Stereospecificity and Kinetic Mechanism of Human Prenylcysteine Lyase, an Unusual Thioether Oxidase* 210

Prenylated proteins contain either a 15-carbon farnesyl or a 20-carbon geranylgeranyl isoprenoid covalently attached to cysteine residues at or near their C terminus. The cellular abundance of prenylated proteins, as well as the stability of the thioether bond, poses a metabolic challenge to cells. A lysosomal enzyme termed prenylcysteine lyase has been identified that degrades a variety of prenylcysteines. Prenylcysteine lyase is a FAD-dependent thioether oxidase that produces free cysteine, an isoprenoid aldehyde, and hydrogen peroxide as products of the reaction. Here we report initial studies of the kinetic mechanism and stereospecificity of this unusual enzyme. We utilized product and dead end inhibitors of prenylcysteine lyase to probe the kinetic mechanism of the multistep reaction. The results with these inhibitors, together with those of other experiments, suggest that the reaction catalyzed by prenylcysteine lyase proceeds through a sequential mechanism. The reaction catalyzed by the enzyme is stereospecific, in that the pro-S hydride of the farnesylcysteine is transferred to FAD to initiate the reaction. With (2R,1′S)-[1′-2H1]farnesylcysteine as a substrate, a primary deuterium isotope effect of 2 was observed on the steady state rate. However, the absence of an isotope effect on an observed pre-steady-state burst of hydrogen peroxide formation implicates a partially rate-determining proton transfer after a relatively fast C–H (C–D) bond cleavage step. Furthermore, no pre-steady-state burst of cysteine was observed. The finding that the rate of cysteine formation was within 2-fold of the steady-statek cat value indicates that cysteine production is one of the primary rate-limiting steps in the reaction. These results provide substantial new information on the catalytic mechanism of prenylcysteine lyase.

Protein prenylation is a type of posttranslational lipid modification involving the covalent attachment of either a 15-car-bon farnesyl or a 20-carbon geranylgeranyl isoprenoid via thioether linkages to a conserved cysteine of many cellular proteins (1,2). The majority of prenylated proteins participate in biological regulatory events occurring at the cytoplasmic surface of cell membranes. Most prenylated proteins contain the so-called CaaX motif, in which the "C" is the modified cysteine, the "a" residues are usually aliphatic amino acids, and the X residue can be one of several amino acids. Following prenylation, most CaaX-type proteins are subject to two additional processing steps: the three C-terminal residues are proteolytically removed, and the new C-terminal prenylcysteine is subject to carboxyl methylation (3)(4)(5).
Unlike many other posttranslational modifications of proteins, prenylation is a stable modification of proteins (6,7). Prenylated proteins comprise up to 2% of total cellular protein and have an approximate half-life of 20 h (6, 7); however, accumulation of prenylcysteines arising from degradation of prenyl proteins is expected to place a major burden on cells. Since free prenylcysteines and analogs have a number of pharmacological effects on cells (8 -11), the mechanisms for cellular disposal of the prenylcysteines produced during the normal turnover of prenylated proteins are important to establish.
Studies aimed at elucidating the metabolism of prenylated proteins resulted in the identification of a lysosomal enzyme that catalyzes the degradation of prenylcysteines and their methyl esters (12,13). This enzyme has been termed prenylcysteine lyase (Pcly). 1 The basic features of the chemical mechanism through which Pcly catalyzes its reaction have been elucidated (14). The main features are as follows: (i) the products of the reaction are free cysteine and the C-1 aldehyde of the isoprenoid moiety (farnesal or geranylgeranial); (ii) the enzyme utilizes molecular oxygen as a co-substrate; (iii) Pcly uses a noncovalently bound FAD cofactor in an NAD(P)Hindependent manner; and (iv) a stoichiometric amount of hydrogen peroxide is produced during the course of the reaction. The last point presumably indicates that no oxygen atom from the molecular oxygen is incorporated into the farnesal product. A chemical mechanism for Pcly that can account for the above observations has been proposed (14). In the proposed mechanism ( Fig. 1), the reaction is initiated by a hydride transfer from C-1 of the isoprenoid moiety of FC to FAD, producing reduced flavin and a probably short-lived sulfur-stabilized carbocation intermediate. Nucleophilic attack of a water molecule on the carbocation results in the formation of a hemithioacetal intermediate, which collapses to the isoprenoid aldehyde with simultaneous C-S bond breakage to produce cysteine. In addition, the hydride of the reduced flavin is transferred to molecular oxygen, resulting in hydrogen peroxide formation and reoxidation of the flavin, thus allowing the enzyme to participate in subsequent turnovers.
