(Received for publication, September 19, 1996, and in revised form, April 9, 1997)
From the Departments of Biochemistry and Molecular
Biology and § Pathology and Laboratory Medicine, Indiana
University School of Medicine, Indianapolis, Indiana 46202-5122
A human liver carboxylesterase (hCE-2) that catalyzes the hydrolysis of the benzoyl group of cocaine and the acetyl groups of 4-methylumbelliferyl acetate, heroin, and 6-monoacetylmorphine was purified from human liver. The purified enzyme exhibited a single band on SDS-polyacrylamide gel electrophoresis with a subunit mass of approximately 60 kDa. The native enzyme was monomeric. The isoelectric point of hCE-2 was approximately 4.9. Treatment with endoglycosidase H caused an increase in electrophoretic mobility indicating that the liver carboxylesterase was a glycoprotein of the high mannose type. The complete cDNA nucleotide sequence was determined. The authenticity of the cDNA was confirmed by a perfect sequence match of 78 amino acids derived from the hCE-2 purified from human liver. The mature 533-amino acid enzyme encoded by this cDNA shared highest sequence identity with the rabbit liver carboxylesterase form 2 (73%) and the hamster liver carboxylesterase AT51p (67%). Carboxylesterases with high sequence identity to hCE-2 have not been reported in mouse and rat liver. hCE-2 exhibited different drug ester substrate specificity from the human liver carboxylesterase called hCE-1, which hydrolyzes the methyl ester of cocaine. hCE-2 had higher catalytic efficiencies for hydrolysis of 4-methylumbelliferyl acetate, heroin, and 6-monoacetylmorphine and greater inhibition by eserine than hCE-1. hCE-2 may play an important role in the degradation of cocaine and heroin in human tissues.
Carboxylesterases hydrolyze ester groups of drugs and toxins, and thus play an important role in their metabolism and detoxication. In humans, cocaine and heroin are metabolized in serum and liver tissue to de-esterified metabolites that are excreted in urine. Two different human liver carboxylesterases (here called hCE-1 and hCE-2)1 were identified and shown to hydrolyze cocaine (1). hCE-1 catalyzes the hydrolysis of the methyl ester of cocaine producing benzoylecgonine and methanol, and the ethyl transesterification of cocaine with ethanol to form cocaethylene and methanol. Benzoyl ester hydrolysis of cocaine is catalyzed by liver hCE-2 (1, 2) and serum cholinesterase (3, 4) and produces ecgonine methyl ester and benzoic acid. Hydrolysis of cocaine by serum cholinesterase and nonenzymatic hydrolysis of the methyl ester of cocaine (1) have complicated the assessment of the relative roles of tissue carboxylesterases involved in cocaine metabolism. Heroin is metabolized by liver hCE-2 and hCE-1 carboxylesterases and serum cholinesterase (5). All three esterases cleave the 3-acetyl group of heroin producing 6-monoacetylmorphine. However, only hCE-2 hydrolyzes 6-monoacetylmorphine to morphine with high catalytic efficiency (5). Hence, the contribution of hCE-2 to the metabolism of cocaine or heroin is only recently recognized and its relative role in drug ester clearance has not been fully elucidated.
Human liver hCE-2 and hCE-1 enzymes belong to the family of
carboxylesterases with a subunit size of about 60 kDa that exhibit broad substrate specificity for ester hydrolysis. These enzymes have
been purified from the microsomal or cytosolic subcellular fractions
(6) of a variety of tissues from various mammalian species. In the rat,
Parkinson and co-workers (7) described a group of four different
esterases: hydrolase A is expressed in liver, testis, lung, prostate,
and pancreas; hydrolase B is expressed in liver, kidney, small
intestine, lung, spleen, heart, and brain (7, 8); hydrolase C is
expressed in rat liver and kidney and is very similar to hydrolase B in
sequence (9); and hydrolase S is a secreted form of this family of
esterases (10). The cDNA sequences for all 4 rat hydrolases have
been published and confirm the suggestion of Mentlein et al.
