Biochemical and molecular analysis of carboxylesterase-mediated hydrolysis of cocaine and heroin

Article abstract of DOI:10.1111/j.1476-5381.2010.00700.x

Background and purpose: Carboxylesterases (CEs) metabolize a wide range of xenobiotic substrates including heroin, cocaine, meperidine and the anticancer agent CPT-11. In this study, we have purified to homogeneity human liver and intestinal CEs and compared their ability with hydrolyse heroin, cocaine and CPT-11. Experimental approach: The hydrolysis of heroin and cocaine by recombinant human CEs was evaluated and the kinetic parameters determined. In addition, microsomal samples prepared from these tissues were subjected to chromatographic separation, and substrate hydrolysis and amounts of different CEs were determined. Key results: In contrast to previous reports, cocaine was not hydrolysed by the human liver CE, hCE1 (CES1), either as highly active recombinant protein or as CEs isolated from human liver or intestinal extracts. These results correlated well with computer-assisted molecular modelling studies that suggested that hydrolysis of cocaine by hCE1 (CES1), would be unlikely to occur. However, cocaine, heroin and CPT-11 were all substrates for the intestinal CE, hiCE (CES2), as determined using both the recombinant protein and the tissue fractions. Again, these data were in agreement with the modelling results. Conclusions and implications: These results indicate that the human liver CE is unlikely to play a role in the metabolism of cocaine and that hydrolysis of this substrate by this class of enzymes is via the human intestinal protein hiCE (CES2). In addition, because no enzyme inhibition is observed at high cocaine concentrations, potentially this route of hydrolysis is important in individuals who overdose on this agent.

Full text of DOI:10.1111/j.1476-5381.2010.00700.x

British Journal of Pharmacology (2010), 160, 1916–1928
2010 The Authors
©
Journal compilation © 2010 The British Pharmacological Society All rights reserved 0007-1188/10
www.brjpharmacol.org
RESEARCH PAPER
Biochemical and molecular analysis of
carboxylesterase-mediated hydrolysis of cocaine
and heroinbph_700 1916..1928
1
1
1
2
1
1
2
MJ Hatfield , L Tsurkan , JL Hyatt , X Yu , CC Edwards , LD Hicks , RM Wadkins and
PM Potter
1
1
2
Department of Molecular Pharmacology, St. Jude Children’s Research Hospital, Memphis, TN, USA, and Department of
Chemistry and Biochemistry, University of Mississippi, University, MS, USA
Background and purpose: Carboxylesterases (CEs) metabolize a wide range of xenobiotic substrates including heroin,
cocaine, meperidine and the anticancer agent CPT-11. In this study, we have purified to homogeneity human liver and
intestinal CEs and compared their ability with hydrolyse heroin, cocaine and CPT-11.
Experimental approach: The hydrolysis of heroin and cocaine by recombinant human CEs was evaluated and the kinetic
parameters determined. In addition, microsomal samples prepared from these tissues were subjected to chromatographic
separation, and substrate hydrolysis and amounts of different CEs were determined.
Key results: In contrast to previous reports, cocaine was not hydrolysed by the human liver CE, hCE1 (CES1), either as highly
active recombinant protein or as CEs isolated from human liver or intestinal extracts. These results correlated well with
computer-assisted molecular modelling studies that suggested that hydrolysis of cocaine by hCE1 (CES1), would be unlikely to
occur. However, cocaine, heroin and CPT-11 were all substrates for the intestinal CE, hiCE (CES2), as determined using both
the recombinant protein and the tissue fractions. Again, these data were in agreement with the modelling results.
Conclusions and implications: These results indicate that the human liver CE is unlikely to play a role in the metabolism of
cocaine and that hydrolysis of this substrate by this class of enzymes is via the human intestinal protein hiCE (CES2). In addition,
because no enzyme inhibition is observed at high cocaine concentrations, potentially this route of hydrolysis is important in
individuals who overdose on this agent.
British Journal of Pharmacology (2010) 160, 1916–1928; doi:10.1111/j.1476-5381.2010.00700.x
Keywords: carboxylesterase; cocaine; heroin; CPT-11; inhibitor; molecular modelling
Abbreviations: BChE, butyrylcholinesterase; CE, carboxylesterase; clogP, calculated logP; CPT-11, irinotecan, 7-ethyl-10-[4-
(
1-piperidino)-1-piperidino]carbonyloxycamptothecin; DMSO, dimethyl sulfoxide; hBChE, human BChE;
hBr3, human brain CE, CES3; hCE1 (CES1), human carboxylesterase 1, CES1; hiCE (CES2), human intestinal
carboxylesterase, CES2, hCE2; i, fractional inhibition; IEC, ion-exchange chromatography; kcat, turnover
constant; K
i
, inhibition constant; K , Michaelis constant; o-NPA, o-nitrophenyl acetate; TBAP, tetrabutyl
m
ammonium phosphate; Vmax, maximal reaction velocity
Introduction
and CPT-11 (Pindel et al., 1997; Danks et al., 1999; Zhang
et al., 1999; Tabata et al., 2004). CEs tend to be expressed in
tissues likely to be exposed to such agents, for example, the
liver, gut, kidney, lung. Although numerous CEs have been
identified in rodents and other mammals, currently only
three such enzymes have been characterized from humans.
Human liver CE (hCE1; CES1) is predominantly expressed in
the liver and is efficient at metabolizing small esterified sub-
strates (Wadkins et al., 2001). Human intestinal CE (hiCE,
hCE2, CES2) is expressed in the epithelial lining of the gut
and the liver, and has been demonstrated to be efficient in the
hydrolysis of the anticancer prodrug CPT-11 (Humerickhouse
Carboxylesterases [carboxylic ester hydrolase (CE), EC 3.1.1.1]
are responsible for the detoxification of numerous xenobiot-
ics including heroin, cocaine, meperidine and lidocaine, as
well as the activation of the anticancer agents, capecitabine
Correspondence: Dr Philip M. Potter, Department of Molecular Pharmacology,
St. Jude Children’s Research Hospital, 262 Danny Thomas Place, Memphis, TN
3
8105-2794, USA. E-mail: philpotter@stjude.org
Received 22 September 2009; revised 13 November 2009; accepted 21
December 2009