The mechanism proposed for Pcly (i.e. oxidative cleavage of a thioether bond without net oxidation at sulfur) appears to be unprecedented in biology. The finding that cysteine is produced distinguishes it from another enzyme that might act on prenylcysteines, namely carbon-sulfur-␤-lyase, which would produce a prenylthiol (15). Furthermore, the lack of a NAD(P)H requirement suggests that Pcly uses a mechanism distinct from cytochrome P450 and flavin-containing monooxygenases that catalyze S-oxidation of FC and other thioethers (16 -18). In addition, Pcly has essentially no sequence similarity to known enzymes, which also attests to its unusual mechanism of action (13). Another interesting facet of Pcly is its possible similarity in biological function to another lysosomal enzyme called palmitoyl-protein thioesterase that catalyzes the degradation of thioesters derived from S-acylated proteins (19). This enzyme has been implicated in a lethal neurodegenerative lysosomal storage disorder, infantile neuroceroid lipofuscinosis (20). These observations stress the importance of understanding both the biology and the mechanism of Pcly and indicate an important biological role for enzymes involved in the metabolism of lipidated proteins.
In the present work, we report studies on the mechanism of this unusual enzyme, including steady-state and pre-steadystate kinetics as well as an analysis of the stereochemistry of hydride transfer. From these studies, we propose a kinetic model for Pcly.

EXPERIMENTAL PROCEDURES
Materials-Farnesylcysteine (FC) and [ 35 S]FC were synthesized as previously described (13,21). Farnesal was also synthesized as previously described (14). (E,E)-Farnesol (trans, trans) was purchased from Sigma. The deuterium-labeled FCs, (2R,1ЈS)-[1Ј-2 H 1 ]FC and (2R,1ЈR)-[1Ј-2 H 1 ]FC, were synthesized from (E,E)-farnesol by known methods with some modifications (14,22,23). Details of the synthetic procedures are provided in the supplemental information. The purity (Ն95% unless indicated otherwise) of the stable intermediates was established by GC and/or 1 H NMR spectroscopic analyses, and their identities were validated by correspondence of the NMR spectral data with values given in the literature references. All deuterated compounds were purified by flash column chromatography on silica gel (24) except the FCs, which were purified by recrystallization.
Production and Purification of Recombinant His 6 -Pcly-Recombinant baculovirus containing a hexahistidine-tagged Pcly was produced as previously described (14). Several spinner cultures of Sf9 cells (18liter total volume) in HyClone HyQ CCM-3 serum-free media were seeded at a density of 1.5 ϫ 10 6 cells/ml and then infected at a multiplicity of infection of about 1 with the human His 6 -Pcly virus. Cells were harvested 49 h after infection.
Recombinant His 6 -Pcly was purified from extracts of infected Sf9 cells by a combination of nickel affinity and anion exchange chromatography as described (14) with the following modifications. The cell pellet from an 18-liter culture of infected Sf9 cells was resuspended in 250 ml of 50 mM Tris-Cl, pH 7.7, 0.2 mM EGTA, 0.2 mM EDTA, containing a mixture of protease inhibitors. Cells were disrupted using a glass-Teflon homogenizer, and the suspension was centrifuged at 100,000 ϫ g for 1 h. Pcly was extracted from the resulting pellet with 250 ml of 50 mM Tris-Cl, pH 7.7, containing 100 mM NaCl, 0.5% Triton X-100 and the protease inhibitor mixture, followed by centrifugation at 100,000 ϫ g for 1 h. The supernatant was loaded onto a 7-ml column of nickel-nitrilotriacetic acid resin (Qiagen). The column was then washed with 10 column volumes of 50 mM Tris-Cl, pH 7.7, 150 mM NaCl, 8 mM imidazole, and 0.2% CHAPS. Pcly was eluted with a 44-ml linear gradient of imidazole (8 -500 mM) in the above buffer. Fractions containing Pcly activity were pooled and dialyzed against 20 mM Tris-Cl, pH 7.7, containing 0.2% CHAPS (buffer A). The dialysate was loaded onto a QHP 16/10 column (Amersham Biosciences), which had been equilibrated in buffer A. Pcly was eluted with a 120-ml gradient of NaCl from 0 to 500 mM in buffer A. Fractions containing Pcly were pooled, dialyzed as above, and stored at Ϫ80°C.
