Abstract
Methylphenidate is an important stimulant prescribed to treat attention-deficit hyperactivity disorder. It has two chiral centers, but most current commercial formulations consist of the racemic mixture of the threo pair of methylphenidate isomers (d-, l-threo-methylphenidate). The d-isomer is the pharmacologically active component. Numerous studies reported that oral administration of the methylphenidate racemate undergoes first-pass, stereoselective clearance in humans with l-methylphenidate being eliminated faster than d-methylphenidate. Accordingly, the kinetics of hydrolysis of individual enantiomers by purified native and recombinant human liver carboxylesterases CES1A1 and CES2 and a colon isoenzyme CES3 were examined with a liquid chromatography/mass spectrometry assay. The expression of CES1A1, CES2, and CES3 in Sf9 cells and the methods for purification of the three isoenzymes are reported. CES1A1 has a high catalytic efficiency for both d- and l-enantiomers of methylphenidate. No catalytic activity was detected with CES2 and CES3 for either enantiomer. The catalytic efficiency of CES1A1 for l-methylphenidate (kcat/Km = 7.7 mM–1 min–1) is greater than that of d-methylphenidate (kcat/Km = 1.3–2.1 mM–1 min–1). Hence, the catalytic efficiency of CES1A1 for methylphenidate enantiomers agrees with stereoselective clearance of methylphenidate reported in human subjects. Both enantiomers of methylphenidate can be fit into the three-dimensional model of CES1A1 to form productive complexes in the active site. We conclude that CES1A1 is the major enzyme responsible for the first-pass, stereoselective metabolism of methylphenidate.
Methylphenidate is one of the most commonly prescribed psychostimulants for children in the United States with attention-deficit hyperactivity disorder (Findling and Dogin, 1998; Wender, 1998). It is also used in some other neurological conditions, including depression in medically ill elderly, traumatic brain injury, and narcolepsy (Challman and Lipsky, 2000). The pharmacokinetic properties and molecular mechanism of action of methylphenidate have been extensively studied. Methylphenidate has two chiral centers (Fig. 1); therefore, there are four isomers, a d,l-erythro-methylphenidate pair and a d,l-threo-methylphenidate pair. Only the racemic mixture of the threo pair of enantiomers (d,l-threo-methylphenidate) is used therapeutically because it has less adverse effects than the erythro pair (Markowitz et al., 2003). The mechanism of action of methylphenidate is thought to result from the increased level of extracellular dopamine in brain by blocking its reuptake through the presynaptic dopamine transporter (Gatley et al., 1996; Volkow et al., 2001). In a double-blind, four-way randomized crossover study, l-methylphenidate showed no difference versus placebo in improving the sustained attention in humans. Moreover, a 5-mg dose of d-threo-methylphenidate has the same attention improvement effect as a 10-mg dose of d,l-threo-methylphenidate racemic mixture. Hence, the clinical effect of methylphenidate is attributed to the d-enantiomer (Srinivas et al., 1992b). A new formulation of methylphenidate, Focalin (dexmethylphenidate), was introduced, which only contains enantiopure d-threo-methylphenidate (Markowitz et al., 2003).
The predominant metabolic pathway of methylphenidate is de-esterification to form pharmacologically inactive d-or l-ritalinic acid (Patrick et al., 1987). Numerous studies reported that d,l-methylphenidate was stereoselectively metabolized in humans (Lim et al., 1986; Srinivas et al., 1987), rats, and dogs (Egger et al., 1981). In both intravenous and oral administration studies, the racemic d,l-methylphenidate undergoes stereoselective clearance, with plasma concentrations of d-methylphenidate being higher than those of l-methylphenidate (Srinivas et al., 1993). In oral formulations including immediate release, sustained release, and osmotic controlled-released oral delivery system, the area under the curveinf values of l-methylphenidate were 10- to 40-fold lower than those of d-methylphenidate (Srinivas et al., 1993; Modi et al., 2000). After oral administration, the plasma concentrations of d-methylphenidate are immediately higher than those of l-methylphenidate at 0.5 h after oral dose. For intravenous administration, it was only 1.5 h after dosing that the significant differences between plasma concentrations of the two enantiomers were observed. The higher concentration of l-ritalinic acid in both human plasma and urine is present in the first 2 h after oral but not intravenous methylphenidate administration (Srinivas et al., 1990, 1992a). Together, these results suggest that with oral administration, d- and l-methylphenidate undergo stereoselective clearance to ritalinic acid before entering systemic circulation with l-methylphenidate being eliminated faster than d-methylphenidate.
