Isosorbide-2-benzyl carbamate-5-salicylate, a peripheral anionic site binding subnanomolar selective butyrylcholinesterase inhibitor

Article abstract of DOI:10.1021/jm9014845

Isosorbide-2-benzyl carbamate-5-benzoate is a highly potent and selective BuChE inhibitor. Meanwhile, isosorbide-2-aspirinate-5-salicylate is a highly effective aspirin prodrug that relies on the salicylate portion to interact productively with human BuChE. By integrating the salicylate group into the carbamate design, we have produced isosorbide-2-benzyl carbamate-5-salicylate, an inhibitor of high potency (150 pM) and selectivity for human BuChE over AChE (666000) and CES2 (23000). Modeling and mutant studies indicate that it achieves its exceptional potency because of an interaction with the polar D70/Y332 cluster in the PAS of BuChE in addition to pseudosubstrate interactions with the active site.

Full text of DOI:10.1021/jm9014845

1190 J. Med. Chem. 2010, 53, 1190–1199
DOI: 10.1021/jm9014845
Isosorbide-2-benzyl Carbamate-5-salicylate, A Peripheral Anionic Site Binding Subnanomolar
Selective Butyrylcholinesterase Inhibitor
Ciaran G. Carolan,† Gerald P. Dillon, Denise Khan, Sheila A. Ryder, Joanne M. Gaynor, Sean Reidy, Juan F. Marquez,
Mike Jones, Valerie Holland, and John F. Gilmer*
School of Pharmacy and Pharmaceutical Sciences, Trinity College, Dublin 2, Ireland, and School of Science, Athlone Institute of Technology,
Westmeath, Ireland. †Present address: European Molecular Biology Laboratory, Hamburg 22607, Germany.
Received October 7, 2009
Isosorbide-2-benzyl carbamate-5-benzoate is a highly potent and selective BuChE inhibitor. Mean-
while, isosorbide-2-aspirinate-5-salicylate is a highly effective aspirin prodrug that relies on the
salicylate portion to interact productively with human BuChE. By integrating the salicylate group into
the carbamate design, we have produced isosorbide-2-benzyl carbamate-5-salicylate, an inhibitor of
high potency (150 pM) and selectivity for human BuChE over AChE (666000) and CES2 (23000).
Modeling and mutant studies indicate that it achieves its exceptional potency because of an interaction
with the polar D70/Y332 cluster in the PAS of BuChE in addition to pseudosubstrate interactions with
the active site.
Introduction
open label study, BuChE but not AChE inhibition was
associated with improvements in speed, attention, and
verbal memory-related functions, indicating that BuChE is
important in some aspects of AD pathophysiology.6 Mean-
while, slower deterioration in attention-related performance
and reduced rates of cognitive decline are apparent in AD
patients who have BuChE variants with reduced enzyme
activity.7 While BuChE inhibits amyloid fibril formation in
vitro, this is not related to its enzymatic actions as a
hydrolase so it is possible to envisage an inhibition strategy
that would leave it putative neuroprotective role intact.8
BuChE is important in drug metabolism and detoxifica-
tion,9 but it is also involved in lipid metabolism and regula-
tion.10 Indeed, serum BuChE activity is positively associated
with obesity, cardiovascular disease, low density lipopro-
tein, hepatic adiposity, and diabetes mellitus type II.10a
BuChE inhibition may constitute a useful strategy for
regulating adiposity.
We reported previously that isosorbide-2-carbamate-5-
esters, e.g., 1, arelow nanomolarselectiveinhibitors of BuChE
relative to AChE.11 The compounds show time dependent
inhibition and can be classified as pseudoirreversible inhibi-
tors. They represent a significant departure from therapeuti-
cally used carbamates (e.g., rivastigmine), which are
principally related to the natural product physostigmine
(Figure 1). We subsequently reported the discovery of an iso-
sorbide-based true aspirin prodrug, isosorbide-2-aspirinate-5-
salicylate, 2.12 The 5-salicylate group in 2 was shown to be an
important factor in ensuring that hydrolysis by human
BuChE takes place at the benzoate ester of the 2-aspirinate
group of 2 rather than at the neighboring acetyl group.
Compound 2 is perhaps the first true aspirin prodrug because
of this productive hydrolysis. Modeling studies suggested that
the hydrolysis pattern of 2 is influenced by interactions
between the 5-ester group and a H-bond network involving
D70/Y332 in the PAS and S79 and W82 in the midgorge
The enzymes acetylcholinesterase (AChE,a EC 3.1.1.7) and
butyrylcholinesterase (BuChE, EC 3.1.1.8) regulate choliner-
gic neurotransmission. AChE activity is dominant in the
normal brain (80%) and the overall CNS distribution points
to a supportive role for BuChE which is mainly glial.1 Never-
theless, specific localization of BuChE may be important.
Individuals with the BuChE silent genotype have been re-
ported to have subtly different cognition to the normal
genotype.2 The importance of BuChE in normal cholinergic
processing has been underlined with the characterization of
the AChE nullizygote mouse.3
In Alzheimer’s disease (AD), the balance of AChE and
BuChE changes. AChE has been reported to be significantly
reduced in the cortex and hippocampus in post mortem AD
brain slices, whereas BuChE levels in these areas is signifi-
cantly increased.4 Dual inhibitors such as the compound
rivastigmine have shown significant clinical utility in con-
ferring cognitive improvements. In rats, administration of
the BuChE selective inhibitor cymserine caused significant
increase in ACh levels (up to 300% of those recorded prior
to administration).5 The inhibitor caused significant im-
provements in the cognitive performance of aged rats using
a water maze model. This was consistent with evidence of
long-term potentiation (LTP) in rat brain slices. In a small
*To whom correspondence should be addressed. Phone: þ353-1-896
2795. Fax: þ353-1-896 2793. E-mail: gilmerjf@tcd.ie.
a Abbreviations: AChE, acetylcholinesterase; ATCI, acetylthiocho-
line iodide; BTCI, butyrylthiocholine iodide; BuChE, butyrylcholines-
terase; CES2, (carboxylesterase-2); DCC, dicyclohexylcarbodiimide;
DTNB, 5,50-dithiobis-(2-nitrobenzoic acid); huBuCHE, human butyr-
ylcholinesterase; ISMN, isosorbide-5-mononitrate; ISMNA, isosorbide
mononitrate aspirinate; MAO, monoamino oxidase; PAS, peripheral
anionic site; SAR, structure-activity relationship, TBAF, tetrabuty-
lammonium fluoride; TBDMS, tert-butyldimethylsilyl; TLC, thin layer
chromatography; TMS, tetramethylsilane.
r
pubs.acs.org/jmc
Published on Web 01/12/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1191
Figure 1. Isosorbide-based inhibitor 1, true aspirin prodrug 2, leading to high potency 2-carbamate-5-salicylate ester type 3.
Table 1. Test Compounds along with IC50 Values for BuChE and %
the phenolic -OH group of the hydroxybenzoic acids before
esterification because of competition between the phenolic
Inhibition of AChE (%I) at 100 μM
BuChEb
cAChE
-OH and isosorbide-OH group. The relevant benzyl pro-
tected acids were obtained as described previously.12,13 The
amines (11-13) were produced by standard reduction of the
nitro series (8-10). A substantial amount of the ortho-nitroso
compound (14) was isolated from the mixture following
reduction of the ortho-nitrobenzoate, and this was character-
ized and tested too. Enzyme inhibition was evaluated using
Ellman’s spectrophotometric determination of cholinesterase
activity with either human plasma or purified BuChE from
human plasma. For AChE measurements, electric eel or
human erythrocyte enzyme was used as specified in Table 1.