Carboxylesterase hydrolysis of cocaine and heroin
MJ Hatfield et al
1917
et al., 2000; Khanna et al., 2000). A third human CE, human
brain CE (hBr3; CES3) has been recently identified, but as yet
little is known about its pattern of expression and/or substrate
specificity (Mori et al., 1999).
(Beaufay et al., 1974; Danks et al., 1998). Protein concentra-
tions were determined using the Bradford method (using Bio-
Rad reagent, Hercules, CA, USA). Data were expressed as
nmoles of o-nitrophenol produced per minute per milligram
of protein (nmol/min/mg) as calculated using the extinction
Cocaine and heroin abuse is a major clinical problem in the
3
-1
-1
2
5 to 49 year-old age group, with drug overdose being the
coefficient for o-nitrophenol (13.6 ¥ 10 M cm ).
fourth leading cause of death, on a par with motor vehicle
accidents (Anderson, 1999). Hence devising alternative treat-
ment options for patients who have overdosed on these
agents is desirable. Although naloxone has significantly
reduced the incidence of heroin-induced fatalities, there are
currently few effective drugs for cocaine overdose. We have
proposed that in patients who have overdosed on cocaine,
administration of a recombinant enzyme (e.g. a CE) that
could metabolize the drug to inactive metabolites, may be
effective in ameliorating the toxicity associated with this
agent (Potter and Wadkins, 2006; Redinbo and Potter, 2005).
However, for any approach to be successful, a detailed analy-
sis of the metabolism of cocaine by different enzymes will be
essential.
Kinetics of nitrophenyl ester hydrolysis
The kinetic parameters K , k and Vmax for a panel of
m
cat
nitrophenyl esters (Table 1) were determined as previously
described (Wadkins et al., 2001). Briefly, assays were
performed in 50 mM 4-(2-hydroxyethyl)-1-piperazineethane-
sulfonic acid (HEPES) pH 7.4 and formation of the products
was determined spectrophotometrically. Product yields were
then calculated using the extinction coefficients for nitrophe-
nol, and kinetic parameters were determined from these
values
Although the metabolism of heroin and cocaine has been
presumed to occur via CE-mediated hydrolysis of the ester
functions, it is clear that butyrylcholinesterase (BChE) can
also hydrolyse these drugs (Gatley, 1990; Lockridge and LaDu,
Kinetics of CPT-11 metabolism
The kinetic parameters for CPT-11 hydrolysis were deter-
mined as previously described (Wadkins et al., 2001).
1
980). However, the relative contribution of CEs and BChE
towards metabolism of each of these substrates is unclear, in
part due to the lack of availability of the purified proteins, and
also due to the lack of specific reagents that could inhibit the
different enzyme isoforms.
Analysis of cocaine hydrolysis
To assess the hydrolysis of cocaine to benzoylecgonine and
ecgonine methyl ester, the drug (1 mM) was incubated with
either media harvested from Sf21 cells (150 mL) or purified
CEs (20 mg) in 50 mM HEPES pH 7.4 at 37°C. After 60 min,
samples were analysed by high-performance liquid chroma-
tography (HPLC) for metabolite formation. For kinetic
studies, 0–5 mM of cocaine was incubated with hiCE (CES2)
As hCE1 (CES1), hiCE (CES2) and BChE are all expressed in
the liver, the determination of which enzyme can metabolize
particular substrates has been difficult. In this article, we have
purified recombinant hiCE (CES2) to homogeneity, assessed
its hydrolytic activity towards several different substrates
including heroin and cocaine, and determined the ability of
specific inhibitors to modulate enzyme activity. Furthermore,
we have correlated these results using human liver and
intestinal specimens and corroborated our experiments using
computer-assisted molecular modelling. These studies
confirm that cocaine is exclusively metabolized by hiCE
(250 U/reaction, 10 min) at 37°C. One unit of enzyme activity
is equivalent to 1 nmol of nitrophenol produced per min.
Reactions were terminated by addition of cold acidified
methanol and drug concentrations were determined using
isocratic HPLC analysis. Samples (20 mL) were separated using
a 3.9 ¥ 300 mm 4 mm Nova-Pak C18 reverse-phase column
(
Waters Corporation, Milford, MA, USA) and a mobile phase
containing 25% (v/v) acetonitrile in 50 mM NH PO
50 mM tetrabutyl ammonium phosphate (TBAP), pH 2.2.
(CES2) and not by hCE1 (CES1).
4
H
2
4
,
1
Metabolites were eluted at a flow rate of 1.0 ml/min, and
identified by ultraviolet detection (235 nm). As ecgonine
methyl ester absorbs poorly at this wavelength, hydrolysis of
cocaine to this product was monitored by the formation of
benzoic acid (the acid released from enzymatic hydrolysis).
Under these conditions, cocaine, benzoylecgonine and
Methods
CE activity assay
CE activity was measured by following o-nitrophenyl acetate
(o-NPA) hydrolysis at 420 nm as previously described
Table 1 Purification of hiCE (CES2) to homogeneity
-1
Sample
Volume (mL)
Total activity (U)
Total protein (mg)
Specific activity (U mg )
Recovery (%)
Crude media
After ultrafiltration
After IEC
After ultrafiltration
After Sephacryl S200
1100
520
146
0.3
5
472 200
472 700
116 400
96 900
300
260
4.00
3.20
0.80
1 550
1 840
29 300
30 700
52 100
100
100
25
21
9
41 700
A typical purification profile for hiCE (CES2), isolated from serum-free media, harvested from Spodoptera frugiperda Sf21 cells expressing a secreted form of the
enzyme. Carboxylesterase (CE) activity was determined using o-NPA as a substrate and 1 U is equivalent to 1 nmol of o-nitrophenol produced per minute.
British Journal of Pharmacology (2010) 160 1916–1928

Carboxylesterase hydrolysis of cocaine and heroin
1
918
MJ Hatfield et al
benzoic acid eluted at 4.3, 2.8 and 7.4 min, respectively. Drug
concentrations were calculated using System Gold software
volumes of 50 mM HEPES pH 7.4 buffer containing 125 mM
NaCl. Elution of hiCE (CES2) was accomplished with 50 mM
HEPES pH 5.5 buffer containing 130 mM NaCl. The hiCE
(CES2)-containing fractions were pooled, concentrated by
ultrafiltration and applied to a 1.5 ¥ 170 cm Sephacryl S-200
HR column (column volume ~300 mL; Amersham Biosciences
Inc., Piscataway, NJ, USA), pre-equilibrated with 50 mM
HEPES pH 7.4. hiCE (CES2) was eluted from the column using
a flow rate of 6 mL·h- and was estimated to be greater than
(Beckmann, Brea, CA, USA) by comparison with known
amounts of standard compounds. The sensitivity of detection
for this system was 2, 2 and 1 ng/mL for cocaine, benzoylecgo-
nine and benzoic acid, respectively. All reactions were cor-
rected for non-enzymatic hydrolysis by subtraction of the
levels of products when cocaine was incubated with buffer
alone.
1
98% pure by sodium dodecyl sulphate-polyacrylamide gel
electrophoresis (SDS-PAGE).
Analysis of heroin hydrolysis
Enzymatic hydrolysis of heroin to 6-monoacetylmorphine
was examined under the following conditions. Drug was
added to a pre-warmed (37°C) mixture containing purified
enzyme and 50 mM HEPES buffer pH 7.4. Typically, 0–5 mM
heroin was incubated for 2 min with hCE1 (CES1) (25 U/
reaction) or hiCE (CES2) (10 U/reaction), at 37°C. All reac-
tions were terminated by addition of equal volume of ice-
cold acidified methanol (containing 0.5% 1 M HCl), and
proteins were pelleted by centrifugation for 5 min at 14 500¥
g at 4°C.
Chromatography of human intestinal microsomal extracts. Ex-
tracts, typically containing ~10 mg of total protein (equiva-
lent to ~10 000 U of CE activity), were applied to a 1 ¥ 170 cm
Sephacryl S-200 HR column equilibrated in 150 mM NaCl,
50 mM HEPES pH7.4, 0.5% Triton X-100 and samples were
eluted using the same buffer. Routinely 3 ml fractions were
collected, concentrated and then assayed for the desired
enzyme activity. Western analysis for hCE1 (CES1) and hiCE
(CES2) was undertaken using specific antibodies for these
proteins using chemiluminescent detection.
The supernatants from the above reactions were then analy-
sed by reverse phase HPLC using a similar system to that for
cocaine. However, heroin and its metabolites were separated
and quantitated using a gradient profile as follows: 100%
buffer A for 1.5 min changing to 100% buffer B over 0.5 min
for 5.5 min [where buffer A is 5% (v/v) acetonitrile in 50 mM
Data analysis and statistical procedures
Enzyme inhibition assays. CE enzyme inhibition was deter-
mined using the previously described, selective CE inhibitors
(
Wadkins et al., 2004; Wadkins et al., 2005). Compounds
analysed were benzil (1), 1,4-bisbenzil (2), 1-(3,4-dimethyl-
phenyl)-2-phenyl-ethane-1,2-dione (3), N,N’-(2,3,5,6-
tetramethyl-1,4-phenylene)bis(4-fluorobenzene sulfonamide)
NH
4
H
2
PO
4
, 150 mM TBAP, pH 2.2 and buffer B is 25% (v/v)
PO , 150 mM TBAP, pH 2.2].
acetonitrile in 50 mM NH
4
H
2
4
Using a flow rate of 1 ml/min, heroin, 6-acetylmorphine and
morphine eluted at 7.0, 5.6 and 4.1 min, respectively, and
were detected at 235 nm. Concentrations were determined in
an identical fashion to that described for cocaine earlier, and
(
(
4) or N,N’-1,4-phenylenebis(3-bromobenzene sulfonamide)
5) (see the chemical structures of these compounds in
Table 4) using cocaine, heroin, o-NPA or CPT-11 as substrates.
Briefly, the formation of product in the presence of inhibitor
at concentrations ranging from 1 nM to 100 mM was deter-
mined (Wadkins et al., 2004; Wadkins et al., 2005). Routinely
at least eight concentrations of inhibitor were used. To deter-
-
1
-1
the sensitivity of this system was 3 ng·mL , 2 ng·mL and
4
-
1
ng·mL for heroin, 6-acetylmorphine and morphine, respec-
tively. All reactions were corrected for non-enzymatic
hydrolysis by subtraction of the levels of products when
cocaine was incubated with buffer alone.
mine the K values (inhibition constants), data were fitted to
i
the following equation (Webb, 1963):
Prediction of logP values
[
I]{[S](1−β) + K
s
(α − β)}
i =
Calculated logP (clogP) values were obtained using Chem-
Silico Predict v2.0 software (ChemSilico LLC, Tewksbury, MA,
USA).
[
I]{[S]+ αK}+ K{α[S]+ αK }
s
i
s
Experimental design
Expression and purification of recombinant hiCE (CES2). The
purification of hiCE (CES2) was achieved by overexpression of
a secreted form of the protein from Spodoptera frugiperda Sf21
cells using baculoviral expression vectors (Morton and Potter,
where i = fractional inhibition, [I] = inhibitor concentration,
[S] = substrate concentration, a = change in affinity of the
substrate for the enzyme in the presence of the inhibitor
(where a > 0), b = change in the rate of enzyme substrate
complex decomposition in the presence of the inhibitor
(where 1 > b > 0), Ks is the dissociation constant for the
enzyme substrate complex (assuming negligible commitment
2
000). All purification steps were carried out at 4°C. Media
harvested from infected Sf21 cells was buffer exchanged with
0 mM HEPES pH 7.4 buffer using ultrafiltration with Ultracel
5
PL-30 membranes (Millipore, Bedford, MA) and the sample
was then applied to a 2.5 ¥ 6 cm macro-prep diethylamino-
ethyl (DEAE) column equilibrated in the same buffer. Follow-
ing washing with 1 column volume (~30 mL) of 50 mM
HEPES pH 7.4, the sample was then washed with 5 column
to catalysis) and K is the inhibitor constant. The data sets
i
were analysed using Prism 4.0 software (GraphPad, La Jolla,
CA, USA) and Perl Data Language, and the mode of enzyme
British Journal of Pharmacology (2010) 160 1916–1928

