Crystal Structures of Human Cholinesterases in Complex with Huprine W and Tacrine: Elements of Specificity for Anti-Alzheimer's Drugs Targeting Acetyl- and Butyrylcholinesterase.

Abstract

The multifunctional nature of Alzheimer's disease calls for multi-target directed ligands (MTDLs) able to act on different components of the pathology, like the cholinergic dysfunction and amyloid aggregation. Such MTDLs are usually based on cholinesterase inhibitors (e.g., tacrine or huprine) coupled to another active molecule aimed at a different target. To aid in the design of these MTDLs we report the crystal structures of human acetylcholinesterase (hAChE) in complex with FAS-2 and a hydroxylated derivative of huprine (huprine W), and of human butyrylcholinesterase (hBChE) in complex with tacrine. Huprine W in hAChE and tacrine in hBChE reside in strikingly similar positions highlighting the conservation of key interactions, namely, π-π/cation-π interactions with Trp86(82), and hydrogen bonding with the main chain carbonyl of the catalytic histidine. Huprine W forms additional interactions with hAChE which explains its superior affinity: the isoquinoline moiety is associated with a group of aromatic residues (Tyr337, Phe338 and Phe295 not present in hBChE) in addition to Trp86; the hydroxyl group is hydrogen bonded to both the catalytic serine and residues in the oxyanion hole; and the chlorine substituent is nested in a hydrophobic pocket interacting strongly with Trp439. There is no pocket in hBChE able to accommodate the chlorine substituent.

Full text

Biochem. J. (2013) 453, 393–399 (Printed in Great Britain) doi:10.1042/BJ20130013 393
Crystal structures of human cholinesterases in complex with huprine W
and tacrine: elements of specificity for anti-Alzheimer’s drugs targeting
acetyl- and butyryl-cholinesterase
Florian NACHON*1, Eug´
enie CARLETTI∗, Cyril RONCO†, Marie TROVASLET∗, Yvain NICOLET‡, Ludovic JEAN†and
Pierre-Yves RENARD†
*D´
epartement de Toxicologie, Institut de Recherche Biom´
edicale des Arm´
ees, 24 Avenue des Maquis du Gr´
esivaudan, BP87, 38702 La Tronche, France, †Normandie Universit´
e,
COBRA, UMR 6014 and FR 3038; Universit´
e Rouen, INSA Rouen; CNRS, 1 rue Tesni`
ere, 76821 Mont-Saint-Aignan, Cedex, France, and ‡Laboratoire de Cristallographie et
Cristallogen`
ese des Prot´
eines, Institut de Biologie Structurale ‘J.P. Ebel’, CEA, CNRS, Universit´
e Joseph Fourier, 41 rue J. Horowitz, 38027 Grenoble, France
The multifunctional nature of Alzheimer’s disease calls for
MTDLs (multitarget-directed ligands) to act on different
components of the pathology, like the cholinergic dysfunction
and amyloid aggregation. Such MTDLs are usually on the basis
of cholinesterase inhibitors (e.g. tacrine or huprine) coupled with
another active molecule aimed at a different target. To aid in
the design of these MTDLs, we report the crystal structures
of hAChE (human acetylcholinesterase) in complex with FAS-
2 (fasciculin 2) and a hydroxylated derivative of huprine (huprine
W), and of hBChE (human butyrylcholinesterase) in complex
with tacrine. Huprine W in hAChE and tacrine in hBChE reside
in strikingly similar positions highlighting the conservation of
key interactions, namely, π-π/cation-πinteractions with Trp86
(Trp82), and hydrogen bonding with the main chain carbonyl
of the catalytic histidine residue. Huprine W forms additional
interactions with hAChE, which explains its superior affinity:
the isoquinoline moiety is associated with a group of aromatic
residues (Tyr337,Phe
338 and Phe295 not present in hBChE) in
addition to Trp86; the hydroxyl group is hydrogen bonded to both
the catalytic serine residue and residues in the oxyanion hole;
and the chlorine substituent is nested in a hydrophobic pocket
interacting strongly with Trp439. There is no pocket in hBChE that
is able to accommodate the chlorine substituent.
Key words: acetylcholinesterase, Alzheimer’s disease, bu-
tyrylcholinesterase, huprine, inhibitor, tacrine, X-ray structure.
INTRODUCTION
AD (Alzheimer’s disease), which is the most common cause of
senile dementia, is a major public health issue with devastating
economic and human impacts. As of 2011, the estimated number
of patients needing treatment is 7–8 million in Europe and 4–
5 million in the U.S.A., with a total of 24 million worldwide.
Considering the aging of the world’s population, the situation
is expected to worsen, with the predicted burden of AD victims
reaching 42 million by 2020 [1].
AD results from a neurodegenerative process occurring in
the central nervous system. It is clinically characterized by
a loss of memory and cognition that is associated with
deterioration of the basal forebrain cholinergic neuronal network,
resulting in a reduction in the level of the neurotransmitter
acetylcholine [2]. AD is histologically characterized by aberrant
proteinaceous deposits: (i) from Aβ(β-amyloid) peptide outside
neurons; and (ii) from microtubule-associated Tau protein inside
neurons. The aetiology of AD is not completely understood,
yet it is clearly complex and multifactorial. Although several
treatment strategies have been proposed [3,4], most current
therapeutic options involve restoring acetylcholine levels in the
brain. hAChE [human AChE (acetylcholinesterase)] inhibitors
donepezil (Aricept®), rivastigmine (Exelon®) and galanthamine
(Reminyl®) are currently approved anti-AD drugs [5]. In previous
studies, BChE (butyrylcholinesterase) received attention owing to
its role in modulating acetylcholine levels in cholinergic neurons
under normal conditions [6] and when AChE activity decreases
[7,8]. Consequently, both enzymes are important targets in AD
treatment [9–11].
