Comparison of the structure and activity of glycosylated and aglycosylated human carboxylesterase 1

Article abstract of DOI:10.1371/journal.pone.0143919

Human Carboxylesterase 1 (hCES1) is the key liver microsomal enzyme responsible for detoxification and metabolism of a variety of clinical drugs. To analyse the role of the single N-linked glycan on the structure and activity of the enzyme, authentically

Full text of DOI:10.1371/journal.pone.0143919

RESEARCH ARTICLE
Comparison of the Structure and Activity of
Glycosylated and Aglycosylated Human
Carboxylesterase 1
1
,2,6
3,4
1,2
1,2
Victoria Arena de Souza
, David J. Scott , Joanne E. Nettleship , Nahid Rahman
,
5
¤
3,6
1,2
Michael H. Charlton , Martin A. Walsh *, Raymond J. Owens
*
1
UK OPPF-UK, The Research Complex at Harwell, Rutherford Appleton Laboratory Harwell Oxford,
Oxfordshire, United Kingdom, 2 Division of Structural Biology, Henry Wellcome Building for Genomic
Medicine, University of Oxford, Roosevelt Drive, Oxford, United Kingdom, 3 The Research Complex at
Harwell, Rutherford Appleton Laboratory Harwell Oxford, Oxfordshire, United Kingdom, 4 School of
Biosciences, University of Nottingham, Sutton Bonington Campus, Sutton Bonington, Leicestershire, United
Kingdom, 5 Chroma Therapeutics Ltd., 93 Innovation Drive Milton Park, Abingdon, United Kingdom,
6
Diamond Light Source, Harwell Science and Innovation Campus, Didcot, United Kingdom
¤
*
Current address: InhibOx, Ltd, Oxford Centre for Innovation, New Road, Oxford, OX11BY, United Kingdom
martin.walsh@diamond.ac.uk (MAW); ray@strubi.ox.ac.uk (RJO)
OPEN ACCESS
Citation: Arena de Souza V, Scott DJ, Nettleship JE,
Rahman N, Charlton MH, Walsh MA, et al. (2015)
Comparison of the Structure and Activity of
Glycosylated and Aglycosylated Human
Abstract
Human Carboxylesterase 1 (hCES1) is the key liver microsomal enzyme responsible for
detoxification and metabolism of a variety of clinical drugs. To analyse the role of the single
N-linked glycan on the structure and activity of the enzyme, authentically glycosylated and
aglycosylated hCES1, generated by mutating asparagine 79 to glutamine, were produced
in human embryonic kidney cells. Purified enzymes were shown to be predominantly tri-
meric in solution by analytical ultracentrifugation. The purified aglycosylated enzyme was
found to be more active than glycosylated hCES1 and analysis of enzyme kinetics revealed
that both enzymes exhibit positive cooperativity. Crystal structures of hCES1 a catalytically
inactive mutant (S221A) and the aglycosylated enzyme were determined in the absence of
any ligand or substrate to high resolutions (1.86 Å, 1.48 Å and 2.01 Å, respectively). Super-
position of all three structures showed only minor conformational differences with a root
mean square deviations of around 0.5 Å over all Cα positions. Comparison of the active
sites of these un-liganded enzymes with the structures of hCES1-ligand complexes showed
that side-chains of the catalytic triad were pre-disposed for substrate binding. Overall the
results indicate that preventing N-glycosylation of hCES1 does not significantly affect the
structure or activity of the enzyme.
Carboxylesterase 1. PLoS ONE 10(12): e0143919.
doi:10.1371/journal.pone.0143919
Editor: Andreas Hofmann, Griffith University,
AUSTRALIA
Received: July 13, 2015
Accepted: November 11, 2015
Published: December 11, 2015
Copyright: © 2015 Arena de Souza et al. This is an
open access article distributed under the terms of the
Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any
medium, provided the original author and source are
credited.
Data Availability Statement: The coordinates and
structural factors for hCES1, hCES1 S221A and
hCES1 N79Q have been deposited in the Protein
Data Bank under accession numbers 5a7f, 5h7g and
5a7h, respectively.
Introduction
Funding: The work described was supported by the
following: Medical Research Council UK Grant no.
MR/K018779/1 (RJO). Nuffield Department of
Medicine, University of Oxford (VAS). Diamond Light
Source Ltd. (VAS). Chroma Therapeutics Ltd. (VAS).
Wellcome Trust Grant no. 075491/Z/04 (VAS, RJO).
The funders had no role in study design, data
Carboxylesterases are a family of enzymes that act on a variety of both exogenous (e.g. cocaine,
heroin) and endogenous (e.g. acyl-CoA esters) substrates. They are defined by their ability to
hydrolyze ester, amide, or thioester bonds to their corresponding alcohol, amine or thiol and
free acid in a diverse range of chemically distinct compounds [1]. Genes coding for five carbox-
ylesterases have been identified in the human genome (hCES1-5) [2], with CES 1, 2 and 3
PLOS ONE | DOI:10.1371/journal.pone.0143919 December 11, 2015
1 / 12