Butyrylthiocholine was used as a substrate for BuChE and
acetylthiocholine as AChE substrate. Evaluations were made
following 30 min incubation time, as this was shown before to
allow optimal interaction between isosorbide-based inhibitors
and BuChE11 and because it was intended to kinetically
characterize the more interesting compounds anyway.
IC50 nM
(95% CI)
%I at
100 μM
compd
ratiod
5
0.15 (0.13-0.21)
16 (14.3-18.8)
323 (251-420)
18 (15-21)
32 (27-37)
117 (100-136)
27 (24-31)
86 (72-101)
297 (251-351)
39 (33-45)
1.2 (1.1-4.2)
9.6 (7.9-11.7)
6.4 (4.2-9.7)
2609 (2204-3089)
1057 (871-1282)
47 ( 7 (52e)
nsa
ns
>6.6 ꢀ 105
>6250
>309
6
7
8
ns
ns
>5555
>3125
>854
9
10
11
12
13
14
15
18
19
20
21
27 ( 2
ns
ns
>3703
>1166
336
33 ( 9
ns
>2564
>83333
>10416
>15625
>38
26 ( 2
16 ( 3
ns
ns
12 ( 5
>95
The test compounds were potent and selective inhibitors of
BuChE (Table 1). The ortho-hydroxybenzoate (5) is one of
the most potent cholinesterase inhibitors discovered (IC50 of
151 pM) and the most highly selective over AChE (BuChE/
AChE > 500000). Previously reported isosoride-based car-
bamates were also highly selective toward BuChE.11 Interest-
ingly some isosorbide-diesters are moderately potent mixed-
mode inhibitors of AChE but exceptionally good substrates
for BuChE. Kinetic and modeling data indicate that these
compounds bind in the AChE gorge distal to the catalytic
subsite and evidently do not interact favorably for turnover.14
The diesters and analogous carbamate compounds appear to
fit BuChE very well but are probably too large for processing
in the smaller AChE catalytic site.
Compound5 is 30-fold more potent than the corresponding
unsubsituted-5-benzoate 1 using the 30 min IC50 values. This
increase in potency was gratifying because it followed directly
from the recognition that the salicylate group in the aspirin
prodrug series (2) promotes productive interactions with
BuChE. For example, isosorbide-2-aspirinate-5-salicylate is
a far more efficient aspirin prodrug than its 5-benzoate
analogue, which is hydrolyzed largely at the acetyl group,
producing the isosorbide-2-salicylate-5-benzoate.13
physostigmine
19 (15-24)f
53 (34-71)f
2.7
a ns means that the estimated inhibition value was not significantly
different from zero. b Measurement made with human BuChE (95% CI =
95% confidence interval). c Measurement made with electric eel AChE.
d A selectivity ratio was estimated using a value of 100 μM for AChE
because the IC50 value was not reached at less than this. e Measurement
made with human erythrocyte AChE rather than electric eel. f IC50
values (nM).
region of BuChE.13 Because the salicylate group was appar-
ently required for specific processing of 2, we hypothesized
that introduction of similar 5-ester groups would increase
inhibitory potency in the parallel isosorbide-carbamate class.
One such salicylate-carbamate is a subnanomolar, highly
specific inhibitor of BuChE over related esterases. It appears
to achieve high potency because of binding to D70/Y332 in
the PAS.
Chemistry, Enzyme Inhibition, Selectivity, and Stability
The key intermediate 4 was obtained in two steps from
isosorbide mononitrate (ISMN).11 It was possible to produce
some of the target compounds by direct esterification of 4 with
the appropriate acid chloride or with the carboxylic acid and a
coupling agent. This was the case for the nitro-substituted
benzoates (8-10) and the toluate (15). However, in preparing
the salicylate 5 and isomers (6-7), it was necessary to protect
The isomeric meta- and para-hydroxybenzoates 6 and 7
were significantly less potent than 5 (16 and 323 nM, re-
spectively). Equally, as reported, the meta and para-hydroxy

1192 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Carolan et al.
Scheme 1. Showing Preparation of 2-Benzylcarbamate-5-Sub-
stituted Benzoates
Figure 2. Inhibition of BuChE enzyme (E) by carbamates CX
begins with a reversible interaction forming the complex E.CX,
which is analogous to the Michaelis complex (Kc) in enzyme
substrate processing. This is followed by carbamylation, which
has a unimolecular rate constant k3. The overall carbamylation
event is characterized by the second-order rate constant ki, which
can be estimated from k3/Kc. k4 represents the rate of recovery of
enzyme activity. Decarbamylation kinetics following inhibition by
isosorbide benzyl- and butyl-carbamate inhibitors was shown to be
very slow previously and can be ignored if the initial phase of the
reaction is monitored.15
carbamylation and deprotection (Scheme 2). Isosorbide-2-
TBDMS (15) was obtained from ISMN in two steps and
esterified with benzyl-protected salicylic acid. The 2-OH was
deprotected (TBAF), carbamylated with the appropriate
primary or secondary carbamyl chloride, and the salicylate-
OH group deprotected (H2, Pd/C). The dimethyl and ethyl-
(i) BnOPhC(O)OH, DCC, DMAP, DCM, RT, 8 h; (ii) Pd/C, H2,
EtOAc, 12 h, RT; (iii) NO2PhC(O)Cl, Et3N, DCM, RT, 8 h or NO2PhC-
(O)OH, DCC, DMAP, DCM, RT, 8 h; (iv) H2, Pd/C, MeOH:EtOAc,
24 h, RT.
carbamate (18, 19) were potent inhibitors of BuChE (IC50
<
10 nM), but the morpholine and diethylcarbamate com-
pounds (20, 21) were micromolar inhibitors. There was a
270-fold difference in potency between the ethyl (18, IC50 9.6 nM)
and diethyl carbamate (20, IC50 2.6 μM). This appears to be
due to steric reasons rather than loss of interactions with the
carbamate NH because the dimethyl compound (19) was equipo-
tent (IC50 6.4 nM) with 18.
Scheme 2. Showing Preparation of 5-Salicylate-2-alkyl Carba-
mates
We also studied the effect of 5 on two closely related
carboxylesterase enzymes human carboxylesterase-1 (CES1)
and carboxylesterase-2 (CES2) present in hepatocyte and
intestinal enterocyte microsomes respectively using para-
nitrophenyl acetate (pNPA) as substrate.15 CES-1 and -2 have
highhomologyand sequence identitywithBuChE (40-50%),
and they too are involved in drug metabolism and prodrug
activation. However, they are markedly less efficient at pro-
cessing positively charged esters such as choline esters and
therefore classical amino-inhibitors tend to be selective to-
ward the cholinesterases. Compound 5 was not an inhibitor of
p-NPA turnover by human hepatic CES-1 but it was moder-
ately potent toward CES-2 (IC50 = 3.42 ( 1.18 μM: BuChE/
CES-2 selectivity 22800).
(i) Benzylprotected salicylic acid, DCC, DMAP, DCM, RT, 8 h; (ii)
TBAF, DCM, RT, 15 min; (iii) carbamoyl chloride, pyr, 80 ꢀC, 24 h; (iv)
H2, Pd/C, MeOH:EtOAc, 24 h, RT.
Another factor likely to influence the distribution and
potential utility of 5 is the vulnerability of its 5-salicylate ester
group toward nonspecific hydrolysis. The stability of 5 was
therefore evaluated in human plasma solution (33-50%),
mouse plasma (33%), and mouse liver homogenate. Remain-
ing ester (1 ꢀ 10-4 M) as a function of time was determined
using HPLC. There was no evidence of hydrolysis in any of
these matrices over the course of 1 h.
benzoates are not as effective as aspirin prodrugs as the ortho
compounds.13 This general pattern in the inhibitor class
was followed by the analogous nitro- and aminobenzoates
(8-10). The ortho-toluate (15) was also a highly potent
compound.
