New potent acetylcholinesterase inhibitors in the tetracyclic triterpene series

Article abstract of DOI:10.1021/jm070536w

A new highly selective inhibitor of acetylcholinesterase (AChE) was discovered by high-throughput screening. Compound 1 was synthesized from a natural product, the N-3-isobutyrylcycloxobuxidine-F 2. A new extraction protocol of this compound is described.

Full text of DOI:10.1021/jm070536w

J. Med. Chem. 2007, 50, 5311-5323
5311
New Potent Acetylcholinesterase Inhibitors in the Tetracyclic Triterpene Series
Thibault Sauvaˆıtre,† Mireille Barlier,† Denyse Herlem,† Nohad Gresh,§ Ange`le Chiaroni,† Daniel Guenard,*,† and
Catherine Guillou*,†
Institut de Chimie des Substances Naturelles, Bt 27, CNRS, AVenue de la Terrasse, 91198 Gif-sur-YVette, France, and Laboratoire de
Pharmacochimie Mole´culaire et Cellulaire, CNRS FRE 2718, U648 INSERM, UniVersite´ Rene´ Descartes (Paris V), 45 rue des Saints-Pe`res,
75006 Paris, France
ReceiVed May 8, 2007
A new highly selective inhibitor of acetylcholinesterase (AChE) was discovered by high-throughput screening.
Compound 1 was synthesized from a natural product, the N-3-isobutyrylcycloxobuxidine-F 2. A new extraction
protocol of this compound is described. The hemisynthesis and optimization of 1 are reported. The analogs
of 1 were tested in vitro for the inhibition of both cholinesterases (AChE and BuChE). These compounds
selectively inhibited AChE. Extensive molecular docking studies were performed with 2 and AChE employing
Discover Biosym software to rationalize the binding interaction. The results suggested that ligand 2 binds
simultaneously to both catalytic and peripheral sites of AChE.
Introduction
(Namenda, Merz Pharma, and Forest Labs)8 has also been
launched on the market. In addition to its catalytic activity,
AChE exerts secondary noncholinergic functions,9,10 related to
its peripheral anionic site (PAS),11,12 in differentiation,13 cell
adhesion,14,15 and in mediating the processing and deposition
of Aâ peptide.16-19 It was postulated that AChE binds through
its peripheral site to the nonamyloidogenic form of Aâ protein
acting as a chaperone protein and inducing conformational
change to the amyloidogenic form with subsequent amyloid
fibril formation.18,19 In vitro, AChE directly promotes the
assembly of Aâ peptide into amyloid fibrils forming stable
AChE-Aâ complexes.17 A renewed interest in the search for
AChE inhibitors thus has occurred.
In the continuation of our ongoing research20 on new dual-
binding site AChE inhibitors, we describe herein our efforts to
explore the structure-activity relationships of a potent inhibitor
1 (EeAChE IC50 ) 29 nM) identified by high-throughput
screening of our Institut’s chemical and natural products
libraries. Compound 1 is a triterpene-type steroidal alkaloid
(Figure 1).
While advances in medicine have increased the life expect-
ancy of the population, the number of people affected by age-
related pathologies has also consequently increased. One of the
most common afflictions of age is Alzheimer’s disease (ADa).
AD1 is a progressive neurodegenerative disorder of the central
nervous system (CNS) that is characterized by profound memory
impairments, emotional disturbance, and also personality changes.
The etiology of this disease is not known. Much effort is devoted
to elucidate the relationships among the neuropathological
hallmarks of the disease, including extracellular amyloid
plaques,2 intracellular neurofibrillary tangles containing abnor-
mally hyperphosphorylated tau protein,3 and loss of neurons in
the nucleus basalis of Meynert and the hippocampus. AD is
characterized by a pronounced alteration of the cholinergic
system and other neurotransmitter systems (glutamate and
serotonine). The cholinergic hypothesis postulates that memory
impairments in patients with AD result from a deficit of
cholinergic function in the brain.4
Current therapeutic approaches are aimed at (i) limiting the
effects of oxidative damage by protecting the nerve cells; (ii)
reducing the levels of toxic substances by decreasing the
production or aggregation of â-amyloid (Aâ) peptide or
increasing its removal by inhibiting the phosphorylation of the
tau protein; (iii) activating other neurotransmitter systems to
indirectly compensate the cholinergic function deficit; and (iv)
restoring the native levels of acetylcholine in the CNS by using
acetylcholinesterase (AChE) inhibitors.
Interestingly, steroidal alkaloids and triterpenoid alkaloids
from Buxus semperVirens L,21,22 Buxus balearica Willd,21,22
Buxus hyrcana P,23 Buxus paillosa S,24 Sarcococca saligna,21,25
Sarcococca hookerina,26 and Sarcococca coriacea27 are known
to be inhibitors of butyrylcholinesterase (BuChE; in micromolar
range) and weak inhibitors of AChE.
Chemistry
With respect to the latter strategy, AChE inhibitors are the
major and most developed class of drugs approved for AD
symptomatic treatment. These include donepezil (Aricept, Pfizer,
Inc., and Eisai Corp.),5 rivastigmine (Exelon, Novartis),6 and
galanthamine (Reminyl renamed Razadyne, Janssen, and Johnson
and Johnson).7 An NMDA receptor antagonist, memantine
Compound 1 is a triterpene-type steroidal alkaloid synthesized
from N-3-isobutyrylcycloxobuxidine-F 2 (Figure 1) isolated
from Buxus balearica Willd using a new extraction protocol.
In the previous protocol, N-3-isobutyrylcycloxobuxidine-F 2 and
N-3-benzoylcycloxobuxidine-F 3 were isolated in the same
fraction. The separation of these two compounds was difficult
and required long, tedious chromatographies and countercurrent
separations.28
* To whom correspondence should be addressed. Phone: +33 (0)1 69
82 30 75 (C.G. and D.G.). Fax: +33 (0)1 69 07 72 47 (C.G. and D.G.).
E-mail: catherine.guillou@icsn.cnrs-gif.fr (C.G.); daniel.guenard@
icsn.cnrs-gif.fr (D.G.).
In the new protocol, the crude alkaloids were isolated from
a dichloromethane/methanol extract of the air-dried leaves
macerated in aqueous ammonia solution. The crude alkaloids
were separated at different pH values. The fractions obtained
at pH 5.8 were combined, dried, and evaporated to give pure
† Institut de Chimie des Substances Naturelles, CNRS.
§ Laboratoire de Pharmacochimie Mole´culaire et Cellulaire, CNRS.
a Abbreviations: AChE, acetylcholinesterase; BuChE, butyrylcholinest-
erase; AD, Alzheimer’s disease; Aâ, â-amyloid; CNS, central nervous
system; PAS, peripheral anionic site.
10.1021/jm070536w CCC: $37.00 © 2007 American Chemical Society
Published on Web 09/29/2007

5312 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 22
SauVaˆıtre et al.
Figure 1. Structures of compounds 1-3.
Scheme 1. Synthesis of Compound 1a
dihydrooxazines 9a-i (minor) and compounds 8a-i (major)
were purified on alumina to give compounds 8a-i; during this
process, 1,3-dihydrooxazines 9a-i were hydrolyzed to com-
pounds 8a-i.
The 1,3-oxazinane derivatives with cyclopropane and open
cyclopropane rings were synthesized following the reactions
depicted in Schemes 3 and 4 using compounds 5 and 11 as key
intermediates, which were converted into the corresponding 1,3-
oxazinanes 10 and 12 by reaction with isobutyraldehyde.
An isomer of 1 that possesses an R,â-ethylenic ketone
function (i.e., 14, instead of 1, which contains an γ,δ-ethylenic
ketone function) was obtained by reaction of 2 with BF3‚OEt2
and subsequent cyclization in the presence of tetraetylammonium
hydroxide at 300 °C under reduced pressure (Scheme 5).
The syntheses of analogs with modified D rings are shown
in Schemes 6 and 7.
a Reagents and conditions: (a) 240 °C, 0.05 mmHg (90%); (b)
Et4N+OH-, 240 °C, 0.05 mmHg (74%); (c) NaOH (20% mol), ethylene
glycol 210 °C (83%).
N-3-isobutyrylcycloxobuxidine-F 2 after trituration with acetone.
Compound 2 (100 g) was thus obtained from 11 kg of air-dried
leaves.
Thus, treatment of 1 with acetyl or pivaloyl chloride and
pyridine gave the corresponding acetate 15 and pivalate 16
(Scheme 6).
Compound 129 was first prepared in two steps from 2.
Pyrolysis of 2 at 240 °C under 0.05 mmHg afforded a mixture
(82/18) of 1 and 4, which was then subjected to a second
pyrolysis (240 °C, 0.05 mmHg) in the presence of tetraethyl-
ammonium hydroxide to yield only 1 (74%). Compound 1 could
also be obtained in one step from 2 by heating in ethylene glycol
in the presence of sodium hydroxide (83% yield; Scheme 1).
A subsequent single-crystal X-ray diffraction analysis con-
firmed the structure of compound 1 (Figure 2).
The synthesis of the N-demethylated analog of 1 (i.e., 19)
was achieved following the strategy summarized in Scheme 7.
N-3-Isobutyrylcycloxobuxidine-F 2 was converted into its
N-oxide 17 by oxidation with m-chloroperbenzoic acid. Sub-
sequent treatment of 17 with a mixture of ferrous sulfate
heptahydrate31 and iron chloride hexahydrate provided N-
demethylated N-3-isobutyrylcycloxobuxidine-F 18, which was
then submitted to pyrolysis (240 °C) at 0.03 mmHg to give 19.
Synthetic routes to tetrahydropyrimidine analogs of 1,3-
dihydrooxazines 26a,b and 33 were also developed (Schemes
8 and 9). The 16-hydroxy function was first protected as an
acetate (i.e., 20a,b, 27). Oxidation of the alcohol group at
position 29 with Dess-Martin-periodinane afforded the corre-
sponding aldehydes (i.e., 21a,b, 28). Subsequent deprotection
of the hydroxy function furnished 22a,b and 29, which were
submitted to reductive amination to give benzylamines 24a,b
and 31. The benzyl group was then removed by hydrogenolysis
of 24a,b and 31 with ammonium formate as hydrogen donor
and palladium on charcoal to afford amines 25a,b and 32.
Treatment of the terminal amines with triethylamine in n-butanol
provided the expected pyrimidines 26a,b and 33.
The new 1,3-dihydrooxazines 9a-i were synthesized from
cycloxobuxidine-F 5 obtained by hydrolysis of N-3-isobutyryl-
cycloxobuxidine-F 2 (Scheme 2).30 Preparation of amides 7a-i
was carried out by coupling the appropriate anhydrides 6a-i
with cycloxobuxidine-F 5. All anhydrides were readily obtained
following the reported methodology,30 except 6a-c and 6f,
which were commercially available. Subsequent pyrolysis of
7a-i under reduced pressure afforded mixtures of 1,3-dihy-
drooxazines 9a-i (minor) and compounds 8a-i (major) having
an open cyclopropane ring. These resulting mixtures were finally
converted into the corresponding 1,3-dihydrooxazines 9a-i by
heating under the above conditions in the presence of tetra-
ethylammonium hydroxide (Scheme 2). The mixtures of 1,3-
Figure 2. ORTEP representation of the X-ray structure of compound 1 drawn by the Platon program. Dashed line represents the hydrogen bonding.