The specific hydrolases responsible for methylphenidate metabolism in humans have not been identified. The stereoselective clearance of methylphenidate indicates that methylphenidate hydrolysis is primarily enzyme-catalyzed. Zhang et al. (2003) reported that a porcine liver carboxylesterase hydrolyzes l-methylphenidate with higher efficiency than d-methylphenidate. Therefore, we propose that human liver carboxylesterase isoenzymes stereoselectively catalyze the hydrolysis of methylphenidate enantiomers.
Human liver carboxylesterases (E.C.3.1.1.1) play an important role in metabolism and detoxification of ester drugs, such as cocaine, heroin, and CPT-11 (Satoh et al., 2002). Carboxylesterases belong to a multigene family, based on the phylogenetic analysis of the sequences of the isoenzymes. The human isoenzymes are classified into three main families: CES1, CES2, and CES3 (Satoh and Hosokawa, 1998). Two major isoenzymes were identified in human liver, CES1A1 (Shibata et al., 1993; Brzezinski et al., 1994) and CES2 (Pindel et al., 1997). A third carboxylesterase, CES3, was initially reported in the NEDO human cDNA sequencing project (gi:7019977), and it is expressed in human liver and colon (Sanghani et al., 2004). There is very limited information on the substrate specificity of CES3. Whereas CES1A1 and CES2 share broad and overlapping substrate specificity, they do exhibit very different catalytic efficiencies for selected substrates. For example in cocaine hydrolysis, CES1A1 hydrolyzes only the methyl ester and CES2 hydrolyzes only the benzoyl ester (Brzezinski et al., 1994; Pindel et al., 1997). CES1A1 may also hydrolyze methylphenidate methyl ester, because both methylphenidate and cocaine have a methyl ester linkage.
In this study, an improved procedure is described for expression and purification of CES1A1, CES2, and CES3 from Sf9 insect cells. CES1A1 and CES2 are also purified from human liver (Humerickhouse et al., 2000; Sanghani et al., 2003). The kinetics of d- and l-methylphenidate hydrolysis by the purified native and recombinant CES1A1, CES2, and CES3 isoenzymes is compared.
Materials and Methods
Human liver cDNA was purchased from BD Biosciences Clontech (Palo Alto, CA). The PCR kit was purchased from Roche (Indianapolis, IN), and primers were ordered from Invitrogen (Carlsbad, CA). The baculovirus DNA and transfer vector were from BD PharMingen (San Diego, CA). Sf9 insect cells and cell culture media were from Invitrogen (Carlsbad, CA). d- and l-Methylphenidate and ritalinic acid were purchased from Sigma-Aldrich (St. Louis, MO). The internal standard ritalinic acid-d5 was purchased from Isotech Inc. (Miamisburg, OH). Unless otherwise indicated, all general chemicals were purchased from Sigma-Aldrich or Fisher Scientific Co. (Pittsburgh, PA).
Cloning of CES1A1. The CES1A1 cDNA was amplified from human liver Marathon-Ready cDNA by PCR. The primer sequence 5′ACGATGTGGCTCCGTGCCTTT3′ (sense) and 5′TCACAGCTCTATGTGTTCTGTCTG3′ (antisense) was based on the published cDNA sequence of human liver carboxylesterase 1 (Munger et al., 1991). The PCR reaction proceeded through 35 cycles, each cycle consisting of 1 min at 95°C for denaturation, 30 s at 62°C for annealing, and 3 min at 72°C for extension. The PCR product was incorporated into pCR-Blunt II –TOPO vector and transformed into TOP 10 Escherichia coli cells (Invitrogen). After plasmid isolation by using Quantum Prep Plasmid Miniprep kit (Bio-Rad, Hercules, CA), the plasmid DNA was sequenced with vector sequencing primers M13 forward primer, 5′GTTTTCCCAGTCACGAC; M13 reverse primer, 5′CAGGAAACAGCTATGAC); and internal primers (internal 1, 5′GTGGTGACCATTCAATATCGCC; internal 2, 5′GCTGACTCTCCAAAGATGG; internal 3, 5′TGCAGTTGATGAGCTATCCA; internal 4, 5′GACAGTGTCGTCTGTTCCTCC. The cDNAs for CES2 and CES3 (Sanghani et al., 2004) were similarly cloned.
