A Novel Mechanism-based Inhibitor (6′-Bromo-5′,6′-didehydro-6′-deoxy-6′-fluorohomoadenosine) That Covalently Modifies Human Placental S-Adenosylhomocysteine Hydrolase*

Most inhibitors ofS-adenosylhomocysteine (AdoHcy) hydrolase function as substrates for the “3′-oxidative activity” of the enzyme and convert the enzyme from its active form (NAD+) to its inactive form (NADH) (Liu, S., Wolfe, M. S., and Borchardt, R. T. (1992) Antivir. Res. 19, 247–265). In this study, we describe the effects of a mechanism-based inhibitor, 6′-bromo-5′,6′-didehydro-6′-deoxy-6′-fluorohomoadenosine (BDDFHA), which functions as a substrate for the “6′-hydrolytic activity” of the enzyme with subsequent formation of a covalent linkage with the enzyme. Incubation of human placental AdoHcy hydrolase with BDDFHA results in a maximum inactivation of 83% with the remaining enzyme activity exhibiting one-third of the k cat value of the native enzyme. This partial inactivation is concomitant with the release of both Br− and F− ions and the formation of adenine (Ade). The enzyme can be covalently labeled with [8-3H]BDDFHA, resulting in a stoichiometry of 2 mol of BDDFHA/mol of the tetrameric enzyme. The 3H-labeled enzyme retains its original NAD+/NADH content. Tryptic digestion and subsequent protein sequencing of the [8-3H]BDDFHA-labeled enzyme revealed that Arg196 is the residue that is associated with the radiolabeled inhibitor. The partition ratio of the Ade formation (nonlethal event) to covalent acylation (lethal event) is approximately 1:1. From these experimental results, a possible mechanism by which BDDFHA inactivates AdoHcy hdyrolase is proposed: enzyme-mediated water addition at the C-6′ position of BDDFHA followed by elimination of Br− ion results in the formation of homoAdo 6′-carboxyl fluoride (HACF). HACF then partitions in two ways: (a) attack by a proximal nucleophile (Arg196) to form an amide bond after expulsion of F− ion (lethal event) or (b) depurination to form Ade and hexose-derived 6-carboxyl fluoride (HDCF), which is further hydrolyzed to hexose-derived 6-carboxylic acid (HDCA) and F− ion (nonlethal event).

Palmer and Abeles (11) elucidated the mechanism by which AdoHcy hydrolase catalyzes the conversion of AdoHcy to Ado and Hcy. The first step involves oxidation of the 3Ј-hydroxyl group (3Ј-oxidative activity) of AdoHcy by the enzyme-bound NAD ϩ (E⅐⅐NAD ϩ ) to form E⅐NADH and 3Ј-keto-AdoHcy, which undergoes ␤-elimination of Hcy to form 3Ј-keto-4Ј,5Ј-didehydro-5Ј-deoxyAdo. Michael-type addition of water to this tightly bound intermediate (hydrolytic activity) affords 3Ј-keto-Ado, which is reduced by E⅐NADH to yield Ado and E⅐NAD ϩ .
In recent years, significant efforts have been made in designing potent and selective inhibitors of AdoHcy hydrolase (12)(13)(14)(15)(16)(17)(18). Most inhibitors of AdoHcy hydrolase are Ado analogs, which inhibit the enzyme by serving as substrates for the 3Ј-oxidative activity of the enzyme and converting it from the active form (NAD ϩ ) to the inactive form (NADH) (6). This type of inhibitor of AdoHcy hydrolase has been defined as a type I mechanism-based inhibitor (5). McCarthy et al. (15,19) designed and synthesized (Z)-4Ј,5Ј-didehydro-5Ј-deoxy-5Ј-fluoroadenosine (ZDDFA), an analog of a tightly bound enzyme reaction intermediate, as a potential type II mechanism-based inhibitor. Type II mechanism-based inhibitors are defined as inhibitors that are activated by the enzyme and irreversibly inactivate the enzyme through covalent modification (5). Yuan et al. (20) have shown that ZDDFA is not a type II mechanismbased inhibitor but is a "pro-inhibitor" that is converted into Ado-5Ј-carboxaldehyde. Enzyme inactivation by ZDDFA does not involve covalent modification; instead, ZDDFA is converted into Ado-5Ј-carboxaldehyde by the enzyme's hydrolytic activity, which then inactivates the enzyme by the type I mechanism.