Acetylcholinesterase Inhibitors in the Triterpene Series
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 22 5313
Scheme 2. Synthesis of Compounds 7a-i, 8a-i, and 9a-ia
a Reagents and conditions: (a) H2SO4 (2 N), MeOH, reflux (94%); (b) (R1CO)2O, 6, MeOH; (c) 0.03 mmHg, 235 °C to 270 °C; (d) 0.03 mmHg, 235 °C
to 270 °C, Et4N+OH-; (e) chromatography on alumina.
Scheme 3. Synthesis of Compound 10a
Scheme 4. Synthesis of Compound 12a
a Reagents and conditions: (a) i-PrCHO, 1,4-dioxane, 50 °C (30%).
Results and Discussion
AChE and BuChE Inhibitory Activity. To determine the
potential interest of the newly synthesized compounds for the
treatment of AD, their AChE inhibitory activity was assayed
by the method of Ellman32 using AChE from Electrophorus
electricus, Torpedo californica, bovine erythrocytes, and human
recombinant with tacrine and galanthamine as reference com-
pounds. Moreover, to further study the biological profiles of
the novel compounds, their BuChE inhibitory activity on human
a Reagents and conditions: (a) (i) H2SO4, MeOH, 2 h, 90 °C; (ii) H2SO4,
1 h, 80 °C (60%); (b) i-PrCHO, 1,4-dioxane, 50 °C, 12 h (42%).
serum butyrylcholinesterase (hBuChE) was also determined by
the same method. Recent studies have shown that, in AD

5314 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 22
Scheme 5. Synthesis of Compound 14a
SauVaˆıtre et al.
suggest that the oxazine and the amidine functions are essential
for anticholinesterase activity.
The presence of the cyclopropane ring increases the AChE
inhibition potency in the oxazinane (i.e., 10 is 3-fold more active
than 12) and pyrimidine series (26b is 2-fold more active than
33).
A pyrimidine with an open cyclopropane ring (i.e., 33) is
two times less-active than the corresponding oxazine (i.e., 9d),
whereas pyrimidines with a cyclopropane ring (26a,b) are
equipotent or slightly more potent compared to oxazines 1 and
9d. The influence of the cyclopropane ring on the conformation
of the terpene can be seen by the superposition of the two
molecules 26a and 33; the total curvature of the molecules is
not the same, leading to a 3 Å difference at the level of carbon
1′, which could explain the modified activity (Figure 3). Similar
conclusions could be observed concerning the influence of the
double bond.
a Reagents and conditions: (a) BF3‚Et2O, CH2Cl2 (86%); (b) Et4N+OH-,
300 °C, 0.03 mmHg, 4 h (62%).
Modifying the group at position 1′ of the oxazine and the
pyrimidine rings affected the affinity for AChE and BuChE.
An order of binding to AChE can be established among the
various substituents: CH3 < C2H5 < CH(CH3)2 e CH(CH3)-
(C2H5) >> C(CH3)3 > C6H11 >> CH2-Ph > Ph. The best
substituent in this position is (S)-MeCHEt (i.e., 9d, IC50 ) 13
nM). From this list it can be seen that a certain degree of
hydrophobicity must be present, while steric hindrance appears
as soon as the chain is longer than three linear carbons. It should
also be noticed that the presence of a hydrogen on carbon 2′
seems to be necessary, perhaps due to a possible mesomeric
form with the vicinal CN double bond.
Scheme 6. Synthesis of Compounds 15 and 16a
Transformation of the alcohol function in position 16 into
an ester (i.e., 15, IC50 ) 380 nM; 16, IC50 ) 110 nM),
demethylation of the nitrogen atom in position 20 (i.e., 19, IC50
) 106 nM; 1, IC50 ) 31 nM), and isomerization of the γ,δ-
unsaturated ketone (i.e., 1) into an R,â-unsaturated ketone (i.e.,
14, IC50 ) 110 nM) are all detrimental to the AChE inhibitory
activity.
a Reagents and conditions: (a) RCOCl, pyridine, rt, 2-3 h.
Scheme 7. Synthesis of Compound 19a
Regarding the BuChE inhibitory activity, the novel com-
pounds reported in this study are weak inhibitors of hBuChE
(IC50 values from 0.37 µM to 7.4 µM; i.e., 9h, IC50 ) 0.37
µM; 4, IC50 ) 7.4 µM in the most active compounds, Table 1).
So, some of the novel compounds are weak inhibitors of
hBuChE, but in the range of previously reported steroidal
alkaloids.21-25
On the basis of the preceding results, steroidal alkaloids 1,
9a-d, 16, 19, 26a, 26b, and 33, which were the most potent
inhibitors in Electrophorus electricus AChE, were also assayed
in Torpedo californica AChE, bovine erythrocyte AChE, and
human recombinant AChE.
The most potent inhibitors of EeAChE and TcAChE are less
active on bAChE and hAChE. Differences in the primary
sequences of cholinesterases from different species could in all
likelihood explain the variations in IC50 for all these compounds.
The primary sequences of EeAChE and TcAChE are very
similar but differ from those of bAChE and hAChE, which are
equivalent (84% overall identity). The increased selectivity
toward EeAChE and especially TcAChE can be explained by
the charge distribution around the peripheral site. For instance,
the anionic residue Glu 75 and hydrophilic residue Gln 76 in
TcAChE are replaced by the more neutral amino acids Thr and
Leu in hAChE. At the opening of the gorge, Asp 273 in EeAChE
is replaced by Val in hAChE and Glu in TcAChE. These
modifications could affect the binding of pyrimidines 26a, 26b,
and 33 because there are more negative charges in TcAChE
a Reagents and conditions: (a) mCPBA, CH2Cl2, rt, 30 min (79%); (b)
(i) FeSO4‚7H2O, FeCl3‚6H2O, MeOH, rt, 1 h and 30 min; (ii) AcOH/AcONa
buffer, CH2Cl2 (76%); (c) Et4N+OH-, 240 °C, 0.03 mmHg (60%).
patients with severe pathology, BuChE increases while AChE
is reduced in specific brain regions;33 this is not in agreement
with a recent study, which shows a decrease of BuChE activity
in vivo.34
The results are summarized in Tables 1 and 2.
The inhibitory potency depends on numerous factors. Ox-
azines (9d-h, 4-32-fold) and pyrimidines (26a,b and 33, 13-
74-fold) compounds are, in general, more active than their
corresponding open precursors (8d-f, 25a,b, 32). When the
oxazine ring (i.e., 1) is reduced into oxazinane (i.e., 12), the
inhibitory potency greatly decreases (76-fold). These results

Acetylcholinesterase Inhibitors in the Triterpene Series
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 22 5315
Scheme 8. Synthesis of Compounds 26a,ba
a Reagents and conditions: (a) Ac2O, pyridine, CH2Cl2; (b) Dess-Martin periodinane, CH2Cl2; (c) NaHCO3, Na2CO3, MeOH, H2O; (d) benzylamine,
MgSO4, CH2Cl2; (e) NaBH3CN, AcOH, MeOH; (f) HCOONH4, Pd/C 30%, MeOH; (g) Et3N, n-BuOH.
than in hAChE at the peripheral site, which increase the
favorable interaction of positively charged pyrimidines with the
protein.
of type 1. From these experiments it seems difficult to determine
the orientation of the molecules within either the active or the
peripheral binding sites. Consequently, molecular modeling was
used to bring some insight concerning enzyme-inhibitor
interactions.
Within the peripheral site, another example of the variation
in charge distribution as a function of the species is given by
the two neutral residues Val 277 and Asn 280 in TcAChE, which
are replaced by two histidines in hAChE. Concerning the gorge
of the site, it should be noted that Phe 330 in TcAChE is
replaced by a Tyr in hAChE and EeAChE. This replacement
could thus be the best explanation for a better binding of
pyrimidines on TcAChE than on EeAChE in comparison to the
binding of oxazine 1.
Molecular Modeling Studies. A molecular modeling study
using Discover Biosym was performed to explore the binding
of 1 to the enzyme TcAChE (see Experimental Section for
details). The crystal structure of the decamethonium-AChE
complex37 was taken as a model because of the similarity of
the length between the nitrogen atoms in both compounds and
the good superposition of the triterpene skeleton with the
decamethonium chain (Figure 6).
Determination of the type inhibition could be interesting,
especially by comparison with another inhibitor 34 that we have
recently studied (Figure 4).20b
The main problem in the study of the interaction of these
compounds is determination of the orientation of the molecule
in the site, that is to say, situation I (ring D and the dimethyl
amino group in the catalytic area) or on the contrary, situation
II (ring A and oxazine ring at this location).
From Figure 5, which shows Lineweaver-Burk plots of
compounds 26a and 34 realized under the same conditions, we
can see that the two systems are different (mixed type for 26a
and competitive for 34), and the type of inhibition by 26a is
closer to that of compounds structurally related to those
mentioned in previous work.35 Another example is given by
decamethonium, an AChE inhibitor interacting with both sites
(see below). Its inhibition type is noncompetitive,36 for a total
length between the two nitrogens identical to that of compounds
It could be observed that the binding of these molecules seems
to be very sensitive to the nature of the substituent on the
oxazine, which is better taken into account by the catalytic site
where steric or ionic constraints are more important (situation
II). However, the charge probably resides more on the nitrogen