As the salicylate 5 exhibited such exceptional potency, we
decided to examine the effect of replacement of the 2-benzyl-
carbamate group with other carbamate groups while main-
taining the 5-salicylate ester1. In our earlier work, we had
shown that in the isosorbide-based compounds, phenyl car-
bamates tend not to be potent BuChE inhibitors, indeed they
are moderately potent AChE inhibitors.11 We therefore
decided to examine a small panel of primary and secondary
alkyl carbamates. In preparing these, it was necessary to
introduce the benzyl protected salicylate first followed by
Enzyme Inhibition Kinetic Studies
Toinvestigatethe reasonsforthe increasedpotency of5 and
its ethyl carbamate analogue 18 relative to the unsubstituted
benzoates, we undertook detailed enzyme inhibition studies
with some of the key compounds including 5, the toluate 15,
18, and related compoundswhoseIC50 values wehad reported

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1193
Figure 3. First-order decay curves for (a) the inhibition of WT BuChE by compound 5 (312 pM-62.5 nM); (b) inhibition of the D70G BuChE
mutant by compound 5 (625 pM-1 μM). A value of kobs, the observed rate constant for inhibition, was calculated from each curve in these plots
and plotted against concentration as shown in the second column below.
previously; thebenzoate1, 2-butylcarbamate-5-benzoate(22),
the unsubstituted isosorbide-2-benzyl carbamate (4), 2-ben-
zylcarbamate-5-nicotinate (23), and 2-benzylcarbamate-5-
biphenylester (24). For comparison, we included also the
clinically used rivastigmine and the classical agent physostig-
mine (eserine) (Figure 1).
Carbamate inhibition of cholinesterases occurs through
initial reversible interaction followed by carbamylation of
the serine of the catalytic triad (Figure 2).16a-c Decay of
enzyme activity follows apparent first-order kinetics when
the inhibitor concentration [CX] is higher than that of
the enzyme [E]. The rate of carbamylation of the enzyme
(kobs) can be estimated by measuring residual enzyme
activity at intervals following mixing of enzyme and inhi-
bitor (Figure 3):
for corresponding ethyl- and butyl- compounds (5/18, 1/22).
Compound 5 exhibited the greatest intrinsic affinity as
reflected in its low nanomolar Kc value, and it had the highest
overall carbamylation efficiency (ki = 201 μM-1 min-1). The
3
Kc valuesfor 5 and theunsubstituted benzoate1 imply that the
simple introduction of the salicylate-OH group increases
enzyme affinity 100 fold. Carbamylation rate constants were
similar across the series apart from two cases, the weak
inhibitor 4 and the potent biphenylester 24, which is probably
explained by its size and slower diffusion speed. Comparison
of the latter with the nicotinate 23 is instructive. These two
compounds are similar in 30 min IC50 values and overall
carbamylation efficiencies (1.5), however, 23 has relative to 24
a high k3 but significantly lower intrinsic affinity (Kc). There-
fore its potency arises because its affinity compensates for its
slower chemistry. Overall though, there was a significant
correlation between Kc and 30 min IC50 (r2 = 0.85). Values
for k5, the decarbamylation reaction, were estimated for the
potent compounds 5 and rivastigmine (Figure 1). Enzyme
recovery was initiated by diluting inhibited enzyme 1000-fold,
and the rate of recovery was monitored over an extended
period of time thereafter. Recovery of enzyme activity
followed first-order curves in each case from which values of
0.32 h-1 and 2.76 h-1 were obtained for 5 and rivastigmine,
respectively. Clearly, the benzyl carbamate adduct resulting
from carbamylation with 5 is stable for relatively long periods
compared to the adduct generated by rivastigmine. This could
be due to its intrinsic stability or to suppression of its hydro-
lysis by the leaving group. Leaving group effects on BuChE
decarbamylation are not well characterized, but the phenom-
enon has been reported for AChE.17
R ¼ R0e-kobs t þ R¥
(1Þ
3
where R, R0, and R¥ are ratios of the inhibited enzyme
activity (vi) to the control activity (v0) at times t, 0, and, t¥,
respectively.16 Plots of kobs versus inhibitor concentration
follows a rectangular pseudo Michaelis Menten curve from
which it was possible to estimate Kc, which characterizes the
initial equilibrium and k3 the unimolecular carbamylation
rate constant:
k3½CXꢁ
kobs
¼
(2Þ
Kcð1 þ ½Sꢁ=KmÞ þ ½CXꢁ
The Kc value can be interpreted as reflecting the affinity
between inhibitor and cholinesterase resulting from topologi-
cal or electronic complementarity: the k3 value, the maximum
velocity of carbamylation (Figure 2). The second-order rate
constant ki (= k3/Kc) is a measure of overall carbamylation
efficiency. We adopted this approach in order to separate out
thesuccessivephases of the reaction and todeterminethe basis
for the differences in 30 min IC50 values potency already
observed (Table 2). Kc values were lowest for the salicylates 5
and 18. They were also lower for the 2-benzyl carbamate than
Modeling of Carbamate Binding to BuChE
The crystalstructuresofthe cholinesteraseenzymes, bothin
their apo- forms and in complex with a variety of ligands, have
been published,18-21 and the structural basis of substrate and
inhibitor processing is well understood. To investigate the
reasons for the marked increase in potency associated with
salicylate compounds, the test compounds were docked to the

1194 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Carolan et al.
Table 2. Kinetic Values for Enzyme Carbamylation of WT Human BuChEa
a Values given ( SEM: † IC50 values from ref 11, ‡ from ref 27.
BuChE active site gorge using Autodock,22 which has
provided accurate results for ligand binding to cholines-
terase enzyme in the past.23,24 Models of the test com-
pounds were flexibly docked in a model of BuChE based
on the X-ray crystal structure, with some key residues
allowing flexibility (depicted in Figures 4 and 5 for 5 and
18). The highest scoring poses in each case could be inter-
preted as Michaelis complexes, with the isosorbide group
close to the base of the gorge, the carbamate in the acyl
pocket, and the serine oriented for attack close to the
carbamoyl carbonyl. The poses were overall similar across
the family of compounds. There are at least two possible
explanations for the ortho effect observed. First, the ortho-
substituted compounds are subject to intramolecular bond-
ing between the phenyl substituent and the carbonyl oxygen
of the ester group. The “ortho effect” is most apparent for
the salicylate compound 5 and the amino-substituted com-
pound 11. In both cases, an intramolecular hydrogen bond
exists between the substituent at the ortho position and the
carbonyl oxygen of the ester group in the enzyme bound
models. The interaction is prominent in the 1H NMR
spectra of the compounds in chloroform. This bond could
act as a preselection mechanism, preorganizing the ligand
so that the phenyl ring is planar with the partner indole
system of W82. As a result, there is a less significant loss of
entropy associated with binding of these ligands compared
to those substituted at the meta and para positions.
Notably, the modeling indicated potentially significant
interactions between the salicylate-OH group and members
of the PAS and H-bonded in the wild type BuChE. The
BuChE PAS plays a key role in initial substrate binding and
steering and in the excess characteristic substrate activation
of BuChE being dynamically linked through the Ω loop to
W82 of the cation-π site.25,26 As shown in Figures 4 and 5,
the ortho-hydroxyl group of compounds 5 and 18 can
hydrogen-bond to the side chains of D70 and Y332 and,
upon 180ꢀ rotation of the benzene ring about its axis, with
the OH group of Y440. The position of the -OH group
ortho to the benzene ring appears optimal for such interac-
tions across the whole 360ꢀ rotation about its axis. The
aromatic PAS of AChE catalyzes the assembly of amyloid
fibrils in vitro; normal substrate interactions there are asso-
ciated with inhibition of turnover. BuChE-selective inhibi-
tors have been reported to reduce Aβ and APP secretion in
vitro and in vivo, but it is not clear if the BuChE PAS is
involved in this.5 The acyl pocket of the enzyme is large but
relatively narrow, framed as it is by residues L286 and V288.