In 1991, Parry et al. (21) showed that an acetylenic analog of adenosine (9-(5Ј,6Ј-dideoxy-␤-D-ribo-hex-5Ј-ynofuranosyl)adenine) was a type II inhibitor of AdoHcy hydrolase. This acetylenic analog inactivated the enzyme by first serving as a substrate for the 3Ј-oxidative activity (converting NAD ϩ to NADH) yielding the 3Ј-keto acetylenic analog. This tightly bound in-termediate was proposed to isomerize to an electrophilic allenic ketone, which reacted with an active site nucleophile.
In this study, we report the first type II mechanism-based inhibitor of AdoHcy hydrolase, 6Ј-bromo-5Ј,6Ј-didehydro-6Ј-deoxy-6Ј-fluorohomoadenosine (BDDFHA), which does not require activation by the 3Ј-oxidative activity, but instead is activated by the enzyme's hydrolytic activity and subsequently inactivates the enzyme by forming a covalent linkage with an amino acid residue at or nearby the active site of the enzyme.

and was [8-3 H]-labeled by NEN Life Science Products.
Assay of AdoHcy Hydrolase Activity-AdoHcy hydrolase activity was assayed in both the synthetic and hydrolytic directions. In the synthetic direction, the rate of formation of AdoHcy from Ado and Hcy was measured. The enzyme was incubated with 0.2 mM Ado and 5 mM Hcy in 500 l of 50 mM potassium phosphate buffer, pH 7.2, containing 1 mM EDTA (buffer A) at 37°C for 5 min. The reaction was terminated by the addition of 25 l of 5 N HClO 4 . After the reaction mixture was kept in ice water for 15 min, the clear supernatant was collected and analyzed for AdoHcy by HPLC using a C18 reverse-phase column (Econosphere C18, 5 m, 250 ϫ 4.6 mm, Alltech, Deerfield, IL) as described earlier (20). In the hydrolytic direction, the assay was performed spectroscopically by measuring the rate of the product (Hcy) formed by reaction with DTNB as reported previously (24). To 800 l of the enzyme solution containing 4.7 g of AdoHcy hydrolase and 4 units of Ado deaminase in buffer A was added 200 l of AdoHcy (500 M) containing 250 M DTNB in buffer A. The reaction mixture was maintained at 37°C for 2 min, and the reaction was monitored at 412 nm continuously using an Hewlett-Packard 8452 diode array UV spectrophotometer (Hewlett-Packard Co., Palo Alto, CA). One unit of the enzyme activity is defined as the amount of enzyme that can synthesize or hydrolyze 1 mol of AdoHcy/min. The protein concentration was determined by the method of Bradford (25), using bovine serum albumin as a standard.
Inactivation of AdoHcy Hydrolase by BDDFHA-Time-dependent inactivation of AdoHcy hydrolase by BDDFHA was measured by incubating various concentrations of the inhibitor (2.5-20 M) with 20 nM AdoHcy hydrolase in buffer A at 37°C for different times (0 -20 min). The activity remaining was determined in the synthetic direction as described above. Pseudo-first-order rate constants (k app ) were obtained by plotting the logarithm of the remaining activity versus time, and the K i and k inact values were obtained by the Kitz and Wilson method (26). Based on the inactivation mechanism shown in Scheme 1, the K i and k inact are expressed in Equations 1 and 2.