5316 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 22
Scheme 9. Synthesis of Compound 33a
SauVaˆıtre et al.
a Reagents and conditions: (a) Ac2O, pyridine, CH2Cl2 (90%); (b) Dess-Martin periodinane, CH2Cl2 (77%); (c) NaHCO3, Na2CO3, MeOH, H2O (84%);
(d) benzylamine, MgSO4, CH2Cl2 (99%); (e) NaBH3CN, AcOH, MeOH (73%); (f) HCOONH4, Pd/C 30%, MeOH (36%); (g) Et3N, n-BuOH (49%).
Table 1. Inhibition of Electrophorus electricus AChE and human BChE Activities for 1 and its Analogs
IC50 (nM)
IC50 (nM)
cmpd
EeAChEa
hBuChEb
cmpd
EeAChEa
hBuChEb
1
2
4
31 ( 4
>1000
12
14
15
16
18
19
25a
25b
26a
26b
30
2380 ( 250
110 ( 10
380 ( 35
110 ( 9
>10 000
1450 ( 160
N.T.c
N.T.
1440 ( 170
N.T.
N.T.
N.T.
>10 000
N.T.
>10 000
>10 000
>10 000
58 ( 7
>10 000
>10 000
9380 ( 1200
105 ( 9
>10 000
7407 ( 910
>10 000
1926 ( 250
>10 000
>10 000
>10 000
>10000
8d
8e
8f
9a
9b
9c
9d
9e
9f
9g
9h
9i
120 ( 13
3270 ( 250
680 ( 55
607 ( 70
225 ( 15
13 ( 2
27 ( 3
102 ( 8
7650 ( 960
400 ( 35
4205 ( 410
820 ( 75
110 ( 11
106 ( 15
291 ( 28
1341 ( 132
18 ( 2
>10 000
>10 000
1535 ( 145
>10 000
375 ( 35
380 ( 45
>10 000
14 ( 2
780 ( 95
368 ( 40
28 ( 4
32
33
tacrine
galanthamine
74 ( 8
360 ( 40
10
a EeAChE from Electrophorus electricus. b BuChE from human serum. c Not tested.
atom of the dimethyl amino group and it is known that cations
are more stabilized by π-stacking in the catalytic area (situation
I; Figure 7).
A previous study of the binding of related molecules to AChE
showed an orientation of this type of molecule according to
situation II. However, molecules had two protonable nitrogen
atoms on ring A and ring D, such that the charge distribution is
different from our molecules. The conclusions reached by these
experiments thus support those of our study.38
To validate the choice of the model (I or II), we decided to
compare the binding of different judiciously chosen analogs of
1 and retain the model, which gives the right order of their
binding to AChE of Torpedo californica, the latter being the
enzyme studied in the X-ray studies.

Acetylcholinesterase Inhibitors in the Triterpene Series
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 22 5317
Table 2. Inhibition of Different Sources of AChE by 1 and Selected
Analogs
not totally accounted for by the energy differences in type I.
This could be due to the simplicity of the force-field used in
this study. Resorting to more elaborate polarizable molecular
mechanics could possibly be envisaged for refinement of the
energy balances39 but is outside the scope of the present work.
We also note that the sole entropy effects encompassed in the
energy balances δE2 are those inherent to the continuum
solvation energies ∆Gsolv. The absolute magnitudes of δE2 are
known to be overestimated in the absence of entropy effects
that account for the reduction of translational and rotational
motions of the ligand upon complex formation.40 Additional
reasons for such overestimations could reside in the above-
mentioned simplicity of the force-field as well as to some
imbalance in the δE1 versus δEsolv values whose difference (in
absolute values) leads to δE2. Free energy calculations with
classical force fields41 could perhaps provide more realistic
estimates of δE2, but are extremely time-consuming and could
also be fraught with uncertainties due inter alia to the lack of
polarization effects, which could be very significant in cation-π
and stacked complexes.
IC50 (nM)
cmpd
EeAChEa
TCAChEb
bAChEc
hAChEd
1
9a
9c
9d
16
19
26a
26b
33
31
680
225
13
110
106
18
14
28
74
385
2792
372
237
724
485
2
2
447
157
756
6116
1176
1364
1385
2226
3277
6486
7351
446
299
5242
351
265
839
517
3108
4368
3832
1030
tacrine
galanthamine
360
a EeAChE from Electrophorus electricus. b TcAChE from Torpedo
californica. c bAChE from bovine erythrocytes. d hAChE from human
recombinant.
According to model II, the inflexible steroid 1 enters the
aromatic gorge through the six membered A ring and the
oxazidine ring. The latter rings are placed at the bottom of the
gorge, which might be due to the apparently greater hydropho-
bicity of these rings in comparison with that of the five-
membered D ring (Figure 9). We cannot exclude a different
binding mode of the pyrimidine relative to the oxazine analogs.
In the pyrimidine series, the nitrogen atoms (calculated pKa 13)
are more basic than the nitrogen atom of the oxazine ring
(calculated pKa 6.5). A protonation of the nitrogen atoms of
the pyrimidine ring could thus occur at physiological pH.
However, due to the strong structural similarity of this series,
this hypothesis does not seem plausible.
Figure 3. Superposition of the two molecules 26a (yellow) and 33
(green).
Compound 1 is completely buried inside the aromatic
gorge. This contributes to the stabilization of the complex
since the steroid backbone of 1 is highly hydrophobic due to
its aliphatic character. The main hydrophobic interactions
between the hydrocarbon skeleton of 1 and the protein were
observed with the residues Tyr121, Phe330, Phe331, and
Tyr334.
Figure 4. Structure of a dual binding site inhibitor in the galanthamine
series.
The selected molecules were 1, 9a, 9c, 16, and 26a.
The results (see details in Experimental Section) show that
for each compound, model I is favored because its energy is
slightly lower than the energy of model II (Figure 8). On the
other hand, total energy is correlated to activity (IC50) in the
case of type II. The correlation between IC50 and δE2_II is
qualitative rather than quantitative. The most active compound
has the strongest gain in energy, but the IC50 differences are
In summary, the type II model seems to be more satisfactory,
but further molecular modeling experiments using other mol-
ecules will be necessary to allow selection of the correct model.
Ideally, an X-ray study of the AChE-inhibitor complex would
be required.
Figure 5. Lineweaver-Burk plots for compounds 26a (A) and 34 (B). 1/Vmax versus 1/[thioacetylcholine] in the presence of various concentrations
of inhibitor (see Experimental Section).

5318 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 22
SauVaˆıtre et al.
Figure 6. Superposition of decamethonium (magenta) and compound
1.
Figure 9. Representation of 1 docked into the binding site of AChE
highlighting the protein residues that form the main interactions with
the different structural units of the inhibitor.
Experimental Section
Chemistry. All reagents were of commercial quality. All
experiments involving water-sensitive compounds were carried out
under argon and scrupulously dry conditions, using anhydrous
solvents. MeOH was distilled from Mg/I2, CH2Cl2 was distilled
from P2O5, ethylene glycol was distilled from sodium hydroxide,
1,4-dioxane was distilled from LiAlH4, and NEt3 and pyridine were
distilled from KOH. All separations were carried out under flash
chromatographic conditions on Merck silica gel 60 (70-230 mesh)
at medium pressure (200 mbar). TLC was performed on Merck
silica gel plates (60F254) with a fluorescent indicator. NMR spectra
were determined on Bruker Avance-300, Bruker AC-400, Bruker
AC-500, or Bruker-600 instruments and using tetramethylsilane
(TMS) as reference. Chemical shifts are reported in parts per million
(ppm) relative to TMS. High-resolution mass spectra (HMRS) were
obtained on a MALDI-TOF spectrometer. Infrared (IR) spectra were
recorded on a Fourier Perkin-Elmer Spectrum BX FT-IR instru-
ments. Elemental analyses were performed by the microanalysis
laboratory of the ICSN, CNRS, Gif-sur-Yvette.
8H-Indeno[5′,4′:4,5]cyclohepta[1,2-f][3,1]benzoxazin-8-one, 10-
[(1S)-1-(Dimethylamino)ethyl]-1,4a,5,7,7a,9,9a,10,11,12,12a,
12b,13,14,14a,14b-hexadecahydro-11-hydroxy-9a,12a,14b-tri-
methyl-3-(1-methylethyl)-, (4aS,7aR,9aR,10S,11R,12aS,12bR,
14aR,14bS; 1). First procedure was completed in the oven:
Compound 2 (100 mg, 0.20 mmol, 1 equiv) is heated at 240 °C
under 0.05 mmHg in a rotating distillation oven with four bulbs.
A product sublimated in 3 h. The sublimate was composed of a
mixture of compound 4 (18%) and the expected dihydro-oxazine
1 (82%). To a solution of the sublimate in 2 mL of dichloromethane
was added 25% tetraethylammonium hydroxide (515 mg, 0.87
mmol, 5 equiv) solution in methanol. The solvent was evaporated
under reduced pressure. The dark red residue was heated at 240
°C under 0.05 mmHg in a rotating distillation oven with four bulbs
during 3 h to give a pale yellow powder of 1 (71.6 mg, 74%) whose
spectroscopic characteristics were consistent with described.30
Second procedure was completed in solution: To a suspension of
Figure 7. The two models of interaction between 1 and AChE.
Conclusion
We describe a novel series of compounds based on the
N-3-isobutyrylcycloxobuxidine-F scaffold that are potent and
selective inhibitors of AChE. N-3-Isobutyrylcycloxobuxidine-F
was isolated in large quantities from the leaves of Buxus
balearica Willd. It is the common precursor of these inhibitors.
Beginning from a high-throughput screening, we were able to
elaborate a new family of AChE inhibitors, with potencies in
the nanomolar range. The oxazine analogs 1 and 9d are very
active on hAChE, whereas the pyrimidine analogs 26a and 26b
are the most potent TcAChE inhibitors. We used these com-
pounds to gain functional and structural insight into the
mechanism of inhibition. The tetracyclic triterpene compounds
presented herein have a rigid conformation allowing interaction
with both peripheral and active sites of AChE, as was shown
by molecular modeling studies. The three-dimensional structures
of their complexes with AChE will need to be solved by X-ray
crystallography.
Figure 8. Energy balances (kcal/mol) after single-point Poisson-Boltzmann calculations of continuum solvation energies, and values of the
experimentally determined inhibitory potencies IC50 (nM) for the molecules selected in our modeling study (see text in the Experimental Section
for a definition of the energy values).