(11) that there are many carboxylesterases expressed in rat liver. Multiple carboxylesterase isozymes also were identified in mouse liver
that are encoded by two gene clusters on chromosome 8 (12). Less is
known about the multiplicity and genetics of carboxylesterases in
rabbit, hamster, pig, or human tissues. Two different esterases were
purified from rabbit tissues and sequenced by Ozols and co-workers (13-15). Coates et al. (16) suggested that in humans there
are many esterase isoenzymes. An enzyme called hCE-1 was recently purified from human liver (2) and alveolar monocytes or macrophages (17). The cDNA (17, 18) and genomic (19) sequences of hCE-1 have
high sequence identity with rat hydrolase A (20), rabbit
carboxylesterase form 1 (14), porcine liver proline--naphthylamidase (21), and mouse carboxylesterase Es-22 (22).
In this study, we describe the purification, molecular cloning, and partial characterization of a human liver carboxylesterase hCE-2 that catalyzes the hydrolysis of 4-methylumbelliferyl acetate to 4-methylumbelliferone, cocaine to ecgonine methyl ester and benzoic acid, heroin to 6-monoacetylmorphine, and 6-monoacetylmorphine to morphine.
Two types of enzyme assays were
performed. Acetylesterase activity was determined using
4-methylumbelliferyl acetate as substrate. The reaction was initiated
by adding 10 µl of enzyme (0.1 µg of the purified enzyme or about
550 µg of the liver extract) to 1 ml of 90 mM
KH2PO4, 40 mM KCl, at pH 7.3 containing 0.5 mM 4-methylumbelliferyl acetate at 37 °C
(2). The production of 4-methylumbelliferone (350 = 12.2 mM
1 cm
1) was monitored with a
Perkin-Elmer Lambda-6 spectrophotometer. Initial rates were linear for
1.5 min. Specific activity was expressed as micromoles of product
formed per minute per milligram of protein. Cocaine benzoyl ester
hydrolase activity was measured by incubating 0.1 ml of enzyme
(adjusted to 5 units/ml based on the 4-methylumbelliferyl acetate
assay) with 3.3 mM cocaine in 50 mM
NaH2PO4, pH 7.4, containing 1 mM
benzamidine, 1 mM EDTA, and 1 mM dithiothreitol
at 37 °C for 1 h in a total volume of 0.4 ml. The reaction was
terminated with an equal volume of 5% trichloroacetic acid, and 0.1 ml
of the internal standard 3,4-dimethylbenzoic acid (10 µg/ml) was added. Precipitated protein was separated by centrifugation at 12,000 × g for 10 min. The supernatant was transferred
to a clean tube and extracted twice with 0.4 ml of methylene chloride.
The two extracts were combined and evaporated to dryness under a stream of nitrogen. The residue was resuspended in 100 µl of the HPLC mobile
phase (14% acetonitrile, 250 mM
KH2PO4, pH 4.0, with 8 µl of diethylamine per
100 ml) and aliquots were injected onto a Waters NovaPack
C18 phase HPLC column (150 mm × 3.9 mm, 4-µm particle size) maintained at 40 °C with a flow rate of 1.2 ml/min. The column eluate was monitored at 235 nm. Benzoic acid concentrations were quantified by comparing peak area ratios of the analyte to the
internal standard. A standard curve was generated with benzoic acid
standards prepared in a trichloroacetic acid-treated protein matrix.
For kinetic analysis with cocaine, purified hCE-2 was incubated with 0.1-5 mM cocaine at 37 °C for 20 min. The linearity of initial reaction rates was verified by incubating enzyme with 0.1-5 mM cocaine for 20, 40, and 120 min at 37 °C. Production of benzoic acid was monitored by HPLC as described above and corrected for nonenzymatic benzoic acid formation. For kinetics with 4-methylumbelliferyl acetate, the substrate concentration ranged from 0.0125 to 0.8 mM. The Km values for substrates were calculated from nonlinear regression analysis of the kinetic data to the Michaelis-Menten equation using the GraFit program (23). The calculation of kcat was based on a molecular mass of 60 kDa.