The benzyl group occupies it especially well (Figure 4).

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1195
Smaller methyl and ethyl groups (compounds 18 and 19) are
easily accommodated but form less productive contacts with
surrounding residues on account of their smaller size (e.g.,
Figure 5). Disubstituted carbamates larger than methyl are
too large to fit in the pocket, and thus compounds 20 and 21
are poor inhibitors of BuChE.
Mutant Studies
Modeling suggested that the increased binding affinity,
reflected in lower Kc values (Table 2) could be attributed to
specific associations between the OH group of 5 and 18 and a
H-bond network encompassing PAS residues D70 and Y332.
We studied the kinetics of inhibition by selected carbamate
inhibitors with D70G and Y332A mutants as well as a D70G/
Y332A double mutant (Table 3). The salicylate compound 5
bound less avidly to the mutant enzymes as reflected in
significant increase in Kc (Kc 9.9/445 nM). Similarly, com-
pound 18 preferentially binds the wild type enzyme compared
to the mutant enzymes, which can also be attributed to loss of
specific associations with the mutated residues. The contrast
with the results from the benzoate compounds 1 and 22, which
lack the important ortho-hydroxyl group, is notable;while
the benzoate 1 does form less favorable contacts with the
mutant enzymes than the wild type enzyme (indicating that
the peripheral site residues are important for general ligand
binding), the difference in Kc values between wild type and
mutant enzymes is much less significant for 1 and 22 than for
the salicylate compounds 5 and 18.
Figure 4. Binding of isosorbide 2-benzyl carbamate 5-salicylate (5)
to huBChE, as predicted by docking the ligand to the crystal structure
using Autodock 4. With the carbamate moiety suitably poised for
interaction with the enzyme’s catalytic machinery, the ligand is in an
orientation in which the -OH group of the 5-salicylate can interact
with Y440, or by rotation about the axis of the aromatic ring, with
D70 and Y332. Such interactions stabilize the ligand in the orienta-
tion indicated, increasing the rate of Michaelis complex formation
and, thus, the rate of overall enzyme inhibition.
To examine the extent to which specific interactions might
be responsible for some of the improvement in inhibitor
binding observed when the inhibitors interact with wild type
compared to mutant enzymes, we examined the free energies
of binding for each ligand to the enzymes. The change in free
energy of initial binding caused by mutations, ΔΔGmut-wt was
calculated according to the following equation:
ΔΔGmut-wt ¼ -RT lnðKcðmutÞ=KcðwtÞÞ
where Kc(mut) and Kc(wt) are the dissociation constants for
Michaelis complex formation in the mutant and wild-type
enzymes respectively. R is the gas constant (1.99 ꢀ 10-3
kcal mol-1
K
-1), and T was taken as 310 K. ΔG of wild
3
3
type BChE was taken as a reference value.
For compound 5, the values of ΔΔG for binding to the
D70G and Y332A mutants were thus found to be -1.17 kcal
3
mol-1 and -1.00 kcal mol-1, respectively. In the double
3
mutant, the calculated value was -2.35 kcal mol-1. The
3
values are almost but not quite additive, suggesting that
conformational changes associated with the double mutation
contribute to reduced affinity for the ligand as well as the loss
of contacts with the ligand shown in the models previously.
For the ethyl carbamate 5-salicylate compound (18), the
values of ΔΔG for binding to the D70G and Y332A mutants
Figure 5. Binding of isosorbide 2-ethyl carbamate 5-salicylate (18) to
huBChE, as predicted by docking the compound to the crystal structure
of the enzyme using the Autodock program. With the carbamate
functionality oriented toward Ser-198 as required for enzyme carba-
mylation, the hydroxyl group of the ligand’s salicylate moiety can
interact with D70 and Y332 at the peripheral site of BuChE, strength-
ening interactions and increasing the rate of enzyme inhibition.
were calculated as -0.89 and -1.06 kcal mol-1, respectively,
3
Table 3. Kinetic Values for Interaction with Mutant BuChE Enzymes
Kc
(μM)
k3
(min-1
)
D70G
Y332A
D70G/Y332A
D70G
Y332A
D70G/Y332A
5
0.067 ( 0.021
2.4 ( 1.3
0.32 ( 0.04
2.7 ( 0.6
0.050 ( 0.006
1.5 ( 0.7
0.42 ( 0.06
1.3 ( 1.1
0.445 ( 0.106
6.4 ( 3.1
2.26 ( 0.40
5.3 ( 0.9
2.7 ( 0.2
3.6 ( 0.9
2.5 ( 0.1
2.6 ( 0.2
2.7 ( 0.2
1.3 ( 0.2
2.9 ( 0.1
1.3 ( 0.6
3.1 ( 0.1
2.0 ( 0.4
2.3 ( 0.2
1.7 ( 0.1
1
18
22

1196 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Carolan et al.
and in the double mutant, -2.10 kcal mol-1. By comparison,
added 1 mol equiv of 2-(benzylaminocarbonyloxy-)-1,4:3,6-
dianhydro-D-glucitol 4 (1 equiv), 1.1 mol equiv of DCC (497.2
mg), and DMAP (294.4 mg). The mixture was stirred overnight
while the reaction vessel was allowed to return to room tem-
perature. The organic solvent was evaporated under vacuum,
DCM (10 mL) was added, and the urea precipitate was collected
by filtration. The filtrate was washed with 1 M HCl (20 mL), 5%
NaHCO3 (20 mL) saturated brine solution (20 mL), and dried
over anhydrous sodium sulfate (1 g). The solution was filtered
into a round-bottom flask and evaporated under vacuum. The
product was dissolved in ethyl acetate/methanol (1:1, 20 mL). A
spatula tip-full of 10% palladium on activated carbon was
added. Air was expelled from the flask and the mixture kept
under an atmosphere of H2 and stirred for 24 h. TLC showed the
formation of a single product. The palladium catalyst by was
removed by filtration and the solvent evaporated under vacuum.
The crude product was dissolved in DCM and purified by flash
chromatography.
3
valuesof-0.60, -0.31, and-1.21 kcal mol-1 werecalculated
3
for the benzoate compound 1. The loss in free energy of
binding of the benzoate compound upon mutation of the
peripheral site residues is much less significant than that for
the salicylate compounds, suggesting that the proposed inter-
actions between the salicylate -OH group and the side chains
of those residues are important. The mutant enzymes have
different conformations compared to the wild type enzyme,
and the hydration state of the active site gorge may change
upon mutation.28 These phenomena may explain some of the
decreased binding to mutant enzymes observed for our
ligands; however, the direct hydrogen-bonding interactions
proposed in our molecular models provide an attractive
explanation for the marked effect of peripheral site mutation
on inhibitor potency.
General Procedure 2 (GP2). To a solution of 2-(benzyl-
aminocarbonyloxy-)-1,4:3,6-dianhydro-D-glucitol 4 (1 equiv)
in anhydrous DCM was added anhydrous triethylamine (1.1
equiv), followed by carboxylic acid chloride (1.1 equiv). The
reaction mixture was left stirring for 12 h under a nitrogen
atmosphere with the absence of light. The reaction mixture was
then washed successively with 1 M HCl, saturated NaHCO3
(aq), water, and brine. The solution was then dried with Na2SO4,
filtered, and solvent removed under reduced pressure.
