Crystal structures of oxime-bound fenamiphos-acetylcholinesterases: Reactivation involving flipping of the His447 ring to form a reactive Glu334-His447-oxime triad

Article abstract of DOI:10.1016/j.bcp.2009.08.027

Organophosphorus insecticides and nerve agents inhibit the vital enzyme acetylcholinesterase by covalently bonding to the catalytic serine residue of the enzyme. Oxime-based reactivators, such as [(E)-[1-[(4-carbamoylpyridin-1-ium-1-yl)methoxymethyl]pyridin-2-ylidene]methyl]-oxoazanium dichloride (HI-6) and 1,7-heptylene-bis-N,N′-2-pyridiniumaldoxime dichloride (Ortho-7), restore the organophosphate-inhibited enzymatic activity by cleaving the phosphorous conjugate. In this article, we report the intermolecular interactions between Mus musculus acetylcholinesterase inhibited by the insecticide fenamiphos (fep-mAChE) and HI-6 or Ortho-7 revealed by a combination of crystallography and kinetics. The crystal structures of the two oxime-bound fep-mAChE complexes show that both oximes interact with the peripheral anionic site involving different conformations of Trp286 and different peripheral-site residues (Tyr124 for HI-6 and Tyr72 for Ortho-7). Moreover, residues at catalytic site of the HI-6-bound fep-mAChE complex adopt conformations that are similar to those in the apo mAChE, whereas significant conformational changes are observed for the corresponding residues in the Ortho-7-bound fep-mAChE complex. Interestingly, flipping of the His447 imidazole ring allows the formation of a hydrogen bonding network among the Glu334-His447-Ortho-7 triad, which presumably deprotonates the Ortho-7 oxime hydroxyl group, increases the nucleophilicity of the oxime group, and leads to cleavage of the phosphorous conjugate. These results offer insights into a detailed reactivation mechanism for the oximes and development of improved reactivators.

Full text of DOI:10.1016/j.bcp.2009.08.027

Biochemical Pharmacology 79 (2010) 507–515
Contents lists available at ScienceDirect
Biochemical Pharmacology
journal homepage: www.elsevier.com/locate/biochempharm
Crystal structures of oxime-bound fenamiphos-acetylcholinesterases:
Reactivation involving flipping of the His447 ring to form a reactive
$
Glu334–His447–oxime triad§,§§,
a
a
b
a
Andreas Ho¨rnberg a, , Elisabet Artursson , Rikard Wa¨rme , Yuan-Ping Pang , Fredrik Ekstro¨m
*
a
˚
Swedish Defence Research Agency, Division of CBRN Defence and Security, S-901 82 Umea, Sweden
b Computer-Aided Molecular Design Laboratory, Mayo Clinic, Rochester, MN, United States of America
A R T I C L E I N F O
A B S T R A C T
Article history:
Organophosphorus insecticides and nerve agents inhibit the vital enzyme acetylcholinesterase by
covalently bonding to the catalytic serine residue of the enzyme. Oxime-based reactivators, such as [(E)-
[1-[(4-carbamoylpyridin-1-ium-1-yl)methoxymethyl]pyridin-2-ylidene]methyl]-oxoazanium dichlor-
ide (HI-6) and 1,7-heptylene-bis-N,N0-2-pyridiniumaldoxime dichloride (Ortho-7), restore the organo-
phosphate-inhibited enzymatic activity by cleaving the phosphorous conjugate. In this article, we report
the intermolecular interactions between Mus musculus acetylcholinesterase inhibited by the insecticide
fenamiphos (fep-mAChE) and HI-6 or Ortho-7 revealed by a combination of crystallography and kinetics.
The crystal structures of the two oxime-bound fep-mAChE complexes show that both oximes interact
with the peripheral anionic site involving different conformations of Trp286 and different peripheral-
site residues (Tyr124 for HI-6 and Tyr72 for Ortho-7). Moreover, residues at catalytic site of the HI-6-
bound fep-mAChE complex adopt conformations that are similar to those in the apo mAChE, whereas
significant conformational changes are observed for the corresponding residues in the Ortho-7-bound
fep-mAChE complex. Interestingly, flipping of the His447 imidazole ring allows the formation of a
hydrogen bonding network among the Glu334–His447–Ortho-7 triad, which presumably deprotonates
the Ortho-7 oxime hydroxyl group, increases the nucleophilicity of the oxime group, and leads to
cleavage of the phosphorous conjugate. These results offer insights into a detailed reactivation
mechanism for the oximes and development of improved reactivators.
Received 1 July 2009
Accepted 27 August 2009
Keywords:
Acetylcholinesterase
Crystallography
Histidine-flip
Oxime
Organophosphorus compound
ß 2009 Elsevier Inc. All rights reserved.
1. Introduction
The protein acetylcholinesterase (AChE, EC 3.1.1.7)1 terminates
cholinergic transmission by hydrolysing the neurotransmitter
acetylcholine. AChE is
susceptible to irreversible inhibition by a class of organopho-
sphates (OPs, e.g., nerve agents and pesticides) which form a
covalent bond to the catalytic Ser203O . The catalytic site (CAS) of
a very efficient enzyme and highly
§
This work was supported by the Swedish Armed Forces Research and
Technology Program and the Swedish Emergency Management Agency. Yuan-
Ping Pang was supported by the United States Army Medical Research & Materiel
Command (W81XWH-08-1-0154) and the University of Minnesota/Mayo/IBM
Collaboration Seed Grant Program.
g
˚
AChE is located close to the bottom of a ꢀ20-A deep and narrow
§§
The pdb entry codes are 2WU3 and 2WU4 for the HI-6ꢁfep-mAChE and Ortho-
active-site gorge which accommodates the conjugate through non-
bonded interactions at the oxyanion hole (main-chain nitrogen
atoms of Gly121, Gly122 and Ala204), the acyl pocket (side chains
of the aromatic residues Trp236, Phe295, Phe297 and Phe338) and
the choline-binding site (side chains of Trp86 and Tyr337; [1]). At
the entrance of the gorge, the peripheral anionic site (PAS; side
chains of Tyr72, Asp74, Tyr124, Trp286 and Tyr341) constitutes a
binding site for various allosteric modulators and reactivators [2–
9]. Even though the interactions of the OP conjugate with AChE are
strengthened by a covalent bond, hydrophobic interactions, and
7ꢁfep-mAChE structures, respectively.
$
The sequence numbering of mAChE is used throughout this paper when amino
acids of any AChE are discussed.
Abbreviations: AChE, acetylcholinesterase; hAChE, Homo sapiens acetylcholinester-
ase; mAChE, Mus musculus acetylcholinesterase; CAS, catalytic active site; HI-6,
[(E)-[1-[(4-carbamoylpyridin-1-ium-1-yl)methoxymethyl]pyridin-2-ylidene]-
methyl]-oxoazanium dichloride; OP, organophosphate; Ortho-7, 1,7-heptylene-bis-
N,N0-2-pyridiniumaldoxime dichloride; PAS, peripheral anionic site; sarin, iso-
propyl methylphosphonofluoridate.
* Corresponding author.
E-mail address: andreas.hornberg@foi.se (A. Ho¨rnberg).
0006-2952/$ – see front matter ß 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.bcp.2009.08.027

508
A. Ho¨rnberg et al. / Biochemical Pharmacology 79 (2010) 507–515
hydrogen bonds, spontaneous reactivation (i.e., cleavage of the
phosphorus conjugate) as well as elimination (aging) of one of the
phosphorous substituents can occur [10,11]. The rates (t1/2) of
these reactions vary from ꢀ3 min to several days depending upon
chemical structures of the phosphorous substituents [12].