For kinetic studies of eserine inactivation the purified hCE-2 was diluted to 0.02 µM in 90 mM KH2PO4, at pH 7.3 containing 0.11-0.44 µM eserine. At 1-25 min, 50-µl aliquots were removed to measure residual acetylesterase activity with 4-methylumbelliferyl acetate as substrate as described above. The first-order rate constant of inactivation, kobs, was calculated by nonlinear regression analysis of the residual activity as a function of time using the "Single Exponential Decay" analysis of the GraFit program (23). The maximal first-order rate constant for the inactivation, kinact, and the eserine dissociation constant, KI, were determined by fitting the eserine dependence of kobs to the Michaelis-Menten equation (24).
Purification of Human Liver Cocaine Benzoyl EsteraseAll buffers utilized in purification procedures were helium purged and contained 1 mM benzamidine, 1 mM EDTA, and 1 mM dithiothreitol. Frozen human liver obtained at autopsy, approximately 70 g, was homogenized in 70 ml of 50 mM HEPES, pH 6.8. The homogenate was centrifuged at 125,000 × g for 35 min at 5 °C. The homogenate was filtered through Miracloth (Calbiochem Corp.) to remove fat and applied to a DEAE-cellulose column (Whatman Biosystems) equilibrated with 50 mM HEPES, pH 6.8. After washing the column extensively with buffer, the carboxylesterase activity was eluted with 250 mM KH2PO4, pH 6.8. The buffer of the sample was exchanged for 75 mM Tris-Cl, pH 7.6, using a Minitan concentrator (30-kDa membrane, Millipore) and the sample was loaded onto a Q-Sepharose fast-flow column (5.5 × 8 cm, Pharmacia). Carboxylesterase activity was eluted with a linear gradient of 75 mM Tris-Cl, pH 7.6, to 250 mM Na2HPO4, pH 7.5. Two peaks of activity were recovered: the first peak contained cocaine methyl esterase activity (hCE-1) and the second peak contained cocaine benzoyl esterase activity (hCE-2) (2). The fractions corresponding to the second peak were pooled and concentrated with a PM30 Centricon concentrator (Amicon). After the buffer was exchanged for 10 mM KH2PO4, pH 7.0, the sample was loaded onto a hydroxylapatite column (2.5 × 25 cm, Bio-Rad) and enzyme was eluted with 30 mM KH2PO4, pH 7.0. Active fractions were pooled, concentrated, and equilibrated with 20 mM Tris-Cl, 0.5 mM NaCl, 1 mM CaCl2, and 1 mM MnSO4, pH 7.4. The sample was applied to a concanavalin A-Sepharose affinity column (1 × 3 cm, Pharmacia Biotech Inc.) and enzyme was eluted with a linear gradient to 0.7 M methylmannoside in buffer. The buffer of the activity pool was exchanged for 50 mM NaH2PO4, pH 7.5, and gel filtered on a Superose 12 column (1.0 × 30 cm, Pharmacia). The purified enzyme was concentrated, sterile-filtered, and stored at 4 °C. Protein concentrations were determined by Bio-Rad protein assay with bovine serum albumin as a standard (25).
SDS-PAGE, IEF, and Gel FiltrationThe carboxylesterase purity was examined at each chromatography step by SDS-PAGE according to the method of Laemmli (26) using the Mighty Small Vertical Slab gel unit (Hoefer Scientific Instruments). The gels were stained for protein with an alkaline silver reagent (27).
Isoelectric focusing of the purified esterase was performed using IsoGel-agarose IEF plates (FMC BioProducts), pH range 3-10. Gels were stained for protein with Coomassie Blue according to the instructions provided by FMC BioProducts and for esterase activity using 0.01% 4-methylumbelliferyl acetate in 0.1 M NaH2PO4, pH 7.0. The fluorescent product 4-methylumbelliferone was visualized under UV light.
Purified enzyme was applied to an AcA 44 gel filtration column in 0.05 M phosphate, pH 7.4, to estimate the molecular size of the esterase. The following molecular mass markers were used to calibrate the column: cytochrome c (12.4 kDa), carbonic anhydrase (31 kDa), bovine serum albumin (66.2 kDa), and yeast alcohol dehydrogenase (150 kDa).