Structure of Aglycosylated Human Carboxylesterase 1
collection and analysis, decision to publish, or
preparation of the manuscript. Chroma Therapeutics
provided support in the form of part of the stipend for
the lead author, VAS and author (MHC) formerly
employed by Chroma Therapeutics was involved in
the conception of the study and contributed to the
writing of the paper. Chroma Therapeutics did not
play any role in the decision to publish the paper.
appearing to be the functionally significant enzymes. All three enzymes show differential tissue
expression, with cells of the monocyte/macrophage lineage being the principal source of
hCES1 outside hepatocytes [3]. hCES1 and 2 are localised to the endoplasmic reticulum (ER)
via the KDEL receptor. The enzymes hydrolyze substrates via a two-step ping pong mechanism
that includes the formation and degradation of an acyl-enzyme intermediate, using water as a
transitional nucleophile [4]. Intense interest in these enzymes stems from their critical role in
Phase 1 metabolism and activation of pro-drugs, notably the anti-cancer agent, CPT-11[5].
In 2003, the first crystal structures of hCES1 in complex with narcotic analogues [6] were
reported, showing that the enzyme forms a hexamer from trimers. Over the past 10 years,
many more structures of hCES1 have been determined with the highest resolution structure
reported to date at 2.0 Å (2h7c [7]). Carboxylesterases comprise three distinct domains; a cen-
tral domain containing the catalytic triad (S221, H467 and E354), a regulatory domain (RD)
and an α/β domain. The active site occupies a 10-15Å deep hydrophobic pocket at the interface
of the three domains with all three catalytic residues arranged such that a proton transfer chain
can be established [6]. It also contains the C-terminal helix of the enzyme. The RD is mainly
helical; containing two disordered loops and has been proposed to regulate substrate binding
and product release [8]. Published structures show that the RD of the enzyme exhibits high
thermal displacement parameters, indicating dynamic mobility within this region [6–8]. The
αβ domain or hydrolase fold, lies adjacent to both the catalytic and regulatory domain. This
domain is common to a number of hydrolytic enzymes of differing catalytic functions and phy-
logenetic origin [9]. Examination of enzyme: substrate complexes has revealed the presence of
two non-selective substrate binding sites in addition to the catalytic site. The ‘Z-site’ which is
located within the regulatory domain of the enzyme [6,10], and that is proposed to control the
trimer–hexamer equilibrium of the enzyme and a ‘side-door’ secondary pore that leads into the
active site from the surface of the enzyme.
Competing Interests: The study is a collaboration
between University of Oxford and Diamond Light
Source Ltd., which is a non-commercial and publicly
funded UK national synchrotron research facility. I
confirm that the part funding by Diamond Light
Source Ltd. does not alter our adherence to all PLOS
ONE policies on sharing data and materials. We have
the following interests: this study received funding
from Chroma Therapeutics Ltd. Michael H. Charlton
was formerly employed by Chroma Therapeutics Ltd.
and is currently employed by Inhibox. There are no
patents, products in development or marketed
products to declare. This does not alter our
adherence to all the PLOS ONE policies on sharing
data and materials, as detailed online in the guide for
authors.
hCES1 has a single N-linked glycosylation site at N79 which is conserved at the equivalent
position in orthologues from other species (www.uniprot.org). Over 20 years ago, Kroetz et al.
provided evidence that N-linked glycosylation was essential for maximal catalytic activity in
hCES1 for simple aromatic and aliphatic esters [11]. The enzyme was expressed in insect cells
using the baculovirus system with the addition of tunicamycin in the culture media to inhibit
GlcNAc phosphotransferase (GPT) and hence produce non-glycosylated enzyme. In addition,
a structural role for terminal sialic acids of N-glycosylated hCES1 in stabilising the trimeric
enzyme has been proposed [9,10], though as an ER resident, the enzyme would not normally
be sialylated. Nevertheless, a role for glycosylation in the activity of hCES1 has become an
established fact in the scientific literature [12]. However some early work on human triacylgly-
cerol hydrolase appears to have been overlooked. This microsomal enzyme which hydrolyses
cytoplasmic triacylglycerol is an isoform of hCES1 and it has been shown that the point muta-
tion, N79A, does not affect activity of the enzyme that was also produced in insect cells [13].
To provide a definitive insight into the importance of N-glycosylation on the structure and
activity of human hCES1, the crystal structure of the enzyme with the mutation, N79Q, has
been determined. The structure-function relationship of this aglycosylated hCES1 was explored
alongside the wild type enzyme and a catalytically inactive mutant (S221A).
Methods
Protein production
DNAs encoding hCES1 and hCES1 S221A optimized for combined human and insect codon
usage were ordered from GeneArt1 (Life Technologies). The hCES1 N79Q mutation was
introduced into hCES1 by PCR strand overlap extension [14]. Genes were inserted into the
PLOS ONE | DOI:10.1371/journal.pone.0143919 December 11, 2015
2 / 12