The complex aetiology of AD has prompted the development of
MTDLs (multitarget-directed ligands) that act simultaneously on
different components of AD pathology, e.g. hAChE and amyloid.
Inhibitors of hAChE have been used as scaffolds to synthesize
such MTDLs [12–15].
Rational design of more potent hAChE and hBChE (human
BChE) inhibitors would benefit from a better understanding
of how the current inhibitors bind to the active sites of these
enzymes. Structural data on hAChE in complex with anti-AD
drugs have been available only recently and have shown how
important it is to analyse complexes formed with the human
enzyme rather than with AChE from other species [16]. Yet, 3D
structures of AChE in complex with inhibitors suitable for the
design of MTDLs are available currently only for mAChE (mouse
AChE) and TcAChE (Torpedo californica AChE) [17–19]. There
is no structural information on hBChE in complex with suitable
inhibitors. In the present paper, we describe the crystal structures
of hAChE in complex with FAS-2 (fasciculin 2) and huprine W
and hBChE bound to tacrine (see Figure 1 for the structures of
these inhibitors). The stabilizing effect of FAS-2 was required to
Abbreviations used: AChE, acetylcholinesterase; AD, Alzheimer’s disease; BChE, butyrylcholinesterase; CHO, Chinese-hamster ovary;
Dm
AChE,
Drosophila melanogaster
AChE; ESRF, European Synchrotron Radiation Facility; FAS-2, fasciculin 2; hAChE, human AChE; hBChE, human BChE;
mAChE, mouse AChE; MTDL, multitarget-directed ligand; PRAD, proline-rich attachment domain;
Tc
AChE,
Torpedo californica
acetylcholinesterase;
WAT, tryptophan amphiphilic tetramerization.
1To whom correspondence should be addressed (email florian@nachon.net).
The atomic co-ordinates and structure factors for the structures of hAChE in complex with huprine W and FAS-2, and hBChE in complex with tacrine,
have been deposited in the PDB under accession codes 4BDT and 4BDS respectively.
c
The Authors Journal compilation c
2013 Biochemical Society
Biochemical Journal www.biochemj.org

394 F. Nachon and others
Figure 1 Chemical structure of tacrine, (7
S
,11
S
)-huprine W and (7
S
,11
S
)-
huprine triazole
Tacrineand huprine W are both used in the present study to create complexes with cholinesterases
for X-ray crystallography. Huprine triazole is described in the Discussion section.
obtain crystals of our preparation of full-length hAChE [20]. Both
of these structurally related ligands are suitable for the design of
MTDLs, and their complex formed with human cholinesterases
provide information that can be used in the design of inhibitors
specific to each enzyme.
EXPERIMENTAL
Chemicals
Huprine W was synthesized as a racemic mixture [21]. FAS-
2 from Bungarus venom was purchased from Latoxan. All other
chemicals including tacrine were purchased from Sigma–Aldrich.
Recombinant hAChE and hBChE
The synthetic gene (GeneArt) coding for the full cDNA of hAChE
was inserted into a pGS vector carrying the glutamine synthetase
gene marker and expressed in CHO (Chinese-hamster ovary)-
K1 cells. The cells were maintained in serum-free Ultraculture
Medium (BioWhittaker) and transfected using jetPEI following
the recommendations of the supplier (Polyplus). Transfected
clones were selected by incubation in medium containing
methionine sulfoximine. The enzymes, secreted into the culture
medium, were purified by affinity chromatography and ion-
exchange chromatography as described previously [20].
Recombinant hBChE (L530stop) is a truncated monomer
containing residues 1–529, but is missing 45 C-terminal residues
that include the tetramerization domain. Four of the nine
carbohydrate attachment sites were deleted by site-directed
mutagenesis. Mutagenesis of Asn486 resulted in glycosylation of
Asn485, an asparagine that is not glycosylated in native hBChE, so
that the recombinant hBChE contains six N-linked glycans [22]
The recombinant hBChE gene was expressed in CHO-K1 cells,
secreted into serum-free culture medium, and purified by affinity
and ion-exchange chromatographies as described previously [22].
The enzymes were concentrated to 14 mg/ml (hAChE) or
7 mg/ml (hBChE) using a Centricon-30 ultrafiltration microcon-
centrator (30000 molecular mass cut-off, Amicon, Millipore) in
10 mM Mes buffer, pH 6.5. The enzyme concentration was de-
termined from its absorbance at 280 nm using a molar absorbance
coefficient of 1.7 for 1 mg/ml hAChE [23] or 1.8 for 1 mg/ml
hBChE [24].
Crystallization of hBChE complexed with tacrine
hBChE was crystallized using the hanging drop vapour diffusion
method as described previously [22]. The mother liquor contained
1 mM tacrine and 0.1 M Mes (pH 6.5) and 2.1 M ammonium
sulfate. A similar set-up was used to make crystals of the hBChE–
huprine W complex, but failed to yield crystals large enough for
diffraction analysis (shower of microcrystals). Crystals of the
hBChE–tacrine complex grew in a couple of weeks at 20◦C.
Crystals were then washed for a few seconds in a cryoprotectant
solution (0.1 M Mes buffer, 1 mM tacrine, 2.3 M ammonium
sulfate and 20%glycerol) before being flash-cooled in liquid
nitrogen for data collection.
Crystallization of hAChE complexed with FAS-2 and huprine W
Purified hAChE (0.1 mM, 6.5 mg/ml) was combined with FAS-
2 (0.2 mM) and huprine W (1 mM) in 10 mM Hepes buffer,
pH 7.4. The ternary complex was crystallized at a concentration of
0.1 mM, using the hanging drop method as described previously
[25]. The mother liquor was 0.1 M Hepes buffer (pH 7.4) and
1.3 M ammonium sulfate. An equal amount of the protein
and the mother liquor were mixed to yield a 3 μl drop. Crystals
grew within a month at 10◦C. The crystals were transferred to
a cryoprotectant solution (0.1 M Hepes buffer, pH 7.4, 1.6 M
ammonium sulfate and 18%glycerol) for few seconds before
they were flash-cooled in liquid nitrogen.