General Procedure 3 (GP3). 2-(Hydroxyl-)-5-O-(o-benzyl-
oxybenzoyl)-1,4:3,6-dianhydro-D-glucitol, 17 (1 equiv), was dis-
solved in anhydrous pyridine and an excess of appropriate
carbamoyl chloride was added. The reaction mixture was heated
to 120 ꢀC for 24 h and kept under an atmosphere of nitrogen.
The mixture was evaporated to dryness under high vacuum to
remove pyridine, and the residue was diluted with DCM (20 mL)
and washed with water (3 ꢀ 20 mL). The organic layer was dried
over anhydrous sodium sulfate (2 g) and evaporated to dryness.
Purification by column chromatography afforded pure title
compound.
Conclusion
Substrates are not always suitable templates for the design
of inhibitors because of the differing geometric and electronic
demands for processing and inhibition. In this study, we have
designed a subnanomolar inhibitor based directly on the as-
pirin prodrug 2. Contributions from peripheral or midgorge
interactions to the enhanced potency were demonstrated as
feasible using modeling and supported by mutant studies.
These kinds of interactions have previously been shown to be
important in reinforcing binding and affinity in AChE in-
hibitor types, but they clearly have utility in the design of
inhibitors for the less explored but potentially important
target BuChE. Compound 5 is one of the most potent
cholinesterase inhibitors discovered, and it is the most selec-
tive. It has significant potential in biochemical and pharma-
cology studies delineating the role and therapeutic value of
BuChE as a target in human disease.
2-(Benzylaminocarbonyloxy-) 5-O-Salicyloyl-1,4:3,6-dianhy-
dro-D-glucitol (5). 2-Benzyloxy-benzoic acid (2.19 mmol, 500
mg) and 2-(benzylaminocarbonyloxy-)-1,4:3,6-dianhydro-D-
glucitol 4 (1.22 g) were reacted according to GP1 to give the
title compound as a white crystalline product after flash chro-
matography using hexane and ethyl acetate (1:1) as eluent (485.5 mg,
Experimental Section
1
Synthesis. H and 13C spectra were recorded at 27 ꢀC on a
Bruker DPX 400 MHz FT NMR spectrometer (400.13 MHz 1H,
100.61 MHz 13C) or a Bruker AV600 (600.13 MHz 1H,
150.6 MHz 13C) in either CDCl3 or (CD3)2CO with tetramethyl-
silane (TMS) as internal standard. In CDCl3, 1H spectra were
assigned relative to the TMS peak at 0.0 ppm, and 13C spectra
were assigned relative to the middle CDCl3 triplet at 77.00 ppm.
In (CD3)2CO, 1H spectra were assigned relative to the
(CD3)2CO peak at 2.05 ppm and 13C spectra were assigned
relative to the (CD3)2CO at 29.5 ppm. Coupling constants were
reported in hertz (Hz). HRMS was performed using a Miromass
mass spectrophotometer with electrospray ionization at the
School of Chemistry, Trinity College Dublin. Elemental ana-
lyses were performed at the Microanalytical Laboratory, De-
partment of Chemistry, University College Dublin. Flash
chromatography was performed on Merck Kieselgel (particle
size: 40-60 μm). Thin layer chromatography (TLC) was per-
formed on silica gel Merck F-254 plates. Compounds were
visually detected by UV absorbance at 254 nm or with potas-
sium permanganate or vanillin staining solution. Enzyme activ-
ity and inhibition assays were performed using an Anthos bt2
plate reader with UV detection at 412 nm. Purity was deter-
mined by elemental analysis or RPHPLC. All test compounds
were >95% pure by HPLC and/or elemental analysis.
1
55.5%); mp 120 ꢀC. H NMR δ (CDCl3): 3.94-4.10 (m, 4H),
4.34-4.42 (d, 2H, J = 6.02 Hz), 4.56 (d, 1H, J = 5.02 Hz), 4.96
(t, 1H, J = 5.27 Hz), 5.12 (t, 1H, J = 5.27 Hz), 5.22 (d, 1H, J =
3.01 Hz), 5.42 (dd, 1H, J = 5.52 and 10.04 Hz), 6.89-6.96 (m,
1H), 7.01 (dd, 1H, J = 1.01 and 8.54 Hz), 7.23-7.40 (m, 5H),
7.51 (m, 1H), 7.89 (dd, 1H, J = 1.5 and 8.03 Hz), 10.60 (s, 1H).
13C NMR ppm (CDCl3): 44.7, 70.3, 73.2, 74.4, 77.9, 80.5, 85.8,
111.4, 117.2, 118.8, 127.1, 127.2, 128.3, 129.5, 135.6, 137.6,
154.7, 161.2, 168.9.
2-(Ethylaminocarbonyloxy-) 5-O-(o-Benzyloxy benzoyl)-1,4:
3,6-dianhydro-D-glucitol. Compound 17 (0.5612 mmol, 200
mg) and ethyl carbamoyl chloride were reacted according to
GP3. Purification by column chromatography, using ethyl
acetate and hexane (3:1, 2:1, 1:1) as eluent, yielded the target
1
compound as a colorless oil (153.8 mg, 64.1%). H NMR δ
(CDCl3): 1.14 (t, 3H, J = 7.28 Hz), 3.21 (m, 2H), 3.87 (dd, 1H,
J = 5.52 and 10.04 Hz), 3.91-4.03 (m, 3H), 4.53 (d, 1H, J =
4.51 Hz), 4.83 (s, 1H), 4.93 (t, 1H, J = 5.02 Hz), 5.12 (s, 1H), 5.19
(s, 2H), 5.37 (dd, 1H, J = 5.52 and 11.54 Hz), 6.98-7.07 (m,
2H), 7.27-7.54 (m, 6H), 7.90 (dd, 1H, J = 2.01 and 8.03 Hz).
13C NMR ppm (CDCl3): 14.7, 35.4, 69.9, 70.1, 73.1, 73.9, 77.8,
80.3, 85.7, 113.2, 119.3, 120.1, 126.7, 127.5, 128.1, 131.8, 133.4,
136.1, 154.7, 157.9, 165.2.
General Procedure (GP1) for the Synthesis of Hydroxybenzoyl
Esters of 2-(Benzylaminocarbonyloxy-)-1,4:3,6-dianhydro-D-glu-
citol. Appropriate benzyloxy-benzoic acid (2.1907 mmol, 500
mg) was dissolved in chloroform (20 mL) in a 100 mL round-
bottom flask and cooled to -20 ꢀC in a Dewar flask. To this was
2-(Ethylaminocarbonyloxy-) 5-O-Salicyloyl-1,4:3,6-dianhydro-D-
glucitol (18). Hydrogenation of its benzyl protected precursor
in the presence of palladium on carbon gave 18 as a white

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1197
1
crystalline product (69.4 mg, 58.6%); mp 118 ꢀC. H NMR δ
(CDCl3): 1.14 (t, 3H, J = 7.28 Hz), 3.19 (m, 2H), 3.92-4.10 (m,
4H), 4.55 (d, 1H, J=4.52Hz),4.84(s,1H),4.97(t,1H,J=5.27Hz),
5.21 (s, 1H), 5.41 (dd, 1H, J = 5.02 and 9.54 Hz), 6.91 (m, 1H),
6.99 (dd, 1H, J = 1.0 and 8.53 Hz), 7.48 (m, 1H), 7.87 (dd, 1H,
J = 1.51 and 8.03 Hz), 10.58 (s, 1H). 13C NMR ppm (CDCl3):
14.7, 35.5, 70.2, 73., 74.4, 77.5, 80.5, 85.9, 111.4, 117.2, 118.9,
129.5, 135.6, 154.6, 161.2, 168.9.