Structure of Aglycosylated Human Carboxylesterase 1
mammalian cell expression vector, pOPINTTGneo [15], by Infusion cloning as described pre-
-
-
viously [16]. For the production of hCES1 and hCES1 S221A, stable HEK293 Gnt1 / [17] cell
lines were generated by selection with G418. hCES1 N79Q was produced by transient expres-
sion in HEK 293T cells (ATCC1 CRL-11268™) [18]. Secreted proteins were purified using an
automated protocol consisting of nickel affinity chromatography using a HiScreen™ Ni FF (GE
Healthcare) column followed by size exclusion chromatography using a HiLoad 16/60Superdex
2
00 column (GE Healthcare) on an ÄKTAxpress unit [19]. Purified proteins were de-glycosy-
0
lated by incubating with PNGase F (Sigma-Aldrich) overnight at 37 C. Samples were analysed
by SDS-polyacrylamide gel electrophoresis under reducing conditions and gels stained using
InstantBlue™ (Expedeon).
Analytical ultracentrifugation
Sedimentation velocity experiments were carried out on a Beckman XL-I analytical ultracentri-
fuge (Beckman-Coulter). Sedimentation velocity was performed at 20°C using 2 channel centre
pieces, with protein loading concentrations of 2 mg/ mL, 1.0 mg/ mL and 0.5 mg/ mL in 20
mM TrisHCl, pH 7.5 containing 200 mM NaCl. Data were obtained at 40,000 rpm, using a
Beckman 50Ti rotor, with the cells scanned radially with interference optics and with absor-
bance optics at a wavelength of 280 nm. Scans were obtained every 10 minutes and data were
analyzed using the program SEDFIT v11.3 (www.analyticalultracentrifugation.com). Sedimen-
tation coefficient distributions were obtained using the c(S) methodology114, and figures were
created in GUSSI 1.0.3. Solution densities and viscosities were measured directly using an
Anton Paar DMA5000 densitometer/viscometer.
Enzyme kinetics
Esterase activity was measured using 4-nitrophenyl acetate (4-NPA) as the substrate [20][20]
at varying concentrations (0–3000 μM). The rate of hydrolysis of 4-NPA was followed by mea-
-
1
-1
suring the production of the nitrophenolate anion at 405 nm (ε 405 = 18000 M cm ) in a
Paradigm Plate Reader (Beckman Coulter). The assay was carried out at 37°C with shaking and
A405 readings taken at 30 second intervals over 30 minutes. Experiments were carried out in
triplicate, results averaged and the amount of product formed plotted against time. Rates of
reaction at different substrate concentrations were calculated from the linear part of the curves.
2
All r values exceeded 0.99. An allosteric sigmoidal substrate-velocity model (nonlinear regres-
sion) was fitted to the data using GraphPad Prism version 6 (www.graphpad.com) and kinetic
parameters calculated using the following equation:
h
V
maxx
Y ¼
h
h
ðK þ x Þ
half
Where Y = initial velocity, x = concentration of substrate, Vmax is the maximum velocity of
the enzyme, h = Hill coefficient and Khalf is the concentration of substrate that produces a half-
maximal enzyme reaction rate.
Crystallization and structure solution
hCES1 and hCES1 S221A were concentrated to 5 mg/ml and hCES1 N79Q to 6 mg/ml in 20
mM Tris pH 7.5, 200 mM NaCl and crystallization trials set up in 200 nL (100nL protein plus
1
2
00 nL reservoir solution) drops in a 96 well format [21][21]. hCES1 crystals were grown at
0°C in 0.1 M MES/imidazole, pH 6.5 containing 0.03 M each of diethylene glycol, triethylene
glycol, tetraethylene glycol, pentaethylene glycol, 10% (w/v) polyethylene glycol 4000, 20%
PLOS ONE | DOI:10.1371/journal.pone.0143919 December 11, 2015
3 / 12

Structure of Aglycosylated Human Carboxylesterase 1
(
w/v) glycerol. Both hCES1 S221A and hCES1 N79Q were crystallized at 4°C in respectively,
0
.1 M Bicine/Trizma1 base, pH 8.5, containing 0.03 M each of diethylene glycol, triethylene
glycol, tetraethylene glycol, pentaethylene glycol, 10% (w/v) PEG 4000, 20% (v/v) glycerol and
.1 M Bis-Tris propane, pH 7.5, containing 20% (w/v) polyethylene glycol 3350, 0.2 M NaI. For
crystals that were harvested from conditions optimised from the Morpheus1 screen [22]
hCES1 and hCES1 S221A), no cryo-protection was needed. For the hCES1 N79Q crystal
0
(
grown from conditions identified from the PACT screen [23], 25% glycerol was added directly
as a cryo-protectant. Diffraction data were collected at Diamond Light Source (DLS), Oxford-
shire, on beamlines I03 and I04. The structure of hCES1 was solved by molecular replacement
with PHASER [24] using the coordinates of the human carboxylesterase PDB entry 2h7c as a
search model [7]. Phases calculated from this initial model were used for manual completion of
the structure using COOT [25] with iterative cycles of refinement with REFMAC [26]. This
high resolution structure of hCES1 was then used as the search model for molecular replace-
ment of the other two structures presented here. For the high resolution structure of hCES1
S221A where data extended to 1.48 Å, anisotropic refinement of individual atomic displace-
ment parameters was feasible. MolProbity4 was used for structure validation. Diffraction data
and refinement statistics are shown in Table 1. SHP was used for structural comparisons [27]
and images for figures prepared with PyMOL (http://pymol.sourceforge.net/). The coordinates
and structural factors for hCES1, hCES1 S221A and hCES1 N79Q have been deposited in the
Protein Data Bank under accession numbers 5a7f, 5h7g and 5a7h, respectively.
Results and Discussion
Production and biophysical characterisation of glycosylated and
aglycosylated human HCES1
Recombinant hCES1 proteins for structural studies have been produced using the Spodoptera
frugiperda (Sf21) [28] insect expression system, infecting the cells with a baculovirus and puri-
fying the enzyme from the supernatant [29–37]. The yields of purified enzymes were reported
as 7.5–12.5 mg/ L cell culture [38]. As an alternative to production in insect cells in culture,
Greenblatt et al., reported using Trichoplusia ni (whole cabbage looper) larvae with a yield of
9
mg protein from ~ 1 kg of infected caterpillar larvae [39]. Here, recombinant hCES1 enzymes
were produced in HEK293 cells using either transient expression (hCES1 N79Q) or from stable
cell lines (hCES1 and S221A). The ER retention motif, (HIEL) was deleted from the sequences
so that proteins were secreted into the cell media and recovered by metal chelate chromatogra-
phy via a C-terminal histidine tag added to the sequences. In order to match the high mannose
glycoforms that would be present on the native enzyme, transient expression in HEK293T cells
was carried out in the presence of the mannosidase inhibitor kifunensine [40]. The HEK
-
-
Gnt1 / mutant cell line in which N-glycosylation is restricted to GlcNAc Man was used for
2
5
generating stable cell lines. The yields of the aglycosylated hCES1 N79Q, glycosylated hCES1
and S221A enzymes were 39 mg/L, 64 mg/L and 68 mg/L respectively showing that N-glycosyl-
ation is not required for expression and secretion of the enzyme consistent with previous
results [11,13]. The calculated monomeric molecular weight of hCES1 is 60.9 kDa (excluding
N-glycosylation) and on size exclusion chromatography the three proteins eluted at around
1
60 kDa consistent with assembly into trimers (Fig 1A). To assess the N-glycosylation states of
the purified proteins, samples were treated with PNGase F and analysed by SDS-polyacryl-
amide gel electrophoresis. As shown in Fig 1B, CES1 migrated as single species above CES1
N79Q consistent with full occupancy of the N79 glycosylation site. Following, de-glycosylation
with PNGase F, the CES1 co-migrated with the N79Q aglycosylated protein confirming
PLOS ONE | DOI:10.1371/journal.pone.0143919 December 11, 2015
4 / 12