Carboxylesterase hydrolysis of cocaine and heroin
MJ Hatfield et al
1919
2
inhibition was determined by evaluating the r values for the
hCE1 (CES1) – M73499 (Munger et al., 1991). Rabbit poly-
clonal antibodies directed towards hiCE (CES2) and hCE1
(CES1) were generated by repeated immunization of rabbits
with recombinant proteins. Antibodies were purified from
sera using IgA affinity chromatography and sensitivity and
specificity were verified by enzyme-linked immunosorbent
assay and Western analyses. Human liver and intestinal
microsomal extracts were purchased from BD Biosciences
curve fits using Akaike’s information criteria (Wadkins et al.,
005). K values were then calculated using the best fit model
2
i
described from these analyses.
In most cases, enzyme inhibition was partially competitive
i.e. where the inhibitor does not affect rate of enzyme/
substrate complex dissociation and only partially hinders sub-
strate binding).
(
(Woburn, MA, USA) and Celsis In Vitro (Baltimore, MD,
Molecular modelling studies. Molecular modelling analyses
were performed using ICM Pro software v3.6 (Molsoft LLC,
La Jolla, CA, USA) running on a dual Xeon computer in a
Windows XP environment. Recent studies indicate that
docking experiments using ICM Pro compare favorably with
FlexX, GOLD and GLIDE (Chen et al., 2006). For hCE1 (CES1),
the coordinates derived from the CE/homatropine crystal
structure were used (1MX5; Bencharit et al., 2003b). For hiCE
USA), respectively.
Results
Expression and purification of recombinant hiCE (CES2)
Media was obtained from S. frugiperda Sf21 cells that
had been infected with baculovirus expressing a secreted
form of the hiCE (CES2) protein (Potter et al., 1998; Khanna
et al., 2000). Initial attempts at purification using isoelectric
focusing (Morton and Potter, 2000) proved unsuccessful and
ion-exchange chromatography was chosen as the preferred
method. A step-wise wash and elution was used to allow
the separation of hiCE (CES2) from two problematic con-
taminants, bovine serum albumin (present in minor
amounts in the culture media) and baculoviral chitinase,
both of which have molecular masses similar to hiCE (CES2)
(CES2), a homology model was developed using the 1MX5
coordinates With Modeller 8 (Sali and Blundell, 1993)
software.
For these analyses, substrate and inhibitor structures were
constructed in ChemDraw (CambridgeSoft, Cambridge, MA,
USA) and subjected to energy minimization using ICM-Pro.
The receptor site, consisting of the residues that encompassed
the enzyme active site gorge, was then selected as a pocket
into which the minimized ligand could be docked. For
cocaine and heroin, binding of both the protonated and the
native form of the molecule was assessed. Docking studies
were performed using essentially the default parameters for
(~65 kDa). The anion-exchange column allowed for
a
15-fold purification followed by an additional 2-fold purifi-
cation by gel filtration, based on yields of CE activity
(Table 1). Typically, we obtained 1 mg of hiCE (CES2) per
liter of insect cell culture media. Multiple batches of hiCE
(CES2) have been made by this protocol with final specific
the ICM Pro software.
A rigid receptor/flexible ligand
approach was adopted that used five potential energy maps
combining hydrophobicity, electrostatics, hydrogen bond for-
mation and two van der Waals parameters. However, minor
modifications included constructing receptor electrostatic
charge maps at a resolution of 0.1 Å and increasing the ‘Thor-
oughness’ parameter to 10. The latter increased the time and
number of poses that were evaluated during the docking
simulations.
-
1
activities ranging from 40 to 55 000 U·mg , with a typical
average of 50 000 U·mg ; however, selected fractions typi-
-
1
-
1
cally showed a specific activity as high as 60 000 U·mg .
Purity of the samples was assessed by SDS-PAGE analysis
(Figure 1) and protein identification by Matrix-assisted laser
desorption/ionization-time of flight time of flight mass
spectrometry.
Materials
Cocaine ((–)-cocaine) and heroin were purchased from Ceril-
liant Corporation (Round Rock, TX) and were dissolved in
acetonitrile. Routinely, solvent concentrations in biochemical
reactions did not exceed 2%, although no effect was seen on
CE activity at 5% acetonitrile. CPT-11 was kindly provided by
Dr J. P. McGovren (Pharmacia, Peapack, NJ, USA) and was
®
dissolved in methanol. The Macro-Prep DEAE support was
from Bio-Rad and Sephacryl S-200 high resolution resin
was obtained from Amersham Biosciences Inc. All other
chemicals were purchased from Fisher (Fair Lawn, NJ, USA),
Organic Research (Cincinnati, OH, USA), Sigma-Aldrich (St.
Louis, MO, USA) or VWR Scientific Products, Inc. (Suwanee,
GA, USA).
hCE1 (CES1) was purified from the media of baculovirus-
infected Spodoptera frugiperda Sf21 cells as previously
described (Bencharit et al., 2003a; Bencharit et al., 2003b;
Morton and Potter, 2000). The GenBank accession numbers
for the cDNAs encoding the CEs used in these studies were
as follows; hiCE (CES2) - Y09616 (Schwer et al., 1997); and
Figure 1 Sodium dodecyl sulphate-polyacrylamide gel electro-
phoresis analysis of different fractions following purification of hiCE
(
CES2). Approximately 3 mg of protein was loaded into each lane and
following electrophoresis, the gel was stained with Coomassie Blue.
Lane 1, Crude culture media; Lane 2, After ultrafiltration; Lane 3, After
ion-exchange chromatography using DEAE; Lane 4, After ultrafiltra-
tion; Lane 5, After chromatography on Sephacryl S200 HR.
British Journal of Pharmacology (2010) 160 1916–1928