Reactivation of the OP-inhibited AChE can be accelerated by
nucleophilic substances (i.e., antidotes) which attack the phos-
phorus atom and cleave the phosphorus conjugate [13]. These
antidotes generally contain a pyridiniumaldoxime group as an
effective nucleophile (Supplementary material online Fig. S1A and
B); reviewed by [14]). Similar to the spontaneous reactivation or
aging reactions, the oxime-mediated reactivation also shows a
high variability for different OPs [12].
Previous crystallographic studies showed that carboxyamino-
pyridinium-containing oximes and bispyridiniumaldoximes form
cation–pi interactions with the side chains of Tyr124 and Tyr72,
respectively, although they all interact with Trp286, when complex-
ing with non-phosphorylated and phosphorylated mAChE [7–9].
In this study, we present the crystal structures of HI-6 and
Ortho-7 in complex with fep-mAChE (termed HI-6ꢁfep-mAChE and
Ortho-7ꢁfep-mAChE, respectively), and biochemical characteriza-
tion using a racemic mixture and an optically pure enantiomer of
fenamiphos (fep-mAChE) that offers mechanistic interpretations of
these structures. We confirm that carboxyaminopyridinium-
containing oximes and bis-pyridiniumaldoximes have distinctively
different interactions with AChE. Moreover, we suggest that a ring
flipping of the His447 imidazole ring leads to the formation of a
Glu334–His447–Ortho-7 triad which activates Ortho-7 before
reaching the transition state according to our published reactiva-
tion mechanism [9].
detector. The total oscillation range of the crystals was 1808 and
data were collected in 18 increments. The data were processed
using XDS [17] and scaled with SCALA [18]. The structures were
solved using rigid body refinement with the mouse apo protein
(1J06 [19]) as a starting model. Further refinement was carried out
by alternating restrained isotropic B-factor refinement, as imple-
mented in the REFMAC5 program [20], with manual rebuilding of
the model after visualization of the 2jFoj ꢃ jFcj and jFoj ꢃ jFcj maps
using the COOT program [21]. The last few rounds of refinement
were performed using the Phenix program [22]. The final round of
refinement was preceded by addition of both fenamiphos
enantiomers into the structure at 0.5 occupancy. To follow the
progress of refinement, the Rfree value was monitored from the
start, with 2% of the reflections included in the test set [23]. The
Spartan program (Wavefunctions Inc.) was used to obtain bond
angles and bond lengths of the fenamiphos-conjugated Ser203. The
quality of the final models was evaluated with the PROCHECK
program [24]. Figures were generated using the PyMol program
[25]. Statistics of data collection and structure refinement are
shown in Table 1.
2.4. Time-dependent reactivation of wild-type and mutants of mAChE
Recombinant human AChE (rhAChE) was cloned, expressed, and
purified with the same procedure as for the mouse enzyme [16].
Mutants were produced by site-directed mutagenesis as described
previously [9]. Purified mAChE wild-type (ꢀ200
mM) was inhibited
by 50 mM fenamiphos for 3 h to an inhibition grade exceeding 90%
(as a comparison, the aging half-time of fenamiphos is ꢀ140 h;
[12]). Excess inhibitor was removed by four consecutive desalting
columns (Zeba desalt spin column, Pierce Technology) equilibrated
in 20 mM HEPES pH 7.0, 50 mM NaCl, 1% PEG750MME. The
2. Materials and methods
2.1. Reactivators and reagents
Table 1
Data collection and refinement statistics of the complexes.
Ortho-7 was made according to a published procedure [15]. HI-
6 was obtained from Dr. John Clement of Defence Research
Establishment, Canada. HEPES (99.5%), PEG750MME, dichloro-
methane (99.9%), 5,50-dithiobis(2-nitrobenzoic acid) (DTNB, 99%),
acetylthiocholine iodide (ATC, 99%) and fenamiphos (99.5%) were
obtained from Sigma–Aldrich (St. Louis, MO, USA). Sodium chloride
(99.5%), dimethyl sulfoxide (DMSO, 99%), formic acid (99%),
sodiumdihydrogenphosphate (99%) and di-sodiumhydrogenpho-
sphate (99%) were purchased from Merck (Darmstadt, Germany).
Magnesium sulphate (99%) and acetonitrile (99.9%) were obtained
from Scharlau Chemie S.A. (Barcelona, Spain). Sodium hydroxide
was purchased from Eka Chemicals AB (Bohus, Sweden).
Data collection
pdb accession code
Space group
HI-6ꢁfep-mAChE
2WU3
Ortho-7ꢁfep-mAChE
2WU4
P212121
P212121
˚
Unit cell dimensions (A) 78.76 ꢂ 111.23 ꢂ 227.70 79.74 ꢂ 112.70 ꢂ 226.69
˚
Resolution range (A)
No. of observations
No. of unique
29.79–2.70 (2.85–2.70)
415,088 (60,606)
55,859 (8,070)
29.22–2.40 (2.53–2.40)
596,715 (86,550)
80,714 (11,636)
observations
Completeness
99.9 (100.0)
7.4 (7.5)
99.9 (100.0)
7.4 (7.4)
Multiplicity
a
Rmerge
0.083 (0.440)
19.2 (4.9)
0.055 (0.403)
25.6 (5.8)
Mean((I)/SD(I))
Refinement
R-factorb/Rfree (%)
17.1/22.6
50.3/55.8
291
18.6/22.3
48.2/55.0
326
c
2.2. Generation of mAChE and crystallization of its oxime-bound
complexes
B-factord (A)
˚
No. water molecules
r.m.s. from ideal values
RecombinantmAChE protein was cloned, expressed, and purified
˚
Bond lengths (A)
0.011
1.395
0.012
1.395
as previously described [16]. In the crystallization setup, 1
mL
Bond angles (8)
purified mAChE (ꢀ20 mg/mL) was mixed with 1
m
L well solution
Ramachandran plot
Most favoured (%)
Additional allowed (%)
Generously allowed (%)
Disallowed (%)
containing 28% PEG750MME and 0.1 mM HEPES pH 7.0–7.2. The
87.2
11.5
0.9
0.3e
90.3
9.5
0.2
mAChE crystals were first incubated with aliquots of well solution
saturatedwithfenamiphos (5ꢂ 2
m
L)for22 h (HI-6ꢁfep-mAChE)and
25 h (Ortho-7ꢁfep-mAChE). The 2
m
L fenamiphos aliquots were
0.0
added at 1 min intervals. An oxime soaking solution (containing a
few grains of HI-6 or Ortho-7) was added to the fenamiphos-soaked
crystal for 5 and 3 min, for HI-6 and Ortho-7, respectively.
P
P
a
Rmerge = ( j I ꢃ hIi j)/ I, where I is the observed intensity and hIi is the average
intensity obtained from multiple observations of symmetry related reflections.
b
P
P
R-factor =
(jFoj ꢃ jFcj)/ jFoj, where jFoj are observed and jFcj are calculated
structure factors.
c
Rfree uses 2% randomly chosen reflections defined in [23].