Expression and Purification of Recombinant CES1A1, CES2, and CES3. Methods for cloning and expression of CES1A1 are similar to those described previously (Kroetz et al., 1993; Morton and Potter, 2000). The CES1A1 cDNA was excised from pCR-Blunt II –TOPO vector by BamHI and NotI digestion and incorporated into baculovirus expression vector pAcMP3 (BD PharMingen, San Diego, CA), which was digested by the same enzymes. The CES1A1 construct (1.3 μg) and 0.5 μg of linearized baculovirus DNA were cotransfected into 2 × 106Sf9 insect cells by calcium phosphate precipitation at 27°C for 4 h. The medium was collected, and cell debris was removed by centrifugation at 500g for 5 min at 4°C. The supernatant contained CES1A1 recombinant virus, which was isolated and titered by plaque assay as described previously (Gonzalez et al., 1991). The recombinant baculovirus was similarly prepared for CES2 and CES3 (Sanghani et al., 2004).
Two liters of Sf9 cells (1.8–2 × 106 cells/ml) were transfected with CES1A1, CES2, or CES3 recombinant baculovirus with a multiplicity of infection of about 1. After incubation for 63 to 66 h at 27°C, the cells were harvested by low-speed centrifugation (500g) and rinsed with phosphate-buffered saline. The cell suspension was sonicated and cell debris was removed by centrifugation at 100,000g for 1 h. The supernatant was applied to a concanavalin A affinity chromatography column (Sigma-Aldrich) equilibrated with 10 column volumes of 20 mM Tris, 0.2 M NaCl, 1 mM CaCl2, 1 mM MgCl2, and 1 mM MnCl2, pH 7.4. Bound protein was eluted with a methyl-α-d-mannopyranoside gradient from 0 to 0.2 M in 200 ml of buffer, and fractions were assayed for esterase activity with 4-methylumbelliferyl acetate. The active fractions were pooled and concentrated and loaded onto a preparative polyacrylamide gel electrophoresis unit (model 491; Bio-Rad). The stacking gel contained 4% polyacrylamide, and the resolving gel contained 6% polyacrylamide. The proteins were electrophoresed at a constant current of 30 mA. Fractions with carboxylesterase activity were pooled, concentrated, and ethylene glycol was added to a final concentration of 10%. The purified enzyme was stored at 4°C. The recombinant CES2 and CES3 (Sanghani et al., 2004) were similarly expressed and purified.
Purification of Native Human Liver Carboxylesterase CES1A1 and CES2. All buffers used in purification were purged with helium and contained 1 mM benzamidine and 1 mM dithiothreitol. Approximately 80 g of frozen human liver obtained at autopsy was homogenized in 180 ml of 50 mM HEPES, pH 6.8. The DEAE ion exchange chromatography and concanavalin A affinity chromatography procedures were described previously (Humerickhouse et al., 2000). The recovered activity pool was applied to a preparative polyacrylamide gel electrophoresis unit and eluted as described above for recombinant enzymes. Two peaks of activity were recovered: the first peak contained CES2 and the second peak contained CES1A1. The purified enzymes were concentrated and 10% ethylene glycol was added. The enzymes were stored at 4°C.
Carboxylesterase Activity and Protein Assays. Carboxylesterase activity was determined by incubating 5 μl of enzyme (purified or homogenate supernatant) with 0.5 mM 4-methylumbelliferyl acetate in 90 mM KH2PO4, 40 mM KCl, pH 7.4, at 37°C in the total volume of 1.0 ml. Formation of the hydrolysis product 4-methylumbelliferone was monitored with a spectrophotometer at 350 nm. Rates of hydrolysis (micromoles per minute) were calculated by linear regression of absorbance versus time using an extinction coefficient of 12.2 cm–1 mM–1 for 4-methylumbelliferone. The specific activity was expressed as micromoles per minute per milligram of protein. Protein concentrations were determined by Bio-Rad protein assay with bovine serum album as standard.