Pcly Stereochemistry-To determine the stereochemistry of the Pcly reaction, three separate incubations were set up in which Pcly used either unlabeled FC, (2R,1ЈS)-[1Ј-2 H 1 ]FC, or (2R,1ЈR)-[1Ј-2 H 1 ]FC as a substrate. Reaction mixtures were prepared in glass vials that were prewashed in hexane and contained the appropriate farnesylcysteine (500 M), 12.5 M (for the reactions with unlabeled FC and the Risomer) or 20 M (for the reaction with the S-isomer) Pcly, 6 mM Zwittergent 3-14, 10 mM Tris-Cl, pH 7.7, and 15% glycerol. The mixtures were incubated for 1.5 h (for the reactions with unlabeled FC and the R-isomer) or 2.5 h (for the reaction with the S-isomer) at 37°C, whereupon NaBH 4 (70 mM final concentration from a stock in 14 N NaOH) was added, and the incubation continued for 30 min at room temperature. Formaldehyde (70 mM) was added to quench the NaBH 4 , followed by the addition of 0.4 M HCl after 30 min.
The labeled farnesols from the Pcly incubation solutions were extracted with heptane (10 -15 l). GC analysis of aliquots in splitless mode on a Restek Rtx-5 fused silica capillary column (5% diphenyl, 95% dimethyl polysiloxane stationary phase) showed peaks for (Z,E)-and (E,E)-farnesols (1:2 ratio; R t 13.4 and 13.7 min). Low resolution field ionization GC-mass spectrometry analyses with the same column and conditions using a VG 70-VSE instrument provided the following intensity data for the molecular ion regions and isotope contents: farnesol reference: m/z (relative intensity) (Z,E): 220 (2. The deuterium contents were calculated using an in-house program based on a standard method (25). The small quantities of the farnesol samples analyzed introduced scatter into low m/z data, which results in errors in the isotope content of about Ϯ1-2%. The anomalously high d 1 content in the (Z, E)-farnesol from the R,S-labeled substrate is attributed to a smaller quantity and background interference.
Enzyme Kinetics Steady State-Pcly activity was routinely measured by two methods. Unless otherwise noted, assays were performed in air-saturated buffers at 25°C in which the oxygen concentration was 242 M (26). The most frequently used assay involved measuring hydrogen peroxide produced in the Pcly reaction using the H 2 O 2 Amplex Red kit (Molecular Probes, Inc., Eugene, OR), as described previously (14). Briefly, Pcly was incubated in assay buffer containing 50 mM sodium phosphate, pH 7.4, 1 unit/ml horseradish peroxidase, and 200 M Amplex Red reagent at 25°C (100-l final volume). The reactions were initiated with FC. In the presence of horseradish peroxidase, the Amplex Red reagent reacts with H 2 O 2 to produce the highly fluorescent product resorufin. Fluorescence was measured on a PerkinElmer Life Sciences LS50B fluorimeter using an excitation wavelength of 563 nm and an emission wavelength of 587 nm. Typically, the increase in fluorescence with time was measured for 5 min. A background fluorescence rate, determined for a control reaction lacking FC, was subtracted from experimental values. The actual reaction rates were calculated either from a resorufin or hydrogen peroxide standard curve.
The second assay involved following the conversion of [ 35  quenched with 10 l of isopropyl alcohol/ammonium hydroxide/H 2 O (60:30:10), and 17.5 l of this mixture was spotted onto a silica TLC plate (PE SILG; Whatman, Clifton, NJ). The plates were developed for 2 h in the quenching buffer mentioned above. After drying, the 35 Slabeled compounds were located by autoradiography and excised from the plates, and radioactivity was quantitated by scintillation counting.