Because k 2 and k 3 can be estimated individually as described below, k 4 can be solved according to equation (2). Time-dependent inactivation of the enzyme was also examined in the presence of the substrate Ado (100 M) by incubation of AdoHcy hydrolase (4 M) with 100 M of BDDFHA for various times. Determination of F Ϫ and Br Ϫ Ions Released from BDDFHA-F Ϫ and Br Ϫ ions released from the reaction of AdoHcy hydrolase and BDDFHA were determined by ion-exchange chromatography. The enzyme (28 M) was incubated with 300 M BDDFHA in buffer A at 37°C for various times (0, 5, 10, 20, 30, and 60 min). At each time point, an aliquot of the reaction mixture containing 900 g of protein (500 l) was withdrawn and added to a test tube containing 1.5 ml of 97% ethanol. The denatured protein was removed by filtration through an Amicon Centricon-3 microconcentrator (3000 M r cutoff), and the filtrate and washing (1 ml) were combined and lyophilized. The residue was dissolved in 100 l of water, 30 l of which was injected onto an ion-exchange column (Anion/R, 250 ϫ 4.1 mm, 10 m, Alltech) in an HPLC system equipped with an conductivity detector (690 Ion Chromatograph, Omega Metrohm, Ltd., Herisau, Switzerland) for halide ion analysis. F Ϫ and Br Ϫ ions were eluted isocratically with 1.5 mM p-hydroxybenzoic acid, 2% methanol (pH 8.5, adjusted by LiOH). The flow rate was 2.0 ml/min. Control experiments were performed using bovine serum albumin and authentic halide ions to assess the recovery from the above procedure. Recoveries routinely were 69 -72% in the control experiments. Quantitation of halide ions was achieved by comparison of the ion peak areas with that of known quantities of authentic ions using a standard curve. The first-order rate constants for the release of Br Ϫ ion (equal to k 2 ) and F Ϫ ion (approximately k 3 ϩ k 4 ) were obtained by fitting the data to an exponential function.
Determination of Ade Formation-AdoHcy hydrolase and BDDFHA were incubated in buffer A under the same conditions as described for F Ϫ and Br Ϫ ion determination, except that at each time point the reaction was stopped by addition of 10 l of 5 N HClO 4 . The precipitate was removed by centrifugation, and the supernatant was used for Ade analysis by HPLC using a reverse-phase C18 column as described previously (22). The peak assigned to Ade was confirmed by coinjection with authentic Ade on the HPLC system as well as by mass spectral analysis using chemical ionization with ammonia as the reagent gas. The first-order rate constant (k 3 ) was obtained by fitting the data to an exponential function.
Determination of E⅐NAD ϩ and E⅐NADH-Quantitation of E⅐NAD ϩ and E⅐NADH was performed using a fluorescence technique (20). The conversion process of the E⅐NAD ϩ to E⅐NADH induced by inhibitors was monitored by UV spectroscopy as described previously (22). The enzyme (2 mg) in 0.8 ml of buffer A was incubated with 500 M BDDFHA or 200 M ZDDFA at 25°C for 30 min. The inhibitor-induced NADH formation was monitored at 340 nm using an Hewlett-Packard 8452 diode array UV spectrophotometer.
Stoichiometries of Binding and Covalent Labeling-The stoichiometry of binding of BDDFHA to AdoHcy hydrolase was determined by incubating the enzyme with [8-3 H]BDDFHA. The enzyme (0.5 mg) in 250 l of buffer A was incubated with 500 M [8-3 H]BDDFHA (187 Ci/mol) at 37°C for 2 h. The reaction mixture was then passed through a Sephadex G-25 spin column to remove unbound inhibitor. The protein concentration, 3 H radioactivity, and remaining enzyme activity in the filtrate were determined, and the ratio of bound BD-DFHA (mole) to the inactivated enzyme (mole of subunit) was calculated as the stoichiometric amount of the total binding (covalent binding ϩ noncovalent binding). For determination of the stoichiometry of covalent labeling, the filtrate from the spin column as described above was denatured with 8 M urea and gel-filtered again through the spin column to remove noncovalently bound [8-3 H]BDDHFA. The radioactivity and the protein content in the filtrate were used to calculate the stoichiometry of the covalent labeling.