2.3. Collection, processing, and refinement of diffraction data
d
B-factor is the mean factor for protein, monomer A/monomer
B of the
asymmetric unit.
e
X-ray diffraction data were collected at the MAXlab synchro-
tron, beamline I911-5 (Lund, Sweden) on a MAR research CCD
The residues located in disallowed regions are Ser203 and Ser541 in monomer A
and Asp494 in monomer B (see also Section 3).

A. Ho¨rnberg et al. / Biochemical Pharmacology 79 (2010) 507–515
509
inhibited protein was aliquoted, frozen, and stored at ꢃ20 8C until
use. Control AChE was handled simultaneously and with an
identical methodology, except that the fenamiphos solution was
exchanged for DMSO. The protein was thawed immediately prior
to use and the assay was initiated by addition of HI-6 or Ortho-7 at
a concentration of 0.1 or 1 mM. The reaction was terminated by
of dichloromethane (5ꢂ 80 mL). The dichloromethane solution
containing the optically pure enantiomer was dehydrated by
anhydrous MgSO4 and the filtered solution was concentrated by
evaporation to a volume of about 100
trated enantiomer was determined with an internal standard by
NMR to 220 and 129 g, for enantiomer-1 (E1) and enantiomer-2
mL. The amount concen-
m
4000-fold and 20,000-fold dilutions for the reactions with 100
m
M
(E2), respectively (E1 being the first peak eluted from the chiral
column). Exact concentration of dissolved fenamiphos enantiomer
in the solution was determined by HPLC using a standard curve
with a known concentration of fenamiphos racemate.
and 1 mM reactivator concentration, respectively. The activity was
measured by using the Ellman assay [26] and the efficiency of the
reactivation was calculated as follows:
Inhibition potency of racemic and E1 on mAChE and hAChE was
determined by activity measurements of residual activity.
Assuming a pseudo-first-order reaction kinetics, the inhibition
constant (ki) was calculated as follows:
vir
ð
v0
v0
=
vrÞ ꢃ vi
E ¼ 100
ꢃ
vi
where E represents the percentage of restored enzyme activity, v0
the control enzyme activity without inhibitor, vir the enzyme
activity after reactivation, vi the residual enzyme activity of the
inhibited enzyme, and vr is the enzyme activity of the control
sample in the presence of a reactivator. The experiments were
performed on a PerkinElmer Lambda18 spectrophotometer.
1
v0
ki ¼
ln
½OPꢄt vt
where [OP] is the initial concentration of fenamiphos, v0 the
reaction rate at time zero and vt is the reaction rate at time t [12].
The non-linear regression curve fit was calculated using the one-
phase exponential decay equation as implemented in GraphPad
Prism version 3.00 for Windows (GraphPad Software, Inc.).
2.5. Separation of enantiomers of fenamiphos and determination of
inhibitory potency
The enantiomers of fenamiphos were separated by using a
chiral column (ChromTech Ltd., Cheshire, U.K.). This column, with
an i.d. of 150 mm ꢂ 4.0 mm, has chiral selector alpha1-acid
3. Results
3.1. General description of the HI-6ꢁfep-mAChE and Ortho-7ꢁfep-
glycoprotein (AGP) coated on spherical 5
m
m silica particles. The
mAChE crystal structures
separation was performed using the High-Performance Liquid
Chromatography (HPLC) system Waters Alliance 2695 Separations
Module coupled to a Waters 2996 Photodiode Array Detector
(Waters Corporation, Milford, USA).
The overall structures of both HI-6ꢁfep-mAChE and Ortho-7ꢁfep-
mAChE are similar to the apo structure of mAChE [19]. The initial
omit electron density maps show a positive difference electron
g
˚
Racemic fenamiphos (1 M in DMSO) was diluted to 1 mM in 5%
density, located ꢀ1.6 A from Ser203O visible at a contour level of
acetonitrile and injected into the HPLC using a loop of 20
m
L.
12s (Fig. 1A and B). This electron density accounts for the
Isocratic elution using a mobile phase composed of acetonitrile and
formic acid (0.01%) resulted in near baseline separation of the two
enantiomers. To obtain sufficient material for the subsequent
experiments, several hundred injections of the racemate were
performed. Pooled enantiomer (ꢀ220 mL of each enantiomer) was
neutralized by NaOH (4 M) and 33 mL saturated NaCl was added
prior to extraction of the separated enantiomer in an organic phase
phosphorous atom of the fenamiphos adduct. The substituents of
the fenamiphos are clearly visible in the electron density.
Although, features of the electron density maps suggest that the
S enantiomer of the fenamiphos phosphorus atom is dominant for
˚
HI-6ꢁfep-mAChE at a resolution of 2.7 A and Ortho-7ꢁfep-mAChE at
˚
a resolution 2.4 A, it is not possible to conclusively assign the
stereochemistry. Therefore, both enantiomers were modelled into
Fig. 1. Electron density maps of HI-6ꢁfep-mAChE (A) and Ortho-7ꢁfep-mAChE (B) calculated after simulated annealing using models where the conjugated Ser203 and oximes
were omitted. The jFoj ꢃ jFcj electron density maps are contoured at 3 (blue) and 12 (red).
s
s

510
A. Ho¨rnberg et al. / Biochemical Pharmacology 79 (2010) 507–515
the final structure with an occupancy of 0.5 each. This is consistent
with inhibition kinetics studies of fenamiphos enantiomers (see
below) that show a 3–7-fold difference in potency between the two
enantiomers, suggesting that both are present within the crystal
structure. Close to 1.0 in occupancy of fenamiphos (i.e., 100% of the
molecules in the crystals are conjugated with fenamiphos) is
supported by the fact that no negative electron density (contour
towards Trp86 (ethoxyl and isopropylamino substituents for the S
and R enantiomer, respectively). The well resolved acyl loop adopts
a conformation similar to that in the apo mAChE, which results in
short distances to the conjugate. For example, the C2 atom of
isopropylamino moiety for the S isomer is found at distances of 4.0,
e
z
3.2, and 3.8 A to the acyl pocket residues Phe295C 1, Phe297C , and
˚
e
Phe338C 2, respectively.
level of 3
s
) is visible for any atom of the fenamiphos conjugate in a
Interestingly, the conjugate is tilted towards Trp86 compared to
an HI-6 complex with mAChE inhibited by isopropyl methylpho-
sphonofluoridate (sarin) (2WHP [9]; Fig. 2B), presumably to reduce
the clash between the fenamiphos conjugate and the acyl pocket.
The tilt of the phosphorous conjugate is primarily due to a shift of
the main chain of the Ser203-containing loop. Moreover, the
torsion angles of the isopropylamino and ethoxyl substituents
jFoj ꢃ jFcj electron density map. The electron density map is
generally of better quality for monomer A than for monomer B of
the asymmetric unit and thereby descriptions of the structures are
based on the coordinates from monomer A.
For both complexes, similar to other mAChE structures [7–
9,16,19,27–29], the loop region containing residues 258–265 is
disordered and thereby excluded from the models. Residues 490–
498 are poorly defined, resulting in unusual backbone torsion
angles. For Ortho-7ꢁfep-mAChE, residues 493–497 were omitted at
the end of the refinement. The isopropyl- and ethyl substituents of
fenamiphos in monomer B of Orthoꢁfep-mAChE are also excluded
in the final structure.
result in
a markedly different conformation of fenamiphos
conjugate relative to that of Ortho-7ꢁfep-mAChE (see below).