Acetylcholinesterase Inhibitors in the Triterpene Series
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 22 5319
Table 3. Weight of Alkaloid in Function of the pH Extraction
compound 2 (50 mg, 0.1 mmol, 1 equiv) in 5 mL of freshly distilled
ethylene glycol was added sodium hydroxide (0.8 mg, 0.02 mmol,
0.2 equiv). After heating at 215 °C during 6 h, the mixture was
then made basic with 60 mL of a 10% ammonia solution (pH )
10) and extracted with dichloromethane (3 × 40 mL). The organic
layer was dried over Na2SO4 and filtered, and the solvent was
evaporated to give a pale yellow powder of 1 (40 mg, 83%) whose
spectroscopic characteristics were consistent with those described.30
weight (g)
residue
pH
1.74
5
4.71
5.8
6.5
7.3
8.2
9.1
3.19
1.80
4.06
6.70
Mp 281.5 °C; [R]25D ) +71 (c 0.8, CHCl3); IR (CHCl3) υmax(cm-1
)
1
3.25
3467, 2966, 1694, 1668, 1466, 1376, 1148, 1037; H NMR (600
MHz, CDCl3) δ 0.66 (3H, s, H-18), 0.71 (3H, s, H-30), 0.84 (3H,
d, J ) 6.6 Hz, H-21), 1.08, 1.09 (6H, 2d, J ) 6.9 Hz, H-3′, H-4′),
1.20 (3H, s, H-28), 1.28 (1H, m, H-6â), 1.30 (1H, m, H-7R), 1.47
(1H, dd, J ) 2.5, 13.8 Hz, H-15R), 1.59 (1H, dd, J ) 7.5, 13.4 Hz,
H-6R), 1.63 (1H, ddd, J ) 3.7, 10.2, 10.9 Hz, H-8â), 1.74 (1H, m,
H-2â), 1.83 (1H, bd, J ) 8.3 Hz, H-5R), 1.89 (1H, m, H-7â), 1.89
(1H, dd, J ) 9.0, 10.2 Hz, H-9R), 1.96 (1H, m, H-15â), 1.96 (1H,
m, H-17R), 2.03 (1H, m, H-19â), 2.20 (6H, bs, NB-CH3), 2.31
(1H, d, J ) 15.8 Hz, H-12â), 2.31 (1H, m, H-2R), 2.37 (1H, dq, J
) 6.9 Hz, H-2′), 2.52 (1H, bd, J ) 15.8 Hz, H-12R), 2.57 (1H, dq,
J ) 6.6, 11.2 Hz, H-20), 3.20 (1H, dd, J ) 6.0, 11.7 Hz, H-3R),
3.31 (1H, d, J ) 15.5 Hz, H-19R), 3.73 (1H, d, J ) 10.2 Hz,
H-29R), 4.04 (1H, d, J ) 10.2 Hz, H-29â), 4.05 (1H, ddd, J )
2.5, 6.9, 9.7 Hz, H-16â), 5.56 (1H, m, H-1); 13C NMR (125 MHz,
CDCl3) δ 9.59 (CH3, C-30), 9.94 (CH3, C-21), 17.75 (CH3, C-18),
18.77 (CH3, C-28), 19.73, 19.80 (CH3, C-3′, C-4′), 24.61 (CH2,
C-6), 30.75 (CH2, C-2), 33.16 (C, C-4), 33.47 (CH2, C-7), 34.27
(CH, C-2′), 37.27 (CH2, C-19), 42.44 (CH2, C-15), 46.14 (C, C-14),
46.97 (C, C-13), 49.44 (CH, C-5), 49.67 (CH, C-8), 49.98 (CH2,
C-12), 50.23 (CH, C-9), 55.46 (CH, C-17), 55.60 (CH, C-3), 61.99
(CH, C-20), 74.72 (CH2, C-29), 78.27 (CH, C-16), 121.81 (CH,
C-1), 137.10 (C, C-10), 162.57 (C, C-1′), 211.55 (C, C-11); ES-
MS m/z 485.3 [M + H]+ (95), 486.3 (100), 487.4 (50), 503.4 (2);
HRES-MS m/z calcd for C30H49N2O3, 485.3743; found, 485.3721;
Anal. Calcd for C30H48N2O3: C, 74.34%; H, 9.98%; N, 5.78%; O,
9.90%. Found: C, 74.13%; H, 9.91%; N, 5.67%; O, 9.97%.
N-3-iso-Butyrylcycloxobuxidine-F (2) or N-[(3â,4R,9â,16R,
20S)-20-(Dimethylamino)-16-hydroxy-4-(hydroxymethyl)-4,14-
Dimethyl-11-oxo-9,19-cyclopregnan-3-yl]-2-methylpropan-
amide (2). First Part: Extraction Procedure for Total Alkaloids
of Buxus balearica Willd. Dried and pulverized leaves (1.5 kg)
from Balearica buxus Willd were alkalized with 600 mL of a 30%
ammonia solution. A first extraction was achieved after stirring
with 10 L of solvents (CH2Cl2-EtOH 9:1) at room temperature
overnight. After filtration, five successive soakings were carried
out with 10 L of solvents (CH2Cl2-EtOH 9:1), while stirring at
room temperature during 3 h between each filtration.
total: 25.4 g and
a loss of about 1.9 g
The pH of each aqueous layer containing the total alkaloids
dissolved in the buffer solution was increased by one pH unit with
a 10% ammonia solution. The aqueous layers were washed
successively with 8 × 500 mL of dichloromethane until exhaustion
of alkaloids in the aqueous layer. The combined organic extracts
were washed successively with 1 L of a 10% ammonia solution,
with 1 L of water and 1 L of brine, then dried over Na2SO4 and
filtered, and the solvent was evaporated to give N-3-isobutyryl-
cycloxobuxidine-F 2 (3.19 g, 0.8%) at pH ) 5.8. This leads to the
weight balance shown in Table 3.
This process was also carried out with 11 kg of Balearica buxus
Willd leaves. Trituration in acetone gave N-3-iso-butyrylcycloxo-
buxidine-F 2 (100 g, 0.9%), whose spectroscopic characteristics
are identical to those described.30 Mp 257 °C; [R]25 ) +69 (c
D
1.01, CHCl3); IR (CHCl3) υmax (cm-1) 3434, 1690, 1654, 1513,
1096; 1H NMR (600 MHz, CDCl3) δ 0.53 (3H, s, H-30), 0.80 (3H,
s, H-18), 0.83 (3H, d, J ) 6.6 Hz, H-21), 0.85 (1H, dddd, J ) 3.6,
11.7, 12.6, 14.5 Hz, H-6â), 1.00 (1H, d, J ) 3.8 Hz, H-19â), 1.13,
1.15 (6H, 2d, J ) 6.8 Hz, H-3′, H-4′), 1.18 (3H, s, H-28), 1.25
(1H, ddd, J ) 4.2, 13.0, 13.5 Hz, H-1R), 1.40 (1H, m, J ) 4.0,
10.9, 12.6, 14.5 Hz, H-7R), 1.46 (1H, dd, J ) 3.0, 14.0 Hz, H-15R),
1.49 (1H, ddd, J ) 3.6, 7.5, 14.5 Hz, H-7â), 1.52 (1H, d, J ) 3.8
Hz, H-19R), 1.55 (1H, ddd, J ) 4.0, 12.7, 13.0 Hz, H-2â), 1.65
(1H, ddd, J ) 3.6, 4.2, 4.2 Hz, H-2R), 1.67 (1H, ddd, J ) 3.6, 4.0,
14.5 Hz, H-6R), 1.94 (1H, dd, J ) 7.3, 10.8 Hz, H-17R), 1.98 (1H,
dd, J ) 7.5, 10.9 Hz, H-8â), 1.98 (1H, dd, J ) 9.5, 14.0 Hz, H-15â),
2.02 (1H, dd, J ) 3.6, 11.7 Hz, H-5R), 2.20 (6H, bs, NB-CH3),
2.26 (1H, d, J ) 17.1 Hz, H-12â), 2.36 (1H, dq, J ) 6.8 Hz, H-2′),
2.38 (1H, ddd, J ) 3.6, 4.0, 13.5 Hz, H-1â), 2.48 (1H, d, J ) 17.1
Hz, H-12R), 2.58 (1H, dq, J ) 6.6, 10.8 Hz, H-20), 2.92 (1H, dd,
J ) 3.2, 12.5 Hz, H-29a), 3.30 (1H, dd, J ) 10.4, 12.5 Hz, H-29b),
3.94 (1H, ddd, J ) 4.2, 9.0, 12.7 Hz, H-3R), 4.05 (1H, ddd, J )
3.0, 7.3, 9.5 Hz, H-16â), 4.37 (1H, dd, J ) 3.2, 10.4 Hz, OH),
5.47 (1H, d, J ) 9.0 Hz, NAH); 13C NMR (125 MHz, CDCl3) δ
9.87 (CH3, C-21), 11.15 (CH3, C-30), 17.79 (CH3, C-18), 18.36
(CH2, C-6), 19.38, 20.09 (CH3, C-3′, C-4′), 20.77 (CH3, C-28),
24.26 (CH2, C-7), 27.49 (CH2, C-2), 27.65 (CH2, C-1), 30.48 (CH2,
C-19), 34.41 (C, C-9), 35.68 (CH, C-2′), 37.72 (C, C-10), 41.12
(CH, C-5), 41.37 (CH, C-8), 42.76 (CH2, C-15), 44.45 (C, C-13),
44.50 (C, C-4), 47.09 (C, C-14), 50.64 (CH, C-3), 51.45 (CH2,
C-12), 55.79 (CH, C-17), 61.98 (CH, C-20), 64.11 (CH2, C-29),
78.27 (CH, C-16), 178.45 (C, C-1′), 211.47 (C, C11); ES-MS m/z
503.4 [M + H]+ (100), 504.4 (10); HRES-MS m/z calcd for
C30H51N2O4, 503.3849; found, 503.3852; Anal. Calcd for
C30H50N2O4: C, 71.67%; H, 10.02%; N, 5.57%; O, 12.73%.
Found: C, 71.51%; H, 10.14%; N, 5.34%; O, 12.87%.
Collected organic layers were evaporated under reduced pressure
to give a crude and dark green extract (240 g). This crude product
was washed successively with 7 L of dichloromethane and washed
10 times with 1 L of hydrochloric acid. The aqueous layers were
made basic with a 10% ammonia solution (pH ) 10), then extracted
with dichloromethane (6 × 2 L). The combined organic extracts
were evaporated under reduced pressure to give a brown powder
(105 g, 7%).
This process was also carried out with 11 kg of Buxus Balearica
Willd leaves.
Second Part: Extraction of N-3-Isobutyrylcycloxobuxidine-
F. Procedure for the Preparation of Solutions: A total of 4 L of
a 1.0 M AcOH/AcONa buffer solution at pH ) 5 were prepared
by dissolving glacial acetic acid (17.136 g, 0.2348 mol) and sodium
acetate (99.296 g, 0.7151 mol) 1 L of water.
The alkaloids were dissolved four times in the buffer solution at
pH ) 5.
The solution of total alkaloids (27.3 g) in 500 mL of dichloro-
methane was washed with 4 × 1 L of the buffer solution at pH )
5. This organic layer was washed successively with 1 L of a 10%
ammonia solution, with 1 L of water and 1 L of brine, then dried
over Na2SO4 and filtered, and the solvent was evaporated to give
a residue (1.74 g).
N-[(2R,3S,3aR,5aR,9S,10S,12aR,12bS)-3-[1-(Dimethylamino)-
ethyl]-2-hydroxy-10-(hydroxymethyl)-3a,10,12b-trimethyl-5-
oxo-1,2,3,3a,4,5,5a,6,8,9,10,10a,11,12,12a,12b-hexadecahydrobenzo-
[4,5]cyclohepta[1,2-e]inden-9-yl]-2-methylpropanamide (4). First
procedure in solution: A suspension of compound 2 (250 mg, 0.49
mmol, 1 equiv) in 20 mL of freshly distilled ethylene glycol on
sodium hydroxide pellets (concentration about 12 g/L) was stirred
for 6 h at 160 °C under inert atmosphere. The mixture was then
made basic with 50 mL of a 10% ammonia solution (pH ) 10)
and extracted with dichloromethane (3 × 40 mL). The organic layer
was washed with brine (30 mL), dried over Na2SO4, and filtered,
and the solvent was evaporated. The crude product was purified