Isolation and Identification of [8-3 H]BDDFHA-labeled
Peptides-Enzyme (2 mg) was dissolved in 0.8 ml of buffer A and incubated with 1 mM of [8-3 H]BDDFHA (189 Ci/mol) at 37°C for 3 h. The reaction mixture was then gel-filtered through a Sephadex G-25 spin column to remove the free [8-3 H]BDDFHA. The enzyme-inhibitor complex was denatured with 8 M urea and passed through the spin column again to remove any noncovalently bound [8-3 H]BDDFHA and to change the buffer to buffer A containing 1 M urea. A freshly prepared solution of L-1-tosylamido-2phenylethyl chloromethyl ketone-trypsin was added to give an enzymeto-substrate ratio of 1:50 (by weight). After incubation for 5 h at 37°C, a second addition of the same amount of trypsin was made, and the digestion was continued for another 5 h. After lyophilization, the trypsin digest was dissolved in 0.1% trifluoroacetic acid, and the peptides were analyzed on a Vydac C18 protein and peptide column (Vydac 218 TP54, C18, 5 m, 0.46 ϫ 25 cm). The solvent system consisted of solvent I (0.1% trifluoroacetic acid) and solvent II (80% CH 3 CN, 20% H 2 O, 0.07% trifluoroacetic acid). The initial conditions were 2% solvent II with a linear gradient to 70% solvent II over 120 min at a flow rate of 0.5 ml/min. The UV absorbance of the eluted peptides was monitored at 220 nm. The radioactivity in the fractions collected (0.5 ml) was measured by liquid scintillation counting. Peptide peaks containing significant radioactivity were collected and concentrated by Speed-Vac and rechromatographed on the same column using the initial conditions of 20% solvent II with a linear gradient to 60% solvent II over 60 min. Detection of peptides and measurement of radioactivity were the same as described above except that fractions were collected manually peak by peak.
The isolated peptides were sequenced by automated Edman degradation on an Applied Biosystems 473A protein sequencer in the Biotechnology Laboratory at Kansas State University, Manhattan, KS. At each sequencing cycle, the washings from the conversion flask and eluate from the HPLC column were collected and combined for determination of radioactivity.

Inactivation of AdoHcy Hydrolase by BDDFHA-When
AdoHcy hydrolase (4 M) was incubated with a large molar excess of BDDFHA (100 M), time-dependent loss of activity was observed as shown in Fig. 1. However, the enzyme was not completely inactivated, and the maximal inactivation was 83% of the original enzyme activity. The enzyme was substantially protected from inactivation by BDDFHA by inclusion of the substrate Ado at a concentration equal to the concentration of BDDFHA (100 M) (Fig. 1). Kinetic analysis using the Kitz and Wilson method gave a K i value of 3.9 M, and k inact value of 0.04 min Ϫ1 for BDDFHA inactivation of human AdoHcy hydrolase. The enzyme inactivation by BDDFHA was irreversible, i.e. the activity of the inactivated enzyme could not be recovered by dialysis or gel filtration.  Fig. 2, when a large molar excess of BDDFHA was incubated with AdoHcy hydrolase, Br Ϫ ion (retention time of 6.78 min) and F Ϫ ion (retention time of 2.60 min) were observed on the ion-exchange chromatogram, with maxima of about 1.1 mol of Br Ϫ ion and 0.9 mol of F Ϫ ion/mol of inactivated enzyme subunit. The rate of Br Ϫ ion formation (k 2 ) was faster than that of F Ϫ ion formation (assumed approximately equal to k 3 ϩ k 4 ). The first-order rate constant for Br Ϫ ion release was estimated to be 0.27 min Ϫ1 (Fig. 2, inset), and the rate for F Ϫ ion formation was 0.098 min Ϫ1 , indicating that Br Ϫ ion was released first followed by F Ϫ ion.