To determine the steady state kinetic parameters k cat and K m , the concentration of FC was varied, and the assays were performed in air-saturated buffer as described above. Initial velocity data were fit to the Michaelis-Menten equation (Equation 1) using Kaleidagraph software (Abelbeck Software), where is the initial velocity, V m is the maximum velocity, and K m is the Michaelis constant of FC. Primary deuterium isotope effects were measured using the substrates ( When the oxygen concentration was to be varied, the hydrogen peroxide coupled assay was performed. The oxygen was removed from the initial enzyme reaction mixture by preparing all of the reagents in a 830-ABC series anaerobic chamber (PLAS Labs, Inc., Lansing, MI), in which the main chamber had been alternatively filled and evacuated with ultrapure nitrogen. All of the reagents except FC were added to an anaerobic quartz semimicroseptum cell. The cells were then taken out of the chamber, and different O 2 /N 2 gas mixtures were bubbled through the solutions for 20 -30 min at 25°C to obtain final oxygen concentrations of 57, 346, 577, and 1.16 mM. Reactions were initiated by adding a saturating concentration of FC to the cuvette through a gas-tight syringe and reading the fluorescence increase with time as described above.
Inhibitor Kinetics-Inhibition analysis was performed by varying inhibitor and FC concentrations at a fixed concentration of oxygen (air saturation, 242 M). Data were analyzed by Lineweaver-Burk plots, and initial rate data were fit to Equation 2, which describes noncompetitive inhibition, using KinetAsyst software II (Intellikinetics). The best fits were determined by the relative fit errors. Subsequently, the lines (Fig. 4) are the fit of the data to Equation 2.
The nomenclature is that of Cleland (27). is the initial velocity; V m is the maximum velocity; A is the substrate concentration; I is the inhibitor concentration; K a is the apparent Michaelis constant for A, and K is and K ii are the slope and intercept inhibition constants, respectively. Pre-steady-state Kinetics-Pre-steady-state experiments were performed on an Applied Photophysics SX.18MV stopped-flow spectrophotometer using a 0.2-cm path length cell at 25°C. The production of hydrogen peroxide was monitored by fluorescence (excitation wavelength 563 nm, 590 nm cut-off emission filter) using the coupled assay described above. Enzyme and substrates were diluted 2-fold in assay buffer. Concentrations indicated throughout or in the figure legends are the final concentrations after dilution. The time course of fluorescence can be described by a single exponential equation with a steady-state term (Equation 3), where F t is the fluorescence at time t, ⌬F is the amplitude of the burst, k obs is the observed first-order rate constant governing the burst phase, and is the linear rate of increase in fluorescence during steady state. The maximum value of k obs, denoted k burst, was calculated according to Equation 4, where K app is the concentration of varied substrate at one-half of the maximum value of k burst .

Chemical Synthesis of (2R,1ЈR)-and (2R,1ЈS)-[1Ј-2 H 1 ]
Farnesylcysteines-Farnesols stereoselectively labeled with deuterium in the pro-R and pro-S positions at C-1 were prepared by asymmetric reductions of farnesal-d 1 with (S)-and (R)-Alpine-Boranes (Fig. 2) (28, 29). The enantiomeric purities of the resulting (R)-and (S)-[1Ј-2 H 1 ]farnesols were 97 and 95% based on 1 H NMR analyses of the corresponding camphanate esters (23,30). Chromatographic purification of the labeled farnesols proved difficult due to persistent contamination by an impurity presumed to be pinanol arising from oxidation of the pinylborane reagent. An effective purification of the S-enantiomer was ultimately accomplished by selective acetylation of farnesol, chromatography, and acetate cleavage. The labeled FC diastereomers were obtained by conversion to the farnesyl chlorides with N-chlorosuccinimide-dimethyl sulfide reagent (31) followed by S-alkylation of cysteine (32). Since both reactions are expected to proceed by S N 2 displacements with inversion of configuration (28,33), the overall stereochemical outcome is retention affording (2R,1ЈR)-and (2R,1ЈS)-[1Ј-2 H 1 ]FCs. The diastereo purities of 91 and 95%, respectively, determined from their 1 H NMR spectra in ethanol-d 6 , reveal that some racemization occurred, probably in forming and/or handling the labile chloride intermediate.