The tilt of the phosphorus atom in HI-6ꢁfep-mAChE positions one of
e
the substituents close to Glu202O 1 and the other substituent close
e
to His447N 2. For the S enantiomer, the distance between the
e
ethoxy oxygen and Glu202O 1 atoms is 3.0 A, whereas the
˚
˚
Typically, the C-terminal end is not visible beyond residue 543
in crystal structures of mAChE [7–9,16,19,27–29]. For HI-6ꢁfep-
mAChE, a continuous electron density is present at the C-terminal
end and defines Ala544, Thr545, and Glu546 in the final model.
isopropylamino nitrogen atom is at a distance of 3.4 A from the
e
His447N 2 atom. Conversely, the ethoxy oxygen of the R isomer is
e
N 2
˚
positioned at 3.2 A distance from His447
and the isopropyla-
˚
e
mino nitrogen atom is close to Glu202O 1 (3.2 A).
The binding of HI-6 to the peripheral site of mAChE is facilitated
by a side chain rotation of Trp286 (ꢀ908 rotation about x1 and a
3.2. Crystal structure of HI-6ꢁfep-mAChE
ꢀ1008 rotation about
x2). This structural change allows the indole
˚
The structure of HI-6ꢁfep-mAChE was refined at 2.7 A resolution
ring of Trp286 and the phenol ring of Tyr124 to sandwich, via
cation–pi interactions, the carboxylamino-pyridinium ring of HI-6.
The carboxy oxygen atom of HI-6 is hydrogen bonded to the main-
with an Rfactor of 17 (Rfree = 23) (Table 1). The final structure
contains a well-defined fenamiphos conjugate as well as most
atoms in the HI-6 molecule (Fig. 2A). The two fenamiphos
enantiomers have hydrogen bonds between their –P55O oxygen
atom and the oxyanion hole made of the amide protons of Gly121,
Gly122, and Ala204. The corresponding hydrogen bond distances
˚
chain nitrogen of Ser298 (2.7 A). Additional interactions between
HI-6 and the protein are: an electrostatic interaction between
Asp74 and two positive pyridinium rings of HI-6 and cation–pi
interactions between the oxime-pyridinium ring and the phenol
˚
˚
˚
are 2.5 and 2.6 A (Gly121), 2.5 and 2.5 A (Gly122), and 3.2 and 3.2 A
(Ala204) for the S and R enantiomers, respectively. One alkyl
fenamiphos substituent is accommodated within the acyl pocket.
For the S enantiomer, this corresponds to the isopropylamino
moiety, whereas the R enantiomer has the ethoxyl moiety in the
acyl pocket. The other alkyl fenamiphos substituent is directed
ring of Tyr341. The interaction between Asp74 and HI-6 positions
˚
the Od1 atom of Asp74 at a 3.1 A distance from the ether oxygen
atom of HI-6. These two atoms are probably engaged in a water-
bridged hydrogen-bond network since Asp74 is deprotonated at
pH 7. There is a strong electron density feature at a hydrogen
bonding distance from the main-chain nitrogen of Phe295
Fig. 2. (A) The fenamiphos conjugate and the HI-6 molecule of HI-6ꢁfep-mAChE covered by the electron density at 1
(lime) is tilted towards Trp86 compared to the conjugate of HI-6ꢁsarin-mAChE (brown).
s
in a 2jFoj ꢃ jFcj map. (B) The conjugate of HI-6ꢁfep-mAChE

A. Ho¨rnberg et al. / Biochemical Pharmacology 79 (2010) 507–515
511
(NPhe295). Modelling the oxime moiety in this density results in a
residual positive electron density and poor connectivity between
the oxime oxygen atom and the pyridinium ring, similar to a
previous report [9]. Attempts to model the oxime moiety (O1, N1
and C1 atoms) were unsuccessful and the oxime moiety was
omitted in the final model. Instead, a water molecule accounts for
the electron density close to NPhe295. Taken together, these
observations suggest high mobility of the HI-6 oxime-pyridinium
ring, consistent with our reported crystallographic and micro-
second molecular dynamics simulation studies of HI-6ꢁsarinnon-
aged-mAChE [9].
the final electron density map clearly delineates both fenamiphos
and Ortho-7 (Fig. 3A) and defines structural changes of the acyl
pocket loop (residues 288–299), and the side chains of Trp286,
Tyr337, Phe338, and His447.
The overall conformation of the fenamiphos conjugate in Ortho-
7 is slightly different from that of HI-6ꢁfep-mAChE. For example,
the conjugate is not tilted towards Trp86 in Ortho-7ꢁfep-mAChE
(see below). The ethoxy substituent of the R enantiomer in Ortho-
˚
7ꢁfep-mAChE interacts mainly with Trp236 (3.6 A, C5 atom of
e
fenamiphos to Trp236C 2
)
and Phe338 (3.5 A, C5 atom of
fenamiphos to Phe338C ), whereas the isopropylamino substituent
˚
z
˚
is directed towards Trp86 (ꢀ5 A distance, C3 atom of fenamiphos
e
3.3. Crystal structure of Ortho-7ꢁfep-mAChE
to Trp86C 3). The S enantiomer positions the isopropylamino
substituent close to Trp236 and Phe338 and the ethoxy substituent
close to Trp86 at similar distances as those in the R enantiomer
(Fig. 3B). The interactions of fenamiphos with the oxyanion hole
˚
The 2.4-A resolution crystal structure of Ortho-7ꢁfep-mAChE
was refined to a final Rfactor of 19 (Rfree = 22) (Table 1). Gratifyingly,
Fig. 3. (A) The fenamiphos conjugate and the Ortho-7 molecule of Ortho-7ꢁfep-mAChE covered by the electron density at 1
s
in a 2jFoj ꢃ jFcj map. (B) The fenamiphos
substituents (isopropylamino- and ethoxy group) in Ortho-7ꢁfep-mAChE (cyan) compared to the corresponding atoms in HI-6ꢁfep-mAChE (lime). (C) Disruption of the acyl
pocket in Ortho-7ꢁfep-mAChE (cyan), where both the acyl loop (containing Phe295 and Phe297) and Phe338 has moved compared to the apo protein (grey). Residues Tyr337
and His447 have been omitted for clarity. (D) The position of the attacking pyridinium ring of Ortho-7ꢁfep-mAChE compared to the corresponding pyridinium ring in HI-6ꢁfep-
mAChE. The figure is in similar view as in (C). The black dashed lines represent incompatibly short C-to-C distances, while the red dashed line represents a hydrogen bond.

512
A. Ho¨rnberg et al. / Biochemical Pharmacology 79 (2010) 507–515
Table 2
are similar in the Ortho-7- and HI-6-containing fep-mAChE
complexes; the –P55O oxygen atom of the R enantiomer is found
Time dependent reactivation of fenamiphos-inhibited mAChE in conditions
simulating the crystallization setup.
˚
2.9, 2.6, and 2.9 A away from the main-chain nitrogen atoms of
HI-6 (% reactivated)
Ortho-7 (% reactivated)
Gly121, Gly122, and Ala204, respectively (2.7, 2.5, and 3.3 for the S
enantiomer, respectively).
100
100
m
m
M, 1 h
M, 3 h
4.5 ꢅ 1.2
11.1 ꢅ 0.7
10.3 ꢅ 2.6
27.3 ꢅ 3.4
5.8 ꢅ 1.1
11.6 ꢅ 1.7
5.7 ꢅ 3.0
Interestingly, a marked distortion of the acyl pocket loop
1 mM, 1 h
1 mM, 3 h
(residues 288–299) is observed in Ortho-7ꢁfep-mAChE (Fig. 3C).