Carboxylesterase hydrolysis of cocaine and heroin
1
920
MJ Hatfield et al
Table 2 A comparison of kinetic parameters of hiCE (CES2) and hCE1 (CES1) with various nitrophenyl esters
Substrate
clogP
Enzyme
K
m
Ϯ s.e.m.
mM)
V
max Ϯ s.e.m.
k
cat/K
m
-1
m
Ratio of kcat/K [hiCE
(CES2)/ hCE1 CES1]
-1
-1
(
(nmol min )
(s mM )
o-Nitrophenyl acetate
1.12
1.16
1.86
2.46
2.75
3.01
hiCE (CES2)
hCE1 (CES1)
79 Ϯ 7.6
73 Ϯ 7.4
0.94 Ϯ 0.02
0.15 Ϯ 0.00
67 Ϯ 6
7.9 Ϯ 0.7
8.5
6.3
p-Nitrophenyl acetate
hiCE (CES2)
hCE1 (CES1)
976 Ϯ 60
822 Ϯ 73
4.17 Ϯ 0.14
0.81 Ϯ 0.02
24 Ϯ 1
3.8 Ϯ 0.3
p-Nitrophenyl propionate
p-Nitrophenyl butyrate
p-Nitrophenyl trimethylacetate
p-Nitrophenyl valerate
hiCE (CES2)
hCE1 (CES1)
201 Ϯ 18
249 Ϯ 9
9.3 Ϯ 0.6
2.05 Ϯ 0.03
261 Ϯ 7
32 Ϯ 0.6
8.2
hiCE (CES2)
hCE1 (CES1)
117 Ϯ 5
169 Ϯ 8
12 Ϯ 0.4
1.65 Ϯ 0.04
580 Ϯ 5
37 Ϯ 1
15.7
14.7
8.1
hiCE (CES2)
hCE1 (CES1)
19.8 Ϯ 1.1
29 Ϯ 5
0.77 Ϯ 0.02
0.8 Ϯ 0.1
220 Ϯ 9
15 Ϯ 0.6
hiCE (CES2)
hCE1 (CES1)
71 Ϯ 11
27 Ϯ 2.7
10.5 Ϯ 1.2
0.73 Ϯ 0.04
421 Ϯ 27
52 Ϯ 2.5
Assessment of substrate hydrolysis was determined spectrophotometrically by monitoring the formation of nitrophenol at 420 nm. Kinetic parameters were then
calculated from these data sets.
s.e.m., standard error of the mean; CE, carboxylesterase.
To determine the oligomerization state of hiCE (CES2), we
calibrated a Sephacryl S200 column using low molecular mass
protein standards and applied a sample of the purified CE.
These studies indicated that hiCE (CES2) had
a pre-
dicted molecular mass of ~66 kDa. Furthermore, in co-
chromatographic studies, it was observed that hiCE (CES2)
and bovine serum albumin (66 kDa) eluted together off this
Sephacryl column. As the hiCE (CES2) cDNA contains an
open reading frame that would give rise to a protein of pre-
dicted molecular mass of 61 807 Da (Schwer et al., 1997), we
conclude that under the conditions employed here, the hiCE
(CES2) protein purifies as a monomer.
Kinetic parameters of recombinant hiCE (CES2) with nitrophenyl
ester substrates
Having isolated recombinant hiCE (CES2) and having
homogenous hCE1 (CES1) at hand (Bencharit et al., 2003b;
Morton and Potter, 2000), we evaluated the kinetic param-
eters of these proteins with a variety of different esterase
Figure 2 A graph demonstrating the linear relationships between
the clogP values of the para-substituted nitrophenyl esters, and their
observed K
m
constants for hiCE (CES2) (
), and hCE1 (CES1) (
).
2
The correlation coefficients (r ) for the data sets are indicated on the
graph.
substrates. The K
m
, Vmax and kcat/K
m
were therefore deter-
mined for various nitrophenyl esters using hiCE (CES2), and
were compared with those for hCE1 (CES1) (Table 2). As can
be seen, the K
m
values demonstrated considerable variation
hiCE (CES2) was ~16-fold more efficient at p-nitrophenyl
butyrate hydrolysis than hCE1 (CES1). As indicated in
Table 2, for all substrates analysed, hiCE (CES2) was more
efficient at substrate turnover as compared with hCE1
(CES1).
As the active sites of the CEs are lined with aromatic amino
acids (Bencharit et al., 2003b; Potter and Wadkins, 2006), we
hypothesized that more hydrophobic molecules would pre-
ferentially localize within these domains. Therefore, we com-
with no obvious correlation with either enzyme. However,
the
1
V
max values for hiCE (CES2) ranged from 0.77 to
2.0 nmol·min- for the different substrates, and these
1
values were up to 6-fold higher than that observed for
hCE1 (CES1). Peak velocities were seen with p-nitrophenyl
butyrate and p-nitrophenyl propionate for hiCE (CES2) and
hCE1 (CES1), respectively. However, maximal catalytic effi-
ciency for hCE1 (CES1) was demonstrated with the valerate
ester. This is a consequence of the much lower Km constants
that are observed with this substrate, resulting in higher
pared the affinity constants (K ) for the para- series of
m
substituted nitrophenyl esters for the two CEs, with their
clogP values (calculated water/octanol partition coefficients).
As shown in Figure 2, excellent correlations were seen with
k
cat/K
m
values.
It was also noted that with the larger, more bulky sub-
2
strates, such as p-nitrophenyl butyrate or p-nitrophenyl tri-
methylacetate, hiCE (CES2) was much more proficient at
substrate hydrolysis. This is exemplified by the significantly
data obtained from both enzymes, with r for the curve fits of
0.79 and 0.89 for hiCE (CES2) and hCE1 (CES1), respectively.
These results confirm that the affinity of the compounds for
the CEs is directly related to their clogP (or hydrophobicity).
larger kcat/K
m
values for these esters (Table 2). For example,
British Journal of Pharmacology (2010) 160 1916–1928

Carboxylesterase hydrolysis of cocaine and heroin
MJ Hatfield et al
1921
Table 3 Kinetic parameters for the metabolism of heroin, cocaine, and CPT-11 for hiCE (CES2) and hCE1 (CES1)
-1
(s-1 mM-1)
Substrate
Enzyme
K
m
Ϯ s.e.m. (mM)
V
max Ϯ s.e.m. (pmol min )
kcat/K
m
Ratio of kcat/K
m
hiCE (CES2)/
hCE1 (CES1)
Heroin
Cocaine
CPT-11
hiCE (CES2)
hCE1 (CES1)
hiCE (CES2)
hCE1 (CES1)
hiCE (CES2)
hCE1 (CES1)
1440 Ϯ 630
445 Ϯ 73
202 Ϯ 45
ND
3.9 Ϯ 0.3
8130 Ϯ 910
10 660 Ϯ 450
589 Ϯ 33
ND
1.6 Ϯ 0.12
0.36 Ϯ 0.02
5.9
9.7
0.44
-
0.41
0.0044
0.65
–
a
a
93
82.8 Ϯ 9.6
The hydrolysis of the respective substrates was determined using HPLC separation and detection of the products, from which the kinetic parameters were
calculated.
a
ND, not determined. No hydrolysis was apparent for this substrate.
s.e.m., standard error of the mean; CE, carboxylesterase.
This is consistent with the notion that the longer alkyl chain
esters localize within the hydrophobic active site gorges
present within the protein.
be noted however, that the kcat/K values for heroin catalysis
m
[5.9 and 9.7 s- mM for hiCE (CES2) and hCE1 (CES1),
1
-1
respectively] were at least 10-fold higher than that observed
for cocaine with hiCE (CES2) (0.44 s- mM ). Indeed, based
upon these kinetic analyses, the catalytic efficiency of heroin
hydrolysis by CEs was ~12- to 40-fold more efficient than
cocaine metabolism by the latter enzyme. Again, no inhibi-
tion of hiCE (CES2) enzyme activity was observed in the
presence of high concentrations of heroin (Figure 3C).
1
-1
Metabolism of cocaine by CEs
Having demonstrated that hiCE (CES2) could hydrolyse large,
bulky esters, we evaluated the ability of this enzyme to
metabolize cocaine and heroin. As before, we also determined
the ability of recombinant hCE1 (CES1) to hydrolyse these
compounds. As indicated in Table 3 and displayed in Figure 3,
hiCE (CES2) readily hydrolysed cocaine to yield benzoic acid
Inhibition of hiCE (CES2) and hCE1 (CES1) with
CE-specific inhibitors
(
2
and ecgonine methyl ester) with K
02 mM and 589 pmol·min , respectively. In contrast, no
m
and Vmax values of
As the inhibition of CE-mediated hydrolysis of drugs may be
used to modulate the levels of toxic metabolites (such as
morphine from heroin), we determined the ability of a series
of recently identified CE inhibitors to alter heroin, cocaine or
CPT-11 hydrolysis. The benzil based compounds (1-3) are
generic CE inhibitors, whereas the bis-benzene sulfonamides,
-
1
hydrolysis of the drug to either benzoylecgonine, or benzoic
acid, was observed with hCE1 (CES1) (Figure 3A). Indeed, no
enzyme-mediated metabolism of cocaine was seen, even after
prolonged incubation (~20 h) with hCE1 (CES1) (data not
shown). Under these same conditions using hiCE (CES2), all
of the cocaine was hydrolysed to benzoic acid (data not
shown). We conclude therefore, that hiCE (CES2) is the
primary CE responsible for cocaine metabolism in vivo.
Analysis of the curve fits for benzoic acid formation versus
substrate concentration (Figure 3B), indicated that no inhibi-
tion of hiCE (CES2) activity was observed with high concen-
trations of cocaine (up to 5 mM). This is in contrast to that
seen for BChE where cocaine concentrations greater than
4
and 5, demonstrate specificity for hiCE (CES2). As indicated
in Table 4, inhibitors 1-3 displayed very low K values towards
i
hiCE (CES2) and hCE1 (CES1) for all substrates. With com-
pounds 4 and 5, which are hiCE (CES2)-specific inhibitors
(Wadkins et al., 2004), no inhibition of metabolism of o-NPA
or heroin was observed with hCE1 (CES1). However, potent
inhibition of hiCE (CES2)-mediated metabolism was seen for
all substrates, with K values as low as 35 nM with heroin.
i
These studies confirm the applicability of using inhibitors to
modulate these reactions and indicate that the selectivity of
the sulfonamides towards hiCE (CES2) may have utility in
modulating either CPT-11 or heroin hydrolysis in vivo.
3
1
00 mM resulted in reduced product formation (Stewart et al.,
977). This suggests that in situations where individuals over-
dose with very high levels of drug, hiCE (CES2) may play a
significant role in detoxification of cocaine.
The metabolism of the anticancer agent CPT-11 was
included in these studies since previous reports have indi-
cated that this substrate is hydrolysed much more effectively
by hiCE (CES2) than by hCE1 (CES1) (Humerickhouse et al.,
Metabolism of heroin and cocaine by human intestinal and
liver extracts
To confirm the results observed with the recombinant
protein, we evaluated the ability of human intestinal and liver
microsomal extracts to hydrolyse the drugs. We chose these
tissues because they are rich in hiCE (CES2) and hCE1 (CES1),
respectively, and demonstrate significantly different ratios of
the two human CEs. This is exemplified by the Western analy-
sis of total microsomal extracts shown in Figure 4A and the
hydrolytic activities of these samples (Figure 4B). As can be
seen, the intestinal microsomes contained significantly more
hiCE (CES2) than hCE1 (CES1); the converse being true for
the liver sample.
2
000). Our studies are in agreement with these results
(Table 3), confirming the specificity and hydrolytic activity of
the recombinant purified hiCE (CES2) protein.
Metabolism of heroin by CEs
In contrast to the results obtained when using cocaine as a
substrate, both human CEs did hydrolyse heroin. The effi-
ciencies of hydrolysis of this drug to 6-acetylmorphine were
similar, as exemplified by the kcat/K values (Table 3). It should
m
British Journal of Pharmacology (2010) 160 1916–1928