SDS-Polyacrylamide Gel Electrophoresis and Western Blot Analysis. The stacking gel contained 4% acrylamide with 0.1% SDS, and the resolving gel contained 10% acrylamide with 0.1% SDS. Protein samples were boiled with β-mercaptoethanol and electrophoresed at 10-mA constant current at room temperature. Gels were stained with Coomassie Blue. For Western blot analysis, the gel was electroblotted overnight onto a nitrocellulose membrane using a Bio-Rad mini transfer-blot cell. The membrane was then blocked for 1 h with 1% bovine serum albumin in Tris-buffered saline to reduce the background, rinsed briefly with TTBS, and incubated for 1 h with 1:1000 dilution of rabbit anti-CES1A1 antibody in TTBS. After washing three times with TTBS, the membrane was incubated for 1 h with 125I-labeled protein A (9.7 mCi/mg) diluted 1:2500 in TTBS. The membrane was washed three times with TTBS. Immunoreactive protein was visualized by autoradiography.
Deglycosylation of CES1A1, CES2, and CES3 by Endoglycosidase H. Ten micrograms of each CES1A1, CES2, and CES3 was mixed with 15 mU endoglycosidase H in 50 μl of 50 mM citrate buffer with 1 mM EDTA, 1 mM benzamidine, and 1 mM dithiothreitol, pH 6.5. The reaction mixture was incubated at 37°C for 70 h. Untreated isoenzymes were similarly incubated without endoglycosidase H. The activities of both treated and untreated isoenzymes were measured with 4-methylumbelliferyl acetate as the substrate. Both treated and untreated isoenzymes were analyzed by SDS-PAGE to determine their molecular weight change.
Assay of Carboxylesterase-Mediated Methylphenidate Hydrolysis to Ritalinic Acid by LC/MS. The CES1A1-catalyzed hydrolysis reaction of l-methylphenidate to ritalinic acid was performed by incubating 20 μg/ml (100 μg/ml for d-methylphenidate) purified native or recombinant enzyme with varying amount of the methylphenidate enantiomers (4, 8, 16, 32, 64, 128, and 256 μM) in 20 mM HEPES, 10% ethylene glycol (v/v) at pH 7.0 for 20 min at 25°C. The total reaction volume was 0.5 ml. The incubation time was 20 min for CES1A1 and 24 h for CES2 and CES3 at 25°C. A control reaction was prepared without added carboxylesterase. The reaction was stopped by adding 10% trichloroacetic acid and ritalinic acid-d5 (4 μM) was added as the internal standard. Precipitated protein was removed by centrifugation, and 0.55 ml of supernatant was recovered for ritalinic acid assay. For the hydrolysis of methylphenidate by CES2 and CES3, 100 μg/ml purified native and recombinant CES2 or CES3 was added to 128 μM d- or l-methylphenidate.
A Strata X33 μm polymeric sorbent solid phase extraction cartridge (Phenomenex, Torrance, CA) was equilibrated sequentially with 1 ml of methanol and 1 ml of 2% aqueous phosphoric acid. After loading 0.5 ml of the reaction supernatant onto the solid phase extraction cartridge, it was washed with 1 ml of 5% methanol and 2% acetic acid (v/v) in water. The sample was eluted with 0.4 ml of 50% methanol, 2% acetic acid. The recovery was about 90% for all analytes. Ritalinic acid and methylphenidate were separated by liquid chromatography with a Luna C18 column (150 × 4.6 mm, 3 μm; Phenomenex) (Doerge et al., 2000). Thirty microliters of the extract was loaded onto the column. The gradient elution was performed by increasing methanol concentration from 20 to 80% in 0.1% aqueous acetic acid over 6.5 min. The flow rate was 0.8 ml/min, and the entire effluent was introduced into the MS atmospheric pressure chemical ionization probe for ionization. Positive ions were acquired in selected ion monitoring mode. The MS conditions (QP8000A single quadrapole; Shimadzu, Kyoto, Japan) were as follows: probe temperature, 400°C; curved desolvation line temperature, 250°C; and nebulizer gas flow, 2.5 l/min. The following [M+H]+ ions were measured: d-methylphenidate (m/z 234), l-methylphenidate (m/z 234), ritalinic acid (m/z 220), ritalinic acid-d5 (m/z 225). The standard curve for ritalinic acid was constructed by adding varying concentrations of ritalinic acid from 0.0625 to 64 μM and a fixed amount of ritalinic acid-d5 internal standard (4 μM) to the reaction buffer without enzyme. The area ratios of different amounts of the ritalinic acid to the corresponding internal standard ritalinic acid-d5 were plotted versus the ritalinic acid concentration. The standard curve was linear from 0.0625 to 64 μM with an R2 value greater than 0.99. The limit of detection was 0.01 μM and the limit of quantification was 0.025 μM. The standard curve was used to quantify the methylphenidate hydrolysis product ritalinic acid formed in enzyme assays.