from male donors were obtained from BD Biosciences (0.5 mL
of a 20 mg mL-1 solution, lot no. 36170). The sample obtained
was diluted in phosphate buffer pH 7.4, producing a stock
3
solution of 2 mg mL-1. To minimize repeated freezing and
3
thawing, the sample was stored in 500 μL aliquots at -80 ꢀC
until use. Pooled human intestinal microsomes (lot no. 84921)
were obtained as a 20 mg mL-1 solution and diluted with
phosphate buffer pH 7.4 to a stock concentration of 80
3
μg mL-1. They were stored in 500 μL aliquots at -80 ꢀC until
2-(Dimethylaminocarbonyloxy-) 5-O-(o-Benzyloxy-benzoyl)-
1,4:3,6-dianhydro-D-glucitol. Compound 17 (0.5612 mmol,
200 mg) and dimethyl carbamoyl chloride (2.2448 mmol,
241.4 mg, 0.2067 mL) was reacted according to GP3. Purifica-
tion by column chromatography, using ethyl acetate and hexane
(3:1, 2:1, 1:1) as eluent, yielded the target compound as a clear oil
(186.2 mg, 77.6%). 1H NMR δ (CDCl3): 3.88 (dd, 1H, J = 6.03
and 10.04), 3.93-4.05 (m, 3H), 4.57 (d, 1H, J = 4.52 Hz), 4.97 (t,
1H, J = 5.02 Hz), 5.11 (s, 1H), 5.19 (s, 2H), 5.37 (dd, 1H, J =
6.03 and 11.55 Hz), 6.97-7.07 (m, 2H), 7.27-7.55 (m, 6H), 7.91
(dd, 1H, J = 1.50 and 8.03). 13C NMR ppm (CDCl3): 35.5, 35.9,
69.8, 70.1, 73.2, 74.0, 78.5, 80.3, 85.7, 113.1, 119.3, 120.1, 126.7,
127.4, 128.1, 131.8, 133.4, 136.1, 154.9, 157.9, 165.3.
2-(Dimethylaminocarbonyloxy-) 5-O-Salicyloyl-1,4:3,6-dia-
nhydro-D-glucitol (19). Hydrogenation of its benzyl protected
precursor in the presence of palladium on carbon gave 19 as a
clear oil (73.8 mg, 62.3%). 1H NMR δ (CDCl3): 2.90 (d, 6H, J =
15.1), 3.95-4.08 (m, 4H), 4.57 (d, 1H, J = 4.52 Hz), 4.99 (t, 1H,
J = 5.02 Hz), 5.19 (s, 1H), 5.41 (q, 1H, J = 5.52 and 10.04 Hz),
6.91 (m, 1H), 7.00 (d, 1H, J = 8.53 Hz), 7.48 (m, 1H), 7.88 (dd,
1H, J = 1.51 and 7.53 Hz), 10.59 (s, 1H). 13C NMR ppm
(CDCl3): 35.5, 35.9, 70.1, 73.4, 74.4, 78.1, 80.5, 85.9, 111.4,
117.2, 118.9, 129.5, 135.6, 154.8, 161.3, 168.9. HRMS (M þ 23):
C16H19NO7Na requires 360.1054; found 360.1059.
Determination of Enzyme Activity and Inhibition. Cholines-
terases. First, 1 M solutions of each inhibitor were prepared in
10 mL of acetonitrile/distilled water (1:1) [25 μL of a 1 M
inhibitor solution in 250 μL of test solution (see below) gave
an inhibitor concentration of 100 mM]. All materials were
purchased from Sigma Aldrich Ireland. Then butyrylthiocho-
line iodide (BTCI)/BTCI (15.9 mg) was dissolved in 10 mL of
phosphate buffer pH 8.0 [25 μL of this solution in 250 μL of test
solution will give a concentration of 0.5 mM]. ATCI (14.5 mg)
was dissolved in 10 mL of phosphate buffer pH 8.0 [25 μL of this
solution in 250 μL of test solution will give a concentration of
0.5 mmol]. For human plasma BuChE, human blood samples
were collected by venipuncture into Li-Heparin Sarstedt
Monovette tubes (9 mL). Plasma was obtained by centrifugation
at 10000 rpm for 5 min. Plasma was stored at 2-6 ꢀC. A plasma
solution for the activity/inhibition assay was prepared by dilut-
ing 1 mL of plasma to 20 mL with phosphate buffer pH 8.0.
HuBuChE activity was measured in replicate samples using the
basic methods of Ellman et al.,29 modified in order to allow use
of a 96-well plate reader. The total volume of test solution in
each well was 250 μL. This consisted of 25 μL of plasma
solution, 175 μL of phosphate buffer pH 8.0, 25 μL of DTNB
solution [0.5 mmol], and 25 μL of MeCN/distilled water (1:1).
The 96-well plate was incubated for 30 min before 25 μL of BTCI
solution [0.5 mmol] was added, and the reaction was measured
at 412 nm over 5 min using an Anthos bt2 plate reader. For the
determination of AChE inhibition: Electric eel AChE (Type
VI-S, 1.17 mg) was diluted to 1 mL with phosphate buffer pH
8.0. An aliquot of this (5 μL) was further diluted to 5 mL (final
activity was 0.05-1.15 units mL-1). Enzyme solutions were
freshly prepared and used immediately. Then 22 μL of this was
diluted to 215 μL with phosphate buffer (pH 8). Test substance
(5 μL of 2 mmol in MeCN) was added and the solution
incubated at 37 ꢀC for 30 min. To this was added DTNB
(27 μL, final concentration 0.3 mmol) and then ACTI (27 μL,
final concentration 0.5 mmol).
3
use. The substrate, para-nitrophenyl acetate, was obtained from
Sigma Ireland.
All assays were carried out in 96-well plates, with the hydro-
lysis of substrate being monitored by the liberation of chromo-
genic p-nitrophenol at 405 nm with an Anthos Lit-2 absorbance
plate reader. The final volume in each well was always 250 μL. It
comprised the substrate, dissolved in methanol or acetone to a
final assay concentration of 3 mM, 10 μL of the stock micro-
somal solution, and, where necessary, inhibitors. The remainder
of the solution in the wells was 50 mM Tris-HCl buffer pH 7.4.
For all assays, the samples to be run included blank reactions in
which the substrate alone was dissolved in buffer, as well as
positive controls;enzyme and substrate acting in the absence of
any inhibitor. The average values for the blank reactions were
always subtracted from the test results in order to account for
spontaneous hydrolysis of substrate.
When screening for inhibition, compounds were added to
each well at a final concentration of 10 μM. Inhibitors were
incubated with enzyme at 37 ꢀC for 30 min prior to addition of
substrate. The results given are based on four independent
measurements.
For IC50 measurements, each 500 μL aliquot of microsomes
was diluted to 3 mL prior to use; 25 μL of this diluted solution
were added to each well for testing. Inhibitors (50 nM-200 μM)
were incubated with the enzyme in Tris-HCl buffer pH 7.4 for
30 min prior to addition of substrate and subsequent analysis.
For each concentration of inhibitor, four independent measure-
ments were made. The rates of spontaneous substrate hydro-
lysis, given by blank reactions run alongside each assay, were
taken into account in calculations. IC50 values were obtained
directly from sigmoidal concentration response curves to which
the data was fit by nonlinear regression (GraphPad Prism5).
Enzyme Kinetic Analyses. Wild type human BuChE, obtained
as a lyophilized powder, was solubilized in phosphate buffer pH
8.0 for use; D70G and recombinant Y332A and D70G/Y332A
mutant enzymes obtained were the kind gift of Dr. Oksana
Lockridge, University of Nebraska. The enzymes were used at a
final concentration of 0.05 U mL-1, while DTNB and BTCI
were used at final concentrations of 0.3 and 0.5 mmol, respec-
tively. All solutions were prepared in phosphate buffer pH 8.0.