Carboxylesterase hydrolysis of cocaine and heroin
1
922
MJ Hatfield et al
Figure 3 (A) Metabolism of cocaine by different carboxylesterases (CEs). Cocaine (1 mM) was incubated with the different samples for 1 h.
The yields of benzoylecgonine and benzoic acid were determined by HPLC. Essentially, only hiCE (CES2) was able to hydrolyse cocaine. Control
–
50 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer pH 7.4 only; Sf21 (m) – media harvested from uninfected Sf21
cells; hCE1 (m) – media harvested from Sf21 cells infected with baculovirus expressing hCE1 (CES1) (~250 units of CE); hiCE (m) – media
harvested from Sf21 cells infected with baculovirus expressing hiCE (CES2) (~75 units of CE); hCE1 (e) – reaction containing pure hCE1 (CES1)
(
20 mg = ~2000 units of CE); hiCE (e) – reaction containing pure hiCE (CES2) (20 mg = ~500 units of CE). (B) The concentration versus velocity
curve for hiCE (CES2) with cocaine. No evidence of enzyme inhibition by high concentrations of substrate was observed in these assays. (C)
The concentration versus velocity curve for hiCE (CES2) with heroin. Similar to that seen for cocaine, no substrate-mediated enzyme inhibition
was observed.
Chromatography of extracts was performed on Sephacryl
S-200 HR and the CE activity of individual fractions was
determined. In addition, the presence of hiCE (CES2) or hCE1
between fractions 8–11 is not known, the data presented in
Figure 4 indicate that this protein did not contribute to
cocaine hydrolysis.
(
CES1) in each fraction was assessed by Western analysis and
As a comparison, we performed similar studies using a
human liver microsomal extracts that demonstrated high
expression of hCE1 (CES1) and low levels of hiCE (CES2) (see
Figure 4A). Little or no cocaine hydrolysis was observed with
either the initial extract (Table in Figure 4B) or from any of
the fractions following chromatography (Figure 4C, bottom
graph, red line), even though large amounts of hCE1 (CES1)
were present in these samples (Figure 4D). Although hiCE
(CES2) was present in the liver sample, the actual level of this
enzyme was very low (see Figure 4A) and the cocaine hydroly-
sis products were probably below the limit of detection for
these assays.
hydrolysis of cocaine and heroin was monitored by HPLC.
Figure 4C, D demonstrate the activity profile of the fractions
obtained from this chromatographic technique.
As can be seen with the intestinal microsomal sample,
two peaks of esterase activity were observed (fractions 8–11
and 14–19) with the latter correlating with the presence of
hiCE (CES2) (see the Western analyses under the graph –
Figure 4D). Interestingly, the earlier peak did not correspond
to hCE1 (CES1) or BChE (data not shown), suggesting that
this represents another unknown esterase in this sample.
Assessment of cocaine and heroin metabolism indicated
that although the latter was effectively converted to
Overall, the results from these studies are consistent with
the in vitro experiments using recombinant protein, that dem-
onstrate that hiCE (CES2) is responsible for conversion of
cocaine to benzoic acid and ecgonine methyl ester.
6
-acetylmorphine by both fractions, only hiCE (CES2) was
capable of hydrolyzing the former drug (Figure 4C, bottom
graph). Although the identity of this enzyme that eluted
British Journal of Pharmacology (2010) 160 1916–1928

Carboxylesterase hydrolysis of cocaine and heroin
MJ Hatfield et al
1923
British Journal of Pharmacology (2010) 160 1916–1928

Carboxylesterase hydrolysis of cocaine and heroin
1
924
MJ Hatfield et al
Figure 4 Chromatographic separation of carboxylesterases (CEs) in human intestinal and liver microsomal extracts using Sephacryl S-200 HR
resin. (A) Western analysis demonstrating the amounts of hiCE (CES2) and hCE1 (CES1) present in the samples prior to chromatography.
Standards present on the right hand side of the image represent 50 ng of each pure protein. (B) Table demonstrating the levels of CE activity,
as well as the cocaine and heroin hydrolysis of the intestine and liver microsomal extracts prior to chromatography. (C) Elution profile for
intestinal (blue line) and liver (red line) microsomal extracts demonstrating levels of CE activity (top graph), 6-acetylmorphine (6-AM)
formation (heroin hydrolysis; middle graph) and benzoic acid (BA) production (cocaine hydrolysis; bottom graph). (D) Western analysis of
these same fractions analysed in panel C, using antibodies specific for hiCE (CES2) or hCE1 (CES1). The images are aligned such that the signals
correspond to the fraction number indicated on the abscissa axis.
Molecular analysis of cocaine and heroin binding to CEs
and Glu) were essentially in the same position and orienta-
tion in both proteins (Figure 5B). However, as can be seen in
Figure 5A, one loop of amino acids (residues 303-320 in hCE1)
was displaced as compared with hiCE (CES2). It is unclear
whether movement of this domain affects either the biologi-
cal activity and/or docking studies, but due its proximity to
the active site, it is likely to play a role in substrate hydrolysis.
Following docking of different molecules the distances
between the catalytic serine residue and the carbonyl carbon
atoms were measured. As can be seen in Table 5 and Figure 5,
the carbonyl carbon atom of the benzyl ester in cocaine
localized 2.8 Å from the serine Og atom for hiCE (CES2)
Having demonstrated that hiCE (CES2) was markedly more
efficient at cocaine hydrolysis than hCE1 (CES1), we sought to
determine whether molecular modelling studies could
provide insight into the alignment of the substrates within
the enzyme active sites. Therefore, we docked the drug (and a
variety of other compounds) into the active sites of both
enzymes. As the crystal structure for hiCE (CES2) is currently
not available, we generated a homology model using the
coordinates for hCE1 (CES1) when bound to homatropine.
Interestingly, when the hCE1 (CES1) and derived hiCE
(CES2) structures were overlaid, the catalytic residues (Ser, His
British Journal of Pharmacology (2010) 160 1916–1928