Structure of Aglycosylated Human Carboxylesterase 1
Table 1. Data collection and refinement statistics.
hCES1 wild type
hCES1 S221A
hCES1 N79Q
Data Collection
X-Ray Source
Wavelength (Å)
Space Group
DLS, I04
0.9795
R3: H
DLS, I03
0.97625
R3: H
DLS, I04
0.9795
R3: H
Unit Cell Parameters
a, b, c [Å]
114.72, 114.72, 117.78
90, 90, 120
115.39, 115.39, 128.14
90, 90, 120
115.47, 115.47, 127.28
90, 90, 120
78.63–2.01 (2.06–2.01)
98.3 (98.0)
α, β, γ [°]
Resolution range [Å]
Completeness (%)
50.00–1.86 (1.88–1.86)
53.94–1.48 (1.52–1.48)
98.4 (91.8)
a
97.0 (71.5)
b
Rmerge (%)
6.5 (43.3)
0.785
4.7 (69)
6.3 (61)
c
CC1/2 (High resolution shell)
0.612
0.64
<
I>/σ(<I>)
17.4 (1.8)
2.8 (2.3)
11.1 (1.0)
9.7 (1.3)
Multiplicity
2.9 (2.0)
3.0 (3.0)
Data Refinement
No. of reflections
No. of unique reflections
132,450
46, 879
15.83
303,901
104,388
12.96
125,184
41,414
17.9
d
R-factor (%)
e
Rfree (%)
18.76
17.19
22.16
R.m.s. deviations
Bond lengths (Å)
Bond angles (°)
0.013
1.63
21
0.012
1.51
19.4
31
0.012
1.4694
19.2
Wilson B-factor (Å2)
Mean B-factor (Å2)
18.7
43.37
a
Values in parentheses refer to data in the highest resolution shell for each protein.
b
R
merge = ∑hkl
∑
j
|Ihkl,j − hIhkli|/∑hkl
j hkl,j
∑ I .
c
CC values are the half-set correlation coefficients as described by Karplus & Diederichs [49]
d
R-factor = ||Fobs| − |Fcalc||/∑|Fobs|;
e
R
free = R-factor for a selected subset (5%) of the reflections that were not included in prior refinement calculation.
doi:10.1371/journal.pone.0143919.t001
removal of the single N-glycan. As expected, the mobility of the N79Q protein was not altered
by PNGase treatment (Fig 1B).
The oligomeric states of the hCES1 and N79Q proteins were investigated by analytical ultra-
centrifugation in sedimentation velocity experiments. The results showed that all three forms
of hCES1 were mainly trimeric with an apparent molecular weight of 152 kDa (Fig 1C). A
small amount of monomer was also identified (64 kDa), as well as some aggregation at a higher
0
2
molecular weight. The corrected sedimentation coefficient (S 0;w) of the trimer was calculated
to be 7.5 S, and the frictional ratio (f/f ) was 1.52. Together the results confirm that hCES1
o
behaves overwhelmingly as a trimer in solution with no evidence for significant hexamer for-
mation and that glycosylation plays no role in trimer assembly.
Analysis of enzyme activity
hCES1 native and N79Q hydrolysed the substrate 4-nitrophenyl acetate (4-NPA) demonstrat-
ing their functional enzymatic activity; as expected the S221A mutant was inactive. Activity
increased linearly with enzyme concentration (Fig 2A and 2B) and the specific activities for
-
1
-1
-1
-1
hCES1 and for hCES1 N79Q were 73 µmol min μg , 98 µmol min μg respectively.
PLOS ONE | DOI:10.1371/journal.pone.0143919 December 11, 2015
5 / 12

Structure of Aglycosylated Human Carboxylesterase 1
Fig 1. Purification of human carboxylesterases. (a) Size exclusion profiles of purified hCES1 (blue trace) hCES1 N79Q (green trace) and hCES1 S221A
(red trace) enzymes from media of transfected HEK cells. Samples were run on a HiLoad 16/60Superdex 200 column (GE Healthcare) in 200 mM NaCl, 20
mM Tris-HCl, pH 7.5. The peak corresponds to a molecular weight of approximately 160 kDa as estimated from the elution volumes of globular proteins of
known molecular weight: Aprotinin (6.5 kDa) Ribonuclease A (13.7 kDa) Carbonic Anhydrase (29 kDa), Ovalbumin (44 kDa), Conalbumin (75 kDa) Aldolase
(
(
158 kDa) Ferritin (440 kDa) and Blue Dextran 2000. (b) SDS-polyacrylamide gel of purified CES1 (lanes 1 and 2) and CES1 N79Q (lanes 3 and 4) untreated
lanes 1 and 3) and treated with PNGaseF (lanes 2 and 4). (c) The sedimentation velocity distribution for hCES1 N79Q. Data for hCES1 and hCES1 S221A
gave the same profiles.
doi:10.1371/journal.pone.0143919.g001
Glycosylation does not appear to markedly affect activity consistent with other reports for
recombinant hCES1 [41].
In contrast to previous results, e.g. [41–43], the activity of both enzymes did not follow clas-
sical Michaelis-Menten kinetics. Initial reaction velocities (V ) plotted against substrate con-
0
centration (S) for both hCES1 and the N79Q mutant gave sigmoidal rather than hyperbolic
profiles indicative of positive cooperativity (Fig 2C and 2D). Fitting the data to an allosteric
model by non-linear regression gave values for the Hill coefficient (h) for both enzymes >1.00
(Table 2). It is interesting to note that native pig and human liver microsomal carboxylesterases
have previously been reported to exhibit positive cooperativity [44,45]. This phenomenon is
typical of enzymes comprising more than one identical subunit and can be interpreted by refer-
ence to one of two models namely the Monod-Wyman-Changeux (MWC) concerted model
[46] and the Koshland-Nemethy-Filmer (KNF) sequential model [47]. Both models postulate
that each subunit can exist in two different conformational states nominally termed R (relaxed)
and T (tense) but differ in their assumptions about subunit interaction and the pre-existence of
both states. In the MWC model, all subunits are in presumed to be in the same conformation
with T and R states in equilibrium in the absence of substrate [46]. One state is assumed to
have a higher binding affinity for the substrate than the other such that on binding of a ligand
(
substrate or effector) the equilibrium is shifted in favour of the more active state. The KNF
model postulates that substrate binding involves a process of induced fitting and that this con-
formational change affects other active sites such that the protein transitions from one state
into a more active one. However, unlike in the MWC model, each subunit can undergo confor-
mational change independently. It is not a priori possible to distinguish between these two
models to explain the behaviour of CES1, though the observation that substrate binding does
not appear to be associated with a conformational change in the active site would favour the
PLOS ONE | DOI:10.1371/journal.pone.0143919 December 11, 2015
6 / 12