14.7 ꢅ 0.4
a
˚
The main-chain movement of the loop is up to 6.0 A (C of residue
Pro290) relative to that of the apo mAChE. Moreover, two acyl-
pocket residues, Phe295 and Phe297, undergo side-chain con-
concentrations, E1 is estimated to be 3–7-fold more potent than E2
(Supplementary material online Fig. S4).
formational changes (ꢀ788 and ꢀ1378 for rotation of
x1 for Phe295
and Phe297, respectively) to accommodate a substituent of the
fenamiphos conjugate. Consequently, the conjugate in Ortho-
7ꢁfep-mAChE is not tilted towards Trp86, and slightly different
torsion angles of the fenamiphos conjugate result in a more
extended conformation of the fenamiphos substituents than that
in HI-6ꢁfep-mAChE.
The short distances between the fenamiphos conjugate and the
side chains of Phe295 and Phe297, which presumably cause the tilt
of the conjugate observed in HI-6ꢁfep-mAChE, led us to investigate
the inhibition of Phe295Ala and Phe297Ala mutants of rhAChE by
fenamiphos. The human enzyme was used to compare with a
previous report [30] and due to similar reactivation kinetics of
human and rat AChEs [31], high similarity among human, rat and
mouse AChEs (89% sequence identity between the human and rat
enzymes; 98% sequence identity between the rat and mouse
enzymes), and the same active-site residues in the three enzymes.
Compared to wild-type, the inhibitions of Phe295Ala and Phe297Ala
by E1 increased by ꢀ3-fold (ki = 45.2 ꢅ 1.1 ꢂ 103 Mꢃ1 minꢃ1) and
decreased by ꢀ5-fold (ki = 3.5 ꢅ 0.2 ꢂ 103 Mꢃ1 minꢃ1), respectively.
Moreover, a comparison of the enantiomers at nearly equimolar
concentrations shows that E1 is more potent than E2 for both mutants
(Fig. S4). Thus, the stereochemical preference is independent of the
residue substitutions in the acyl pocket.
To compare the efficacies of HI-6 and Ortho-7 in reactivating
the fenamiphos-inhibited mAChE at the same pH as used during
the crystal soaking, mAChE was inhibited by fenamiphos and the
amount of reactivation was then measured after 1- and 3-h
incubations with HI-6 or Ortho-7. At an oxime concentration of
1 mM, HI-6 has a higher efficacy than Ortho-7 (27.3 ꢅ 3.4% vs.
14.7 ꢅ 0.4% restored activity after 3 h) (Table 2). On the other hand, at
a lower oxime concentration (0.1 mM), their efficacy on fep-mAChE is
similar (11.6 ꢅ 1.7% and 11.1 ꢅ 0.7% for Ortho-7 and HI-6, respec-
tively).
Similar to our previous crystal structures of Ortho-7 [7,8], one of
the oxime-pyridinium rings is sandwiched between Trp286 and
e
Tyr72 with the oxime oxygen hydrogen bonding to Glu285O 1
˚
(2.5 A). To form the sandwich, the side chain of Trp286 rotates
ꢀ908 about
x
1 and ꢀ1508 about
x
2. The other oxime pyridinium of
˚
Ortho-7 is close to the fenamiphos adduct (3.5 A to the
˚
isopropylamino substituent of the R enantiomer and 3.6 A to the
S enantiomer) and positioned nearly parallel to Tyr337 (Fig. 3D).
The structural change of the acyl loop results in hydrophobic
interactions between the heptylene linker of Ortho-7 and Leu294.
Interestingly, the oxime oxygen atom at the catalytic site forms a
e
hydrogen bond to His447N 2 (2.8 A), positioning the nucleophilic
˚
˚
oxime oxygen atom 5.7–5.9 A away from the phosphorus atom.
This hydrogen bond is possible due to a side chain rotation of
His447 (an imidazole ring flip: ꢀ1408 about x1 and ꢀ1808 about
x
2) and Tyr337 (ꢀ308 about x1 and ꢀ608 about
x2) as well as
z
˚
retraction of Phe338 (3.0 A for the C atom relative to that of the
apo mAChE structure) (Fig. 3D). A residual and positive electron
density feature was observed near His447, which probably
represents a low-occupancy conformation of His447 found in
the apo mAChE structure. However, the apo-like conformation is
estimated to have an occupancy below 10% and is therefore not
included in the final structure. In addition, two residual and
positive electron density features close to the fenamiphos
conjugate could not be unambiguously identified (Supplementary
material online Fig. S2). These electron density features are visible
4. Discussion
4.1. Structural requirement for the reactivation of fep-mAChE
at 4.5
˚
s
in the final jFoj ꢃ jFcj electron density map and positioned
We recently proposed a mechanism for the reactivation of
sarinnon-aged-mAChE by HI-6 in which a reactive tetrad (Glu334–
His447–Water–HI-6) is formed to deprotonate the hydroxyl group
of HI-6 in the HI-6ꢁsarinnon-aged-mAChE complex at the Michaelis–
Menten state [9]. In this mechanism, His447 plays a key role in
reactivation since the imidazole ring not only activates the oxime
oxygen atom by accepting the hydroxyl proton, but also donates a
3.0 A away from the C2 and the C5 atoms of the fenamiphos
conjugate. These features may represent low-occupancy water
molecules and suggest that a fraction of the conjugate is aged (note
that no residual negative electron density is observed for the
phosphorus atom). Nevertheless, the non-aged fenamiphos adduct
to Ser203 is the dominating species with >90% occupancy.
g
proton to Ser203O at the transition state. In this context, the
3.4. Biochemical characterization
structure of Ortho-7ꢁfep-mAChE is of particular interest since a
significant structural change of His447 is clearly defined by the
electron density map. Indeed, the Ortho-7ꢁfep-mAChE crystal
structure shows an analogous hydrogen bond network among
Glu334, His447 and the oxime oxygen of Ortho-7 (Supplementary
material Fig. S5A). Interestingly, flipping the His447 imidazole ring
relative to the apo mAChE is necessary to establish a reactive triad of
Glu334–His447–Ortho-7. The imidazole ring flip is sometimes
difficult to deduce because of similar electron densities of nitrogen
and carbon [32]. Moreover, in Ortho-7ꢁfep-mAChE, both the flipped
and unflipped imidazole ring conformations result in short C-to-C
distances. However, analysis of the hydrogen bonding distances and
angles for the two conformations indicates that the unflipped
imidazole ring conformation is less favourable (Supplementary
material Fig. S5B). Itisconceivablethat the Glu334–His447–Ortho-7
To assign the stereochemistry of the phosphorous conjugate in
HI-6ꢁfep-mAChE and Ortho-7ꢁfep-mAChE, a biochemical character-
ization study using an optically pure fenamiphos enantiomer was
performed. The enantiomerwasobtained byisocraticelutionusinga
chiral column, resulting in an optical purity of >98% and ꢀ96%, for
enantiomer-1 (E1) and enantiomer-2 (E2), respectively (Supple-
mentary material online Fig. S3). The bimolecular rate constant for
the inhibition (ki) of rhAChE was higher for E1 than for racemic
fenamiphos (16.8 ꢅ 0.8 ꢂ 103 vs. 5.5 ꢅ 0.7 ꢂ 103 Mꢃ1 minꢃ1) and a
comparable result was obtained for mAChE (3.0 ꢅ 0.6 ꢂ 103 vs.