Stereochemistry of the Pcly Reaction-In the mechanism proposed previously for Pcly (14) (See Fig. 1), a hydride from the C-1 of the isoprenoid moiety of FC is transferred to FAD, forming a sulfur-stabilized carbocation intermediate and reduced flavin. It is expected that either the pro-S or pro-R hydride will be transferred to FAD, whereas the other will be retained in the product. Pcly was incubated with unlabeled FC or with each of the deuterated isomers (see "Experimental Procedures"). After incubation, the reactions were treated with NaBH 4 to reduce the farnesal product to the corresponding alcohol, farnesol. The deuterium content in the trans, transfarnesol was determined by GC-mass spectrometry analysis. The results, summarized in Table I, reveal that Pcly action on (2R,1ЈS)-[1Ј-2 H 1 ]FC yields an isoprenoid product with essentially no deuterium retention. On the other hand, the isoprenoid product of Pcly action on (2R,1ЈR)-[1Ј-2 H 1 ]FC retains essentially all of the deuterium. These results indicate that the Pcly reaction is stereospecific and that the pro-S hydrogen of the FC substrate is transferred during the reaction.
Steady-state Kinetics and Isotope Effects-Initial steadystate kinetic analysis of Pcly revealed typical Michaelis-Menten behavior using FC as a substrate (Fig. 3). The kinetic parameters are summarized in Table II. Pcly is a very slow enzyme, with a k cat of 0.008 s Ϫ1 . Whereas this turnover number is low compared with many other oxidases (34 -39), it is similar to those of the mammalian CaaX prenyltranferases that attach the isoprenoid moiety to the Cys residue (40,41). The K m value determined for FC using the recombinant enzyme (3 M) is similar to that obtained for Pcly purified from native tissue (12). Kinetic parameters were also determined for Pcly action on the two deuterated isomers (Fig. 3; Table II (Table II). The observation of a primary kinetic deuterium isotope effect with the 1ЈS-isomer of FC not only confirmed the stereochemistry result discussed above but also suggested that hydride transfer is partially rate-limiting in steady-state catalysis.
Pcly utilizes molecular oxygen as a co-substrate in its reaction (14). To examine the oxygen dependence of the reaction, assays were conducted over a range of oxygen concentrations from 0.057 to 1.16 mM (see "Experimental Procedures"). Interestingly, no concentration dependence was observed over this range of oxygen concentrations (data not shown), although this range of oxygen concentrations encompassed most of the reported K m values for other flavin-dependent oxidases (42). These data indicate that Pcly has a very low K m for oxygen. Perhaps important in this regard is the fact that prenylcysteines are recognized with such high affinity by Pcly and that K m values for oxygen of flavin-dependent oxidases are generally an order of magnitude lower than those for organic substrates (38 -39, 42). Since the K m value for FC is quite low (3 M), it is perhaps not surprising that the K m for oxygen is also low. From a practical standpoint, the finding that the oxygen K m is Ͻ50 M allowed subsequent kinetic analysis to be per-formed in air-saturated buffers, which contained a saturating (for Pcly) oxygen concentration of 242 M.
Kinetic Studies of Pcly Utilizing Product and Dead End Inhibitors-A detailed knowledge of the kinetic mechanism of an enzyme is very important for several reasons. Inhibitor design, future pre-steady-state kinetic analysis, and detailed structure-function studies are all dependent on an understanding of the basic steady state kinetic mechanism. For a bisubstrate enzyme such as Pcly, there are two basic mechanisms: (i) sequential (which can involve either random or compulsory ordered addition of substrates), in which an obligate ternary complex forms, and (ii) ping-pong, in which the first product is released prior to binding of the second substrate (43). Classically, initial velocity studies in which both substrate concentrations are varied would be used to distinguish between a ping-pong mechanism and a sequential mechanism (44 -47). However, since we could not stably achieve low enough oxygen concentrations to perform this type of a study with Pcly, an alternate technique involving analysis of product and dead end inhibition patterns was employed.
The action of Pcly on FC generates three products: cysteine, hydrogen peroxide, and farnesal (Fig. 1). We first examined cysteine and hydrogen peroxide as product inhibitors of Pcly. Surprisingly, cysteine did not inhibit Pcly up to the highest achievable concentration of 100 mM (data not shown), indicating a very weak affinity of this product for the enzyme. Whereas analysis of hydrogen peroxide inhibition can sometimes be complicated, since this oxidizing agent can irreversibly inhibit enzymes (44), this phenomenon did not appear to be the case for Pcly. In fact, hydrogen peroxide did not inhibit Pcly up to a concentration of 5 mM (data not shown), indicating very weak affinity for this product as well.