Structure of Aglycosylated Human Carboxylesterase 1
Fig 2. Enzyme activity of human carboxylesterases. Plots of initial rates of reaction against enzyme
concentration assayed as described in the Methods section at a substrate concentration of 750 μM 4-NPA (a)
CES1 (b) CES1 N79Q. Plots of initial reaction rates against substrate concentration for the hydrolysis of
4-NPA by (c) hCES1 and (d) hCES1 N79Q (3.4 nM each enzyme). The molarity of the enzyme was
calculated assuming 100% trimer with a molecular weight of 182.7 kDa.
doi:10.1371/journal.pone.0143919.g002
concerted model. It has been suggested that the so-called Z-site, shown to be occupied by sub-
strates in some hCES1-ligand complexes is an allosteric binding site [6]. Thus transitions from
low to high substrate binding affinity may be modulated by substrate binding to a non-catalytic
site on the enzyme. Most interestingly, analysis of the kinetics of natural pig microsomal car-
boxylesterase 1 according to a concerted model of cooperativity suggested that each subunit of
the trimeric enzyme had two substrate binding sites [45]. Overall the enzyme assay results con-
firm that N-glycosylation does not affect binding of the 4-NPA substrate and are consistent
with recent results for the hCES1 expressed in E.coli as an insoluble protein and subsequently
refolded in vitro [41]. Analysis of enzyme kinetics show that recombinant hCES1 exhibits posi-
tive cooperativity, though the molecular mechanism for this remains unknown.
Table 2. Enzyme kinetic data.
Parameter
CES1
CES1 N79Q
-
1
Vmax (μMmin )
9.25 +/- 1.30
1.46 +/- 0.26
1123 +/- 280.2
0.9915
10.92 +/- 2.01
1.57 +/- 0.4
1090 +/- 343.3
0.9819
h
Khalf (μM)
2
Goodness of Fit (r )
Values are given +/- standard errors from the analysis of V0 vs S using the allosteric model in GraphPad
Prism version 6
doi:10.1371/journal.pone.0143919.t002
PLOS ONE | DOI:10.1371/journal.pone.0143919 December 11, 2015
7 / 12

Structure of Aglycosylated Human Carboxylesterase 1
Comparison of the structures of glycosylated and aglycosylated human
hCES1
All three proteins readily crystallized in the space group, R3: H, a new space group for hCES1
which has most commonly crystallized in the orthorhombic space group P2 2 2 . The struc-
1
1 1
tures of hCES1, S221A and N79Q mutants were solved by molecular replacement using hCES1
in complex with Coenzyme A determined to 2.0 Å resolution (PDB code: 2h7c) as the initial
search model and refined to 1.86 Å, 1.48 Å and 2.01 Å respectively (Table 1). These represent
the highest resolution structures of hCES1 solved to date in the absence of any substrates. In all
three structures, the asymmetric unit contained one monomer of hCES1, containing, two disul-
phide bridges were present, C87- C116 and C274—C285 as seen in other hCES1 structures.
Superposition of the monomeric structures showed no differences between the glycosylated
and aglycosylated hCES1 (Fig 3A). The individual monomers superimposed with root mean
square deviations (rmsd) of 0.47 Å (hCES1) and 0.3 Å (S221A) for all C atoms (Table 3). In
α
the crystal structure of hCES1 and hCES1 S221A, an N-acetylglucosamine adduct was observed
attached to the side chain of residue N79. (Fig 3B). Space group symmetry was used to generate
the biological trimer that is seen in solution (Fig 3C).
Superimposition of the hCES1 structures onto examples of ligand-bound hCES1 complexes
showed that ligand-binding does not change the overall conformational of the protein. In the
case of the 2H7C, the root mean square deviation (rmsd) for all Cα atoms was 0.5 Å. (Table 3).
The orientation of the side-chains of the catalytic triad (S221, E354, H467) (Fig 3D) in bound
and unbound enzymes was also identical showing that the enzyme is pre-disposed for substrate
binding.
Previously, it had been noted that trimer formation in hCES1 was mediated by two charge
clamps across the trimer interface, from R186 and E183 of one monomer to E72 and K78
respectively, of the adjacent monomer [48]. Examination of the interfaces of the hCES1 struc-
tures presented here, and other structures of hCES1 (1mx1, 1yah, 2h7c), confirms that the
E183 and K78 pairing does indeed form a salt bridge between two hCES1 monomer subunits
molecules. However, the E72:R186 salt bridge was not apparent in the structures reported here,
and in fact is only observed in the interface formed between monomer A and B of the trimer in
PDB entry 1ya8. E72 takes up an alternate conformation in all other CES1 structures that have
been reported. The acquisition of high resolution data, reported here, facilitated identification
of a second salt bridge between residues K275 and E292. Examination of structures for which
experimental data have been deposited with the PDB (2hrq, 3k9b, 1ya8, 1ya4 and 1yah)
showed this interaction to be conserved. Therefore, we propose that the K275:E292 salt bridge
together with the K78:E183 pairing, plays a potential role in stabilisation of the hCES1 trimer
(Fig 3E).
Conclusions
A comparison of authentically glycosylated human hCES1 produced in HEK cells with an agly-
cosylated version of the enzyme shows that preventing glycosylation at N79 does not affect the
synthesis, activity or structure of the enzyme. In the hydrolysis of the model substrate 4-NPA,
both hCES1 and aglycosylated hCES1 showed positive cooperatvity suggesting that substrate
binding elicits conformation changes in the protein. However no major differences were
observed in a comparison of crystal structures of hCES1 reported here and published ones with
substrates/inhibitors bound into the active site. Further experiments comparing the solution
behaviour of the proteins with/without substrates may be more revealing. More generally, the
expression of hCES1 in mammalian cells has produced approximately fivefold higher yields
than previously reported for insect cell expression. Crystal structures of hCES1 expressed in
PLOS ONE | DOI:10.1371/journal.pone.0143919 December 11, 2015
8 / 12