Glycoprotein AnalysisPurified protein (5 µg) was incubated without and with 3 milliunits of endoglycosidase H for 30 h at 37 °C (28). The reaction products were analyzed on a 7.5% polyacrylamide gel containing 0.1% SDS. Protein bands were visualized with a silver stain (27).
Peptide Mapping and Amino Acid SequencingPurified esterase
(20 µg) was treated overnight with 1.0 M urea in 0.1 M Tris-Cl, pH 8.5, at 20 °C. The denatured protein was
quickly thawed and incubated with 1 µg of trypsin (Boehringer Mannheim) overnight at 37 °C (29). For digestion with
Staphylococcus aureus V8 protease (Boehringer Mannheim),
purified esterase (20 µg) was treated overnight with 1.0 M urea in 25 mM
(NH4)2CO3, pH 7.8, at
20 °C.
After rapid thawing, 1 µg of V8 protease was added and the mixture
was incubated for 5 h at 25 °C (29). The proteolytic reactions
were stopped with 20 µl of 10% trifluoroacetic acid. Peptide
mixtures were separated on an Aquapore reverse-phase HPLC column (ABI;
2.1 × 30 mm) using an Applied Biosystems 130A HPLC with a
gradient of 0.1% trifluoroacetic acid to 70% acetonitrile in 0.1%
trifluoroacetic acid (30). Amino acid sequencing was performed by Edman
degradation on a Porton automated amino acid sequencer (Beckman
Instruments). A search of the peptide sequences was performed in the
GenBank protein data base.
We used the information for
alignment of hCE-2 peptides with rabbit carboxylesterase form 2 (15) to
design a strategy for the human liver enzyme. The following degenerate
oligonucleotides were synthesized based on amino acid regions of the
rabbit enzyme: 114-120 (primer 1, 5-CC(A/G/C/T)GT(A/G/C/T)ATGGT(A/G/C/T)TGGAT(A/C/T)CA-3
) and 464-780
(primer 2, 5
-TT(A/G/C/T)CG(A/G/C/T)GC(A/G)AA(A/G)TT(A/G/C/T)GCCCA-3
; and
primer 3, 5
-TT(C/T)CT(A/G/C/T)GC(A/G)AA(A/G)TT(A/G/C/T)GCCCA-3
). These primers were used to amplify human liver poly(A)+
mRNA (CLONTECH). The following amplification
protocol was employed for 30 cycles: denaturing at 94 °C for 1 min,
primer annealing at 50 °C for 1 min, and primer extension at
72 °C for 3 min. The cDNA fragments obtained by PCR were
purified, cleaved to form blunt-ends, and subcloned into M13 mp18 (Life
Technologies, Inc.) for sequencing. A 1095-bp PCR product was
end-labeled with [
-32P]ATP (U. S. Biochemical Corp.)
and used as a probe to screen a
gt10 human liver cDNA library
(CLONTECH) (31). The hybridization was done in
5 × Denhardt's solution, 5 × SSPE, 50% formamide, 0.1 mg/ml denatured salmon sperm DNA, 0.1% SDS, and radiolabeled probe
(2 × 106 cpm/ml) at 42 °C overnight. The final
wash of nitrocellulose filters was performed in 0.1 × SSC, 0.1%
SDS at 65 °C for 30 min. Positive clones were purified and the
cDNA inserts were subcloned into M13mp18 digested with
EcoRI.
The 5 end was obtained in 2 sequential PCR amplifications of the human
liver 5
-RACE ready templates (CLONTECH) with 2 nested sequence-specific primers: 5
-AAGGAAGCCATGCCAAAAACAA-3
and
5
-CACCGTGAATCCATACCATCA-3
. The 5
-RACE was performed according to the
manufacturer's protocol. Thirty cycles of PCR were performed as
follows: 1 min 94 °C (denaturation), 1 min at 60 °C (annealing),
and 5 min at 72 °C (extension). The PCR product was treated with
Klenow fragment of DNA polymerase, and subcloned into
HincII-digested M13mp18 for sequencing.