We then tested farnesal as a product inhibitor of Pcly. The experiment was performed with FC as the variable substrate and oxygen concentration fixed at air saturation in the presence of a range of farnesal concentrations. Analysis of the data via a double-reciprocal plot of 1/ versus 1/[FC] (Fig. 4A) revealed a primarily noncompetitive inhibition pattern for farnesal versus FC. The data do not completely rule out a potentially more complex pattern, termed mixed inhibition, since the lines do not intersect exactly on the 1/[FC] axis. However, both noncompetitive and mixed inhibition patterns would clearly indicate that the binding sites of FC and farnesal do not overlap. Thus, farnesal does not bind to the same form of the enzyme as FC (i.e. the unliganded oxidized state).
The prenyl alcohol farnesol can be considered a dead end inhibitor of the Pcly reaction, since this compound is an analog for the product farnesal. Farnesol was examined for its inhibitory properties under conditions at which the oxygen concentration was fixed at air saturation, and FC concentration was varied. Analysis of the data via a double reciprocal plot of 1/ versus 1/[FC] (Fig. 4B) demonstrated that farnesol is also a noncompetitive inhibitor versus FC. The noncompetitive inhibition pattern indicates that farnesol binds to the enzyme after the substrate FC binds. Thus, as with farnesal, farnesol does not bind to the unliganded enzyme in the oxidized form.
Pre-steady-state Kinetics-The discovery of a steady-state isotope effect in the Pcly reaction suggested that hydride transfer is one of the slow steps in the Pcly reaction. To further elucidate the individual steps in its reaction cycle, we undertook a pre-steady-state analysis of the enzyme. Pcly was preincubated with the components needed for hydrogen peroxide detection, and then increasing concentrations of FC were added. As can be seen in Fig. 5A, a rapid burst of hydrogen peroxide production upon the addition of FC was detected. The amplitude of the observed bursts (i.e. the amount of hydrogen   1.2 Ϯ 0.3), as would be expected if Pcly exhibited typical burst behavior. The rate constants for the exponential phase, k obs , were obtained by fitting the data to a single exponential equation that contained a steady-state term (Equation 3); this analysis revealed a hyperbolic dependence of k obs on FC concentration (Fig. 5B). The maximum value of k obs, denoted k burst, was calculated according to Equation 4. The value of k burst is about 70-fold greater than the steady state turnover rate (0.62 s Ϫ1 versus 0.008 s Ϫ1 ) (Table III). This observation implies the hydride transfer step, which is a necessary component of k burst , is not, in fact, even partially ratelimiting in the reaction. Further support for this conclusion comes from analysis of a similar pre-steady-state experiment using (2R,1ЈS)-[1Ј-2 H 1 ]FC as a substrate. This analysis, the results of which are also shown in Table III, indicated that there is no primary deuterium isotope effect on k burst. The implications of this finding are discussed below.
The steady-state rate of hydrogen peroxide formation by Pcly correlates well with the rate of cysteine formation (14), indicating that the enzyme produces a stoichiometric amount of hydrogen peroxide during the reaction. To determine whether there was a burst of cysteine formation similar to that of hydrogen peroxide formation, we performed a pre-steady-state analysis of the formation of this product also (See "Experimental Procedures"). Since the steady-state rate of Pcly is so slow, we were able to perform the analysis by hand quenching at the times indicated (see legend to Fig. 6). However, only a linear increase in the rate of cysteine formation was observed, with no apparent rapid burst (Fig. 6). The rate of linear increase (0.0035 s Ϫ1 ) is within a factor of 2 of the steady state k cat determined by measuring hydrogen peroxide formation (0.008 s Ϫ1 ; Table II). This observation indicates that cysteine formation is one of the primary rate-determining steps in the Pcly reaction.

DISCUSSION
Whereas the biochemical mechanisms and biological consequences of protein prenylation have been established, little is known about the fate of prenylated proteins. Understanding the metabolism of this important class of proteins, particularly the fate of the prenylcysteine moiety they contain, is of considerable importance, since free prenylcysteines can disrupt a variety of cellular signaling processes (8 -11). Cellular degradation of prenylcysteines is thought to be primarily catalyzed by Pcly, a lysosomal thioether oxidase. In the present study, we conducted stereochemical and kinetic analyses of this enzyme to determine how it achieves this unusual type of catalysis.
