of the available enzyme became irreversibly inactivated under the initial rate assay conditions employed and in the less than saturating concentration of HCTL used in the determination of K I. There was no apparent enzyme turnover as indexed by the lack of H2O2 production in incubations of lysyl oxidase with HCTL, indicating that HCTL is not a productive substrate for lysyl oxidase. Indeed, simple alkylamines stemming from secondary carbon atoms, as in 1-ethylpropylamine, were also found not to be substrates for lysyl oxidase (data not shown). Moreover, neither N-acetyl-HCTL, α-amino-γ-butyrolactone (structures shown in Fig. 1), nor dl-homocysteine inhibited lysyl oxidase at concentrations ≤1 mm, indicating the essentiality of the α-amino group and the five-membered lactone structure in the inhibition by these compounds. The inhibitory effect of HCTL on LO was also confirmed by assays against protein substrates of lysyl oxidase, yielding IC50values of 30 ± 3 μm and 220 ± 20 μm, using tropoelastin and collagen, respectively, as substrates. In comparison to the value of 23.0 ± 3.6 μm obtained with 1,5-diaminopentane as substrate, spatially extended interactions between the enzyme and its collagen substrate previously described (34Nagan N. Kagan H.M. J. Biol. Chem. 1994; 269: 22366-22371Abstract Full Text PDF PubMed Google Scholar) may account for the significantly decreased sensitivity noted with this substrate. The specificity of the inhibition of selected copper-dependent amine oxidases by the selenium and sulfur lactones was investigated in view of the apparent similarity among the mechanisms of action of these enzymes (35Klinman J.P. J. Biol. Chem. 1996; 271: 27189-27192Abstract Full Text Full Text PDF PubMed Scopus (79) Google Scholar). As shown (Fig.4), HCTL and SeHCL are each highly selective for lysyl oxidase among these catalysts. Although LO lost all activity at 100 μm HCTL and at 10 μm SeHCL, the activities of plasma amine oxidase and diamine oxidase were only slightly affected at these concentrations of these lactones. HCTL can covalently derivatize proteins via aminolysis of the thiol ester by protein α- and/or ε-amino groups (13Vidal M. Sainte-Marie J. Philippot J. Bienvenue A. Biochimie. 1986; 68: 723-730Crossref PubMed Scopus (22) Google Scholar, 14McCully K.S. Cancer Res. 1976; 36: 3198-3202PubMed Google Scholar, 36Benesch R. Benesch R.E. J. Am. Chem. Soc. 1956; 78: 1597-1599Crossref Scopus (59) Google Scholar), resulting in the gain of free sulfhydryl groups due to the opening of the thiolactone ring. To clarify whether this is relevant to the inhibition of LO by HCTL, the native enzyme was modified by HCTL as described and then reacted with [1-14C]iodoacetamide, using the isotopically labeled alkylating reagent to detect sulfhydryl groups newly introduced by reaction of the enzyme with the thiolactone. Autoradiography of SDS-PAGE analyses of the modified enzyme (Fig.5) indicated that the low background level of [1-14C]iodoacetamide associated with the native enzyme was not significantly changed by the modification with HCTL. Prior reduction of the enzyme disulfide bonds by treatment of the native enzyme with 2-mercaptoethanol or dithiothreitol markedly increased incorporation of the isotope, consistent with the alkylation of the newly available cysteine residues. These data indicate that HCTL does not introduce free SH functions and are consistent with previous observations that each of the cysteine residues of native lysyl oxidase exists in disulfide linkage (15Williams M.A. Kagan H.M. Anal. Biochem. 1985; 149: 430-437Crossref PubMed Scopus (29) Google Scholar). The preparation of lysyl oxidase used in this experiment included a co-purified band at ∼24