Stoichiometry of Binding and Covalent Modification-After
Formation of Ade-Incubation of AdoHcy hydrolase with BDDFHA resulted in the formation of Ade as shown in Fig. 3. Two main peaks appeared on the HPLC chromatogram. The peak with retention time of 14.84 min was unreacted BDDFHA, and the peak with retention time of 9.69 min was Ade as determined by its same retention time as authentic Ade upon coinjection on HPLC. Its chemical ionization mass spectrometric analysis gave an ion (MH ϩ ) at m/z 136, which is consistent with the mass of Ade. The first-order rate constant of Ade release (k 3 ) from BDDFHA was estimated to be 0.056 min Ϫ1 which is slower than the rates of both Br Ϫ and F Ϫ ion release. The maximal Ade formation was approximately 0.  mol of Ade/mol of the inactivated enzyme subunit. Effect of BDDFHA on E⅐NAD ϩ -Inactivation of AdoHcy hydrolase by BDDFHA did not involve the reduction of E⅐NAD ϩ to E⅐NADH as determined by a fluorescence method (data not shown) and by a UV spectroscopic method as shown in Fig. 4. Incubation of AdoHcy hydrolase with BDDFHA resulted in about 60% inactivation of the enzyme in 20 min as shown in Fig. 1, whereas no NADH formation was observed when the reaction was monitored at 340 nm for NADH as shown in Fig.  4, line a. By contrast, incubation of the enzyme with ZDDFA, a type I mechanism-based inhibitor, resulted in complete inactivation of the enzyme and conversion of E⅐NAD ϩ to E⅐⅐NADH in 3 min (Fig. 4, line c). When the enzyme was first incubated with BDDFHA, and then the enzyme⅐BDDFHA complex (with 30% remaining activity) was incubated with ZDDFA, the enzyme was completely inactivated and about half of the E⅐NAD ϩ was converted to E⅐NADH as shown in Fig. 4,  again. The radiolabeled protein was then subjected to tryptic digestion, and the peptides were separated by reverse-phase HPLC. Fig. 5 shows the HPLC chromatogram of the tryptic digested enzyme modified by [8-3 H]BDDFHA. The radioactivity was distributed among three major fractions (a, b, and c) as indicated by the bars below the chromatogram. Fraction a (about 25% of total radioactivity) was from incompletely digested large protein fragments, and further digestion of this fraction by trypsin resulted in the formation of fractions b and c (data not shown). Fractions b and c (about 15 and 55% of the total radioactivity, respectively) were collected separately and rechromatographed twice on the same column with different elution conditions. As shown in Fig. 5, inset, fraction c was purified to one major and several minor components with radioactivity associated only with the major component (fraction c'). However, rechromatography of fraction b was not successful enough to obtain completely separated radioactive component required for protein sequencing. The identity of this labeled peptide(s) in fraction b remains unknown. The HPLC-purified radiolabeled peptide (approximately 200 pmol) from fraction cЈ was subjected to N-terminal protein sequencing. Results from the first 13 cycles of the protein sequencing revealed that fraction cЈ had an amino acid sequence of 1 Ser-Lys-Phe-Asp-Asn-Leu-Tyr-Gly-X-X-Glu-Ser-X 13 , in which X represents unidentified amino acid residues. The tritium radioactivity was found to be only associated with the unidentified residue X in cycle number 10. By comparing the amino acid sequencing deduced from the cDNA sequencing encoding human placental AdoHcy hydrolase with the sequence obtained from fraction cЈ, it was found that fraction cЈ contained a fragment of the hydrolase protein from Ser 187 to Leu 199 with a sequence of Ser-Lys-Phe-Asp-Asn-Leu-Tyr-Gly-Cys-Arg-Glu-Ser-Leu. The amino acid residue associated with radioactivity was then identified as Arg 196 . DISCUSSION Pharmacological modulation of intracellular methylation can be achieved through feedback inhibition of methyltransferase activity by AdoHcy (2). Intracellular AdoHcy concentrations can be elevated by decreasing AdoHcy hydrolase activity (27). The tryptic digested enzyme was applied to a Vydac C18 protein and peptide column (Vydac 218 TP54, C18, 5 m, 0.46 ϫ 25 cm). Elution was carried out with a gradient of acetonitrile in 0.1% trifluoroacetic acid from 2 to 70% in 120 min. For rechromatography (inset), the elution gradient was from 20 to 60% acetonitrile in 60 min. The bar graphs indicate the radioactivity in fractions from the HPLC chromatography. Inset, profile of the rechromatography of fraction cЈ.