Kinetic Analysis of Methylphenidate Hydrolysis by CES1A1, CES2, and CES3. For steady-state kinetic analysis with purified carboxylesterase isoenzymes, the ritalinic acid (m/z 220) area and the corresponding internal standard ritalinic acid-d5 (m/z 225) area from each methylphenidate hydrolysis assay were obtained from the LC/MS chromatogram. The ratio of ritalinic acid to ritalinic acid-d5 was used to calculate the amount of ritalinic acid produced in each assay based on the standard curve. The amount of enzymatic ritalinic acid production was obtained by subtracting the control nonenzymatic production from the total ritalinic acid production. This difference was used to calculate the velocity of the hydrolysis reaction based on the enzyme added (20 μg/ml CES1A1 for l-methylphenidate hydrolysis, and 100 μg/ml CES1A1 for d-methylphenidate hydrolysis) and incubation time (20 min). Each substrate concentration was examined in duplicate assays. All data were fit to the Michaelis-Menten equation, and kinetic parameters were calculated by nonlinear regression analysis with GraFit software (Erithacus Software Horley, Surrey, UK).
Results
Human liver carboxylesterase CES1A1 cDNA (Shibata et al., 1993) was cloned from human liver cDNA by PCR, inserted into a baculovirus transfer vector, and cotransfected with baculovirus DNA into insect cells. Sf9 cells infected with CES1A1 recombinant virus produced high carboxylesterase activity, 78 μmol/min total activity per liter of cell culture (Table 1), with 4-methylumbelliferyl acetate as substrate, compared with the uninfected Sf9 cells, 26 μmol/min total activity per liter of cell culture. Western blot analysis confirmed the expression of recombinant CES1A1 in Sf9 cells (data not shown). The enzyme was purified to >95% homogeneity by concanavalin A affinity chromatography and preparative nondenaturing polyacrylamide gel electrophoresis with a yield of 2.5 mg of protein per liter of Sf9 cells at a concentration of 2 × 106 cells/ml (Table 1). The recombinant CES1A1 had the same subunit mass as native CES1A1 (60 kDa) on SDS-PAGE (Fig. 2A). The specific activity of purified recombinant CES1A1 (7 μmol/min/mg) is consistent with the native CES1A1 purified from human liver (Brzezinski et al., 1994).
CES2 was similarly cloned and expressed in Sf9 cells. Infected cells generated 680 μmol/min total activity per liter of cells above the uninfected cells activity (Table 1). The enzyme was purified by concanavalin A affinity chromatography and preparative nondenaturing polyacrylamide gel electrophoresis. The yield was 1.9 mg protein per liter of Sf9 cells at a concentration of 2 × 106 cells/ml (Table 1). Recombinant CES2 had the same subunit mass as native CES2 (60 kDa) in SDS-PAGE (Fig. 2B). The recombinant CES2 had about the same specific activity with 4-methylumbelliferyl acetate (120 μmol/min/mg) as native CES2 purified from human liver (Pindel et al., 1997). CES3 was expressed and purified to homogeneity by method similar to CES1A1 and CES2 (Sanghani et al., 2004). The purified recombinant CES3 had a specific activity of 1.5 μmol/min/mg with 4-methylumbelliferyl acetate as substrate. CES3 also had a subunit molecular weight of about 60 kDa (Fig. 2A). The amino acid sequence identities of all expressed carboxylesterases were confirmed by matrix-assisted laser desorption ionization-MS analysis (data not shown).