X-ray data collection and processing, structure determination
and refinement
Diffraction data were collected at 100 K on the ID14-eh2
and ID14-eh4 beam-lines at the ESRF (European Synchrotron
Radiation Facility). All datasets were processed with XDS (X-ray
Detector Software) [26], intensities of integrated reflections were
scaled using XSCALE and structure factors were calculated using
XDSCONV. The structures were solved with the CCP4 suite [27]
using the recombinant hBChE structure (PDB code 1P0I) or the
hAChE structure (PDB code 2X8B) as starting models. The initial
models were refined by iterative cycles of model building with
Coot [28], then restrained and TLS refinement with Phenix [29].
RESULTS
X-ray data and quaternary arrangement of hAChE
hAChE was co-inhibited by FAS-2 and huprine W in solution. The
addition of FAS-2 was required to obtain diffracting crystals of
the full-length form of hAChE. The co-crystals of the ternary
complex were obtained under conditions identical to those
used previously for hAChE–FAS-2 binary complexes [20,25].
The crystals belong to space group H32 and diffract to 3.1 Å. The
unit structure consists of three pairs of canonical hAChE dimers.
Data and refinement statistics are shown in Table 1.
The long C-terminal helix of the T-splice variants of
cholinesterases contains the WAT (tryptophan amphiphilic
tetramerization) sequence involved in a four-to-one association
with one PRAD (proline-rich attachment domain) [30]. Yet,
the WAT domain has never been observed in crystal structures
of cholinesterases because it was either truncated by design
[16,22,25,31,32] or no matching electron density was visible
c
The Authors Journal compilation c
2013 Biochemical Society

Structures of human cholinesterases with huprine and tacrine 395
Table 1 Crystallographic and refinement statistics
Values in brackets refer to the highest resolution shell.
Parameters Huprine W–FAS-2–AChE Tacrine–BChE
PDB code 4BDT 4BDS
Space group
H
32
I
422
Unit cell parameters (A
˚)
a
,
b
151.6 155.7
c
246.4 127.9
Resolution (A
˚) 58–3.1 (3.2–3.1) 41–2.1 (2.4–2.1)
Completeness (%) 98.1 (93.7) 99.8 (99.9)
R
sym (%) 9.1 (66.4) 6.9 (42.4)
I
/σ
I
16.0 (3.0) 18.4 (4.8)
Number of unique reflections 19677 45776
Redundancy 3.1 (3.1) 6.4 (6.5)
R
cryst (%) 15.9 17.5
R
free (%) 21.9 20.9
RMSD bond length (A
˚) 0.008 0.009
RMSD bond angles (◦) 1.291 1.135
Mean
B
factor (A
˚2) 48.2 40.4
Non-hydrogen atoms
Total 5057 4708
Non-solvent 4938 4430
Solvent 119 278
Figure 2 Crystal packing of hAChE in a complex with FAS-2 and huprine W
Three canonical dimers (red and green pairs) interact via extensive contacts between bound
FAS-2 (yellow and magenta, surface shown) to form a non-physiological hexamer. The long
C-terminal helices of each dimer cross at a 60◦angle and intertwine at the centre of the hexamer.
Huprine W is represented as spheres (grey, carbon; blue, nitrogen; red, oxygen; green, chlorine).
[20,33]. In the structure shown in Figure 2, we were fortunate
to model six previously unreported turns of the WAT domain
including residues 544–567. Although the side chains remain
poorly resolved, the main chain was clearly identified. The
C-terminal helices of the canonical dimers cross each other at
Leu546,ata60
◦angle. This arrangement is difficult to reconcile
with the X-ray structure of the four-to-one complex formed
between an isolated portion of WAT and PRAD, where the
helices run parallel to form a left-handed superhelix structure
around an antiparallel left-handed PRAD helix [34]. In particular,
the crossed helices do not allow the construction of a plausible
tetramer model as that proposed by Dvir et al. [34], thus strongly
indicating that this arrangement is not representative of the
conformation of the helices in the physiological tetramer.
Figure 3 X-ray crystal structures of hAChE in a complex with FAS-2 and
huprine W, and hBChE in a complex with tacrine
Top panel, hAChE in a complex with FAS-2 and huprine W. Bottom panel, hBChE in a complex
with tacrine. Key residues are represented as sticks, with nitrogen atoms in dark blue, sulfur
atoms in yellow and oxygen atoms in red. Conserved water molecules are represented as red
spheres and chlorine atoms as magenta spheres. The
F
o−
F
celectron density maps, calculated
by omitting the ligand, are represented by green mesh (3 σ).
The three canonical dimers interact via FAS-2 molecules,
forming a peculiar non-physiological hexamer with a central
pore. The extremities of the C-terminal helices from each dimer
are intertwined in the pore. A hexameric arrangement for the
same recombinant hAChE (at high concentration) was observed
by electron microscopy (G. Effantin, E. Carletti and J.P. Colletier,
unpublished work).
Interactions of huprine W in the active site of hAChE
A7.5σpeak in the initial Fo–Fcelectron density was present in the
active site of hAChE, which could be unambiguously modelled
as the (7S,11S) configuration of huprine W (Figure 3, top panel).
Met33 of FAS-2 blocks the entrance of the active site gorge by
binding to the peripheral site residues Trp286,Tyr
72,Asp
74 and
Tyr341 (not shown). FAS-2 does not penetrate deep enough into
the gorge to overlap with the binding site of huprine W. Huprine
W is 8.0 Å from FAS-2, with the closest atoms being the amine
nitrogen of huprine W (N2) and Met33-CZ of FAS-2. The absence
of overlap explains the existence of the ternary complex.