To obtain the full-length cDNA of hCE-2, two primers were
synthesized, 5-ATGAGCGCGGTGGCCTGTGGG-3
, based on sequence of the 5
-RACE PCR product, and 5
-CTACAGCTCTGTGTGTCTCTC-3
, based on the
sequence of partial cDNA clone and used for PCR. First strand cDNA was synthesized from the poly(A)+ mRNA by
reverse transcription using random hexamers as primers. The hCE-2
cDNA was amplified by 30 cycles of PCR as follows: denaturing at
94 °C for 1 min, primer annealing at 60 °C for 1 min, and primer extension at 72 °C for 3 min. The 1650-bp cDNA was treated with Klenow fragment of DNA polymerase, subcloned into the HincII
restriction site of M13mp18 vector, and the sequenase was
determined.
Single-stranded M13 DNA sequencing was done by the dideoxy termination method of Sanger et al. (32) using the Sequenase version 2 DNA sequencing kit (U. S. Biochemical Corp.). The search for sequence homology of hCE-2 with other proteins was performed in the GenBank protein data base. Alignment of the deduced protein sequence to other carboxylesterases, calculation of the percent identity, and generation of a phylogenetic tree were performed using the progressive alignment method of Feng and Doolittle (33) including the set of programs: FORMAT, SCORE, PREALIGN, TREE, and BLEN.
A carboxylesterase called hCE-2 was purified from frozen human liver obtained at autopsy. The extraction buffer did not contain detergents and enzyme in the 125,000 × g supernatant was purified to homogeneity using ion-exchange, hydroxylapatite, lectin affinity, and gel filtration chromatography (Table I). The hCE-2 was separated from the carboxylesterase that cleaves the methyl ester group of cocaine (hCE-1) by chromatography on Q-Sepharose (2). Two types of assay were performed to follow carboxylesterase activity throughout purification procedures (Table I). The rapid spectrophotometric assay with 4-methylumbelliferyl acetate as a substrate measures hydrolytic activity of both hCE-1 and hCE-2. The assay with cocaine as a substrate measures the specific cocaine benzoyl esterase activity that is catalyzed by hCE-2 but not hCE-1 (2). A 690-fold purification was obtained using the cocaine benzoyl esterase assay and the final yield of hCE-2 was 5% as calculated from the DEAE-cellulose step. The apparent increase in activity from the liver extract to DEAE-cellulose (Table I, cocaine assay) is undoubtedly due to underestimation of the activity in the liver extract because of substrate depletion by hydrolysis of cocaine to benzoylecgonine (2). The progress of hCE-2 purification was monitored by SDS-PAGE (Fig. 2A). Purified enzyme exhibited a single band on SDS-PAGE with a subunit mass of approximately 60 kDa. The subunit mass was calculated by comparison of the distance of hCE-2 migration with the four reference proteins. The native size of hCE-2 was estimated to be about 80 kDa based on a calibrated AcA 44 gel filtration column (not shown). Hence, we conclude that the native enzyme is a monomer. The purity of the carboxylesterase also was examined by isoelectric focusing (Fig. 2C). Multiple esterase activity bands appeared using 4-methylumbelliferyl acetate as the substrate that coincided with multiple protein bands. The apparent pI values ranged from 4.8 to 5.0. The purified hCE-2 was digested with endoglycosidase H, which hydrolyzes Asn-linked high mannose oligosaccharides (28). Analysis of the treated enzyme by SDS-PAGE revealed a decrease in subunit mass (Fig. 2B) after cleavage of carbohydrate residues.
|
The Km for 4-methylumbelliferyl acetate hydrolysis
was 0.15 mM for hCE-2 and 0.8 mM for hCE-1
(Table II). The catalytic efficiency,
kcat/Km, for hCE-2 was 30 times greater than that for hCE-1. The Km of hCE-2
for cocaine was 0.39 mM and
kcat/Km was 18.4 mM1 min
1. Kinetic constants for
heroin and 6-monoacetylmorphine hydrolysis by both liver
carboxylesterases reported by Kamendulis et al. (5) are
shown in Table II. hCE-1 was inhibited only 25% by 1 mM
eserine. hCE-2 was quite sensitive to irreversible inhibition by
micromolar concentrations of eserine. The KI value for eserine inhibition was 0.10 ± 0.01 µM and the
maximal rate of inactivation was 0.15 ± 0.004 min
1.