Cleavage of the C␣-H bond at carbon-1 of the isoprenoid moiety during the Pcly reaction is supported by the deuterium isotope effect studies using the substrate (2R,1ЈS)-[1Ј-2 H 1 ]FC (Fig. 1). The studies demonstrating that the Pcly reaction is stereospecific, in which the pro-S hydride is transferred to FAD, also support this conclusion. Stereospecific removal of the pro-S hydrogen at C-1 of the farnesyl chain indicates tight binding of the FC substrate in the Pcly active site, in accord with the submicromolar apparent K m values measured previously (0.69 M for FC and 0.84 M for geranylgeranylcysteine) (14). The small primary kinetic isotope effect (k H /k D ϭ 2) determined for degradation of FC bearing deuterium in the pro-S position under steady-state conditions as measured by H 2 O 2 production appears to indicate at least one rate-determining hydrogen (deuterium) transfer in the overall mechanism. However, the absence of a detectable isotope effect on the presteady-state rate is inconsistent with the expectation of an initial rate-limiting cleavage of the isoprenoid C-H bond adjacent to sulfur. Noncompetitive inhibition kinetics with the product farnesal and with farnesol under steady-state conditions suggest that these very similar compounds do not bind to the free enzyme in its oxidized form. These new experimental findings provide important insights into the catalytic process.
The mechanisms of flavoprotein-catalyzed redox reactions vary widely and may involve free radicals, radical ion pairs, carbanions, carbocations, and covalent adducts with the heterocyclic cofactor (48,49). In the case of Pcly, no evidence is presently available to distinguish a direct hydride transfer from FC to FAD in the oxidative half-reaction from more complex scenarios involving single-electron transfers and short lived radical pair intermediates such as those proposed for monoamine oxidase reactions (49,50). However, the stereospecificity of hydrogen removal from C-1 of the farnesyl chain seems inconsistent with a conformationally unrestricted FC free radical intermediate. It seems more likely that the hydrophobic polyene chain is fixed in the active site with the pro-S hydrogen positioned near N-5 of the oxidized FAD cofactor (or an active site mediator) by analogy with UDP-N-acetylenolpyruvylglucosamine reductase (51). In addition, alignment of the C-H bond with the -orbitals of the adjacent 2,3 double bond would stabilize the transition state for an initiating hydride (or hydrogen atom) transfer on the way to a planar conjugated thiocarbenium ion (or ␣-thioallylic radical). In view of the precedent for hydride transfer mechanisms in stereospecific flavin monooxidase reactions (51,52) and in the absence of contravening evidence, we opt for a minimal kinetic scheme portraying a hydride transfer oxidation (Fig. 1).
The experimental results outlined above can be accommodated by the six-step reaction pathway shown in Fig. 7. However, it seems inevitable that this minimal kinetic mechanism will be modified in the future to take into account new experimental findings. In the mechanism proposed, FC binding to the enzyme (step 1) is followed by hydride transfer from C-1 of the isoprenoid moiety of FC to FAD, forming reduced flavin and FC containing a sulfur-stabilized carbocation (see first intermediate in Fig. 1). This thiocarbenium ion reacts rapidly with a water molecule to form the hemithioacetal intermediate (designated FCЈ in Fig. 7). Thus, in this case, the water molecule participates as a nucleophile as the hydride is being abstracted in step 2 of the kinetic mechanism or immediately afterward. The second substrate (molecular oxygen) now reacts with enzyme⅐FCЈ, forming a transient ternary complex (not shown) that serves to reoxidize the flavin and generate hydrogen per- oxide. The hemithioacetal then collapses to enzyme-bound farnesal and cysteine, followed by an ordered release of products in which farnesal is liberated before cysteine. The noncompetitive nature of both inhibitors in this study (farnesal and farnesol) demonstrates that neither compound inhibits Pcly by binding to the enzyme in its unliganded oxidized form (E⅐Fl oxid in Fig. 7). These hydrophobic compounds must therefore bind to the enzyme after FC, perhaps to an enzyme-bound intermediate. We propose that farnesal and farnesol exert their inhibitory effects by binding to the oxidized enzyme-cysteine binary complex (E⅐Fl oxid ⅐Cys in Fig. 7) after the product farnesal is released.