To monitor enzyme carbamylation over time, a 96-well plate
was used, with all assay mixtures being made up to a final
volume of 250 μL for monitoring of absorbance at 412 nm. At
time 0, 150 μL of inhibitor, at various concentrations, were
added to an equal volume of enzyme. Thereafter, at successive
time-points, 50 μL aliquots were removed from these wells and
added to wells containing 200 μL of DTNB and BTCI in
phosphate buffer pH 8.0. Hydrolysis of substrate was monitored
by examining the changes in absorbance in each well at 412 nm
over the following 30 s. Negative controls were run alongside the
test solutions to take account of nonenzymatic hydrolysis, while
positive controls gauged the rate of the uninhibited reaction. At
least two replicate wells were run for each concentration of
inhibitor. A first-order inhibition constant, kobs was obtained
for each concentration of inhibitor by nonlinear regression to
eq 1. The resulting values were fitted to eq 2 using nonlinear
regression analysis in order to determine Kc and k3.
The decarbamylation rate constants (k5) were determined by
following directly the recovery of the inhibited enzymes subse-
quent to their inhibition by the tested agents. The enzymes were
incubated with each carbamate for at least 1 h at a concentration
Carboxylesterases. Human microsomal samples were used as
enzyme sources for all assays. Pooled human liver microsomes

1198 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Carolan et al.
that produced more than 85% inhibition. A 1000-fold dilution
was made at t = 0 with phosphate buffer pH 8 in order to
minimize reinhibition by excess inhibitor. The enzyme was then
maintained at a temperature of 37 ꢀC for the remainder of the
analysis. Samples were withdrawn at successive time points, and
enzyme activity was measured to determine its recovery. Unin-
hibited enzyme, treated in exactly the same way, was tested at
each time point in order to take account of any enzyme
degradation over time. Because decarbamylation is a first-order
reaction, k5 was computed from the fit of the data points to a
single exponential first-order association curve using the control
as a measure of the fully reactivated enzyme.
central cholinergic pathways and can use butyrylcholinesterase to
hydrolyze acetylcholine. Neuroscience 2002, 110, 627–639.
(4) (a) Carson, K. A.; Geula, C.; Mesulam, M. M. Electron micro-
scopic localization of cholinesterase activity in Alzheimer brain
tissue. Brain Res. 1991, 540, 204–208. (b) Arendt, T.; Br€uckner, M. K.;
Lange, M.; Volker, B. Changes in acetylcholinesterase and butyrylcho-
linesterase in Alzheimer's disease resemble embryonic development;
A study of molecular forms. Neurochem. Int. 1992, 21, 381–396.
(c) Giacobini, E. Cholinergic function and Alzheimer's disease. Int. J.
Geriatr. Psychiatry 2003, 18, S1–S5.
(5) Greig, N. H.; Utsuki, T.; Ingram, D. K.; Wang, Y.; Pepeu, G.;
Scali, C.; Yu, Q. S.; Mamczarz, J.; Holloway, H. W.; Giordano, T.;
Chen, D.; Furukawa, K.; Sambamurti, K.; Brossi, A.; Lahiri, D. K.
Selective butyrylcholinesterase inhibition elevates brain acetylcho-
line, augments learning and lowers Alzheimer beta-amyloid pep-
tide in rodent. Proc. Natl. Acad. Sci. U.S.A 2005, 102, 17213–17218.
(6) Giacobini, E.; Spiegel, R.; Enz, A.; Veroff, A. E.; Cutler, N. R.
Inhibition of acetyl- and butyryl-cholinesterase in the cerebrospinal
fluid of patients with Alzheimer’s disease by rivastigmine: correlation
with cognitive benefit. J. Neural Transm. 2002, 109, 1053–1065.
(7) O’Brien, K. K.; Saxby, B. K.; Ballard, C. G.; Grace, J.; Harrington,
F.; Ford, G. A.; O’Brien, J. T.; Swan, A. G.A.; Fairbairn, F.;
Wesnes, K.; del Ser, T.; Edwardson, J. A.; Morris, C. M.; McKeith,
I. G. Regulation of attention and response to therapy in dementia
by butyrylcholinesterase. Pharmacogenetics 2003, 13, 231–239.
(8) Diamant, S.; Podoly, E.; Friedler, A.; Ligumsky, H.; Livnah, O.;
Soreq, S. Butyrylcholinesterase attenuates amyloid fibril formation
in vitro. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 8628–8633.
(9) (a) Lynch, T. J.; Mattes, C. E.; Singh, A.; Bradley, R. M.; Brady, R.
O.; Dretchen, K. L. Cocaine detoxification by human plasma
butyrylcholinesterase. Toxicol. Appl. Pharmacol. 1997, 145, 363–
371. (b) Morton, C. L.; Wadkins, R. M.; Danks, M. K.; Potter, P. M. The
anticancer prodrug CPT-11 is a potent inhibitor of acetylcholinesterase
but is rapidly catalyzed to SN-38 by butyrylcholinesterase. Cancer Res.
1999, 59, 1458–1463.
(10) (a) Iwasaki, T.; Yoneda, M.; Nakajima, A.; Terauchi, Y. Serum
butyrylcholinesterase is strongly associated with adiposity, the
serum lipid profile and insulin resistance. Intern. Med. 2007, 46,
1633–1639. (b) Li, B.; Duysen, E. G.; Lockridge, O. The butyrylcho-
linesterase knockout mouse is obese on a high-fat diet. Chem.-Biol.
Interact. 2008, 175, 88–91.
(11) Carolan, C. G.; Dillon, G. P.; Gaynor, J. M.; Reidy, S.; Ryder, S.
A.; Khan, D.; Marquez, J. F.; Gilmer, J. F. Isosorbide-2-carbamate
Esters: Potent and Selective Butyrylcholinesterase Inhibitors. J.
Med. Chem. 2008, 51, 6400–6409.
(12) Moriarty, L. M.; Lally, M. N.; Carolan, C. G.; Jones, M.; Clancy, J.
M.; Gilmer, J. F. Discovery of a “true” aspirin prodrug. J. Med.
Chem. 2008, 51, 7991–7999.
(13) Jones, M.; Inkielewicz, I.; Medina, C.; Santos-Martinez, M. J.;
Radomski, A.; Radomski, M. W.; Lally, M. N.; Moriarty, L. M.;
Gaynor, J.; Carolan, C. G.; Khan, D.; O’Byrne, P.; Harmon, S.;
Holland, V.; Clancy, J. M.; Gilmer, J. F. Isosorbide-Based Aspirin
Prodrugs: Integration of Nitric Oxide Releasing Groups. J. Med.
Chem. 2009, 52, 6588–6598.