Many of these nucleoside analogs are type I mechanismbased inhibitors of AdoHcy hydrolase, which inactivate the enzyme by reducing the E⅐NAD ϩ to E⅐NADH in an irreversible manner but not involving covalent linkage between the enzyme and inhibitor. Various attempts have been made to prepare type II mechanism-based inhibitors of this enzyme (14,15,19,21), which would be activated by the enzyme's catalytic activity and inactivate the enzyme through covalent modification. This type of mechanism-based inhibitor, starting with relatively unreactive structures that are activated only by specific targeted enzymes, often have better pharmacological specificity (29). ZDDFA was designed and synthesized as a potential type II mechanism-based inhibitor (15). Through mechanistic studies on the enzyme inhibition by ZDDFA (20) and its homoanalog EDDFHA (30), it has been demonstrated that ZDDFA is not a type II mechanism-based inhibitor, instead, it is a proinhibitor, which inactivates the enzyme by a type I mechanism. More importantly, with these molecular tools (ZDDFA and EDDFHA), we have demonstrated that AdoHcy hydrolase possesses 5Ј-and 6Ј-hydrolytic activities that are independent of the 3Ј-oxidative activity (20,30). This finding has led us to the design of BDDFHA which targets the enzyme's 6Ј-hydrolytic activity to generate a strong electrophilic center (i.e. acyl fluoride analog) in the active site, which could then react with a juxtaposed protein nucleophile to form a covalent adduct with the enzyme. Evidence obtained in this study strongly supports the conclusion that BDDFHA is the first type II mechanismbased inhibitor of AdoHcy hydrolase, which relies only on the enzyme's hydrolytic activity for activation. Scheme 2 contains the proposed mechanism by which BDDFHA inactivates the enzyme. In this mechanism, addition of enzyme-mediated water at the C-6Ј position of BDDFHA followed by elimination of Br Ϫ ion results in the formation of ho-moAdo 6Ј-carboxyl fluoride (HACF). HACF then partitions in two ways: (a) attack by a protein nucleophile (e.g. Arg 196 ) forms an amide bond after expulsion of F Ϫ ion or (b) depurination releases Ade and forms hexose-derived 6-carboxyl fluoride (HDCF) (21,30). HDCF is further hydrolyzed to hexose-derived 6-carboxylic acid (HDCA) and F Ϫ ion. In contrast to the type I mechanismbased inhibitors, inactivation of AdoHcy hydrolase by BDDFHA does not involve reduction of E⅐NAD ϩ to E⅐NADH. No NADH was formed during the enzyme inactivation process, and the inactivated enzyme retains its original NAD ϩ /NADH content.
Evidence consistent with the pathway leading to enzyme inactivation includes the observations that incubation of AdoHcy hydrolase with BDDFHA results in release of both Br Ϫ and F Ϫ ions as well as covalent labeling of the enzyme. The first-order rate of Br Ϫ ion release (k 2 ϭ 0.27 min Ϫ1 ) is greater than the rate of F Ϫ ion release (k app ϭ 0.098 min Ϫ1 ), in agreement with the prediction that Br Ϫ is a better leaving group than F Ϫ . Release of Br Ϫ affords HACF, which contains a strong electrophilic center. Nucleophilic attack at C-6Ј of HACF by Arg 196 results in elimination of F Ϫ ion and covalent linkage between Arg 196 and C-6Ј carboxylic group of HACF, probably via formation of an amide bond. The stoichiometry of covalent labeling of AdoHcy hydrolase by BDDFHA is found to be only 2 mol of inhibitor/mol of the tetrameric enzyme, indicating that two of the four subunits are covalently labeled and the other two are not. The two unlabeled enzyme subunits are still enzymatically viable but with reduced activity. The k cat value for the BDDFHA-inactivated enzyme is 1.2 s Ϫ1 , which is only one-third of that of the native enzyme (k cat ϭ 3.6 s Ϫ1 , in the hydrolytic direction). From these k cat values, the calculated remaining enzyme activity theoretically should be 16.7%, which agrees well with the experimental result (approximately 20%). This remaining enzyme activity can be further inhibited by ZDDFA (Table I) via conversion of the E⅐NAD ϩ to NADH (Fig. 2, line b). Covalent modification of Arg 196 on two of the subunits by BDDFHA appears to cause a conformational change that disrupts the 3Ј-oxidative activity of the two unmodified subunits. This conformational change is transmitted, presumably through intersubunit contact, to the neighboring subunit which loses its catalytic activity toward BDDFHA but retains reduced enzymatic activity toward the substrate AdoHcy or the inhibitor ZDDFA. Similar inactivation of two subunits that affected the other two subunits in a tetrameric enzyme was also observed by Abeles et al. (31) upon inactivation of calf liver AdoHcy hydrolase with 2Ј-deoxyadenosine. Parry et al. (21) also reported that inactivation of AdoHcy hydrolase by an acetylenic analogue of adenosine resulted in covalent modification of two subunits. This acetylenic analog of adenosine, which also appears to be a type II inhibitor, needs to be activated by the enzyme's 3Ј-oxidative activity leading to conversion of two equivalents of NAD ϩ to NADH. Parry et al. (21) reported that the two remaining equivalents of NAD ϩ in the tetramer were "released from the enzyme." These results suggest subunit-subunit interaction involving the binding of the cofactor NAD ϩ .