Huprine W is stabilized in hAChE by interactions similar to
those described for other huprine derivatives in complex with
c
The Authors Journal compilation c
2013 Biochemical Society

396 F. Nachon and others
Figure 4 Views of the active site structure of hBChE in a complex with tacrine
(A) Side view and (B) top view of the active site structure of hBChE in a complex with tacrine. Key residues are represented as sticks with nitrogen atoms in dark blue, oxygen atoms in red, carbon
atoms in green and dummy atoms in grey. A water molecule is represented as a red sphere. Hydrogen bonds are represented by dashes. The
F
o−
F
celectron density maps, calculated by omitting
the ligands, are represented by green mesh (3 σ).
non-hAChE [36,37]. The quinolinium substructure is embedded
in a remarkable group of aromatic rings including Trp236,Phe
295,
Phe338,Tyr
337 and Trp86 (Figure 3). The central aromatic ring of
the quinolinium is facing Trp86 while the lateral aromatic ring is
facing Tyr337. Following complex formation, Tyr337-Cαshifts by
1 Å, its Cα–Cβbond rotates 60◦counterclockwise, and its Cβ–
Cγbond rotates 30◦clockwise compared with its conformation in
the apo structure (PDB code 1B41). The chlorine atom is nested
in a hydrophobic pocket delimited by Tyr337,Pro
446,Tyr
449, Met443
and Trp439. Chlorine interacts the strongest with Trp439 (3.4 Å) and
induces a 1.3 Å translation of the indole ring compared with the
apo structure. This translation is accompanied by a 1 Å translation
of Tyr449 and 1.5 Å translation of the short helical portion of the
Cys69–Cys96 -loop (consisting of residues Phe80 to Met85 ) located
behind Trp439 (structure not shown). No other significant change
in the conformations of the other active site residues is observed.
Additional stabilization of the quinolinium substructure is
provided by a hydrogen bond between the aromatic nitrogen and
the main chain carbonyl of His447 (2.8 Å). Unlike in previous
huprine–AChE complexes, the closest water molecule is at least
3.4 Å from the amino group, showing that this group does not
interact strongly with the conserved water molecule network of
the active site gorge. The water molecule network was well defined
in the electron density map despite the low-resolution data. The
hydroxyl group of huprine W makes strong hydrogen bonds with
the γ-hydroxyl of Ser203 and the α-amine of Gly122 (2.3 and
2.9 Å). It occupies a similar position as that of the water molecule
that often bridges the catalytic serine residue to the oxyanion hole
in the apo forms of cholinesterases [38].
X-ray structure of the tacrine–hBChE complex
The structure of tacrine-inhibited hBChE was solved at 2.1 Å
resolution (Table 1). The clear elongated peak near Trp82 in
the initial Fo–Fcelectron density map reveals that tacrine binds
to the catalytic site of hBChE in a very similar way to how
huprines and tacrine bind to AChE [32,37,39] (Figure 3, bottom
panel). The key interactions are aromatic stacking with Trp82,
hydrogen bonding between the aromatic nitrogen N7 and the main
chain carbonyl of His438 (3.2 Å), and hydrogen bonding between
the amino group N15 and two water molecules belonging to the
active site water network (both 3.1 Å). These water molecules
are firmly anchored to the protein by hydrogen bonds to Asp70,
Ser79 and Thr120 (2.7, 2.8 and 2.8 Å). The partially saturated ring
of tacrine is remarkably embedded in the water network. Unlike
in AChE, additional stabilization by aromatic stacking involving
the lateral aromatic ring of tacrine is not observed because in
BChE Ala328 replaces Tyr337, the residue involved in the aromatic
stacking interaction in hAChE.
An additional strong positive peak in the initial Fo–Fcelectron
density map is found in the vicinity of the catalytic serine residue
(Figure 4). A similar peak was first observed in the structure of
recombinant hBChE crystallized in the absence of any ligand
[40]. Careful analysis of the density in that case showed that
it could be modelled as a carboxylic acid bound to Ser198 with
an unusually long bond distance of 2.16 Å between Ser198-Oγ
and the carboxylic-C. The best density fit was achieved with a
butyrate molecule. Very recently, a different recombinant hBChE
expressed in insect cells, purified using a new protocol and
crystallized under completely different conditions, displayed a
similar carboxylic acid bound to the serine residue [41]. In that
study, the density was modelled as a β-alanine rather than a
butyrate. Assignment was on the basis of the observation that
one terminal heavy atom was very close to a water molecule
and thus more likely to be a hydrogen bond donor or acceptor.
In the structure shown in Figure 4, the shape of the density is
not linear as was observed for the putative β-alanine or butyrate.
Rather, it is consistent with a small cyclic molecule. After various
attempts with substituted aromatic or saturated six- and five-
atom ring molecules, a remarkably good fit was obtained with
formyl-proline (Figure 4). Yet, we stress that this interpretation
remains a guess, especially knowing that the source of this putative
molecule remains unknown (culture medium, protein degradation
or bacterial contaminant?). The carboxylic carbon of the formyl-
proline is not covalently bonded to the serine residue (2.9 Å). The
serine residue adopts two alternative conformations. One oxygen
atom is partly engaged in the oxyanion hole and makes strong
hydrogen bonds with the Oγof Ser198 and the α-amine of Gly117
(2.7 and 2.7 Å). In addition, the formyl oxygen is 2.9 Å from a
water molecule connected to Thr120 (2.7 Å).
Weaker peaks of electron density close to Tyr332 in the initial
Fo−Fcelectron density map were modelled as a string of dummy
c
The Authors Journal compilation c
2013 Biochemical Society

Structures of human cholinesterases with huprine and tacrine 397
atoms. Obviously, this refinement model is speculative, but we
believe that it is more satisfying to show these unidentified
densities in the PDB model, especially when they are in regions
as important as the active site.