The pre-steady-state data can also be explained by the proposed mechanism. The rapid burst of a stoichiometric amount of hydrogen peroxide, but not of cysteine, dictates that hydrogen peroxide must be formed and released prior to cysteine in the Pcly reaction, indicating that a ternary complex containing reduced enzyme, the hemithioacetal (FCЈ), and oxygen must be formed in step 3. The finding that the rate of cysteine formation is within 2-fold of the steady state turnover rate requires that cysteine production and/or one of the subsequent steps in the second half-reaction shown in Fig. 7 are rate-limiting in the reaction. Thus, in our sequential mechanism, the rapid "burst" of hydrogen peroxide formation is followed by slower steps. Whereas the majority of our kinetic data is compatible with the mechanism outlined in Fig. 7, one possible inconsistency needs to be addressed. We observe a steady state isotope effect of 2 when (2R,1ЈS)-[1Ј-2 H 1 ]FC is the substrate. However, in Fig. 7, the hydride transfer step, which should be the step most sensitive to isotopic substitution, directly precedes hydrogen peroxide formation, which is the step that is monitored in the burst experiment. Thus, it seems logical that a deuterium isotope effect should appear on the burst of hydrogen peroxide production, but no such effect was observed. A V max isotope effect on the overall reaction, however, is a combination of conformational changes, bond-breaking and bond-making steps, and product dissociation (53). Therefore, if the oxidative or hydride transfer step (k 2 ) is not rate-limiting, the observed isotope effect will be less than the maximum, or intrinsic isotope effect. Whereas the upper limit for an isotope effect on the rate of C-H bond cleavage is 15, most enzymes have values between 6 and 10 (54). Instrinsic isotope effects have been measured for D-amino acid oxidase and have values ranging from 4 to 7 (55). Therefore, by analogy, the isotope effect value of 2 determined for the Pcly reaction implies that another step (such as decomposition of the hemithioacetal intermediate, release of farnesal, and/or release of cysteine) must be the primary rate-limiting step.
The data presented here clearly show that the deuterium isotope effect occurs at a step after hydrogen peroxide production (step 3) and hence after hydride transfer. A plausible chemical mechanism can be formulated to explain this apparent anomaly. After abstraction by the flavin, the deuterium atom is transferred to an active site base that is very well sequestered in the active site such that no exchange with solvent can occur. This same base then functions as the general acid that catalyzes the breakdown of the hemithioacetal shown in Fig. 1. Hence, transfer of the deuterium results in an isotope effect that is manifest not on the first half-reaction that produces hydrogen peroxide but rather on the step that produces the cysteine and farnesal products. Whereas this explanation is logical, formal proof would require a demonstration that the deuterium atom is incorporated into the cysteine product. This will be quite difficult to demonstrate experimentally, since the cysteine thiol deuterium would exchange rapidly with solvent protons upon release of the product. However, precedent exists for kinetically sequestered hydrogens on other reduced flavin intermediates generated by hydride transfer (see Ref. 51 and references therein).
Whereas the model outlined in Fig. 7 is not the only type of mechanism that could be drawn for Pcly, we believe it is the simplest one consistent with all the available data. Although Pcly appears to be a unique enzyme in the reaction that it catalyzes (i.e. oxidative cleavage of a thioether bond with no net oxidation at sulfur), its mechanism shares some properties with other oxidases. Whereas it is more common for an oxidase to display a ping-pong kinetic mechanism (44,47,56,57), a sequential mechanism, like the one proposed here for Pcly, is not unprecedented. For example, L-amino acid oxidase (58), urate oxidase (59), and D-amino acid oxidase (46) exhibit sequential kinetic mechanisms.
In summary, evidence is provided in support of a sequential kinetic mechanism for Pcly. More detailed pre-steady-state studies are likely to yield significant findings on how this novel enzyme achieves catalysis. Although the proposed hemithioacetal intermediate seems likely, in the absence of more direct evidence, its structure is hypothetical, and other possibilities (e.g. with farnesylcysteine covalently attached to the enzyme or the flavin) cannot be excluded. In addition, substrate specificity studies will further our understanding of the Pcly active site. Equally important are ongoing efforts to understand the cellular role of Pcly in the metabolism of prenylcysteines (i.e. the determination of whether Pcly has an essential physiological function). , the maximum value, k burst , can be determined.
b The hydrogen peroxide concentration is obtained from a standard resorufin curve, in which the amplitude of the burst in fluorescence units can be converted to a concentration of hydrogen peroxide produced.