The bimolecular rate constant for inactivation would be 1.5 min
1 µM
1.
|
The purified enzyme was cleaved with two proteases: trypsin that
cleaves after Lys and Arg residues and S. aureus V8 protease that cleaves after Glu residues at pH 7.8. After separation of peptides
by reversed phase HPLC, the amino acid sequences of 10 tryptic and
S. aureus V8 protease fragments (Fig. 3) were determined. The sequences of these peptides (78 amino acids) had highest identity (about 70%) with rabbit carboxylesterase form 2 (15). Alignment of
known mammalian carboxylesterase sequences revealed regions that are
highly conserved. Degenerate primers were designed and used to amplify
cDNA of related carboxylesterases from human liver mRNA based
on two amino acids regions that exhibit highest identity (Fig. 3). The
1.1-kilobase PCR product was subcloned into M13mp18 and 12 individual
clones were sequenced. The sequences of four clones were unrelated to
any known carboxylesterases, six clones matched the human
carboxylesterase hCE-1 (17, 18), and two clones (1095 bp) exhibited a
sequence highly homologous to the rabbit carboxylesterase form 2 (15)
and perfectly matched the sequence of peptides from the purified hCE-2
(Fig. 3). The 1095-bp cDNA fragment corresponded to nucleotides
367-1461 of the cloned hCE-2 (Fig. 3) and was used to screen a human
liver cDNA library. Three clones with insert sizes of 1000, 750, and 380 bp were isolated. The 1000- and 750-bp sense fragments both
corresponded to sequence starting at position 1007 of the cloned hCE-2
(Fig. 3), overlapped the 1095-bp cDNA fragment by 455 bases, and
spanned the 3-end containing a stop codon (TAG) followed by the
untranslated region (Fig. 3). The sequence of the 380-bp antisense
fragment, corresponded to nucleotides 1007 to 1386. Screening did not
yield any full-length cDNA clones or partial clones at the 5
-end.
To obtain the remaining sequence of hCE-2 cDNA, sequence-specific
primers were created (Fig. 3) for the 5
-end of the 1095-bp cDNA.
These primers were successively utilized in the amplification of human
liver 5
-RACE ready templates. Polymerase chain reaction yielded a
395-bp product which overlapped the 1095-bp cDNA by 29 bases and
included the ATG initiator codon (Fig. 3).
To obtain a full-length cDNA, sequence-specific primers were
designed according to the nucleotide sequence of the 5- and 3
-ends of
hCE-2. mRNA from human liver was isolated and amplified by reverse
transcription PCR. The 1650-bp PCR product was cloned into M13mp18 and
both strands were sequenced. The sequence matched those of the 5
-RACE
and cDNA clones. Authenticity of the cDNA was confirmed by
perfect matches with peptides derived from purified preparations of
hCE-2 (Fig. 3). The complete nucleotide sequence and the deduced amino
acid sequence of hCE-2 are shown in Fig. 3 and are deposited in GenBank
(U60553[GenBank]). The enzyme consists of 550 amino acids, including 17 amino
acids of a putative signal peptide.
The complete amino acid sequence of hCE-2 was aligned with
carboxylesterases from human (17, 18), rabbit (14, 15), rat (8, 9, 10,
34), mouse (22, 35, 36), pig (21), and hamster (37). The alignment of
hCE-2 with 4 of the carboxylesterases is shown in Fig. 4. A
phylogenetic tree of carboxylesterase family (Fig. 5) was produced on
the basis of this alignment. The initial version of the tree had two
negative branch lengths, so modifications were made by regrouping
certain members into clusters. The final version of the tree had no
negative branch lengths and low percentage standard deviation of 2.8. The hCE-2 exhibited 48% sequence identity with hCE-1. The highest
sequence identity was obtained with the rabbit liver carboxylesterase
form 2 (73%) (15) and hamster liver esterase AT51p (67%) (37).