5320 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 22
SauVaˆıtre et al.
by column chromatography on alumina with dichloromethane-
methanol (98:2) as eluent followed by trituration in acetone to afford
a colorless powder of 4 (105 mg, 42%), whose spectroscopic
characteristics are identical to those described.30 Second procedure
in the oven: Compound 2 (100 mg, 0.20 mmol, 1 equiv) was heated
at 240 °C under 0.05 mmHg in a rotating distillation oven with
four bulbs. A product sublimated in 3 h. The sublimate was
composed of a mixture of the expected product 4 (18%) and
dihydrooxazine 1 (82%). The sublimate was purified by column
chromatography on alumina with dichloromethane-methanol (98:
2) as eluent followed by trituration in acetone to afford a colorless
powder of 4 (91 mg, 91%), whose spectroscopic characteristics were
(CH3, C-21), 11.4 (CH3, C-30), 17.9 (CH3, C-18), 18.4 (CH2, C-6),
20.9 (CH3, C-28), 24.4 (CH2, C-7), 27.7 (CH2, C-1), 27.7 (CH2,
C-2), 30.6 (CH2, C-19), 34.5 (C, C-9), 37.8 (C, C-10), 41.3 (CH,
C-5), 41.4 (CH, C-8), 42.8 (CH2, C-15), 44.5 (C, C-4), 44.9 (C,
C-13), 47.2 (C, C-14), 51.5 (CH, C-3), 51.7 (CH2, C-12), 55.8 (CH,
C-17), 62.1 (CH, C-20), 64.3 (CH2, C-29), 78.3 (CH, C-16), 127.0,
128.8 (CH, C-Ar ) 3′,4′,6′,7′), 131.9 (C, C-5′), 133.9 (C, C-2′),
168.8 (C, C-1′), 211.6 (C, C-11); ES-MS m/z 537.4 [M + H]+
(100), 538.4 (15); HRES-MS m/z calcd for C33H49N2O4, 537.3692;
found, 537.3679. Anal. (C33H48N2O4‚0.5H2O) C, H, N, O.
N-[(2R,3S,3aR,5aR,9S,10S,10aR,12aR,12bS)-3-[(1S)-1-(Di-
methylamino)ethyl]-1,2,3,3a,4,5,5a,6,8,9,10,10a,11,12,12a,12b-
hexadecahydro-2-hydroxy-10-(hydroxymethyl)-3a,10,12b-tri-
methyl-5-oxobenzo[4,5]cyclohept[1,2-e]inden-9-yl]acetamide (8a).
Compound 7a (221 mg, 0.46 mmol, 1 equiv) was heated at 240 °C
under 0.03 mmHg in a rotating distillation oven with four bulbs.
A product sublimated in 3 h. The sublimate was composed of a
mixture of the expected product 8a (98%) and dihydrooxazine 9a
(2%). The sublimate was purified by column chromatography on
alumina with dichloromethane-methanol (99.7:0.3) as eluent
followed by trituration in acetone to afford a colorless powder of
8a (79.4 mg, 88%). Mp 253 °C; IR (CHCl3) υmax (cm-1) 3435,
consistent and identical to those described.30 Mp 242 °C; [R]25
)
D
-46 (c 1.0, CHCl3); IR (CHCl3) υmax (cm-1) 3433, 1693, 1652,
1513, 1462, 1380, 1097, 1043, 1015; 1H NMR (300 MHz, CDCl3)
δ 0.47 (3H, s, H-30), 0.69 (3H, s, H-18), 0.87 (3H, d, J ) 6.2 Hz,
H-21), 1.19 (6H, 2d, J ) 7.0 Hz, H-3′, H-4′), 1.27 (3H, s, H-28),
1.29 (1H, m, H-6â), 1.50 (1H, dd, J ) 2.3, 13.8 Hz, H-15R), 1.58
(1H, bd, J ) 11.7 Hz, H-7R), 1.64 (1H, m, H-8â), 1.83 (1H, m,
H-6R), 1.88 (1H, m, H-7â), 1.92 (1H, m, H-2â), 1.99 (1H, m,
H-15â), 2.02 (1H, dd, J ) 7.0, 11.0, H-17R), 2.03 (1H, m, H-9R),
2.08 (1H, m, H-19â), 2.24 (6H, bs, NB-CH3), 2.28 (1H, m, H-2R),
2.34 (1H, d, J ) 16.0 Hz, H-12â), 2.42 (1H, m, J ) 7.0 Hz, H-2′),
2.44 (1H, m, H-19â), 2.58 (1H, bd, J ) 16.0 Hz, H-12R), 2.59
(1H, m, H-20), 2.60 (1H, m, H-5R), 3.10 (1H, dd, J ) 4.3, 11.0
Hz, H-29a), 3.35 (1H, d, J ) 13.9 Hz, H-19R), 3.38 (1H, m, H-29b),
4.08 (1H, m, H-3R), 4.09 (1H, m, J ) 6.6 Hz, H-16â), 4.57 (1H,
dd, J ) 4.3, 10.6 Hz, OH), 5.32 (1H, d, J ) 8.9 Hz, NAH), 5.48
(1H, m, H-1); 13C NMR (75.5 MHz, CDCl3) δ 9.9 (CH3, C-21),
10.0 (CH3, C-30), 17.8 (CH3, C-18), 19.0, 19.5 (CH3, C-3′, C-4′),
20.1 (CH3, C-28), 25.0 (CH2, C-6), 30.0 (CH2, C-2), 33.5 (CH2,
C-7), 35.9 (CH, C-2′), 37.6 (CH2, C-19), 42.1 (C, C-4), 42.4 (CH2,
C-15), 43.4 (CH, C-5), 46.2 (C, C-14), 47.1 (C, C-13), 47.7 (CH,
C-3), 49.9 (CH, C-8), 50.0 (CH2, C-12), 50.1 (CH, C-9), 55.3 (CH,
C-17), 62.0 (CH, C-20), 63.8 (CH2, C-29), 78.4 (CH, C-16), 117.9
(CH, C-1), 140.0 (C, C-10), 178.9 (C, C-1′), 212.1 (C, C-11); ES-
MS m/z 503.3 [M + H]+ (100), 504.3 (90); HRES-MS m/z calcd
for C30H51N2O4, 503.3849; found, 503.3852. Anal. (C30H50N2O4‚
0.5H2O) C, H, N.
1
2928, 1694, 1654, 1516, 1462, 1375, 1095, 1043; H NMR (300
MHz, CDCl3) δ 0.46 (3H, s, H-30), 0.67 (3H, s, H-18), 0.85 (3H,
d, J ) 6.2 Hz, H-21), 1.27 (3H, s, H-28), 1.28 (1H, m, H-6â), 1.43
(1H, dd, J ) 2.3, 13.9 Hz, H-15R), 1.58 (1H, bd, J ) 11.9 Hz,
H-7R), 1.57 (1H, m, H-8â), 1.76 (1H, m, H-6R), 1.81 (1H, m,
H-7â), 1.88 (1H, m, H-2â), 1.89 (1H, m, H-15â), 1.92 (1H, m,
H-9R), 2.02 (1H, m, H-17R), 2.04 (1H, s, H-2′), 2.11 (1H, m,
H-19â), 2.22 (6H, bs, NB-CH3), 2.28 (1H, m, H-2R), 2.32 (1H, d,
J ) 15.8 Hz, H-12â), 2.56 (1H, bd, J ) 15.8 Hz, H-12R), 2.58
(1H, m, H-5R), 2.59 (1H, m, H-20), 3.17 (1H, dd, J ) 4.5, 12.7
Hz, H-29a), 3.33 (1H, d, J ) 14.5 Hz, H-19R), 3.40 (1H, m, H-29b),
4.04 (1H, m, H-3R), 4.07 (1H, m, H-16â), 4.53 (1H, dd, J ) 4.5,
11.0 Hz, OH), 5.46 (1H, bs, H-1), 5.50 (1H, d, J ) 8.9 Hz, NAH);
13C NMR (75.5 MHz, CDCl3) δ 9.9 (CH3, C-21), 10.0 (CH3, C-30),
17.8 (CH3, C-18), 18.9 (CH3, C-28), 23.3 (CH, C-2′), 24.9 (CH2,
C-6), 29.9 (CH, C-2), 33.5 (CH2, C-7), 37.5 (CH2, C-19), 42.0 (C,
C-4), 42.4 (CH2, C-15), 43.4 (CH, C-5), 46.2 (C, C-14), 47.0 (C,
C-13), 48.4 (CH, C-3), 49.9 (CH, C-8), 50.0 (CH2, C-12), 50.1
(CH, C-9), 55.3 (CH, C-17), 62.0 (CH, C-20), 63.8 (CH2, C-29),
78.4 (CH, C-16), 117.9 (CH, C-1), 140.0 (C, C-10), 172.0 (C, C-1′),
212.1 (C, C-11); ES-MS m/z 475.3 [M+H]+ (100), 476.3 (60);
HRES-MS m/z calcd for C28H47N2O4, 475.3536; found, 475.3548.
(4aS,7aR,9aR,10S,11R,12aS,12bR,14bS)-10-[1-(Dimethylami-
no)ethyl]-11-hydroxy-3,9a,12a,14b-tetramethyl-1,4a,5,7,7a,9,
9a,10,11,12,12a,12b,13,14,14a,14b-hexadecahydro-8H-indeno-
[5′¢,4′¢:4′,5′]cyclohepta[1′,2′:3,4]benzo[1,2-d][1,3]oxazin-8-one
(9a). To a solution of compound 7a (171 mg, 0.36 mmol, 1 equiv)
in 2 mL of dichloromethane was added 25% tetraethylammonium
hydroxide (1059 mg, 1.8 mmol, 5 equiv) solution in methanol. The
solvent was evaporated under reduced pressure. The dark red residue
was heated at 240 °C under 0.03 mmHg in a rotating distillation
oven with four bulbs. A product sublimated in 3 h. This crude
product was crystallized in acetone to afford a pale yellow powder
of 9a (125 mg, 76%): mp 284 °C; [R]23D ) +123 (c 0.25, CHCl3);
IR (CHCl3) υmax (cm-1) 3331, 2971, 1694, 1672, 1462, 1383, 1039;
1H NMR (300 MHz, CDCl3) δ 0.70 (3H, s, H-18), 0.79 (3H, s,
H-30), 0.87 (3H, d, J ) 6.6 Hz, H-21), 1.24 (3H, s, H-28), 1.29
(1H, m, H-6â), 1.33 (1H, m, H-7R), 1.50 (1H, dd, J ) 2.3, 13.7
Hz, H-15R), 1.63 (1H, m, H-6R), 1.66 (1H, m, H-8â), 1.74 (1H,
m, H-2â), 1.83 (1H, m, H-5R), 1.89 (1H, m, H-7â), 1.92 (3H, s,
H-2′), 1.96 (1H, m, H-9R), 2.02 (1H, m, H-15â), 2.03 (1H, m,
H-17R), 2.09 (1H, m, H-19â), 2.24 (6H, bs, NB-CH3), 2.34 (1H,
d, J ) 16.0 Hz, H-12â), 2.30 (1H, m, H-2R), 2.56 (1H, bd, J )
16.0 Hz, H-12R), 2.60 (1H, dd, J ) 6.6, 11.0 Hz, H-20), 3.25 (1H,
dd, J ) 6.0, 12.0 Hz, H-3R), 3.34 (1H, d, J ) 15.4 Hz, H-19R),
3.79 (1H, d, J ) 10.2 Hz, H-29R), 4.06 (1H, d, J ) 10.2 Hz,
H-29â), 4.09 (1H, m, H-16â), 5.60 (1H, m, H-1); 13C NMR (75.5
MHz, CDCl3) δ 9.9 (CH3, C-30), 10.0 (CH3, C-21), 17.8 (CH3,
C-18), 18.8 (CH3, C-28), 21.0 (CH3, C-2′), 24.6 (CH2, C-6), 30.6
N-3-Benzoylcycloxobuxidine-F (3) or N-[(3â,4R,5R,9â,16R,
20S)-20-(Dimethylamino)-16-hydroxy-4-(hydroxymethyl)-4,14-
dimethyl-11-oxo-9,19-cyclopregnan-3-yl]benzamide (7g). To a
solution of cycloxobuxidine-F 5 (166 mg, 0.38 mmol, 1 equiv) in
10 mL of methanol was added benzoic anhydride (95 mg, 0.46
mmol, 1.1 equiv). After stirring for 4 h at room temperature, the
mixture was neutralized with a few drops of a 10% sodium
bicarbonate solution (pH ) 8), then made basic with 50 mL of a
10% ammonia solution (pH ) 10), and extracted with dichloro-
methane (3 × 40 mL). The organic layer was washed with brine
(30 mL), dried over Na2SO4, and filtered, and the solvent was
evaporated. The crude product was purified by column chroma-
tography on alumina with dichloromethane-methanol (98:2) as
eluent followed by trituration in acetone to afford a colorless powder
of 7g (156 mg, 76%), whose spectroscopic characteristics are
identical to those described.30 Mp 277 °C; [R]25 ) +52 (c 1.00,
D
CHCl3); IR (CHCl3) υmax (cm-1) 3438, 1645, 1628, 1518, 1487,
1
1384, 1040; H NMR (300 MHz, CDCl3) δ 0.65 (3H, s, H-30),
0.85 (3H, s, H-18), 0.88 (3H, d, J ) 6.6 Hz, H-21), 0.92 (1H, m,
H-6â), 1.03 (1H, d, J ) 3.8 Hz, H-19R), 1.24 (3H, s, H-28), 1.34
(1H, m, H-1R), 1.42 (1H, m, H-7R), 1.52 (1H, dd, J ) 2.6, 14.1
Hz, H-15R), 1.55 (1H, m, H-7â), 1.60 (1H, d, J ) 3.8 Hz, H-19â),
1.70 (1H, m, H-2â), 1.75 (2H, m, H-6R, H-2R), 1.99 (1H, m,
H-17R), 2.03 (1H, m, H-15â), 2.07 (1H, m, H-8â), 2.14 (1H, dd,
J ) 3.3, 12.1 Hz, H-5R), 2.26 (6H, bs, NB-CH3), 2.32 (1H, d, J )
17.2 Hz, H-12â), 2.47 (1H, ddd, J ) 3.0, 3.3, 13.6 Hz, H-1â),
2.54 (1H, d, J ) 17.2 Hz, H-12R), 2.63 (1H, dq, J ) 6.6, 10.8 Hz,
H-20), 3.12 (1H, dd, J ) 3.4, 12.7 Hz, H-29a), 3.40 (1H, dd, J )
10.4, 12.7 Hz, H-29b), 4.12 (1H, ddd, J ) 3.0, 7.3, 10.0 Hz, H-16â),
4.24 (1H, ddd, J ) 4.3, 8.9, 12.5 Hz, H-3R), 4.44 (1H, dd, J )
3.4, 10.4 Hz, OH), 6.05 (1H, d, J ) 8.9 Hz, NAH), 7.48, 7.76
(5H, m, HAr ) 3′,4′,5′,6′,7′); 13C NMR (75.5 MHz, CDCl3) δ 10.0