DISCUSSION
Huprine substituents change the conformation of mammalian AChE
Since mAChE and hAChE share identical active site residues, a
comparison of the complexes they form with different huprine
derivatives provides insight into huprine specificities. Figure 5
(top panel) shows that the central three-ring scaffold of huprine
in the mAChE structure is superimposable on the three rings
of huprine W in the hAChE structure (RMSD=0.3 Å). The
presence of the triazole at position 9 vice ethoxyl in huprine
W affects the conformation of Tyr337,Tyr
341 and Asp74 (Figure 5,
top panel, where the residue numbering for mAChE is the same
as that for hAChE). In the hAChE–FAS-2–huprine W structure,
Tyr341 is hydrogen-bonded to Asp74 ,andTyr
337 has sufficient
room to point towards the hydroxyl of the ethoxyl substituent
at position 9. In the mAChE structure [37], Tyr337 must rotate
by 60◦clockwise away from the triazole substituent and towards
Tyr341 . In response, Tyr341 rotates away around its Cα–Cβbond
by 90◦counterclockwise. The rotation of Tyr341 is favoured by
an aromatic stacking interaction with the triazole ring. However,
the hydrogen bond between Tyr341 and Asp74 is lost. The loss of the
hydrogen bond changes the orientation of Asp74, which pushes a
nearby water molecule by 1.1 Å towards the amino group of
huprine allowing the formation of a strong hydrogen bond (2.8 Å)
in the mAChE structure. Thus the absence of an equivalent water
hydrogen bond in the hAChE–huprine W structure seems related
to the nature of the substituent at position 9. These conformation
changes appear independent of the presence of FAS-2 because
no rotation of Tyr341 and subsequent loss of the hydrogen bond
are observed in the structures of apo mAChE/hAChE and FAS2-
mAChE/hAChE [16,18,31].
From this analysis, it is apparent that the conformational
adaptations of active site residues involved in the interaction with
huprines are extremely difficult to predict from the apo structure
alone. Because main chain translations occur, dedicated molecular
docking software, even those implementing flexibility of the
residue side chains, could not predict the various conformational
changes induced by the binding of different huprines. The
prediction of ligand binding to such a complex and flexible active
site as that of hAChE necessitates a more elaborate strategy.
Determination of the crystal structures for representative members
of an inhibitor family seems to be a necessary first step in order to
identify correctly the rules for active site and inhibitor interactions.
To that end, we are examining the crystal structures of hAChE and
hBChE bound to various inhibitors.
Once the basic rules of engagement are known, it should be
possible to accurately dock other derivatives into the active site,
either manually or by using docking software. Water molecules
that are an integral part of the active site architecture would
be conserved at this stage. Further refinement of the binding
conformation then could be performed by means of molecular
dynamics simulations. This phase would let the active site residues
adapt their conformation to the specificities of each derivative.
We described this strategy for huprine W binding to hAChE
in an earlier report [21]. The process predicted the binding
conformation and energy with remarkably accuracy, the RMSD
between the predicted ligand co-ordinates and experimentally
measured co-ordinates being only 0.4 Å.
Figure 5 Superimposition of the active site region of huprine W–hAChE
and huprine triazole–mAChE, and huprine W–hAChE, huprine X-
Tc
AChE, and
tacrine – hBChE
Top panel, huprine W–hAChE (cyan) and huprine triazole–mAChE (yellow). Bottom panel,
huprine W–hAChE (cyan), huprine X–
Tc
AChE (slate) and tacrine–hBChE (green). Key residues
are represented by sticks with oxygen atoms in red, nitrogen atoms in dark blue, sulfur atoms
in yellow and chlorine atoms in magenta. Conserved water molecules are represented as red
spheres and chlorine atoms as magenta spheres. The van der Waals surface of the chlorine atom
of huprine W and Met447-Cεof hBChE is represented as dots. Displayed residues and their
numbering are strictly conserved for hAChE and mAChE.
Selectivity emerging from a hydrophobic pocket
Design of inhibitors that are selective for AChE from various
organisms also requires a better understanding of the inhibitor–
protein interactions than can be obtained by docking and
molecular dynamics simulations alone. This point was described
above by comparing the effect of substituting ethoxy by triazole
on position 9 of huprine. Another example is provided by the
analysis of the hydrophobic chlorine-binding subsite for huprines
complexed with hAChE and TcAChE. Application of the insights
obtained from hAChE and TcAChE can then be used to predict
the consequences of the structure in that region on the binding of
huprines to hBChE, DmAChE (Drosophila melanogaster AChE)
and Culex pipens AChE.
The structures of hAChE in a complex with huprine W, TcAChE
in a complex with huprine X and hBChE in a complex with tacrine
are shown in Figure 5 (bottom panel). The chlorine substituent
on huprine is located in a hydrophobic pocket. For hAChE, this
pocket is defined by Tyr337,Pro
446,Tyr
449, Met443 and Trp439.
For TcAChE, Pro446 is replaced by a bulkier isoleucine residue
(Ile439, note the change in residue numbering), which makes
c
The Authors Journal compilation c
2013 Biochemical Society

398 F. Nachon and others
the hydrophobic pocket hosting the chlorine atom narrower.
Narrowing of the pocket causes a 1 Å shift in the position of
the huprine in TcAChE compared with that in hAChE. Owing
to this shift, the amino group of huprine moves closer to the
water molecule network in TcAChE than in hAChE, allowing
the formation of two hydrogen bonds (3.1 and 3.0 Å), both of
which are absent in the hAChE complex. In addition, Phe330 and
Tyr337 of TcAChE have each adapted their positions to optimize
stacking interactions with the aromatic ring system of the huprine,
resulting in translations of approximately 1.5 Å for each.