Two cocaine carboxylesterases have been identified in human liver
(1). One enzyme, called hCE-1, catalyzes the hydrolysis of cocaine to
benzoylecgonine and methanol and the ethyl transesterification of
cocaine to form cocaethylene, a pharmacologically active metabolite of
cocaine found in individuals succumbing to overdose of cocaine and
ethanol (38, 39). The purification and characterization of this enzyme
was reported recently (2). In this study, we report the purification,
cloning, and characterization of a new human liver carboxylesterase
(hCE-2) that catalyzes the hydrolysis of the benzoyl group of cocaine
and the acetyl groups of heroin (Fig. 1).
The purified hCE-2 enzyme is a 60-kDa monomer (Fig. 2A). Treatment of the purified enzyme with endoglycosidase H indicates that the native liver enzyme is a glycoprotein (Fig. 2B). A partial amino acid sequence of the purified enzyme and the complete cDNA nucleotide sequence for hCE-2 were determined (Fig. 3). Two potential Asn-X-Ser/Thr glycosylation sites were identified at positions 85 and 250 of the deduced amino acid sequence of the cDNA (Fig. 4). The enzyme exhibits multiple protein bands that coincide with multiple activity bands on an isoelectric focusing gel with pI values ranging from 4.8 to 5.0 (Fig. 2C).
The alignment of the deduced hCE-2 amino acid sequence with
carboxylesterases reported in GenBank from different species did not
reveal an identical match indicating that hCE-2 is a new human carboxylesterase. The alignments of hCE-2 with hCE-1, rabbit form 2, and two rat forms are shown in Fig. 4. The amino acid sequence of the
cloned enzyme has high sequence identity to rabbit carboxylesterase form 2 (73%, (15)) and hamster carboxylesterase AT51p (67% (37)). Rabbit esterase form 2 was identified as a protein belonging to the
family of enzymes with a characteristic /
-hydrolase fold where
eight
-sheet strands and several connecting
helices provide a
stable structure for the active site catalytic triad (40, 41). The
sequence identity of hCE-2 with other mammalian carboxylesterases is
substantially lower, an average 47%. The phylogenetic tree calculated
by progressive alignment of 13 carboxylesterase sequences shows
that the enzymes cluster into two main groups (Fig. 5). This tree is similar to that for eight carboxylesterase sequences reported by Shibata et al. (19). The hCE-2 is grouped with
rabbit form 2 and hamster AT51p carboxylesterases. These three enzymes have a 15-residue gap (19), and share a potential
N-glycosylation site (37) with the Asn-X-Ser
sequence that occurs at amino acid 250 in hCE-2 (Fig. 4). This group
diverged relatively early from the other major group of nine
carboxylesterases that includes rat, mouse, pig, and rabbit form 2 carboxylesterases, and hCE-1 (Fig. 5). The mouse Es-male esterase seems
very distantly related to these two groups of carboxylesterases.
Surprisingly, an enzyme with high sequence identity to hCE-2 has not
been identified in either rat or mouse liver.
The hCE-2 sequence, deduced from cDNA (Fig. 3) has amino acids that
are highly conserved among carboxylesterases (Fig. 4). They include two
pairs of Cys residues in positions 69, 97, and 254, 265 that are
thought to participate in the formation of intramolecular disulfide
bonds (42). The active site Ser202 in a
Gly-Glu-Ser-Ala-Gly-Gly sequence and catalytic
His431 in an Ala-Asp-His-Gly-Asp-Glu sequence
could be two members of the esterase catalytic triad, since the
consensus sequence appears in esterases having the /
-hydrolase
fold (40). Esterase activity is thought to be dependent also on an
acidic amino acid (43), which is the third member of the catalytic
triad. Glu319 in the
Gly-(aa)4-Glu-Phe-Gly sequence (Fig. 4) could
serve this function (40).