Structure of Aglycosylated Human Carboxylesterase 1
Fig 3. Crystal structures of human carboxylesterases. (a) Overlay of hCES1 (blue) and hCES1 N79Q (gold) (b) Views of the glycosylation site for hCES1,
S122A hCES1 and the N79Q mutant. The Fo–Fc omit map electron density is shown carved around the glycosylation site (N79) and contoured at a level of 3
σ (c) Trimer of hCES1 generated by the space group symmetry, N-acetylglucosamine sugars attached to N97 of hCES1 are represented as spheres in red
and green (d) Superpostion of the catalytic triad of hCES1 (mauve), hCES1 N79Q (green), hCES1 S221A (yellow) hCES1 tamoxifen complex (PDB id 1ya4,
turquoise). Residues are in stick representation. The bound tamoxifen is coloured turquoise. (e) Cartoon representation of the hCES1 trimer generated by the
space group symmetry with the K78:E183 and K275:E292 salt bridges shown in spheres (K78 in green, E183 in magenta, K275 in cyan and E292 in orange).
doi:10.1371/journal.pone.0143919.g003
PLOS ONE | DOI:10.1371/journal.pone.0143919 December 11, 2015
9 / 12

Structure of Aglycosylated Human Carboxylesterase 1
Table 3. Structure alignment of human hCES1.
RMSD (Å)
hCES1
Wild type
S221A
N79Q
Wild type
S221A
-
0.44
-
0.47
0.30
0.48
0.56
0.68
-
2h7c*
0.50
0.52
0.72
0.48
0.57
0.70
4ab1
1ya8*
The structures were aligned using the PDBeFold server and the RMSD calculated between all Ca-atoms at
best 3D superposition of the query and target structures (*average from all chains in the asymmetric unit).
doi:10.1371/journal.pone.0143919.t003
mammalian cells have been determined at </ = 2.0 Å, representing the highest resolution
structures reported to date and provide crystal systems which could be exploited in the design
of inhibitors which could be used to modulate the metabolism of clinically important drugs.
Acknowledgments
We are grateful to David Krige and Tricia Kirwin-Jones for advice with the hCES1 enzyme
assay. We thank the staff of beamlines I03 and I04 at Diamond Light Source for help with data
collection.
Author Contributions
Conceived and designed the experiments: RJO MAW MHC VAS DJS. Performed the experi-
ments: VAS MAW DJS JEN NR. Analyzed the data: RJO VAS DJS MAW. Contributed
reagents/materials/analysis tools: VAS MAW RJO DJS MHC JEN NR. Wrote the paper: VAS
MAW RJO DJS MHC.
References
1.
Satoh T, Hosokawa M (1998) The mammalian carboxylesterases: from molecules to functions. Annu
Rev Pharmacol Toxicol 38: 257–288. PMID: 9597156
2.
Holmes RS, Wright MW, Laulederkind SJ, Cox LA, Hosokawa M, Imai T, et al. (2010) Recommended
nomenclature for five mammalian carboxylesterase gene families: human, mouse, and rat genes and
proteins. Mamm Genome 21: 427–441. doi: 10.1007/s00335-010-9284-4 PMID: 20931200
3.
Needham LA, Davidson AH, Bawden LJ, Belfield A, Bone EA, Brotherton DH, et al. (2011) Drug target-
ing to monocytes and macrophages using esterase-sensitive chemical motifs. J Pharmacol Exp Ther
339: 132–142. doi: 10.1124/jpet.111.183640 PMID: 21778281
4
.
.
Potter PM, Wadkins RM (2006) Carboxylesterases—detoxifying enzymes and targets for drug therapy.
Curr Med Chem 13: 1045–1054. PMID: 16611083
5
Hatfield MJ, Tsurkan L, Garrett M, Shaver TM, Hyatt JL, Edwards CC, et al. (2011) Organ-specific car-
boxylesterase profiling identifies the small intestine and kidney as major contributors of activation of the
anticancer prodrug CPT-11. Biochem Pharmacol 81: 24–31. doi: 10.1016/j.bcp.2010.09.001 PMID:
20833148
6
7
8
.
.
.
Bencharit S, Morton CL, Xue Y, Potter PM, Redinbo MR (2003) Structural basis of heroin and cocaine
metabolism by a promiscuous human drug-processing enzyme. Nat Struct Biol 10: 349–356. PMID:
12679808
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:
201–214. PMID: 16962139
Bencharit S, Morton CL, Howard-Williams EL, Danks MK, Potter PM, Redinbo MR (2002) Structural
insights into CPT-11 activation by mammalian carboxylesterases. Nat Struct Biol 9: 337–342. PMID:
11967565
PLOS ONE | DOI:10.1371/journal.pone.0143919 December 11, 2015
10 / 12