1.0 ꢅ 0.1 ꢂ 103 Mꢃ1 minꢃ1). Since E2 contained ꢀ4% of E1, the
bimolecular rate constant of E2 was not determined. However, by
testing the inhibition of hAChE by E1 and E2 at nearly equimolar

A. Ho¨rnberg et al. / Biochemical Pharmacology 79 (2010) 507–515
513
g
triad increases the nucleophilicity of the oxime just as the Glu334–
His447–Water–HI-6 tetrad [9]. In excellent agreement with this
finding, the imidazole ring flip that permits the formation of the
Glu334–His447–Ortho-7 triad is also present in our previously
reported structure of tabun-inhibited mAChE in complex with
Ortho-7 [8]. An imidazole ring flip has also been reported to produce
a catalytic triad in glycinamide ribonucleotide transformylase
(GART), where the formation of the Asp144–His108–GAR (glycina-
mide ribonucleotide) triad deprotonates the GAR molecule [33].
In contrast to the conformational change of His447 in Ortho-
7ꢁfep-mAChE as described above, His447 adopts the conformation
found in the apo mAChE structure and no imidazole ring flip is
observed in HI-6ꢁfep-mAChE. Despite extensive model-building
trials, the oxime moiety of HI-6 could not be modelled into the HI-
6ꢁfep-mAChE structure. One plausible explanation for the missing
electron density for the oxime moiety is that the oxime-pyridinium
ring of HI-6 is mobile, consistent with our recently reported
crystallography and microsecond multiple molecular dynamics
simulations (MMDSs) of HI-6ꢁsarinnon-aged-mAChE [9]. Since the
oxime moiety is undetermined, microsecond MMDSs are neces-
sary to confirm the mobility and investigate deprotonation of HI-6
by a putative tetrad or triad in the fenamiphos-containing active
site.
to donate a proton to Ser203O at the transition state [9]. Taken
together, it is conceivable that the higher reactivation rate of HI-6
than Ortho-7 in reactivating fenamiphos-AChE is due to a mobile
oxime-pyridinium and the apo-like conformation of His447, two
factors that facilitate the conversion from the Michaelis–Menten
g
state to the transition state as well as the protonation of Ser203O
at the transition state.
4.4. Modulation of the acyl pocket
Superposition of the HI-6ꢁfep-mAChE and Ortho-7ꢁfep-mAChE
structures shows that the phosphorous conjugate in the HI-6-
˚
containing complex is tilted towards Trp86 (Fig. 3B). The 1.2-A tilt
of the phosphorus atom, in combination with strained torsion
angles of the fenamiphos conjugate in HI-6ꢁfep-mAChE, allows an
apo-like conformation of the acyl loop. In contrast, a striking
feature of Ortho-7ꢁfep-mAChE is the extensive structural change of
the acyl loop (Fig. 3C). The distortion of the acyl loop may be
explained by steric effects induced by the Ortho-7 pyridinium ring
positioned in the catalytic site nearly parallel to Tyr337. In this
position, the pyridinium ring of Ortho-7 is incompatible, due to
short C-to-C distances, with the tilted fenamiphos conjugate
conformation observed in HI-6ꢁfep-mAChE. Thus, one of the
fenamiphos substituents is shifted towards the acyl loop,
presumably inducing the structural change of the loop. The new
position of acyl loop provides space to accommodate the
fenamiphos substituent and, consequently, no tilt of the phos-
phorus conjugate is present in Ortho-7ꢁfep-mAChE. A tilted
conjugated serine residue has previously been described for aged
phosphoroamidates in butyrylcholinesterase. The tilt is due to a
salt bridge between the negatively charged dealkylated aging
product and the histidine corresponding to His447 in AChE (His438
in butyrylcholinesterase) [34].
4.2. Relative reactivation potencies of HI-6 and Ortho-7
Our biochemical characterization shows that HI-6 is a more
efficient reactivator than Ortho-7 at 1 mM, whereas the efficacies
of the two oximes are similar at 100
mM (Table 2). At 1 mM, the
concentration is likely above the dissociation constant (KD) for
both HI-6 and Ortho-7 (KD for HI-6 of fenamiphos-bound hAChE
is 889
mM [12]). For HI-6, the reactivation is reduced by 2-fold at
lower oxime concentration, while for Ortho-7 the effect of the
decreased concentration on efficacy is insignificant, suggesting
It has been suggested that an ethoxy group may fit the acyl
pocket without perturbing the conformations of Phe295 and
Phe297 found in the apo mAChE structure [30]. This proposal was
based on two observations from conjugates with methyl and
ethoxy substituents; (i) inhibition kinetics where an increased size
of the acyl pocket (Phe295Ala) had a small effect (ꢀ1.5-fold) on the
reactivity for the R stereoisomer of non-charged O-ethyl S-3-
(isopropyl-4-methylpentyl) methylphosphonothioate (P[R] ncVX)
which has the ethoxy moiety towards the acyl pocket [30] and (ii)
the lack of stereoselectivity for MEPQ (7-(methylethoxypho-
sphinyloxy)1-methyl-quinolinium iodide), which has methyl
and ethoxy substituents [35]. In the present study, inhibitions of
wild-type hAChE, Phe295Ala and Phe297Ala by separated fena-
miphos enantiomers show that none of the two mutants reverses
the stereochemical preference. Moreover, for the more reactive
enantiomer, the Phe295Ala mutation increases the inhibition
constant by ꢀ3-fold whereas the Phe297Ala mutation decreases
the inhibition by ꢀ5-fold. Thus, in this respect, our findings are
consistent with the reported kinetics of methylphosphonates P[R]
ncVX and MEPQ (i.e., one of the fenamiphos substituents can be
accommodated in the acyl pocket without inducing a structural
change).
On the other hand, inhibition kinetics of the phosphates
ethylparaoxon, diethylfluorophosphate (DEFP) and echothiophate
show a 20–40-fold enhanced inhibition potency in the Phe295Ala
mutant [30,36]. One obvious difference between the methyl
phosphonates discussed above and these phosphates is that the
methyl phosphonates have methyl and ethoxy substituents
whereas the phosphates have two ethoxy substituents. As
previously reported, this indicates that substituents distal to the
acyl loop also can affect the accommodation of the phosphorus
conjugate [30,35], a notion that ideally requires verification by
structural studies of separated enantiomers.
that KD for Ortho-7 is below 100
mM. Under the assumptions that
the stereochemical preferences for fenamiphos by HI-6 and
Ortho-7 are similar and that the stabilities of the phosphorylated
oximes of HI-6 and Ortho-7 are also comparable, these
reactivation results along with the structural consideration
discussed below suggest that Ortho-7 has a higher affinity than
HI-6 for fep-mAChE, whereas HI-6 has a higher reactivation rate
than Ortho-7.
4.3. Structural basis for the higher reactivation rate of
HI-6 than Ortho-7
The tilted conjugate in HI-6ꢁfep-mAChE pushes one of the
fenamiphos substituents towards the HI-6 binding site, resulting
in a translation of the oxime-pyridinium ring of HI-6 towards
Trp286. Superposition of the available HI-6-containing mAChE
structures shows that the position of fenamiphos in HI-6ꢁfep-
mAChE creates incompatibly short C-to-C distances between
fenamiphos and the site favoured by HI-6 in non-conjugated
mAChE and HI-6ꢁsarin-mAChE [7,9]. In contrast to the disordered
electron density of the oxime-pyridinium of HI-6, the well
resolved electron density map of Ortho-7ꢁfep-mAChE shows
extensive and favourable interactions between Ortho-7 and fep-
mAChE, including the hydrogen bond to His447 described above.