In addition to these conserved sequences of carboxylesterases, the
human hCE-2 enzyme has two microsomal targeting sequences: a
hydrophobic 17-residue signal peptide at the N terminus that apparently
targets the protein to the endoplasmic reticulum (Fig. 3) and a
conserved 4-amino acid sequence, HTEL (Fig. 3 and 4), at the C terminus
that presumably functions in the retention of proteins in the
endoplasmic reticulum (44, 45). These sequence comparisons suggest that
hCE-2 is localized in the endoplasmic reticulum. However, as indicated
under "Results," the carboxylesterase protein in this study could
be isolated from the 125,000 × g supernatant without
treatment of the liver homogenates with detergents. Hence, hCE-2 seems
weakly bound to the endoplasmic reticulum or the enzyme could have been
released during storage of autopsy liver at 70 °C. The use of
Triton X-100 (0.5%) in the homogenate buffer did release about 3 times
more carboxylesterase activity measured with 4-methylumbelliferyl acetate.2 We have not attempted to purify
the enzyme from the detergent-treated liver homogenate and characterize
the released enzyme.
The catalytic properties of hCE-2 and hCE-1 are clearly different, as shown in Table II. The catalytic efficiency of hCE-2 for 4-methylumbelliferyl acetate hydrolysis is about 30 times that of hCE-1. With respect to the two ester groups of cocaine (Fig. 1), hCE-1 specifically hydrolyzes the methyl ester while hCE-2 specifically hydrolyzes the benzoyl group. Both esterases hydrolyze the 3-acetyl group of heroin (Fig. 1), but hCE-2 has about 5 times higher catalytic efficiency. Hydrolysis of the 6-acetyl group of 6-monoacetylmorphine is much more efficient with hCE-2 than hCE-1 by almost 1000-fold (5). The two human carboxylesterases can also be differentiated by the sensitivity of hCE-2 but not hCE-1 to inhibition by eserine. In a comparison of active site structures of Torpedo californica acetylcholinesterase and the Geotrichum candidum lipase with sequence alignments of 32 enzymes, Cygler and co-workers (40) suggested that amino acids of many carboxylesterases and lipases corresponding to the region following Met283 of hCE-2 form an important loop over the active site. The deletion of this region in hCE-2 (Fig. 4) is likely to account for some of the differences in substrate specificity relative to hCE-1.
The carboxylesterase reported here is one of the three enzymes that play an important role in metabolism of cocaine and heroin in humans (Fig. 1) (1, 5). They are serum cholinesterase (3) and the organ-specific hCE-1 and hCE-2 carboxylesterases (1, 2). hCE-2 was the first human enzyme reported to hydrolyze 6-monoacetylmorphine to morphine with high catalytic efficiency (5). The Km values of hCE-2 and hCE-1 for cocaine and heroin are much higher than the likely drug concentrations in vivo (46, 47). Hence, these enzymes would exhibit first-order kinetics with respect to cocaine and heroin at physiological concentrations. An analysis of the kinetic properties and tissue content of the two human tissue carboxylesterases hCE-2 and hCE-1, as well as serum cholinesterase, should provide insight into the relative importance of these esterases in the pharmacokinetics of cocaine and heroin elimination. The rat and mouse are frequently used as animal systems to model the pharmacokinetics of cocaine and heroin detoxification in humans. It is important to note that esterases having high sequence identity with hCE-2 have not been described in rats or mice. Hence, use of these rodent models may not be appropriate for the study of cocaine and heroin metabolism in humans. While hCE-2, hCE-1, and related carboxylesterases from other species all exhibit relatively broad substrate specificities for ester hydrolysis, the above mentioned studies with cocaine and heroin do indicate important differences in catalytic efficiencies for certain substrates. The identification of differences in substrate-binding site structures should provide important information about the relative roles of the different carboxylesterase isoenzymes in drug and xenobiotic ester metabolism.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U60553[GenBank].
We thank Kirill M. Popov for helpful suggestions with cloning and Natividad Dumaual for assistance with HPLC analysis.