Acetylcholinesterase Inhibitors in the Triterpene Series
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 22 5321
Table 4. Energy Balances (kcal/mol) after Single-Point Poisson-Boltzmann Calculations of Continuum Solvation Energies, and Values of the
Experimentally Determined Inhibitory Potencies IC50 (nM)a
cmpd
typeb
Etot
Elig
δE1
Esolvtot
Esolvlig
δEsolv
δE2
IC50 (nM)
1
1
I
II
I
II
I
II
I
II
I
II
-10 500.0
-10 495.0
-10 502.0
-10 491.0
-10 483
1.4
1.4
0.4
0.4
24.6
24.6
4
4
1.6
1.6
-120.8
-115.8
-121.8
-110.8
-127.
-1529.2
-1531.0
-1529.4
-1523.3
-1527.5
-1538.3
-1526.3
-1531.4
-1498.1
-1486.3
-17.3
-17.3
-17.5
-17.5
-16.6
-16.6
-17.8
-17.8
-40.6
-40.6
80.4
78.6
80.0
86.3
81.2
70.4
83.3
78.3
134.3
146.1
-40.4
-37.2
-41.0
-24.4
-45.8
-33.6
-42.1
-37.2
-50.3
-40.9
385
385
2790
2790
525
525
372
372
2
9a
9a
15
15
9c
9c
26a
26a
-10 460
-104.
-10 502
-10 492
-10 566.3
-10 568.6
-125.4
-115.4
-184.6
-187.
2
a Given in nanomoles (see text for definition of the energy values). b Type I: oxazine ring at peripheral site; type II: oxazine ring at catalytic site.
(CH2, C-2), 32.9 (C, C-4), 33.5 (CH2, C-7), 37.2 (CH2, C-19), 42.4
(CH2, C-15), 46.1 (C, C-14), 47.0 (C, C-13), 49.3 (CH, C-5), 49.7
(CH, C-8), 50.0 (CH2, C-12), 50.2 (CH, C-9), 55.4 (CH, C-17),
55.8 (CH, C-3), 62.0 (CH, C-20), 74.9 (CH2, C-29), 78.3 (CH,
C-16), 121.6 (CH, C-1), 137.2 (C, C-10), 156.4 (C, C-1′), 211.6
(C, C-11); ES-MS m/z 457.4 [M + H]+ (100), 458.4 (5); HRES-
MS m/z calcd for C28H45N2O3, 457.3430; found, 457.3470; Anal.
(C28H44N2O3) C, H, N.
10-9 to 2 × 10-5 M. Percentage inhibition was calculated relative
to a control sample, and IC50 values displayed represent the mean
( standard deviation for triplicate assays.
Computer-Aided Molecular Modeling. Calculations of mol-
ecules, the different superpositions and the various representations
of molecules and docking were performed using sybyl software
from Tripos (mmff 94 as force field, MOPAC for semiempirical
calculations).
Biochemical Methods. In Vitro AChE Inhibition Assay.
Enzymes EeAChE from electrophorus electricus (reference C
2888), bAChE from bovine erythrocytes (reference C 5021), and
human recombinant hAChE (reference C 1682) were purchased
from Sigma. TcAChE was purified from the electric organ tissue
of T. californica, as described.42,43
The molecular dynamics computations were carried out with the
Discover module of the Accelrys software and the Cff91 force-
field.44 We used the 2 Å resolution X-ray crystal structure of AChE
of Torpedo californica complexed with decamethonium (PDB
access code 1acl37). Throughout the simulations, the protein
backbone was held frozen, and its side chains and the inhibitor
were relaxed. Manual docking, followed by a preliminary round
of energy-minimization (EM), was performed by use of our
computer graphics facilities prior to molecular dynamics. The latter
used the same protocol as in ref 41.45 After 5000 fs initialization
steps at 300 K, 100 steps of molecular dynamics (MD) were
performed. These were done at 300 K during 5000 steps of 1 fs.
Each snapshot was subjected to conjugate gradient energy mini-
mization and stored. The search for the best conformation of the
uncomplexed protein and inhibitors was carried out by the same
protocol. Solvation energies of the selected minima were computed
using the Poisson-Boltzmann (PB) procedure with the Delphi
software (Table 4).46 The solute and solvent dielectric constants
were 4 and 80, respectively.
Etot denotes the energy of the protein-ligand complex. δElig
denotes the lowest energy of the isolated ligand following the above-
mentioned MD and EM protocols. Similarly, Eprot denotes the lowest
energy of the isolated protein using this protocol. δE1 is the
difference between Etot and the sum of Eprot and Elig. Esolvtot is the
PB solvation energy of the protein-ligand complex, while Esolvprot
and Esolvliig are the corresponding solvation energies of the isolated
protein and ligand, respectively. δE2 is the difference between Esolvtot
Inhibition of AChE activity was determined by the spectroscopic
method of Ellman et al.,33 using acetylthiocholine iodide as
substrate, in 96-well microtiter plates. All solutions were brought
to room temperature prior to use. Aliquots of 200 µL of a solution
containing 640 µL of 10 mM DTNB in 0.1 M sodium phosphate,
pH 8.0, 19.2 mL of the same buffer, and 13 µL of a solution of
AChE (100 U/mL) in water, were added to each well, followed by
2 µL of an aqueous solution of the inhibitor. The reaction was
initiated by adding 20 µL of acetylthiocholine iodide (7.5 mM) to
each well and was followed by monitoring the appearance of the
thiolate dianion produced by reduction of DTNB at 412 nm for
120 s at 25 °C in a Molecular Devices Spectra Max 384 Plus plate
reader. Each inhibitor was evaluated at several concentrations in
the range of 10-9 to 2 × 10-5 M. Percentage inhibition was
calculated relative to a control sample (DMSO), and IC50 values
displayed represent the mean ( standard deviation for triplicate
assays.
Kinetic Inhibition Studies. The determination of the type of
inhibition of compounds 26a and 34 was done by plotting the
reciprocal of the rate of the reaction against the reciprocal of the
substrate concentration as Lineweaver-Burk plot using, on a 96-
well microtiter plate, a series of eight concentrations for the inhibitor
(from 2000 to 15 nM by diluting by a factor 2) and eight for the
acetylthiocholine substrate (from 6400 to 80 nM by diluting by a
factor 2) obtained with a Beckman Biomek 3000 and Biomek NX
robots. The kinetics were followed under the same conditions as
in the above experiments.
In vitro BuChE Inhibition Assay. Enzyme hBuChE from
human serum (reference C 9971) was purchased from Sigma.
Inhibition of BuChE activity was determined by the spectroscopic
method of Ellman et al.,33 using butyrylthiocholine iodide as
substrate, in 96-well microtiter plates. All solutions were brought
to room temperature prior to use. Aliquots of 200 µL of a solution
containing 640 µL of 10 mM DTNB in 0.1 M sodium phosphate,
pH 8.0, 19.2 mL of the same buffer, and 13 µL of a solution of
BuChE (100U/mL) in water were added to each well, followed by
2 µL of an aqueous solution of the inhibitor. The reaction was
initiated by adding 20 µL of butyrylthiocholine iodide (7.5 mM)
to each well, followed by monitoring the appearance of the thiolate
dianion produced by reduction of DTNB at 412 nm for 120 s at 25
°C in a Molecular Devices Spectra Max 384 Plus plate reader. Each
inhibitor was evaluated at several concentrations in the range of
and the sum of Esolvprot and Esolvliig. Eprot: -10380.6; Esolvprot
:
-1591.8.
X-ray Crystallography. X-ray Structure Analysis of Com-
pound 1.47 Data were obtained from a small colorless prismatic
crystal (0.50 × 0.35 × 0.17 mm). Empirical formula: C30H48N2O3,
Mw ) 484.70. The compound crystallizes in the triclinic system,
space group P1, chiral, with one molecule in the unit-cell of
parameters a ) 6.135(2), b ) 6.770(3), c ) 17.170(7) Å, R )
92.30(3), â ) 97.87 (3), γ ) 99.35(3)°, V ) 695.6 Å;3 dc ) 1.157
g cm-3, F(000) ) 266, λ (Mo KR) ) 0.710 73 Å, and µ ) 0.074
mm-1. Data were measured with a Nonius Kappa-CCD area-
detector diffractometer, using graphite monochromated Mo KR
radiation, according to the phi and omega scan method, up to θ )
27.0°. A total of 6626 intensity data was collected and reduced to
4963 unique triclinic reflections (Rint ) 0.025).48 Structure was
solved with program SHELXS8649 and refined by full-matrix least-
squares, upon all unique F2 with program SHELXL97.50 All the
hydrogen atoms were located in difference Fourier syntheses and
fitted at theoretical positions. They were assigned an isotropic
displacement parameter equivalent to 1.12 the one of the bonded
atom, 1.15 for those of the methyl groups and treated as riding.