Carboxylesterase hydrolysis of cocaine and heroin
MJ Hatfield et al
1925
Figure 5 Computer-assisted docking of cocaine or heroin into the active sites of hCE1 (CES1) or hiCE (CES2). (A) An overlay of the ribbon
representations of the hCE1 (CES1) (taupe) and hiCE (CES2) (orange/brown) structures used for the docking studies. As can be seen one loop
of hCE1 (CES1) (highlighted in pink) is displaced (marked by the white arrow) and projects towards the active site catalytic amino acids (Ser,
His, Glu). (B) An overlay of the catalytic amino acids in hCE1 (CES1) (taupe) and hiCE (CES2) (orange/brown) with the distances between the
relevant atoms indicated. (C, D) – Docking of cocaine into the active sites of hiCE (CES2) and hCE1 (CES1), respectively. The active site serine
is depicted and the distance from the Og atom to the carbonyl carbon atoms in the drug are displayed. (E, F) – Docking of heroin into the active
sites of hiCE (CES2) and hCE1 (CES1), respectively. The active site serine is depicted and the distance from the Og atom to the carbonyl carbon
atoms in the drug are displayed. In all panels, the nitrogen and oxygen atoms displayed in blue and red, respectively, and distances are marked
in Å.
(
(
Figure 5C), but was localized greater than 7 Å away in hCE1
CES1) (Figure 5D). As the latter distance is too large for
value for heroin with hCE1 (CES1) is ~threefold greater than
that seen for hiCE (CES2) Table 3).
nucleophilic attack, this suggests that hCE1 (CES1) would be
unable to hydrolyse this ester function. This was true for both
the protonated and un-protonated forms of the drug.
However, following docking of heroin, the carbonyl carbon
atoms for the methyl ester in the 3-position localized within
As controls for the docking studies, we also assessed the
binding of a series of related substrates and inhibitors. As
noted in Table 5, the carbonyl carbon atom was present
within 3.7 Å of the serine Og atom for o-NPA, p-nitrophenyl
valerate, CPT-11 and benzil, for both hCE1 (CES1)and hiCE
(CES2). Finally as a negative control, we docked cam-
phorquinone into the enzyme active sites. This molecule con-
tains the ethane-1,2-dione chemotype (similar to benzil), but
is not an inhibitor of either of the human CEs (Hyatt et al.,
3
.9 Å for both CEs (Figure 5E and F).
Indeed, for hCE1 (CES1) the distance to the ester was
smaller than that for hiCE (CES2). This is in agreement with
our experimental results that demonstrate that the kcat/K
m
British Journal of Pharmacology (2010) 160 1916–1928

Carboxylesterase hydrolysis of cocaine and heroin
1
926
MJ Hatfield et al
Table 5 Distance between the catalytic serine Og atom and the
carbonyl carbon atoms for a series carboxylesterase (CE) substrates
and inhibitors
performed using enzymes purified from extracts derived from
human liver. As hCE1 (CES1) and hiCE (CES2) (as well as
BChE and other esterases) are expressed in this organ, poten-
tially, contamination of ‘pure’ CEs with other liver esterases
could confuse the results obtained from these biochemical
experiments. We are unable to assess the level of purity of the
enzymes used in these previous studies and hence we cannot
comment on which protein would be definitively responsible
for cocaine hydrolysis; however, it should be noted that no
specific assays or techniques were used by these authors to
assess the levels of other esterases present in their enzyme
preparations.
We relied on the use of recombinant hCE1 (CES1) or hiCE
(CES2), purified from the culture media of the insect cell line,
S. frugiperda. We have used this approach, rather than purify-
ing the proteins from cell extracts, to avoid any possible cross
contamination with intracellular insect CEs, and to minimize
proteolytic degradation of the mammalian enzymes. As the
only protein contaminants we have observed during the hiCE
Compound
Distance from Ser OgH to C = O
Å)
(
hCE1 (CES1)
hiCE (CES2)
a
Cocaine
4.9 (Bz)
2.6 (Bz)
5.1 (Me)
2.8 (Bz)
3.7 (Me)
3.9 (3-)
6.2 (6-)
3.6 (3-)
5.2 (6-)
3.7
2.8
3.1
3.0, 3.3
6.7 (2-)
6.8 (3-)
3
.9 (Me)
7.1 (Bz)
.1 (Me)
3.2 (3-)
.9 (6-)
2.8 (3-)
.7 (6-)
a
Cocaine (protonated)
3
b
Heroin
3
b
Heroin (protonated)
5
o-NPA
3.3
3.7
NA
o-Nitrophenyl valerate
CPT-11
c
d
Benzil
3.9, 4.8
5.8 (2-)
e
Camphorquinone
5
.0 (3-)
(
CES2) purification have been bovine serum albumin and
The distance between the catalytic serine Og atom and the carbonyl carbon
atoms for the listed compounds were determined following docking into the
active sites of either hCE1 (CES1) or hiCE (CES2). Values were calculated using
ICM Pro software.
Autographa californica chitinase, and neither of these proteins
has esterase activity, it is likely that the enzymes used in our
studies were homogenous. We did not observe hydrolysis of
cocaine by the recombinant hCE1 (CES1) that we used in
these studies, and as this protein is catalytically active towards
a variety of substrates, we cannot confirm the results gener-
ated in previous reports that indicate that the drug is metabo-
lized by this enzyme (Kamendulis et al., 1996; Brzezinski et al.,
a
Distances are indicated to the carbonyl carbon atom of either the benzyl ester
(
Bz) or the methyl ester (Me).
Distances are indicated to the carbonyl carbon atom of the methyl ester at
b
either the 3- or 6- position of the molecule.
c
Docking of CPT-11 into the active site of hCE1 (CES1) was not undertaken as
steric constraints present at the entrance to the active site gorge significantly
impede access of the drug to the catalytic amino acids (Wadkins et al., 2001;
Wierdl et al., 2008).
1
997).
d
Potentially, there may be different isoforms of hCE1
Two distance values are given for benzil as it contains two carbonyl groups,
however, because this molecule is symmetrical, no assignment of these atoms
is included.
Distances are indicated to the carbonyl carbon atom present at the 2- or
(CES1) that result from nucleotide polymorphisms (Zhu
et al., 2008) that could account for the discrepancy between
our results and those of Brzezinski et al. (1997). However, as
we have sequenced the cDNA that encodes the hCE1 (CES1)
protein used in these studies (Danks et al., 1999), deter-
mined that this enzyme is pure by both physical and bio-
chemical methods (Bencharit et al., 2003b; Morton and
Potter, 2000; Wadkins et al., 2001), and obtained the crystal
structure of this protein in complex with a variety of differ-
ent small molecules (Bencharit et al., 2003b; Fleming et al.,
e
3
-position of the molecule.
2
007). As indicated, the measured distances [~6.8 Å for hiCE
(CES2) and 5.0–5.8 Å for hCE1 CES1)] would be too large for
nucleophilic attack by the Og atom towards this molecule. In
summary, the results of these docking studies are in complete
agreement with the obtained experimental data.
2005; Bencharit et al., 2006), our results clearly indicate that
the protein encoded by this cDNA cannot hydrolyse
cocaine. Additionally, our studies implicate hiCE (CES2) as
the principal CE involved in drug hydrolysis. As hiCE
(CES2) is also expressed at relatively high levels in liver, skel-
etal muscle, heart, colon and kidney (Wu et al., 2003), it is
likely that this enzyme will be important in the hydrolysis
of this molecule, regardless of the route of cocaine
administration.
To validate the in vitro studies, we evaluated the esterase
levels and drug hydrolyzing properties of intestinal and liver
microsomal extracts after chromatographic separation. Con-
sistent with our previous results, we determined that hiCE
(CES2) was responsible for cocaine metabolism and that hCE1
(CES1) was inactive towards this substrate (Figure 4). This was
achieved using samples that were either, rich in hiCE (CES2)
and low in hCE1 (CES1) (intestine), or where the converse was
true (liver). Confirmation of the identity of the enzymes
involved was documented using antibodies specific for these
proteins. Hence, our results using these human microsomal
Discussion and conclusions
We have purified recombinant human intestinal CE, hiCE
(CES2), to homogeneity and determined the kinetic param-
eters for this enzyme with a variety of clinically relevant
substrates. This has demonstrated that hiCE (CES2) efficiently
metabolized CPT-11, cocaine and heroin. In contrast, the
human liver CE, hCE1 (CES1), only hydrolysed heroin. Until
recently, the enzymes involved in the metabolism of all of
these agents have not been unambiguously identified. For
example, although hiCE (CES2) has been proposed as the
principal CE responsible for CPT-11 activation in vivo (Hum-
erickhouse et al., 2000; Khanna et al., 2000), other esterases,
for example, BChE can also activate this drug (Dodds and
Rivory, 1999; Morton et al., 1999). Similarly, previous reports
have indicated that cocaine can be metabolized by both hCE1
(CES1)and hiCE (CES2) (Kamendulis et al., 1996; Brzezinski
et al., 1997; Pindel et al., 1997). However, these studies were
British Journal of Pharmacology (2010) 160 1916–1928