Examination of the same region in hBChE reveals that the side
chain of Met437 fills the hydrophobic pocket. As shown in Figure 5
(bottom panel), the van der Waals volume of Met437-Cε(green
dots) overlaps that of the chlorine atom of huprine W bound
to hAChE (magenta dots). Thus the chlorine atom of huprine
would not be able to fit into the active site of hBChE without
a substantial rearrangement of the residues in the preferred
hydrophobic chlorine-binding subsite. This steric restriction is
reflected in the affinity of huprine W for hBChE (IC50 =1200 nM),
where it binds three orders of magnitude more weakly than it
does to hAChE (IC50 =1.1 nM) [21]. This strongly suggests that
huprine W adopts a different orientation when bound to hBChE
than when bound to hAChE. A similar trend is observed for the
IC50 of 6-chlorotacrine, which bears a chlorine atom at the same
position as in huprine W (IC50 =8 nM for hAChE compared with
IC50 =900 nM for hBChE) [42].
By contrast, tacrine is a good fit for the hydrophobic region of
hBChE with the closest atom of tacrine being 3.9 Å from Met437-
Cε(Figure 5, bottom panel). This rationalizes why the inhibition
constant of tacrine for hBChE is only 6-fold lower than that for
TcAChE (Ki=25 nM for hBChE compared with Ki=3.8 nM for
TcAChE) [43].
Another aspect of the hydrophobic chlorine subsite is found
when examining the structure of DmAChE [44]. In this enzyme, a
channel exists connecting the active site gorge to the bulk solvent
in the region of the hydrophobic chlorine-binding pocket. This
channel results from the replacement of Tyr449 in hAChE with an
aspartate residue in the Drosophila enzyme. The X-ray structure
of DmAChE in a complex with benzyl-tacrine (PDB code 1DX4)
shows that position 6 of the tacrine moiety is next to this channel
[32]. It follows that a large substituent at position 6 could avoid
steric hindrance by entering into the channel of DmAChE, whereas
it would not fit in the hydrophobic pocket of hAChE and would
therefore reduce the binding affinity.
The same Tyr→Asp replacement is present in C. pipens AChE
(UnitprotKB Q86GC8) and many other insect AChEs [45]. We
propose that this feature could provide the basis on which to
design specific inhibitors of insect AChE that would not interact
significantly with hAChE.
Implications for the design of MTDLs
The X-ray structure reveals that huprine is closely embedded in
the active site of hAChE. Although close embedding accounts for
great affinity for the enzyme, it limits the number of positions
available on the scaffold to introduce a linker for connecting
a ligand for the second molecular target. In particular, any
substitution on the central tricyclic substructure would lead to
unfavourable steric clashes with active site residues. Substitution
on the amine is an option that has been successfully employed,
but at the price of disrupting the active site water network in
the vicinity of the amine [19]. This is not an issue when the
second ligand targets the peripheral site of the enzyme, because
the synergy of binding to both sites is expected to largely
counterbalance the unfavourable disruption of the water network
[42]. However, this synergy does not exist for ligands that have
no affinity for the peripheral site. Substitution at position 9 of
huprine appears to be the best option as evidenced by the structure
of huprine triazole [37], because the linker can snake outside the
gorge to the bulk solvent without disruption of the active site
structure.
Despite the active site of hBChE being wider than that of
hAChE, tacrine is also tightly embedded, with no apparent
available substitution position that would be devoid of
unfavourable steric effects. The huprine scaffold with substitution
at position 9 appears again to be the best option for hBChE-
specific MTDLs, under the condition to remove the chlorine
substituent at position 3 for the reasons discussed above.
In summary, we presented structures of hAChE and hBChE
in complexes with reversible inhibitors that are of interest for
the design of MTDLs. Comparison of huprine complexes with
mAChE, TcAChE and hAChE illustrates how active site gorge
residues rearrange to better accommodate each derivative, and
highlight the necessity of using full X-ray structure analysis
to develop an accurate understanding of the protein structure
changes that occur on binding, before attempting to make binding
predictions via docking or molecular dynamics procedures. The
structure of hBChE in a complex with tacrine re-enforces this
conclusion.
AUTHOR CONTRIBUTION
Florian Nachon, Yvain Nicolet, Ludovic Jean and Pierre
-
Yves Renard conceptually
developed the project and contributed to writing the paper. Florian Nachon, Cyril Ronco,
Ludovic Jean and Pierre
-
Yves Renard designed and synthesized huprine W. Florian
Nachon, Eug´
enie Carletti, Yvain Nicolet and Marie Trovaslet made the crystals, and
collected and analysed the X
-
ray data.
ACKNOWLEDGEMENTS
We are grateful to Lawrence M Schopfer (Nebraska Medical Center) for the critical reading
of the paper and helpful suggestions before submission. We thank the ESRF for beam-time
under long-term projects (IBS BAG).
FUNDING
C.R. is funded by a Ph.D. fellowship from The French Minist`
ere de l’Enseignement
Sup´
erieur et de la Recherche (MRES). The Institut Universitaire de France (IUF), the
R´
egion Haute Normandie (Crunch program), the Agence Nationale pour la Recherche
[grant numbers ANR 06-BLAN-0163 DETOXNEURO and ANR-09-BLAN-0192 ReAChE]
and the Direction G´
en´
erale de l’Armement [grant number DGA/DSP/STTC PDH-2-NRBC-
3-C-301] are gratefully acknowledged for their financial support.
REFERENCES
1 Jellinger, K. A. (2006) Alzheimer 100: highlights in the history of Alzheimer research.
J. Neural Transm. 113, 1603–1623
2 Daulatzai, M. A. (2010) Early stages of pathogenesis in memory impairment during
normal senescence and Alzheimer’s disease. J. Alzheimer’s Dis. 20, 355–367
3 Holzgrabe, U., Kapkova, P., Alptuzun, V., Scheiber, J. and Kugelmann, E. (2007) Targeting
acetylcholinesterase to treat neurodegeneration. Expert Opin. Ther. Targets 11, 161–179
4 Castro, A., Conde, S., Rodriguez-Franco, M. I. and Martinez, A. (2002) Non-cholinergic
pharmacotherapy approaches to the future treatment of Alzheimer’s disease. Mini-Rev.