Molecular Modeling. Docking of the carbamate compounds
was carried out using AUTODOCK 4, run on Windows XP
using the Cygwin engine. A Lamarckian genetic algorithm
(LGA) was used in all docking simulations. The inhibitors were
built using the builder function of MOE (Chemical Computing
Group, Montreal, Canada) and minimized with MOPAC 7
(PM3 method) interfaced to MOE. The protein structure,
meanwhile, was the crystal structure of human BChE com-
plexed with a choline molecule (PDB Code 1POM).13 All
nonprotein atoms were removed from the model prior to its
use. The missing side chain of residue Gln486 was added in
Swiss-PDB Viewer version 3.7 and the missing residues 1-3,
378-379, and 455 added using the “add residue” command in
the same program. These added residues were minimized within
MOE, with the rest of the macromolecule held fixed. Of the two
conformations of the catalytic serine observed in the crystal
structure noted above, the one where it makes the anticipated
hydrogen bond with His438 was used. His438 was protonated at
the δ position. The substrates and the enzyme were further
prepared for the docking calculation using AutoDockTools
1.4.5. Rotation was allowed for all appropriate bonds in the
ligand; the protein was rigid except for the side chains of residues
Phe-329, Asp-70, Trp-82, and Phe-398, which were allowed to be
flexible during the docking runs. For docking, the 3D affinity
grid box was designed to include the full active site gorge of
human BChE. The number of grid points in the x-, y-, and z-axes
was 64, 64, and 40, respectively, with grid points separated by
˚
0.375 A. A distance dependent dielectric was used for the
modeling of electrostatic interactions. At the beginning of each
docking simulation, a population of 300 individuals was ran-
domly selected. During population evolution, a maximum of
25000000 energy evaluations were permitted per generation and
27000 generations were allowed. Just one individual result
survived from each generation. Crossover rates and mutation
rates were set at 0.8 and 0.02, respectively. In all, 100 docking
runs were performed, and the highest scoring pose was taken as
being indicative of the correctly docked pose.
(14) Carolan, C. G.; Gaynor, J. M.; Dillon, G. P.; Khan, D.; Ryder, S.
A.; Reidy, S.; Gilmer, J. F. Novel isosorbide di-ester compounds as
inhibitors of acetylcholinesterase. Chem.-Biol. Interact. 2008, 175
(1-3), 293–297.
(15) Imai, T. Human carboxylesterase isozymes: catalytic properties and
rational drug design. Drug. Metab. Pharmacokinet. 2006, 21, 173–185.
(16) (a) Main, A. R.; Hastings, F. L. Carbamylation and Binding
Constants for the Inhibition of Acetylcholinesterase by Physostig-
mine. Science 1966, 154, 400–402. (b) Feaster, S. R.; Quinn, D. M.
Mechanism-Based Inhibitors of Mammalian Cholesterol Esterase.
Methods Enzymol. 1997, 286, 231–252. (c) Rampa, A.; Piazzi, L.;
Belluti, F.; Gobbi, S.; Bisi, A.; Bartolini, M.; Andrisano, V.; Cavrini, V.;
Cavalli, A.; Recanatini, M.; Valenti, P. Acetylcholinesterase Inhibitors:
SAR and kinetic studies on ω-[N-methyl-N-(3-alkylcarbamoyloxyphe-
nyl)-methyl]-aminoalkoxyaryl derivatives. J. Med. Chem. 2001, 44,
3810–3820.
Acknowledgment. This work was funded by Science Foun-
dation Ireland (05/RFP/CHE0046) and IRCSET (Embark
Initiative)
Supporting Information Available: Characterization data for
compounds 6, 7, 8, 9, 10, 11, 12, 13, 14, 20, and 21; elemental
analysis data, purity data by HPLC. This material is available
free of charge via the Internet at http://pubs.acs.org.
(17) Bar-On, P.; Millard, C. B.; Harel, M.; Dvir, H.; Enz, A.; Sussman,
J. L.; Silman, I. Kinetic and structural studies on the interaction of
cholinesterases with the anti-Alzheimer drug rivastigmine. Bio-
chemistry 2002, 41, 3555–3564.
(18) Nicolet, Y.; Lockridge, O.; Masson, P.; Fontecilla-Camps, J. C.;
Nachon, F. Crystal structure of human butyrylcholinesterase and
of its complexes with substrate and products. J. Biol. Chem. 2003,
278, 41141–41147.
(19) Sussman, J. L.; Harel, M.; Frolow, F.; Oefner, C.; Goldman, A.;
Toker, L.; Silman, I. Atomic structure of acetylcholinesterase from
Torpedo californica: a prototypic acetylcholine-binding protein.
Science 1991, 253, 872–879.
References
(1) Tago, H.; Maeda, T.; McGeer, P. L.; Kimura, H. Butyrylcholines-
terase-rich neurons in rat brain demonstrated by a sensitive histo-
chemical method. J. Comp. Neurol. 1992, 325, 301–312.
(2) Manoharan, I.; Kuznetsova, A.; Fisk, J. D.; Boopathy, R.; Lockridge,
O.; Darvesh, S. Comparison of cognitive functions between people
with silent and wild-type butyrylcholinesterase. J. Neural Transm.
2007, 114, 939–945.
(3) Mesulam, M. M.; Guillozet, A.; Shaw, P.; Levey, A.; Duysen,
E. G.; Lockridge, O. Acetylcholinesterase knockouts establish

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1199
(20) Nachon, F.; Ehret-Sabatier, L.; Loew, D.; Colas, C.; van Dorsselaer,
A.; Goeldner, M. Trp82 and Tyr332 are involved in two quaternary
ammonium binding domains of human butyrylcholinesterase as
revealed by photoaffinity labeling with [3H]DDF. Biochemistry
1998, 37, 10507–10513.
(21) Harel, M.; Schalk, I.; Ehret-Sabatier, L.; Bouet, F.; Goeldner, M.;
Hirth, C.; Axelsen, P. H.; Silman, I.; Sussman, J. L. Quaternary
ligand binding to aromatic residues in the active-site gorge of
acetylcholinesterase. Proc. Natl. Acad. Sci. U.S.A 1993, 90, 9031–
9035.
(22) Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart,
W. E.; Belew, R. K.; Olson, A. J. Automated docking using a
Lamarckiangenetic algorithm and an empirical binding free energy
function. J. Comput. Chem. 1998, 19, 1639–1662.
(23) Sippl, W. Development of biologically active compounds by
combining 3D QSAR and structure-based design methods. J. Com-
put.-Aided Mol. Des. 2002, 16, 825–830.
(25) Masson, P.; Froment, M. T.; Bartels, C. F.; Lockridge, O. Asp70 in
the peripheral anionic site of human butyrylcholinesterase. Eur. J.
Biochem. 1996, 235, 36–48.
(26) Masson, P.; Xie, W.; Froment, M. T.; Levitsky, V.; Fortier, P. L.;
Albaret, C.; Lockridge, O. Interaction between the peripheral site
residues of human butyrylcholinesterase, D70 and Y332, in binding
and hydrolysis of substrates. Biochim. Biophys. Acta 1999,1433,281–293.
(27) Grieg, N. H.; Sambamurti, K.; Quian-sheng, Y.; Perry, T. A.;
Holloway, H. W.; Haberman F.; Brossi, B.; Ingram D. K.; Lahiri
D. K. Butyrylcholinesterase: its selective inhibition and relevance
to Alzheimer’s disease therapy. In Butyrlcholinesterase, Its Func-
tions and Inhibitors, 1st ed.; Giacobini, E., Ed.; Martin Dunitz: London,
2003; pp 69-90.
(28) Masson, P.; Bec, N.; Froment, M. T.; Nachon, F.; Balny, C.;
Lockridge, O.; Schopfer, L. M. Rate-determining step of butyr-
ylcholinesterase-catalyzed hydrolysis of benzoylcholine and ben-
zoylthiocholine. Volumetric study of wild-type and D70G mutant
behavior. Eur. J. Biochem. 2004, 271, 1980–1990.
(24) Bartolucci, C.; Perola, E.; Pilger, C.; Fels, G.; Lamba, D. Three-
dimensional structure of a complex of galanthamine (Nivalin) with
acetylcholinesterase from Torpedo californica: implications for the
design of new anti-Alzheimer drugs. Proteins 2001, 42, 182–191.
(29) Ellman, G. L.; Courtney, K. D.; Andres, V., Jr.; Feather-Stone,
R. M. A new and rapid colorimetric determination of acetylcho-
linesterase activity. Biochem. Pharmacol. 1961, 7, 88–95.