Structure of Aglycosylated Human Carboxylesterase 1
9.
Ollis DL, Cheah E, Cygler M, Dijkstra B, Frolow F, Franken SM, et al. (1992) The alpha/beta hydrolase
fold. Protein Eng 5: 197–211. PMID: 1409539
1
0. Bencharit S, Morton CL, Hyatt JL, Kuhn P, Danks MK, Potter PM, et al. (2003) Crystal structure of
human carboxylesterase 1 complexed with the Alzheimer's drug tacrine: from binding promiscuity to
selective inhibition. Chem Biol 10: 341–349. PMID: 12725862
1
1. Kroetz DL, McBride OW, Gonzalez FJ (1993) Glycosylation-dependent activity of baculovirus-
expressed human liver carboxylesterases: cDNA cloning and characterization of two highly similar
enzyme forms. Biochemistry 32: 11606–11617. PMID: 8218228
12. Husokawa M (2008) Structure and Catalytic Properties of Carboxylesterase Isozymes involved in meta-
bolic activitation of prodrugs. Molecules 13: 412–431. PMID: 18305428
13. Alam M, Vance DE, Lehner R (2002) Structure-function analysis of human triacylglycerol hydrolase by
site-directed mutagenesis: identification of the catalytic triad and a glycosylation site. Biochemistry 41:
6679–6687. PMID: 12022871
14. Ho SN, Hunt HD, Horton RM, Pullen JK, Pease LR (1989) Site-directed mutagenesis by overlap exten-
sion using the polymerase chain reaction. Gene 77: 51–59. PMID: 2744487
1
5. Nettleship JE, Watson PJ, Rahman-Huq N, Fairall L, Posner MG, Upadhyay A, et al. (2015) Transient
expression in HEK 293 cells: an alternative to E. coli for the production of secreted and intracellular
mammalian proteins. Methods Mol Biol 1258: 209–222. doi: 10.1007/978-1-4939-2205-5_11 PMID:
25447866
1
6. Berrow NS, Alderton D, Sainsbury S, Nettleship J, Assenberg R, Rahman N, et al. (2007) A versatile
ligation-independent cloning method suitable for high-throughput expression screening applications.
Nucleic Acids Res 35: e45. PMID: 17317681
1
7. Reeves PJ, Kim JM, Khorana HG (2002) Structure and function in rhodopsin: a tetracycline-inducible
system in stable mammalian cell lines for high-level expression of opsin mutants. Proc Natl Acad Sci U
S A 99: 13413–13418. PMID: 12370422
18. Aricescu AR, Lu W, Jones EY (2006) A time- and cost-efficient system for high-level protein production
in mammalian cells. Acta Crystallogr D Biol Crystallogr 62: 1243–1250. PMID: 17001101
19. Nettleship JE, Rahman-Huq N, Owens RJ (2009) The production of glycoproteins by transient expres-
sion in Mammalian cells. Methods Mol Biol 498: 245–263. doi: 10.1007/978-1-59745-196-3_16 PMID:
18988030
2
0. Potter PM, Pawlik CA, Morton CL, Naeve CW, Danks MK (1998) Isolation and partial characterization
of a cDNA encoding a rabbit liver carboxylesterase that activates the prodrug irinotecan (CPT-11). Can-
cer Res 58: 2646–2651. PMID: 9635592
2
1. Walter TS, Diprose JM, Mayo CJ, Siebold C, Pickford MG, Carter L, et al. (2005) A procedure for setting
up high-throughput nanolitre crystallization experiments. Crystallization workflow for initial screening,
automated storage, imaging and optimization. Acta Crystallogr D Biol Crystallogr 61: 651–657. PMID:
1
5930615
2. Gorrec F (2009) The MORPHEUS protein crystallization screen. Journal of Applied Crystallography
2: 1035–1042. PMID: 22477774
2
2
4
3. Newman J, Egan D, Walter TS, Meged R, Berry I, Ben Jelloul M, et al. (2005) Towards rationalization of
crystallization screening for small- to medium-sized academic laboratories: the PACT/JCSG+ strategy.
Acta Crystallogr D Biol Crystallogr 61: 1426–1431. PMID: 16204897
2
2
2
2
2
2
4. McCoy AJ, Grosse-Kunstleve RW, Storoni LC, Read RJ (2005) Likelihood-enhanced fast translation
functions. Acta Crystallogr D Biol Crystallogr 61: 458–464. PMID: 15805601
5. Emsley P, Cowtan K (2004) Coot: model-building tools for molecular graphics. Acta Crystallogr D 60:
2126–2132. PMID: 15572765
6. Murshudov GN, Vagin AA, Dodson EJ (1997) Refinement of macromolecular structures by the maxi-
mum-likelihood method. Acta Crystallogr D Biol Crystallogr 53: 240–255. PMID: 15299926
7. Stuart DI, Levine M, Muirhead H, Stammers DK (1979) Crystal structure of cat muscle pyruvate kinase
at a resolution of 2.6 A. J Mol Biol 134: 109–142. PMID: 537059
8. Vaughn JL, Goodwin RH, Tompkins GJ, McCawley P (1977) The establishment of two cell lines from
the insect Spodoptera frugiperda (Lepidoptera; Noctuidae). In Vitro 13: 213–217. PMID: 68913
9. Fleming CD, Edwards CC, Kirby SD, Maxwell DM, Potter PM, Cerasoli DM, et al. (2007) Crystal struc-
tures of human carboxylesterase 1 in covalent complexes with the chemical warfare agents soman and
tabun. Biochemistry 46: 5063–5071. PMID: 17407327
3
0. Hemmert AC, Otto TC, Wierdl M, Edwards CC, Fleming CD, MacDonald M, et al. (2010) Human car-
boxylesterase 1 stereoselectively binds the nerve agent cyclosarin and spontaneously hydrolyzes the
nerve agent sarin. Mol Pharmacol 77: 508–516. doi: 10.1124/mol.109.062356 PMID: 20051531
PLOS ONE | DOI:10.1371/journal.pone.0143919 December 11, 2015
11 / 12