Carboxylesterase hydrolysis of cocaine and heroin
MJ Hatfield et al
1927
samples corroborate our previous studies using the recombi-
nant proteins.
As hiCE (CES2) has been identified as being proficient in the
hydrolysis of heroin and the activation of CPT-11, potentially,
selective CE inhibitors that prevent these hydrolytic reactions
may prove useful in ameliorating the toxicity observed with
these drugs. Therefore, we assessed the ability of a panel of
benzil-based compounds and bis benzene-sulfonamides to
modulate drug hydrolysis. As we have previously reported
that these types of inhibitors act in a partially competitive
fashion (Wadkins et al., 2004; Wadkins et al., 2005), the
potency of enzyme inhibition will be dependent upon the
substrate used for the analyses. Our results indicate however,
that the phenylethane-1,2-dione based compounds that were
chosen for analysis were potent inhibitors of all mammalian
CEs, with all of the substrates tested (o-NPA, cocaine, heroin
and CPT-11), whereas the sulfonamides only demonstrated
In an attempt to understand the differences in cocaine
hydrolysis by these two CEs, we undertook molecular docking
studies using the hCE1 (CES1) crystal structure and a homol-
ogy model for hiCE (CES2) based upon these same coordi-
nates. Initially, we were concerned that as the former enzyme
cannot hydrolyse cocaine, the use of the hCE1 (CES1) coor-
dinates might result in a conformation in hiCE (CES2) that
would preclude informative results from these analyses.
However, upon docking of this drug (and others small mol-
ecules displayed in Table 5), it was apparent that the homol-
ogy model demonstrated sufficient variation in the active site
gorge, such that different conformations of the CE/drug com-
plexes were observed. Gratifyingly, results from these docking
studies placed the active site serine Og atom within 3.9 Å of
the carbonyl carbon atoms for all substrates and inhibitors
with all enzymes, the exceptions being hCE1 (CES1) with
cocaine and camphorquinone with both enzymes. The latter
selectivity for hiCE (CES2). The K values for enzyme inhibi-
i
tion were typically in the nM range, although the benzil
analogues demonstrated greatest potency towards hiCE
(CES2) when using CPT-11 as a substrate. We believe that
selective administration of CE inhibitors after CPT-11 dosing,
may alleviate the delayed diarrhea that results from intestinal
drug activation. Because this toxicity is delayed up to 96 h
following chemotherapy, this may provide a suitable window
in which to inhibit hiCE (CES2). We are currently evaluating
this hypothesis in animal models.
The studies described here demonstrate that hiCE (CES2) is
efficient in hydrolysing cocaine and that inhibition of this
enzyme by benzil-based compounds can be achieved in the
presence of a variety of different clinically relevant substrates.
These results should allow the development of approaches that
may have utility in modulating the toxicity for such agents. For
example, i.v. administration of hiCE (CES2) to rapidly convert
cocaine to its inactive metabolites may be possible. Alterna-
tively, selective hiCE (CES2) inhibitors may reduce drug-
induced toxicity associated with CPT-11 hydrolysis in vivo.
1
,2-dione is not an inhibitor of these CEs (Hyatt et al., 2007),
even though it demonstrates properties suitable for enzyme
inhibition. The lack of CE inhibition and the failure of cam-
phorquinone to dock adjacent to the catalytic residues is
likely to be due to steric hindrance enforced by the methyl
bridge present in this bicyclic structure (Hyatt et al., 2007).
In previous modelling studies, based upon the crystal struc-
ture of homatropine bound in the hCE1 (CES1) active site
(Bencharit et al., 2003b), the tropine ring of cocaine was over-
laid onto the coordinates for the corresponding moiety of the
former molecule. This demonstrated that (–)-cocaine would
be preferentially hydrolysed by hCE1 (CES1). However, in
these experiments, no computational docking studies were
performed and hence the electrostatics of drug binding were
not taken into account. Here, we demonstrate that when
these forces are considered in the context of cocaine docking
into the active sites of the CEs, significant differences in the
interaction of the drug with the respective catalytic residues
were observed. This resulted in models that confirmed our
experimental data, that is, that the carbonyl group present
within the benzyl ester chemotype of cocaine was juxtaposed
to the catalytic serine Og atom in hiCE (CES2) (2.8 Å), but was
up to 4 Å further away in hCE1 (CES1). This would make drug
hydrolysis by the latter enzyme unlikely, again consistent
with the biochemical results presented earlier.
Acknowledgements
We thank Dr J.P. McGovren for the kind gift of CPT-11. This
work was supported in part by NIH Grants CA76202,
CA79763, CA98468, CA108775, DA18116, a Cancer Center
Core Grant CA21765 and by the American Lebanese Syrian
Associated Charities.
Recent analysis of the pharmacokinetics of oral heroin
administration demonstrated that the free drug and
6
-acetylmorphine could not be detected in the arterial blood
Conflicts of interest
following dosing (Girardin et al., 2003). However, appreciable
levels of morphine were detected, suggesting that an esterase
present within the alimentary canal mediated the hydrolysis
of heroin to this metabolite. Because high levels of hiCE
None.
(
CES2) are present within the epithelial lining of the human
References
gut, this suggests that this enzyme may be responsible for
metabolism of heroin when this drug is taken orally. Our
results, presented here, confirm that this enzyme is proficient
in hydrolysing this drug, suggesting that in previous studies
Anderson RN (1999). Deaths: leading causes for 1999. Natl Vital Stat
Rep 49: 1–88.
Beaufay H, Amar-Costesec A, Feytmans E, Thines-Sempoux D, Wibo
M, Robbi M et al. (1974). Analytical study of microsomes and iso-
lated subcellular membranes from rat liver. I. Biochemical methods.
J Cell Biol 61: 188–200.
(Girardin et al., 2003), hiCE (CES2) present in the gut metabo-
lized heroin immediately after absorption. This would
account for the lack of the parent molecule present within the
circulation.
Bencharit S, Morton CL, Hyatt JL, Kuhn P, Danks MK, Potter PM et al.
British Journal of Pharmacology (2010) 160 1916–1928