5322 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 22
SauVaˆıtre et al.
(5) Sugimoto, H. Donepezil hydrochloride: A treatment drug for
Alzheimer’s disease. Chem. Rec. 2001, 1, 63-73.
(6) Jann, M. W. Rivastigmine, a new-generation cholinesterase inhibitor
for the treatment of Alzheimer’s disease. Pharmacotherapy 2000,
20, 1-12.
(7) (a) Marco-Contelles, J.; Carreiras, M. d. C.; Rodriguez, C.; Villarroya,
M.; Garcia, A. G. Synthesis and pharmacology of Galantamine.
Chem. ReV. 2006, 106, 116-133. (b) Robinson, D. M.; Plosker, G.
L. Galantamine extended release. CNS Drugs 2006, 20, 673-681.
(8) Rogawski, M. A.; Wenk, G. L. The neuropharmacological basis for
the use of memantine in the treatment of Alzheimer’s disease. CNS
Drug ReV. 2003, 9, 275-308.
(9) Soreq, H.; Seidman, S. Acetylcholinesterase: New roles for an old
actor. Nat. ReV. Neurosci. 2001, 2, 8-17.
(10) Silman, I.; Sussman, J. L. Acetylcholinesterase: Classical and
nonclassical functions and pharmacology. Curr. Opin. Pharmacol.
2005, 5, 1-10.
(11) Johnson, G.; Moore, S. W. The adhesion function on acetylcho-
linesterase is located at the peripheral anionic site. Biochem. Biophys.
Res. Commun. 1999, 258, 758-762.
(12) Giacobini, E. Cholinesterase: New roles in brain function and
Alzheimer’s disease. Neurochem. Res. 2003, 25, 512-522.
(13) Blasina, M. F.; Faria, A. C; Gardino, P. F.; Hokoc, J. N.; Almeida,
O. M.; de Mello, F. G.; Arruti, C.; Dajas, F. Evidence of non-
cholinergic function of acetylcholinesterase during development of
chicken retina as shown by fasciculin. Cell Tissue Res. 2000, 299,
173-174.
Refinement of 328 parameters converged to R1(F) ) 0.0501
calculated with the 4448 observed reflections having I g 2σ (I)
and wR2(F2) ) 0.1419 calculated with all the 4963 unique data.
Goodness-of-fit was 1.024. The residual electron density was found
between -0.18 and 0.20 eÅ-3. Crystallographic results are given
in the cif file and Supporting Information.
Molecular Conformation. The structure has been elucidated
unambiguously by the X-ray structure analysis. The molecule is
shown in Figure 1, with the known absolute configuration of steroid
compounds and the conventional labeling, drawn with program
PLATON.51 The chain fixed at C17 is stabilized by an intramolecular
hydrogen bond of 2.973 Å established between the hydroxyl group
OH16 and the nitrogen NB. All ring conformations can be deduced
from torsion angles values. So, the new ring including atom O1′ is
in an envelope conformation in C4, this atom being deviated by
0.720(2) Å from the mean plane of the other five atoms. As a
consequence of the double bond C1-C10 of 1.336(3) Å, ring A
appears nearly in a half-chair conformation with atoms C3 and C4
deviated, respectively, by -0.268 (2) and 0.527(2) Å from the mean
plane of the other four atoms. Ring B exhibits the chair conforma-
tion of the cycloheptane, with atoms C5 and C10, respectively,
situated at -1.234(2) and -0.961(2) Å, and atom C8 at 0.701(2),
from the central mean plane of the four other atoms (C6, C7, C9,
C19). Ring C can be seen as a flattened chair in C9, atoms C9 and
C13 being deviated by, respectively, -0.369(2) and 0.717(2) Å
from the mean plane of the other four atoms of this ring. Cycle D
exhibits a nearly half-chair conformation, with the atoms C13 and
C14 apart by 0.289(2) and -0.486(2) Å from the mean plane of
the remaining atoms. In the crystal packing, only Van der Waals
contacts are observed.
(14) Munoz, F. J.; Aldunate, R.; inestrosa, N. C. Peripheral binding site
is involved in the neurotrophic activity of acetylcholinesterase.
NeuroReport 1999, 26, 3621-3625.
(15) Sharma, K. V.; Koenigsberger, C.; Brimijoin, S.; Gigbee, J. W. Direct
evidence for an adhesive function in the noncholinergic role of
acetylcholinesterase in neurite outgrowth. J. Neurosci. Res. 2001,
63, 165-175.
(16) Inestrosa, N. C.; Alarcon, R. Molecular interactions of acetylcho-
linesterase with senile plaques. J. Physiol. (Paris) 1998, 92, 314-
344.
Acknowledgment. We thank the CNRS and ICSN for
financial support. T. Sauvaitre and M. Barlier were supported
by fellowhips from the Institut de Chimie des Substances
Naturelles (ICSN-CNRS). Professor J. L. Lallemand is gratefully
acknowledged for his interest in our work, as well as Dr.
Franc¸oise Gue´ritte and Alain Montagnac, responsible for the
natural products and chemical libraries of ICSN. We would also
like to thank Dr. Franc¸oise Kuong-Huu and Dr. Jordi Molgo
for fruitful discussions and R. H. Dodd for carefully reading
the manuscript and for helpful comments. The molecular
dynamics computations were performed on the computers of
CINES (Montpellier, France), which we thank for support. We
also wish to thank Dr. Jean-Edouard Ombetta (Faculte de
Pharmacie de Tours, France) for his help in the Delphi
computations.
(17) (a) Inestrosa, N. C; Alvarez, A.; Calderon, F. Acetylcholienesterase
is a senile plaque component that promotes assembly of amyloid
â-peptide into Alzheimer’s filaments. Mol. Psychiatry 1996, 1, 359-
361. (b) De Ferrari, G. V.; Canales, M. A; Shin, I.; Weiner, L. V.;
Silman, I.; Inestrosa, N. C. A structural motif that promotes amyloid
â-peptide fibril formation. Biochemistry 2001, 40, 10447-10457.
(18) De Ferrari, G. V.; Canales, M. A.; Shin, I.; Weiner, L. M.; Silman,
I.; Inestrosa, N. C. A structural motif of acetylcholinesterase that
promotes amyloid â-peptide fibril formation. Biochemistry 2001, 40,
10447-10457.
(19) Bartolini, M.; Bertucci, C.; Cavrini, C.; Andrisano, V. â-Amyloid
aggregation induced by human cholinesterase: Inhibition studies.
Biochem. Pharmacol. 2003, 65, 407-416.
(20) (a) Mary, A.; Renko, D. Z.; Guillou, C.; Thal, C. Potent acetyl-
cholinesterase inhibitors: Design, synthesis and structure-activity
relationships of galanthamine bis-ligands. Bioorg. Med. Chem. 1998,
6, 1835-1850. (b) Guillou, C.; Mary, A.; Renko, D. Z.; Gras, E.;
Thal, C. Potent acetylcholinesterase inhibitors: Design, synthesis and
structure-activity relationships of alkylene linked bis-galanthamine
and galanthaminium salts. Bioorg. Med. Chem. Letters 2000, 10,
637-639. (c) Greenblatt, H. M.; Guillou, C.; Guenard, D., Argaman,
A.; Botti, S.; Badet, B.; Thal, C.; Silman, I.; Sussman; J. L. The
complex of a bivalent derivative of galanthamine with Torpedo
acetylcholinesterase displays drastic deformation of the active-site
gorge: Implications for structure-based drug design. J. Am. Chem.
Soc. 2004, 126, 15405-15411.
Supporting Information Available: Detailed experimental
procedures, compound characterization data for 7a-7i, 8b-8i, 9a-
9i, 10-19, 20a-26b, and 27-33 and crystal data and structure
refinement for compound 1. This material is available free of charge
via the Internet at http://pubs.acs.org.
(21) Vincent, D.; Mathou, T. Inhibiting action of boxwood (Buxus
semperVirens and Buxus balearica) on the cholinesterase of serum.
C. R. Hebd. Seances Acad. Sci. 1945, 220, 474-476.
References
(1) Walsh, D. M.; Selkoe, D. J. Deciphering the molecular basis of
memory failure in Alzheimer’s disease. Neuron 2004, 44, 181-193.
(2) Hardy, J.; Selkoe, D. J. The amyloid hypothesis of Alzheimer’s
disease: Progress and problems on the road to therapeutics. Science
2002, 297, 353-356.
(3) Tolnay, M.; Probst, A. Review: tau protein pathology in Alzheimer
disease and related disorders. Neuropathol. Appl. Neurobiol. 1999,
25, 171-187.
(4) (a) Bartus, R. T.; Dean, R. L., 3rd; Beer, B.; Lippa, A. S. The
cholinergic hypothesis of geriatric memory dysfunction. Science 1982,
217, 408-414. (b) Dunnett, S. B.; Fibiger, H. C. Role of forebrain
cholinergic system in learning and memory: relevance to the
cognitive deficits of aging and Alzheimer’s dementia. Prog. Brain
Res. 1993, 98, 413-420. (c) Bartus, R. T. On neurodegenerative
diseases, models and treatment strategies: Lessons learned and
lessons forgotten a generation following the cholinergic hypothesis.
Exp. Neurol. 2000, 163, 495-529.
(22) Vincent, D.; Sero, I.; Laurent, R. Donne´es pharmacologiques sur le
buis. Therapie 1948, 3, 29-32.
(23) (a) Choudhary, M. I.; Shahnaz, S.; Parveen, S.; Khalid, A.; Ayatollahi,
S. A. M.; Atta-ur-Rahman; Parvez, M. New triterpenoid alkaloid
cholinesterase inhibitors from Buxus hyrcana. J. Nat. Prod. 2003,
66, 739-742. (b) Atta-ur-Rahman; Parveen, S.; Khalid, A.; Farooq,
A.; Ayatollahi, S. A. M.; Choudhary, M. I. Acetylcholinesterase
inhibiting triterpenoidal alkaloids from Buxus hyrcana. Heterocycles
1998, 49, 481-488.
(24) Atta-ur-Rahman Parveen, S.; Khalid, A.; Farooq, A.; Choudhary, M.
I. Acetyl and butyrylcholinesterase-inhibiting triterpenoids from
Buxus papillosa. Phytochemistry 2001, 58, 963-968.
(25) Atta-ur-Rahman; Zaheer ul, H.; Feroz, F.; Khalid, A.; Nawaz, S. A.;
Khan, M. R.; Choudhary, M. I. Cholinesterase-inhibiting new
steroidal alkaloids from Sarcococca hookeriana of Nepalese origin.
HelV. Chim. Acta 2004, 87, 439-448.