Med. Chem. 2, 37–50
5 Smith, D. A. (2009) Treatment of Alzheimer’s disease in the long-term-care setting.
Am. J. Health-Syst. Pharm. 66, 899–907
6 Mesulam, M. M., Guillozet, A., Shaw, P., Levey, A., Duysen, E. G. and Lockridge, O.
(2002) Acetylcholinesterase knockouts establish central cholinergic pathways and can
use butyrylcholinesterase to hydrolyze acetylcholine. Neuroscience 110, 627–639
7 Darvesh, S., Hopkins, D. A. and Geula, C. (2003) Neurobiology of butyrylcholinesterase.
Nat. Rev. Neurosci. 4, 131–138
c
The Authors Journal compilation c
2013 Biochemical Society

Structures of human cholinesterases with huprine and tacrine 399
8 Greig, N. H., Utsuki, T., Ingram, D. K., Wang, Y., Pepeu, G., Scali, C., Yu, Q. S., Mamczarz,
J., Holloway, H. W., Giordano, T. et al. (2005) Selective butyrylcholinesterase inhibition
elevates brain acetylcholine, augments learning and lowers Alzheimer β-amyloid peptide
in rodent. Proc. Natl. Acad. Sci. U.S.A. 102, 17213–17218
9 Venneri, A. and Lane, R. (2009) Effects of cholinesterase inhibition on brain white matter
volume in Alzheimer’s disease. NeuroReport 20, 285–288
10 Venneri, A., McGeown, W. J. and Shanks, M. F. (2005) Empirical evidence of
neuroprotection by dual cholinesterase inhibition in Alzheimer’s disease. NeuroReport
16, 107–110
11 Ballard, C. G. (2002) Advances in the treatment of Alzheimer’s disease: benefits of dual
cholinesterase inhibition. Eur. Neurol. 47, 64–70
12 Leon, R., Garcia, A. G. and Marco-Contelles, J. (2013) Recent advances in the
multitarget-directed ligands approach for the treatment of Alzheimer’s disease.
Med. Res. Rev. 33, 139–189
13 Tumiatti, V., Minarini, A., Bolognesi, M. L., Milelli, A., Rosini, M. and Melchiorre, C.
(2010) Tacrine derivatives and Alzheimer’s disease. Curr. Med. Chem. 17,
1825–1838
14 Cavalli, A., Bolognesi, M. L., Minarini, A., Rosini, M., Tumiatti, V., Recanatini, M. and
Melchiorre, C. (2008) Multi-target-directed ligands to combat neurodegenerative
diseases. J. Med. Chem. 51, 347–372
15 Van der Schyf, C. J., Gal, S., Geldenhuys,W. J. and Youdim, M. B. (2006) Multifunctional
neuroprotective drugs targeting monoamine oxidase inhibition, iron chelation, adenosine
receptors, and cholinergic and glutamatergic action for neurodegenerative diseases.
Expert Opin. Invest. Drugs 15, 873–886
16 Cheung, J., Rudolph, M. J., Burshteyn, F., Cassidy, M. S., Gary, E. N., Love, J., Franklin,
M. C. and Height, J. J. (2012) Structures of human acetylcholinesterase in complex with
pharmacologically important ligands. J. Med. Chem. 55, 10282–10286
17 Greenblatt, H. M., Dvir, H., Silman, I. and Sussman, J. L. (2003) Acetylcholinesterase: a
multifaceted target for structure-based drug design of anticholinesterase agents for the
treatment of Alzheimer’s disease. J. Mol. Neurosci. 20, 369–383
18 Bourne, Y., Taylor, P., Radic, Z. and Marchot, P. (2003) Structural insights into
ligand interactions at the acetylcholinesterase peripheral anionic site. EMBO J. 22,
1–12
19 Bourne, Y., Kolb, H. C., Radic, Z., Sharpless, K. B., Taylor, P. and Marchot, P. (2004)
Freeze-frame inhibitor captures acetylcholinesterase in a unique conformation. Proc. Natl.
Acad. Sci. U.S.A. 101, 1449–1454
20 Carletti, E., Colletier, J. P., Dupeux, F., Trovaslet, M., Masson, P. and Nachon, F. (2010)
Structural evidence that human acetylcholinesterase inhibited by tabun ages through
O-dealkylation. J. Med. Chem. 53, 4002–4008
21 Ronco, C., Foucault, R., Gillon, E., Bohn, P., Nachon, F., Jean, L. and Renard, P. Y. (2011)
New huprine derivatives functionalized at position 9 as highly potent acetylcholinesterase
inhibitors. ChemMedChem 6, 876–888
22 Nachon, F., Nicolet, Y., Viguie, N., Masson, P., Fontecilla-Camps, J. C. and Lockridge, O.
(2002) Engineering of a monomeric and low-glycosylated form of human
butyrylcholinesterase: expression, purification, characterization and crystallization. Eur. J.