Carboxylesterase hydrolysis of cocaine and heroin
1
928
MJ Hatfield et al
(
2003a). Crystal structure of human carboxylesterase 1 complexed
Morton CL, Potter PM (2000). Comparison of Escherichia coli, Saccha-
romyces cerevisiae, Pichia pastoris, Spodoptera frugiperda and COS7
cells for recombinant gene expression: application to a rabbit liver
carboxylesterase. Mol Biotechnol 16: 193–202.
Morton CL, Wadkins RM, Danks MK, Potter PM (1999). CPT-11 is a
potent inhibitor of acetylcholinesterase but is rapidly catalyzed to
SN-38 by butyrylcholinesterase. Cancer Res 59: 1458–1463.
Munger JS, Shi GP, Mark EA, Chin DT, Gerard C, Chapman HA (1991).
A serine esterase released by human alveolar macrophages is closely
related to liver microsomal carboxylesterases. J Biol Chem 266:
18832–18838.
Pindel EV, Kedishvili NY, Abraham TL, Brzezinski MR, Zhang J, Dean
RA et al. (1997). Purification and cloning of a broad substrate speci-
ficity human liver carboxylesterase that catalyzes the hydrolysis of
cocaine and heroin. J Biol Chem 272: 14769–14775.
Potter PM, Wadkins RM (2006). Carboxylesterases – detoxifying
enzymes and targets for drug therapy. Curr Med Chem 13: 1045–
1054.
Potter PM, Wolverton JS, Morton CL, Wierdl M, Danks MK (1998).
Cellular localization domains of a rabbit and a human carboxy-
lesterase: influence on irinotecan (CPT-11) metabolism by the
rabbit enzyme. Cancer Res 58: 3627–3632.
Redinbo MR, Potter PM (2005). Mammalian carboxylesterases:
from drug targets to protein therapeutics. Drug Discov Today 10:
313–325.
Sali A, Blundell TJ (1993). Comparitive protein modelling by satisfac-
tion of spatial restraints. J Mol Biol 234: 779–815.
Schwer H, Langmann T, Daig R, Becker A, Aslanidis C, Schmitz G
(1997). Molecular cloning and characterization of a novel putative
carboxylesterase, present in human intestine and liver. Biochem
Biophys Res Comm 233: 117–120.
with the Alzheimer’s drug tacrine. From binding promiscuity to
selective inhibition. Chem & Biol 10: 341–349.
Bencharit S, Morton CL, Xue Y, Potter PM, Redinbo MR (2003b).
Structural basis of heroin and cocaine metabolism by a promiscuous
human drug-processing enzyme. Nat Struct Biol 10: 349–356.
Bencharit S, Edwards CC, Morton CL, Howard-Williams EL, Kuhn P,
Potter PM et al. (2006). Multisite promiscuity in the processing of
endogenous substrates by human carboxylesterase 1. J Mol Biol 363:
2
01–214.
Brzezinski MR, Spink BJ, Dean RA, Berkman CE, Cashman JR, Bosron
WF (1997). Human liver carboxylesterase hCE-1: binding specificity
for cocaine, heroin, and their metabolites and analogs. Drug Metab
Dispos 25: 1089–1096.
Chen H, Lyne PD, Giordanetto F, Lovell T, Li J (2006). On evaluating
molecular-docking methods for pose prediction and enrichment
factors. J Chem Inf Model 46: 401–415.
Danks MK, Morton CL, Pawlik CA, Potter PM (1998). Overexpression
of a rabbit liver carboxylesterase sensitizes human tumor cells to
CPT-11. Cancer Res 58: 20–22.
Danks MK, Morton CL, Krull EJ, Cheshire PJ, Richmond LB, Naeve CW
et al. (1999). Comparison of activation of CPT-11 by rabbit and
human carboxylesterases for use in enzyme/prodrug therapy. Clin
Cancer Res 5: 917–924.
Dodds HM, Rivory LP (1999). The mechanism for the inhibition of
acetylcholinesterases by irinotecan (CPT-11). Mol Pharmacol 56:
1
346–1353.
Fleming CF, Bencharit S, Edwards CC, Hyatt JL, Tsurkan L, Bai F et al.
2005). Structural insights into drug processing by human carboxy-
(
lesterase 1: tamoxifen, mevastatin, and inhibition by benzil. J Mol
Biol 352: 165–177.
Gatley SJ (1990). Rapid stereoselective hydrolysis of (+)-cocaine in
baboon plasma prevents its uptake in the brain: implications for
behavioral studies. J Neurochem 54: 720–723.
Girardin F, Rentsch KM, Schwab MA, Maggiorini M, Pauli-Magnus C,
Kullak-Ublick GA et al. (2003). Pharmacokinetics of high doses of
intramuscular and oral heroin in narcotic addicts. Clin Pharmacol
Ther 74: 341–352.
Hicks LD, Hyatt JL, Stoddard S, Tsurkan L, Edwards CC, Wadkins RM
et al. (2009). Improved, selective, human intestinal carboxylesterase
inhibitors designed to modulate 7-ethyl-10-[4-(1-piperidino)-1-
piperidino]carbonyloxycamptothecin (Irinotecan; CPT-11) toxicity.
J Med Chem 52: 3742–3752.
Humerickhouse R, Lohrbach K, Li L, Bosron W, Dolan M (2000).
Characterization of CPT-11 hydrolysis by human liver carboxy-
lesterase isoforms hCE-1 and hCE-2. Cancer Res 60: 1189–1192.
Hyatt JL, Stacy V, Wadkins RM, Yoon KJ, Wierdl M, Edwards CC et al.
Stewart DJ, Inaba T, Tang BK, Kalow W (1977). Hydrolysis of cocaine
in human plasma by cholinesterase. Life Sci 20: 1557–1563.
Tabata T, Katoh M, Tokudome S, Nakajima M, Yokoi T (2004). Iden-
tification of the cytosolic carboxylesterase catalyzing the 5′-deoxy-
5-fluorocytidine formation from capecitabine in human liver. Drug
Metab Dispos 32: 1103–1110.
Wadkins RM, Morton CL, Weeks JK, Oliver L, Wierdl M, Danks MK
et al. (2001). Structural constraints affect the metabolism of 7-ethyl-
10-[4-(1-piperidino)-1-piperidino]carbonyloxycamptothecin (CPT-
11) by carboxylesterases. Mol Pharmacol 60: 355–362.
Wadkins RM, Hyatt JL, Yoon KJ, Morton CL, Lee RE, Damodaran K
et al. (2004). Identification of novel selective human intestinal
carboxylesterase inhibitors for the amelioration of irinotecan-
induced diarrhea: synthesis, quantitative structure-activity
relationship analysis, and biological activity. Mol Pharmacol 65:
1336–1343.
(
1
2005). Inhibition of carboxylesterases by benzil (diphenylethane-
,2-dione) and heterocyclic analogues is dependent upon the aro-
Wadkins RM, Hyatt JL, Wei X, Yoon KJ, Wierdl M, Edwards CC et al.
(2005). Identification and characterization of novel benzil
(diphenylethane-1,2-dione) analogues as inhibitors of mammalian
carboxylesterases. J Med Chem 48: 2905–2915.
maticity of the ring and the flexibility of the dione moiety. J Med
Chem 48: 5543–5550.
Hyatt JL, Wadkins RM, Tsurkan L, Hicks LD, Hatfield MJ, Edwards CC
et al. (2007). Planarity and constraint of the carbonyl groups in
Webb JL (1963). Enzyme and Metabolic Inhibitors. Volume 1. General
Principles of Inhibition. Academic Press Inc: New York.
Wierdl M, Tsurkan L, Hyatt JL, Hatfield MJ, Edwards CC, Danks MK
et al. (2008). An improved human carboxylesterase for use in
enzyme/prodrug therapy with CPT-11. Cancer Gene Ther 15: 183–
192.
1,2-diones are determinants for selective inhibition of human car-
boxylesterase 1. J Med Chem 50: 5727–5734.
Kamendulis LM, Brzezinski MR, Pindel EV, Bosron WF, Dean RA
(
1996). Metabolism of cocaine and heroin is catalyzed by the same
human liver carboxylesterases. J Pharmacol Expt Ther 279: 713–717.
Khanna R, Morton CL, Danks MK, Potter PM (2000). Proficient
metabolism of CPT-11 by a human intestinal carboxylesterase.
Cancer Res 60: 4725–4728.
Lockridge O, LaDu BN (1980). Diacetylmorphine (heroin) hydrolysis
by human-serum cholinesterase (pseudocholinesterase or butyryl-
cholinesterase). Fed Proc 39: 1754.
Mori M, Hosokawa M, Ogasawara Y, Tsukada E, Chiba K (1999). cDNA
cloning, characterization and stable expression of novel human
brain carboxylesterase. FEBS Lett 458: 17–22.
Wu MH, Chen P, Remo BF, Cook EH Jr, Das S, Dolan ME (2003).
Characterization of multiple promoters in the human carboxy-
lesterase 2 gene. Pharmacogenetics 13: 425–435.
Zhang J, Burnell JC, Dumaual N, Bosron WF (1999). Binding and
hydrolysis of meperidine by human liver carboxylesterase hCE-1. J
Pharmacol Exp Ther 290: 314–318.
Zhu HJ, Patrick KS, Yuan HJ, Wang JS, Donovan JL, DeVane CL et al.
(2008). Two CES1 gene mutations lead to dysfunctional carboxy-
lesterase 1 activity in man: clinical significance and molecular basis.
Am J Hum Genet 82: 1241–1248.
British Journal of Pharmacology (2010) 160 1916–1928