CES1A1, CES2, and CES3 were deglycosylated by treatment with endoglycosidase H. The deglycosylated enzymes had greater mobility than the native enzymes in SDS-PAGE (Fig. 2). Deglycosylation resulted in a removal of approximately 3 kDa of carbohydrate residues. The deglycosylation of CES1A1 and CES3 had no significant effect on catalytic activity (Fig. 2, see legend). The CES2 catalytic activity was decreased by only 19% after deglycosylation (Fig. 2, see legend).
The hydrolysis of d- and l-methylphenidate by human carboxylesterase isoenzymes was examined by an LC/MS assay. The substrate methylphenidate (m/z = 234; MP) was well separated from the product ritalinic acid (m/z = 220; RA) in the LC/MS chromatogram (Fig. 3). The area of ritalinic acid peak was used to generate a standard curve to quantify product formation in the enzyme assay. Methylphenidate was spontaneously hydrolyzed in the assay buffer. The spontaneous RA formation was linear with reaction time for 24 h and methylphenidate concentration (4–256 μM). The first order rate constant for nonenzymatic hydrolysis was the same for d- and l-methylphenidate, 5 × 10–4 min–1. In the enzymatic hydrolysis assay, the spontaneous RA production was subtracted from the total RA production (Table 2). The CES1A1 enzymatic RA production was linear with reaction time for 1 h (most reactions were for 20 min) and enzyme concentration (10–100 μg/ml). The Km and Vmax values of CES1A1 for methylphenidate were determined by measuring the rate of enzymatic formation of ritalinic acid at different methylphenidate concentrations and fitting the data to the Michaelis-Menten equation (Fig. 4 and Table 3). In all cases, the reaction conditions were adjusted so that less than 10% of the substrate was consumed during the time of measurement. The kcat values were obtained by multiplying Vmax by the enzyme subunit molecular mass of 60 kDa.
As shown in Table 2, only CES1A1 had hydrolytic activity with methylphenidate. Recombinant CES1A1 had about 6-fold higher catalytic efficiency with l-methylphenidate than d-methylphenidate (Table 3). The activity of native versus recombinant CES1A1 was similar. Native and recombinant CES2, and recombinant CES3, demonstrated very little or no detectable catalytic activity for d- and l-methylphenidate (Table 2).
Discussion
The expression of recombinant CES1A1, CES2, and CES3 in Sf9 insect cells creates a reproducible source of enzyme to study their structure/function relationships among this family of related carboxylesterase isoenzymes. Although enzymes can be purified from human liver (Humerickhouse et al., 2000), recent studies indicate the presence of genetic variants (Munger et al., 1991; Shibata et al., 1993; Wu et al., 2003) that could influence enzyme expression or catalytic activity. The first successful expression of human carboxylesterase CES1A1 in the Sf9 insect cell system was reported by Kroetz et al. (1993). They found that inhibition of N-glycosylation of the CES1A1 in insect cells prevents formation of active enzyme. Whereas enzyme can be expressed in E. coli, it was not glycosylated and not fully active (J. Zhang and W. F. Bosron, unpublished data). In this study, we expressed recombinant CES1A1, CES2, and CES3 in the Sf9 insect cell system. The expressed CES1A1 and CES2 isoenzymes had the same specific activity as the native enzymes purified from human liver (Table 1). Deglycosylation of the three enzymes by endoglycosydase H (Fig. 2) had minimal effect on catalytic activity. In the reported crystal structure of CES1A1 (Bencharit et al., 2003), the sugar moieties are located at the trimer interface of CES1A1 and presumably stabilize subunit association. Hence, glycosylation could assist the correct folding and subunit association to yield the native, active CES1A1. However, after the proper protein folding is achieved, the removal of the carbohydrate moiety may not change the overall protein structure dramatically. Whether glycosylation has any major effect on CES2 or CES3 folding (both monomers) is not known.