Biochem. 269, 630–637
23 Rosenberry, T. L. and Scoggin, D. M. (1984) Structure of humanerythrocyte
acetylcholinesterase. Characterization of intersubunit disulfide bonding and detergent
interaction. J. Biol. Chem. 259, 5643–5652
24 Lockridge, O. and La Du, B. N. (1978) Comparison of atypical and usual human serum
cholinesterase. Purification, number of active sites, substrate affinity, and turnover
number. J. Biol. Chem. 253, 361–366
25 Kryger, G., Harel, M., Giles, K., Toker, L., Velan, B., Lazar, A., Kronman, C., Barak, D.,
Ariel, N., Shafferman, A. et al. (2000) Structures of recombinant native and E202Q mutant
human acetylcholinesterase complexed with the snake-venom toxin fasciculin-II. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 56, 1385–1394
26 Kabsch, W. (2010) XDS. Acta Crystallogr., Sect. D: Biol. Crystallogr. 66, 125–132
27 Collaborative Computational Project 4 (1994) The CCP4 suite: programs for protein
crystallography. Acta Crystallogr., Sect. D: Biol. Crystallogr. 50, 760–763
28 Emsley, P., Lohkamp, B., Scott, W. G. and Cowtan, K. (2010) Features and development of
Coot. Acta Crystallogr., Sect. D: Biol. Crystallogr. 66, 486–501
29 Adams, P. D., Afonine, P. V., Bunkoczi, G., Chen, V. B., Davis, I. W., Echols, N., Headd,
J. J., Hung, L. W., Kapral, G. J., Grosse-Kunstleve, R. W. et al. (2010) PHENIX: a
comprehensive Python-based system for macromolecular structure solution. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 66, 213–221
30 Simon, S., Krejci, E. and Massoulie, J. (1998) A four-to-one association between peptide
motifs: four C-terminal domains from cholinesterase assemble with one proline-rich
attachment domain (PRAD) in the secretory pathway. EMBO J. 17, 6178–6187
31 Marchot, P., Ravelli, R. B., Raves, M. L., Bourne, Y., Vellom, D. C., Kanter, J., Camp, S.,
Sussman, J. L. and Taylor, P. (1996) Soluble monomeric acetylcholinesterase from
mouse: expression, purification, and crystallization in complex with fasciculin. Protein
Sci. 5, 672–679
32 Harel, M., Kryger, G., Rosenberry, T. L., Mallender, W. D., Lewis, T., Fletcher, R. J., Guss,
J. M., Silman, I. and Sussman, J. L. (2000) Three-dimensional structures of
Drosophila
melanogaster
acetylcholinesterase and of its complexes with two potent inhibitors.
Protein Sci. 9, 1063–1072
33 Ngamelue, M. N., Homma, K., Lockridge, O. and Asojo, O. A. (2007) Crystallization and
X-ray structure of full-length recombinant human butyrylcholinesterase. Acta Crystallogr.,
Sect. F: Struct. Biol. Crystal. Commun. 63, 723–727
34 Dvir, H., Harel, M., Bon, S., Liu, W. Q., Vidal, M., Garbay, C., Sussman, J. L., Massoulie,
J. and Silman, I. (2004) The synaptic acetylcholinesterase tetramer assembles around a
polyproline II helix. EMBO J. 23, 4394–4405
35 Reference deleted
36 Dvir, H., Wong, D. M., Harel, M., Barril, X., Orozco, M., Luque, F. J., Munoz-Torrero, D.,
Camps, P., Rosenberry, T. L., Silman, I. and Sussman, J. L. (2002) 3D structure of
Torpedo californica
acetylcholinesterase complexed with huprine X at 2.1 A
˚resolution:
kinetic and molecular dynamic correlates. Biochemistry 41, 2970–2981
37 Ronco, C., Carletti, E., Colletier, J. P., Weik, M., Nachon, F., Jean, L. and Renard, P. Y.
(2012) Huprine derivatives as sub-nanomolar human acetylcholinesterase inhibitors:
from rational design to validation by X-ray crystallography. ChemMedChem 7, 400–405
38 Koellner, G., Kryger, G., Millard, C. B., Silman, I., Sussman, J. L. and Steiner, T. (2000)
Active-site gorge and buried water molecules in crystal structures of acetylcholinesterase
from
Torpedo californica
.J.Mol.Biol.296, 713–735
39 Harel, M., Schalk, I., Ehret-Sabatier, L., Bouet, F., Goeldner, M., Hirth, C., Axelsen, P. H.,
Silman, I. and Sussman, J. L. (1993) Quaternary ligand binding to aromatic residues in
the active-site gorge of acetylcholinesterase. Proc. Natl. Acad. Sci. U.S.A. 90, 9031–9035
40 Nicolet, Y., Lockridge, O., Masson, P., Fontecilla-Camps, J. C. and Nachon, F. (2003)
Crystal structure of human butyrylcholinesterase and of its complexes with substrate and
products. J. Biol. Chem. 278, 41141–41147
41 Brazzolotto, X., Wandhammer, M., Ronco, C., Trovaslet, M., Jean, L., Lockridge, O.,
Renard, P. Y. and Nachon, F. (2012) Human butyrylcholinesterase produced in insect
cells: huprine-based affinity purification and crystal structure. FEBS J. 279, 2905–2916
42 Camps, P., Formosa, X., Galdeano, C., Munoz-Torrero, D., Ramirez, L., Gomez, E.,
Isambert, N., Lavilla, R., Badia, A., Clos, M. V. et al. (2009) Pyrano[3,2-c]quinoline-6-
chlorotacrine hybrids as a novel family of acetylcholinesterase- and β-amyloid-directed
anti-Alzheimer compounds. J. Med. Chem. 52, 5365–5379
43 Berman, H. A. and Leonard, K. (1992) Interaction of tetrahydroaminoacridine with
acetylcholinesterase and butyrylcholinesterase. Mol. Pharmacol. 41, 412–418
44 Nachon, F., Stojan, J. and Fournier, D. (2008) Insights into substrate and product traffic in
the
Drosophila melanogaster
acetylcholinesterase active site gorge by enlarging a back
channel. FEBS J. 275, 2659–2664
45 Lu, Y., Pang, Y. P., Park, Y., Gao, X., Yao, J., Zhang, X. and Zhu, K. Y. (2012) Genome
organization, phylogenies, expression patterns, and three-dimensional protein models of
two acetylcholinesterase genes from the red flour beetle. PLoS ONE 7, e32288
Received 2 January 2013/15 May 2013; accepted 17 May 2013
Published as BJ Immediate Publication 17 May 2013, doi:10.1042/BJ20130013
c
The Authors Journal compilation c
2013 Biochemical Society