These observations support the notion that Ortho-7 has a higher
affinity than HI-6 for fep-mAChE.
Because of the higher affinity of Ortho-7 for fep-mAChE, one can
speculate that the hydrogen bond between His447 and the oxime
oxygen of Ortho-7 entraps the nucleophile at the Michaelis–
Menten state and hampers the conversion to the transition state.
e
g
Moreover, the His447N 2 atom is 5.4–5.5 A away from Ser203
,
O
˚
which suggests that a conformation change of His447 is necessary

514
A. Ho¨rnberg et al. / Biochemical Pharmacology 79 (2010) 507–515
observed in Ortho-7ꢁfep-mAChE, have been described in the bis-
tacrine-bound Torpedo californica AChE complex (pdb code:
2cmf; [39]).
4.7. Effect of conjugation on the acyl loop conformation
Analogously, non-aged and aged forms of diisopropyl fluor-
ophosphate (DFP) and the aged form of fenamiphos induce
structural changes of the acyl loop upon conjugating to Ser203
[29,37]. The acyl loop of the previously reported non-aged form of
fep-mAChE adopts a conformation found in the apo mAChE
structure; however, due to the low occupancy of the phosphorous
conjugate in fep-mAChE, the reported stereochemistry as well as
the position of the acyl loop [29] need further confirmation.
Despite the apo-like conformation acyl loop in HI-6ꢁfep-mAChE, it
is likely that aging of fenamiphos occur via the same mechanism as
described for tabun, i.e., His447 is involved as a catalyst for aging
and subsequently stabilizes the negatively charged dealkylated
product [40].
5. Conclusion
The data presented herein propose that an imidazole ring flip of
His447 is required to establish a favourable hydrogen bond
network among a triad (Glu334–His447–Oxime) that can increase
the nucleophilicity of the Ortho-7 oxime due to deprotonation of
the oxime group by the triad but can also reduce the reactivation
potency due to entrapment of Ortho-7 at the Michaelis–Menten
state by the triad. For our ongoing structure-based design of novel
nerve-agent antidotes, these findings highlight an array of complex
interrelationship among the catalytic-site architecture, the spatial
arrangements of the phosphorus conjugate, and the attacking
oxime. We suggest that novel reactivators should be designed in
such a way that His447 can participate in a transient Glu334–
His447–(Wat)–Oxime arrangement.
Fig. 4. A superposition of Ortho-7ꢁfep-mAChE (cyan) onto Ortho-7ꢁmAChE (pink),
and Ortho-7ꢁtabun-mAChE (violet) showing different binding of the central linker in
the fenamiphos inhibited complex. The black dashed lines indicate close CꢁꢁꢁC
contacts.
4.5. Difference between Ortho-7ꢁtabun-mAChE and
Ortho-7ꢁfep-mAChE
In the previously reported Ortho-7ꢁtabun-mAChE structure
(2JF0; [8]), the triad (Glu334–His447–Ortho-7) is also observed.
The His447 residue is in a similar position and has a flipped
imidazole ring just as described in Ortho-7ꢁfep-mAChE. However,
in the tabun-conjugated structure, the distance between His447
and the oxime oxygen of Ortho-7 is significantly longer than the
distance in the fenamiphos-conjugated structure (Supplementary
material online Fig. S6). The structural change of the acyl loop in
Acknowledgements
˚
Ortho-7ꢁfep-mAChE allow a 2–3 A adaptive structural change of
We thank Drs. Tomas Ursby and Krister Larsson for assistance at
the beamline I911 at the MAXlab synchrotron, Lund, Sweden.
Phe338, which in turn permits the stronger interaction between
the oxime oxygen and His447. Another difference between the
fenamiphos- and tabun-containing structures is that in Ortho-
7ꢁtabun-mAChE, the electron density map of the oxime-pyridinium
is relatively weak and shows notably short interaction distances to
Tyr337 and Phe338 implying that the oxime-pyridinium is mobile
in the tabun-containing complex.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bcp.2009.08.027.
References
4.6. Flexibility of the acyl loop
[1] Harel M, Quinn DM, Haridasan KN, Silman I, Sussman JL. The X-ray structure of
a transition state analogue complex reveals the molecular origins of the
catalytic power and substrate specificity of acetylcholinesterase. J Am Chem
Soc 1996;118:2340–6.
[2] Radic Z, Reiner E, Taylor P. Role of the peripheral anionic site on acetylcho-
linesterase: inhibition by substrates and coumarin derivatives. Mol Pharmacol
1991;39:98–104.
[3] Shafferman A, Velan B, Ordentlich A, Kronman C, Grosfeld H, Leitner M, et al.
Substrate inhibition of acetylcholinesterase: residues affecting signal trans-
duction from the surface to the catalytic center. EMBO J 1992;11:3561–8.
[4] Barak D, Kronman C, Ordentlich A, Ariel N, Bromberg A, Marcus D, et al.
Acetylcholinesterase peripheral anionic site degeneracy conferred by amino
acid arrays sharing a common core. J Biol Chem 1994;269:6296–305.
[5] Eichler J, Anselment A, Sussman JL, Massoulie J, Silman I. Differential effects of
‘‘peripheral’’ site ligands on Torpedo and chicken acetylcholinesterase. Mol
Pharmacol 1994;45:335–40.
A superposition of three available complexes of Ortho-7-
bound mAChE (Ortho-7 in complex with mAChE, tabun-mAChE
and fep-mAChE) shows a similar interaction between AChE and
the pyridinium ring at the catalytic site, whereas the pyridinium
ring at the peripheral site and the central linker is markedly
different in the fenamiphos-inhibited complex (Fig. 4). The
different binding of the Ortho-7 linker in the fenamiphos-
inhibited complex is due to the structural change of the acyl loop,
which pushes the side chain of Ile294 into the active-site gorge
where it forms hydrophobic interactions with the linker of
Ortho-7. An adaptable central linker is consistent with the
previous suggestion that Ortho-7 can adjust to different
conformations of the active-site gorge [7]. Moreover, the acyl
loop has previously been described as flexible and various
[6] Rosenberry TL, Mallender WD, Thomas PJ, Szegletes T. A steric blockade model
for inhibition of acetylcholinesterase by peripheral site ligands and substrate.
Chem Biol Interact 1999;119–120:85–97.
[7] Ekstro¨m F, Pang YP, Boman M, Artursson E, Bo¨rjegren S. Crystal structures of
acetylcholinesterase in complex with HI-6, Ortho-7 and obidoxime: structural
basis for differences in the ability to reactivate tabun conjugates. Biochem
Pharmacol 2006;72:597–607.
reversible and irreversible ligands are known to induce
structural change in this region [29,37–39]. For example, side-
chain rotations of Phe295 and Phe297, similar to the movement
a

A. Ho¨rnberg et al. / Biochemical Pharmacology 79 (2010) 507–515
515
[8] Ekstrom FJ, Astot C, Pang YP. Novel nerve-agent antidote design based on
crystallographic and mass spectrometric analyses of tabun-conjugated acetyl-
cholinesterase incomplex withantidotes. Clin PharmacolTher 2007;82:282–93.
[9] Ekstrom F, Ho¨rnberg A, Artursson E, Hammarstro¨m LG, Schneider G, Pang YP.