The average yield of the three expressed enzymes in this study is low, only about 2 mg from 1 liter of insect cells. All three carboxylesterases contain the C-terminal four amino acid endoplasmic reticulum retention signal (Satoh and Hosokawa, 1998). Morton and Potter (2000) studied expression of a rabbit liver carboxylesterase in insect cells. Expression of the C-terminal truncated rabbit liver carboxylesterase resulted in secretion of the enzyme into the serum-free cell culture media (Morton and Potter, 2000). The major advantage of the expression of truncated carboxylesterase as a secreted protein was that it increased yield and simplified purification from cell media (Morton and Potter, 2000).
CES1A1 and CES2 have been shown to exhibit different substrate specificity (Satoh et al., 2002). For example, CES1A1 catalyzes the hydrolysis of the methyl ester of cocaine, and CES2 catalyzes the hydrolysis of the benzoyl ester of cocaine (Brzezinski et al., 1994; Pindel et al., 1997). CES1A1 hydrolyzes meperidine, whereas CES2 does not (Zhang et al., 1999). CES2 hydrolyzes aspirin and procaine, whereas CES1A1 does not (Takai et al., 1997). Even in the overlapping substrates such as 4-methylumbelliferyl acetate, heroin, and CPT-11, the catalytic efficiencies are significantly different (Pindel et al., 1997; Humerickhouse et al., 2000). Based on the analysis of the substrates structures, we proposed that CES1A1 preferentially cleaves the ester linkage with small alcohol group and large acyl group (e.g., cocaine methyl ester and meperidine), and CES2 prefers the substrate with small acyl group and large alcohol group, e.g., cocaine benzoyl ester, CPT-11, and heroin (Bosron and Hurley, 2002). We predicted that CES1A1 would hydrolyze methylphenidate, because methylphenidate has a similar methyl ester linkage as cocaine. The kinetics results in this study demonstrate that CES1A1 has high catalytic efficiency, kcat/KM = 1.3–7.7 mM–1 min-1 with d-or l-methylphenidate, whereas CES2 and CES3 are not active with methylphenidate as substrate (Table 3).
A recently resolved crystal structure of CES1A1 with the cocaine analog homatropine revealed that the cocaine methyl ester linkage (small alcohol group) is oriented toward the active site serine of CES1A1 and fits well into the selective small, rigid substrate binding pocket (Bencharit et al., 2003). In an attempt to identify potential structural correlations for the higher catalytic efficiency of d-versus l-methylphenidate, (Table 3), we docked both d- and l-methylphenidate into CES1A1 active site by using the experimentally observed position of homatropine as a guide (Fig. 5). Both enantiomers can be modeled into the active site and interact well with the surrounding residues. The modeled distance between the hydroxyl group of Ser221 and the carbonyl carbon of methylphenidate is within the range for nucleophilic attack from Ser221 (2.6 Å). We were not able to identify any obvious structural differences that might account for the 6-fold difference in the stereoselectivity of hydrolysis from this simple docking experiment, although it is possible that the enzyme might exhibit a preference for a particular orientation of the benzyl versus the piperidine substituents of methylphenidate within the substrate pocket. In our docking experiment we have treated these as interchangeable interactions, but it is likely their different chemical characteristics will influence the actual position of the substrate when bound to the enzyme. A crystal structure of CES1A1 in complex with each methylphenidate enantiomer is needed for a more thorough explanation for the observed stereoselectivity for hydrolysis.
Orally administrated d,l-methylphenidate undergoes significant stereoselective and first-pass clearance (Hubbard et al., 1989; Srinivas et al., 1993). One of the possible explanations is that methylphenidate is stereoselectively metabolized in the gastrointestinal tract or liver before entering the systemic circulation as suggested by Zhang et al. (2003). Accordingly, we studied the stereoselectivity of the three major human carboxylesterase isoenzymes, CES1A1, CES2, and CES3, which are known to be expressed in human liver and gastrointestinal tissues. We found that CES1A1 has a high catalytic efficiency for both d- and l-methylphenidate, whereas CES2 and CES3 have no catalytic activity with either enantiomers (Tables 2 and 3). The catalytic efficiency of CES1A1 for l-methylphenidate is 3.6- to 6-fold greater than d-methylphenidate (Table 3). The Km value of the purified CES1A1 for methylphenidate, 40 to 126 μM, is several orders of magnitude higher than the pharmacological methylphenidate concentration in blood, 25 to 64 nM (Challman and Lipsky, 2000). Therefore, CES1A1 is expected to exhibit first-order kinetics in the hydrolysis of methylphenidate in vivo. We also found that d- and l-methylphenidate have the same first order rate constant (5 × 10–4 min–1) in their spontaneous hydrolysis; thus, their stereospecific clearance is entirely due to their different enzymatic hydrolysis rates.