Evidence in support of the nonlethal pathway observed with BDDFHA derives from the observation that incubation of AdoHcy hydrolase with BDDFHA results in the formation of Ade (0.6 mol/enzyme subunit) (Fig. 3). Moreover, the difference in the stoichiometries of total F Ϫ ion release (0.9 mol/enzyme subunit) and covalent modification (0.5 mol/enzyme subunit) suggests that there must be other pathways that produce F Ϫ ion. The proton on C-5Ј of HACF could be abstracted by a base from the enzyme, which could result in depurination via a retro-Michael addition process to generate Ade and HDCF. SCHEME 2. Possible mechanism by which BDDFHA inactivates AdoHcy hydrolase.
Water attack at C-6 of HDCF results in the release of F Ϫ ion and formation of HDCA. Combination of these pathways produces about equal mol of Br Ϫ and F Ϫ ions, which is approximately equal to the sum of the Ade and covalent incorporation. Based on equation (2), k 4 is calculated to be 0.057 min Ϫ1 , which is equal to k 3 (0.056 min Ϫ1 ). Therefore, the partition ratio of the nonlethal to lethal pathways, or k 3 /k 4 , is approximately 1:1, indicating one lethal event (enzyme inactivation via covalent modification) for every two enzymatic turnovers.
An important criteria for a mechanism-based inhibitor is the formation of a covalent linkage between the enzyme and the inhibitor. Arg 196 was identified as the major nucleophile that attacks the electrophilic center in HACF resulting in formation of a covalent linkage. This most likely is an amide bond between C-6Ј of HACF and the guanidium group of Arg 196 . It is possible that Arg 196 is located at the enzyme active site close to the bound C-5Ј or C-6Ј of the inhibitor. This is supported by the fact that Arg 196 is next to Cys 195 , which was identified as an essential residue at the active site of the enzyme. It may play a role in maintaining the 3Ј-reduction potential for regeneration of the NAD ϩ form of the enzyme from the NADH form (24). Arg 196 is also next to Glu 197 , which was demonstrated to be at the enzyme active site by limited proteolytic studies (32). In addition, Arg 196 is only about 10 amino acid residues from a peptide (Val 175 -Lys 186 ) located in the Ade-ring-binding region (33) and 14 amino acid residues from the NAD ϩ binding region (34). However, it is also possible that Arg 196 is not positioned closely to the C-5Ј and C-6Ј of HACF bound in the enzyme active site, but is located in such a position that allows it to react with HACF dissociating from the active site. This may explain the observation that multiple radiolabeled tryptic peptides (fragments b and c in Fig. 5) were found after incubation of the enzyme with [8-3 H]BDDFHA, indicating that Arg 196 is not the only amino acid residue reacting with C-6Ј of HACF. In addition, results from this study do not tell if Arg 196 is the base that abstracts the C-4Ј proton to form a Michael acceptor in the normal enzymatic catalysis of AdoHcy hydrolase in the Palmer-Abeles mechanism (11).
To our knowledge, BDDFHA is the first type II mechanismbased inhibitor of AdoHcy hydrolase that relies only on the enzyme's hydrolytic activity for activation. This high specificity of enzyme inactivation should diminish cytotoxicity resulting from other metabolic effects induced by most type I mechanism-based inhibitors (23). The antiviral activity and cytotoxicity of this first type II mechanism-based inhibitor are currently under investigation. 2