Structure of HI-6ꢁsarin-acetylcholinesterase determined by X-ray crystallo-
graphy and molecular dynamics simulation: reactivator mechanism and
design. PLoS ONE 2009;4:e5957.
[26] Ellman GL, Courtney KD, Andres Jr V, Feather-Stone RM. A new and rapid
colorimetric determination of acetylcholinesterase activity. Biochem Pharma-
col 1961;7:88–95.
[27] Bourne Y, Kolb HC, Radic Z, Sharpless KB, Taylor P, Marchot P. Freeze-frame
inhibitor captures acetylcholinesterase in a unique conformation. Proc Natl
Acad Sci USA 2004;101:1449–54.
[28] Bourne Y, Radic Z, Sulzenbacher G, Kim E, Taylor P, Marchot P. Substrate and
product trafficking through the active center gorge of acetylcholinesterase ana-
lyzed by crystallography and equilibrium binding. J Biol Chem 2006;281:29256–
67.
[29] Hornberg A, Tunemalm AK, Ekstrom F. Crystal structures of acetylcholines-
terase in complex with organophosphorus compounds suggest that the acyl
pocket modulates the aging reaction by precluding the formation of the
trigonal bipyramidal transition state. Biochemistry 2007;46:4815–25.
[30] Ordentlich A, Barak D, Sod-Moriah G, Kaplan D, Mizrahi D, Segall Y, et al.
Stereoselectivity toward VX is determined by interactions with residues of the
acyl pocket as well as of the peripheral anionic site of AChE. Biochemistry
2004;43:11255–6.
[31] Worek F, Reiter G, Eyer P, Szinicz L. Reactivation kinetics of acetylcholinester-
ase from different species inhibited by highly toxic organophosphates. Arch
Toxicol 2002;76:523–9.
[32] NielsenJE, AndersenKV, HonigB, HooftRWW, Klebe G, VriendG, etal. Improving
macromolecular electrostatics calculations. Protein Eng 1999;12:657–62.
[33] Morikis D, Elcock AH, Jennings PA, McCammon JA. Proton transfer dynamics of
GART: the pH-dependent catalytic mechanism examined by electrostatic
calculations. Protein Sci 2001;10:2379–92.
[34] Carletti E, Aurbek N, Gillon E, Loiodice M, Nicolet Y, Fontecilla-Camps JC, et al.
Structure–activity analysis of aging and reactivation of human butyrylcholi-
nesterase inhibited by analogues of tabun. Biochem J 2009;421:97–106.
[35] Levy D, Ashani Y. Synthesis and in vitro properties of a powerful quaternary
methylphosphonate inhibitor of acetylcholinesterase: a new marker in blood–
brain barrier research. Biochem Pharmacol 1986;35:1079–85.
[36] OrdentlichA,BarakD, KronmanC,ArielN,SegallY,VelanB,etal. Thearchitecture
of human acetylcholinesterase active center probed by interactions with
selected organophosphate inhibitors. J Biol Chem 1996;271:11953–62.
[37] Millard CB, Kryger G, Ordentlich A, Greenblatt HM, Harel M, Raves ML, et al.
Crystal structures of aged phosphonylated acetylcholinesterase: nerve agent
reaction products at the atomic level. Biochemistry 1999;38:7032–9.
[38] Greenblatt HM, Guillou C, Gue´nard D, Argaman A, Botti S, Badet B, et al. The
complex of a bivalent derivative of galanthamine with Torpedo acetylcholi-
nesterase displays drastic deformation of the active-site gorge: implications
for structure-based drug design. J Am Chem Soc 2004;126:15405–11.
[39] Rydberg EH, Brumshtein B, Greenblatt HM, Wong DM, Shaya D, Williams LD,
et al. Complexes of alkylene-linked tacrine dimers with Torpedo californica
acetylcholinesterase: binding of Bis5-tacrine produces a dramatic rearrange-
ment in the active-site gorge. J Med Chem 2006;49:5491–500.
[10] Benschop HP, Keijer JH. On the mechanism of aging of phosphonylated
cholinesterases. Biochim Biophys Acta 1966;128:586–8.
[11] Michel HO, Hackley Jr BE, Berkowitz L, List G, Hackley EB, Gillilan W, et al.
Ageing and dealkylation of Soman (pinacolylmethylphosphonofluoridate)-
inactivated eel cholinesterase. Arch Biochem Biophys 1967;121:29–34.
[12] Worek F, Thiermann H, Szinicz L, Eyer P. Kinetic analysis of interactions
between human acetylcholinesterase, structurally different organopho-
sphorus compounds and oximes. Biochem Pharmacol 2004;68:2237–48.
[13] Hobbiger F. The inhibition of acetylcholinesterase by organophosphorus
compounds and its reversal. Proc R Soc Med 1961;54:403–5.
[14] Kuca K, Jun D, Musilek K, Bajgar J. Reactivators of tabun-inhibited acetylcho-
linesterase: structure–biological activity relationship. Front Drug Des Discov
2007;3:381–94.
[15] Pang YP, Kollmeyer TM, Hong F, Lee JC, Hammond PI, Haugabouk SP, et al.
Rational design of alkylene-linked bis-pyridiniumaldoximes as improved
acetylcholinesterase reactivators. Chem Biol 2003;10:491–502.
[16] Ekstro¨m F, Akfur C, Tunemalm A-K, Lundberg S. Structural changes of phe-
nylalanine 338 and histidine 447 revealed by the crystal structures of tabun-
inhibited murine acetylcholinesterase. Biochemistry 2006;45:74–81.
[17] Kabsch W. Automatic processing of rotation diffraction data from crystals of
initially unknown symmetry and cell constants. J Appl Cryst 1993;26:795–
800.
[18] Kabsch W. Evaluation of single-crystal X-ray diffraction data from a position-
sensitive detector. J Appl Cryst 1988;21:916–24.
[19] Bourne Y, Taylor P, Radic Z, Marchot P. Structural insights into ligand inter-
actions at the acetylcholinesterase peripheral anionic site. EMBO J 2003;22:1–
12.
[20] Murshudov GN, Vagin AA, Dodson EJ. Refinement of macromolecular struc-
tures by the maximum-likelihood method. Acta Crystallogr D Biol Crystallogr
1997;D53:240–55.
[21] Emsley P, Cowtan K. Coot: model-building tools for molecular graphics. Acta
Crystallogr D Biol Crystallogr 2004;D60:2126–32.
[22] Adams PD, Grosse-Kunstleve RW, Hung LW, Ioerger TR, McCoy AJ, Moriarty
NW, et al. PHENIX: building new software for automated crystallographic
structure determination. Acta Crystallogr D Biol Crystallogr 2002;58:1948–54.
[23] Brunger AT. Free R value: a novel statistical quantity for assessing the accuracy
of crystal structures. Nature 1992;355:472–4.
[24] Laskowski RA, MacArthur MW, Moss DS, Thornton JM. PROCHECK: a program
to check the stereochemical quality of protein structures.
1993;26:283–91.
J Appl Cryst
[40] Carletti E, Li H, Li B, Ekstrom F, Nicolet Y, Loiodice M, et al. Aging of
cholinesterases phosphorylated by tabun proceeds through O-dealkylation.
J Am Chem Soc 2008;130:16011–20.
[25] DeLano WL. The PyMOL Molecular Graphics System. Palo Alto, CA, USA:
DeLano Scientific LLC; 2008.