Acetylcholinesterase Inhibitors in the Triterpene Series
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 22 5323
(26) Atta-ur-Rahman; Feroz, F.; Zaheer ul, H.; Nawaz, S. A.; Khan, M.
R.; Choudhary, M. I. New steroidal alkaloids from sarcococca. Nat.
Prod. Res. 2003, 17, 235-241.
(27) Kalauni, S. K.; Choudhary, M. I.; Khalid, A.; Manandhar, M. D.;
Shaheen, F.; Atta-ur-Rahman; Gewali, M. B. New cholinesterase
inhibiting steroidal alkaloids from the leaves of Sarcococca coriacea
of Nepalese origin. Chem. Pharm. Bull. 2002, 50, 1423-1426.
(28) Herlem, D. The`se: Alkaloids from Buxus balearica Willd (Buxace´es)s
Isolation and structural determination; Science Faculty of Paris: Paris,
1967.
(29) Compound 1 has been previously prepared by pyrolysis at 210 °C
under 0.01 mmHg in the presence of tetramethylammonium hydrox-
ide. Herlem-Gaulier, D.; Kuong-Huu-Laine´, F.; Goutarel, R.;
Magdeleine, M. J. Alcalo¨ıdes ste´ro¨ıdiques LXXII. Alcalo¨ıdes du
groupe des cylcloxobuxidines retire´s du Buxus balearica Willd: N-3-
iso-Butyryl cycloxobuxidine-F (bale´abuxidine), N-3-isobutyryl cy-
cloxobuxidine-H, N-3-benzoyl cycloxobuxidine-F et cycloxobuxox-
azine-C. Bull. Soc. Chim. Fr. 1968, 2, 763-773.
phosphomannose isomerase from Candida albicans studied by
polarizable molecular mechanics and quantum mechanics. J. Comput.
Chem. 2007, 28, 938-957.
(40) Wu, H. J.; Roux, B. Calculation of absolute protein-ligand binding
free energy from computer simulation. Proc. Natl. Acad. Sci. U.S.A.
2005, 102, 6825-6830.
(41) (a) Chipot, C.; Pearlman, D. A. Free energy calculations. The long
and winding gilded road. Mol. Simul. 2002, 28, 1-12. (b) Chipot,
C.; Rozanska, X.; Dixit, S. B. Can free energy calculations be fast
and accurate at the same time? Binding of low-affinity, nonpeptide
inhibitors to the SH2 domain of the src protein. J. Comp.-Aided Mol.
Des. 2005, 19, 765-770.
(42) Raves, M. L.; Harel, M.; Pang, Y. P.; Silman, I.; Kozikowski, A. P.;
Sussman, J. L. 3D structure of acetylcholinesterase complexed with
the nootropic alkaloid, (-)-huperzine A. Nat. Struct. Biol. 1997, 4,
57-63.
(43) Sussman, J. L.; Harel, M.; Frolow, F.; Varon, L.; Toker, L.; Futerman,
A. H.; Silman, I., Purification and crystallization of a dimeric form
of acetylcholinesterase from Torpedo californica subsequent to
solubilization with phosphatidylinositol-specific phospholipase C. J.
Mol. Biol. 1988, 203, 821-823.
(44) Accelrys Inc., 9685 Scranton Rd., San Diego, CA 92121-3752.
(45) Plocki, S.; Tavare`s-Camarinha, F.; Ahamada-Himidi, A.; Aoun, D.;
Dong, C. Z.; Massicot, F.; Huet, J.; Adolphe-Pierre, S.; Chau, F.;
Ombetta, J. E.; Gresh, N.; Godfroid, J. J.; Heymans, F. Inhibition of
secretory phospholipase A2. Molecular modeling, design, and
synthesis of less lipophilic derivatives of 4,5-dihydro-3-(4-tetradecyl-
oxybenzyl)-1,2,4H oxadiazol-5-one specific for group II enzyme. Eur.
J. Org. Chem. 2005, 2747-2757.
(30) Shaw, J. T.; Corbett, W. L.; Layman, D. L.; Cuny, G. D.; Kerschner,
J. Fused s-triazino heterocycles. XV. 13H-4,6,7,13a,13c-pentaaza-
benzo[hi]chrysene, 13H,4,7,13a,13c- tetraazabenzo [hi]chrysene and
7H-3,7,10,10b-tetraazacyclohepta[de]naphtalene, three new ring sys-
tems J. Heterocycl. Chem. 1988, 25, 1837-1840.
(31) Mary, A.; Zafiarisoa, D.; Guillou, C.; Thal, C. Selective N-
demethylation of galanthamine via a nonclassical Polonovski reaction.
Tetrahedron Lett. 1997, 38, 5151-5152.
(32) Ellman, G. L.; Courtney, K. D.; Andres, V.; Featherstone, R. M. A
new and rapid colorimetric determination of acetylcholinesterase
activity. Biochem. Pharmacol. 1961, 7, 88-95.
(33) Giacobini, E. Butyrylcholinesterase its functions and inhibitors;
Martin Dunitz: 2003.
(46) Honig, B.; Sharp, K.; Yang, A. S. Macroscopic models of aqueous
solutions: Biological and chemical applications. J. Phys. Chem. 1993,
97, 1101-1109.
(34) Kuhl D. E.; Koeppe R. A.; Snyder, S. E.; Minoshima, S., Frey, K.
A.; Kilbourn, M. R. Ann. Neurol. 2006, 59, 13-20.
(47) CCDC 651096 contains the Supplementary Crystallographic data for
compound 1. These data can be obtained free of charge via
www.ccdc.cam.ac.uk or from the Cambridge Crystallographic Data
Centre, 12 Union Road, Cambridge CB2 1EZ, U.K. (fax: +44-
(0)1223-336033, E-mail: deposit@ccdc. cam. ac.uk).
(48) Otwinowski, Z.; Minor, W. Processing of X-ray Diffraction Data
Collected in Oscillation Model, Methods in Enzymology, Volume
276: Macromolecular Crystallography, part A; Carter, C. W., Jr.,
Sweet, R. M., Eds.; Academic Press: New York, 1997; pp 307-
326.
(49) Sheldrick, G. M. SHELXS86. Acta Cryst. A 1990, 46, 467-473.
(50) Sheldrick, G. M. SHELX97, Program for the Refinement of Crystal
Structures; Univ. of Go¨ttingen: Germany, 1997.
(51) (a) Spek, A. L.; Platon, A multipurpose crystallographic tool; Utrecht
University: Utrecht, The Netherlands. (b) Spek, A. L. Single-crystal
structure validation with the program Platon. J. Appl. Cryst. 2003,
36, 7-13.
(35) Khalid, A.; Zaheer-ul-Haq Anjum, S.; Riaz Khan, M.; Atta-ur-
Rahman, Choudhary, M. I. Kinetic and structure-activity relationship
studies on pregnane-type steroidal alkaloids that inhibit cholinest-
erases. Bioorg. Med. Chem. 2004, 12, 1995-2003.
(36) Robaire, B.; Kato, G. Effect of edrophonium, eserine, decamethonium,
D-tubocurarine, and gallamine on the kinetics of membrane-bound
and solubilized Eel acetylcholinesterase. Mol. Pharmacol. 1975, 11,
722-734.
(37) Harel, M.; Schalk, I.; Ehret-Sabattier, L.; Bouet, F.; Goeldner, M.;
Hirth, C.; Axelsen, P.; Silman, I.; Sussman, J. L. Quaternary ligand
binding to aromatic residues in the active-site gorge of acetyl-
cholinesterase. Proc. Nat. Acad. Sci. U.S.A. 1993, 90, 9031-9035.
(38) Khalid, A.; Azim, M. K.; Parveen, S.; Atta-ur-Rahman, Choudhary,
M. I. Structural basis of acetylcholinesterase inhibition by triter-
penoidal alkaloids. Biochem. Biophys. Res. Comm. 2005, 331, 1528-
32.
(39) For a recent application, see: Roux, C.; Gresh, N.; Perera, L. E.;
Piquemal, J.-P.; Salmon, L. Binding of 5-phospho-D-arabinono-
hydroxamate and 5-phospho-D-arabinonate inhibitors to zinc
JM070536W