Structure of Aglycosylated Human Carboxylesterase 1
3
3
3
3
3
1. Kroetz DL, McBride OW, Gonzalez FJ (1993) Glycosylation-dependent activity of baculovirus-
expressed human liver carboxylesterases: cDNA cloning and characterization of two highly similar
enzyme forms. Biochemistry 32: 11606–11617. PMID: 8218228
2. Bencharit S, Morton CL, Hyatt JL, Kuhn P, Danks MK, Potter PM, et al. (2003) Crystal structure of
human carboxylesterase 1 complexed with the Alzheimer's drug tacrine: from binding promiscuity to
selective inhibition. Chem Biol 10: 341–349. PMID: 12725862
3. Wadkins RM, Hyatt JL, Edwards CC, Tsurkan L, Redinbo MR, Wheelock CE, et al. (2007) Analysis of
mammalian carboxylesterase inhibition by trifluoromethylketone-containing compounds. Mol Pharma-
col 71: 713–723. PMID: 17167034
4. Parkinson EI, Jason Hatfield M, Tsurkan L, Hyatt JL, Edwards CC, Hicks LD, et al. (2011) Require-
ments for mammalian carboxylesterase inhibition by substituted ethane-1,2-diones. Bioorg Med Chem
19: 4635–4643. doi: 10.1016/j.bmc.2011.06.012 PMID: 21733699
5. Crow JA, Bittles V, Herring KL, Borazjani A, Potter PM, Ross MK (2012) Inhibition of recombinant
human carboxylesterase 1 and 2 and monoacylglycerol lipase by chlorpyrifos oxon, paraoxon and
methyl paraoxon. Toxicol Appl Pharmacol 258: 145–150. doi: 10.1016/j.taap.2011.10.017 PMID:
22100607
3
3
3
3
6. Nishi K, Huang H, Kamita SG, Kim IH, Morisseau C, Hammock BD (2006) Characterization of pyre-
throid hydrolysis by the human liver carboxylesterases hCE-1 and hCE-2. Arch Biochem Biophys 445:
1
15–123. PMID: 16359636
7. Satoh T, Hosokawa M, Atsumi R, Suzuki W, Hakusui H, Nagai E (1994) Metabolic activation of CPT-
1, 7-ethyl-10-[4-(1-piperidino)-1- piperidino]carbonyloxycamptothecin, a novel antitumor agent, by car-
1
boxylesterase. Biol Pharm Bull 17: 662–664. PMID: 7920428
8. Morton CL, Potter PM (2000) Comparison of Escherichia coli, Saccharomyces cerevisiae, Pichia pas-
toris, Spodoptera frugiperda, and COS7 cells for recombinant gene expression. Application to a rabbit
liver carboxylesterase. Mol Biotechnol 16: 193–202. PMID: 11252804
9. Greenblatt HM, Otto TC, Kirkpatrick MG, Kovaleva E, Brown S, Buchman G, et al. (2012) Structure of
recombinant human carboxylesterase 1 isolated from whole cabbage looper larvae. Acta Crystallogr
Sect F Struct Biol Cryst Commun 68: 269–272. doi: 10.1107/S1744309112003326 PMID: 22442219
4
4
0. Chang VT, Crispin M, Aricescu AR, Harvey DJ, Nettleship JE, Fennelly JA, et al. (2007) Glycoprotein
structural genomics: solving the glycosylation problem. Structure 15: 267–273. PMID: 17355862
1. Boonyuen U, Promnares K, Junkree S, Day NP, Imwong M (2015) Efficient in vitro refolding and func-
tional characterization of recombinant human liver carboxylesterase (CES1) expressed in E. coli. Pro-
tein Expr Purif 107: 68–75. doi: 10.1016/j.pep.2014.11.006 PMID: 25462813
4
4
4
4
2. Greenblatt HM, Otto TC, Kirkpatrick MG, Kovaleva E, Brown S, Buchman G, et al. (2012) Structure of
recombinant human carboxylesterase 1 isolated from whole cabbage looper larvae. Acta Crystallogr
Sect F Struct Biol Cryst Commun 68: 269–272. doi: 10.1107/S1744309112003326 PMID: 22442219
3. 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. PMID: 11455023
4. Takahashi S, Katoh M, Saitoh T, Nakajima M, Yokoi T (2008) Allosteric kinetics of human carboxyles-
terase 1: species differences and interindividual variability. J Pharm Sci 97: 5434–5445. doi: 10.1002/
jps.21376 PMID: 18383336
5. Suzuki-Kurasaki M, Yoshioka T, Uematsu T (1997) Purification and characterization of guinea-pig liver
microsomal deacetylase involved in the deacetylation of the O-glucoside of N-hydroxyacetanilide. Bio-
chem J 325 (Pt 1): 155–161. PMID: 9224641
4
4
4
6. Monod J, Wyman J, Changeux JP (1965) On the Nature of Allosteric Transitions: A Plausible Model. J
Mol Biol 12: 88–118. PMID: 14343300
7. Koshland DN G Filmer D (1966) Comparison of experimental binding data and theoretical models in
proteins containing subunits. Biochemistry 5: 365–385. PMID: 5938952
8. Fleming CD, Bencharit S, Edwards CC, Hyatt JL, Tsurkan L, Bai F, et al. (2005) Structural insights into
drug processing by human carboxylesterase 1: tamoxifen, mevastatin, and inhibition by benzil. J Mol
Biol 352: 165–177. PMID: 16081098
49. Karplus PA, Diederichs K (2012) Linking crystallographic model and data quality. Science 336: 1030–
1033. doi: 10.1126/science.1218231 PMID: 22628654
PLOS ONE | DOI:10.1371/journal.pone.0143919 December 11, 2015
12 / 12