The absorption of orally dosed methylphenidate is complete and rapid through gastrointestinal tract (Faraj et al., 1974). However, the bioavailability of the intact drug is low and variable, due to the extensive first-pass, stereoselective metabolism (Markowitz et al., 2003). We were unable to find evidence for stereoselective absorption of the drug in gastrointestinal tissues. The three carboxylesterase isoenzymes examined in this study are expressed both in liver and gastrointestinal tissues. For CES1A1, the relative abundance is liver ≫ stomach > colon (Satoh et al., 2002). CES2 is relatively more abundant in small intestine and colon than CES1A1 (Satoh et al., 2002), and CES3 is mainly expressed in liver and colon (Sanghani et al., 2004). However, CES2 and CES3 do not have hydrolytic activity with methylphenidate. Hence, the expression of CES1A1 in liver and stomach is consistent with presystemic stereoselective metabolism of methylphenidate.
Methylphenidate is usually prepared as a chiral drug composed of d,l-threo-methylphenidate racemic mixture. d-Methylphenidate has been shown to be more potent than its l-isomer (Maxwell et al., 1970; Patrick et al., 1987). Some in vivo studies supported that the higher clinical activity of d-methylphenidate is due to its higher potency for the dopamine transporter (Ding et al., 1997: Challman and Lipsky, 2000). However, the decreased catalytic efficiency of CES1A1 for d-versus l-methylphenidate may also contribute to the increased potency of d-methylphenidate in vivo. The decreased catalytic efficiency of CES1A1 for d-versus l-methylphenidate is consistent with the higher plasma concentration of d-methylphenidate after oral or intravenous administration of the racemic drug (Hubbard et al., 1989; Srinivas et al., 1993). Several pharmacokinetic and pharmacodynamic studies of methylphenidate demonstrated the significant interindividual variation in dosage and clinical effectiveness in children with attention-deficit hyperactivity disorder (Hungund et al., 1979; Shader et al., 1999). It is possible that these variations are related to the interindividual variability in the hydrolysis of methylphenidate in patients (Srinivas et al., 1992b). Therefore, the pharmacogenetic studies of CES1A1 could be useful in the effort to predict the effectiveness, dosage requirement, and toxicity of methylphenidate and other drugs metabolized by CES1A1.
Acknowledgments
We thank Drs. Hiremagalur N. Jayaram and Paresh C. Sanghani for helpful discussion of the manuscript. We also thank the INGEN Protein Expression Core (Department of Biochemistry and Molecular Biology, Indiana University, Indianapolis, IN) for assistance in enzyme expression and purification.
Footnotes
-
This research was funded, in part, by National Institute on Drug Abuse Grant R01DA006836.
-
DOI: 10.1124/jpet.104.067116.
-
ABBREVIATIONS: CPT-11, 7-ethyl-10-[4-(1-piperidino)-1-piperidino]carbonyloxycamptothecin; CES1A1 (gi: 179927), carboxylesterase 1A1 or hCE-1 or HU1a; CES2 (gi: 7262375), carboxylesterase 2 or hCE-2 or hiCE; CES3 (gi: 7019977), carboxylesterase 3; PCR, polymerase chain reaction; TTBS, Tween Tris-buffered saline; PAGE, polyacrylamide gel electrophoresis; LC/MS, liquid chromatography/mass spectrometry; MS, mass spectrometry; MP, methylphenidate; RA, ritalinic acid.
- Received February 17, 2004.
- Accepted April 8, 2004.
- The American Society for Pharmacology and Experimental Therapeutics