Homodimeric bis-quaternary heterocyclic ammonium salts as potent acetyl- and butyrylcholinesterase inhibitors: A systematic investigation of the influence of linker and cationic heads over affinity and selectivity

ARTICLE

pubs.acs.org/jmc

Homodimeric Bis-Quaternary Heterocyclic Ammonium Salts as

Potent Acetyl- and Butyrylcholinesterase Inhibitors: A Systematic

Investigation of the Influence of Linker and Cationic Heads

over Affinity and Selectivity

Ana Conejo-García,^{†,^}Leonardo Pisani,^{‡,^}Maria del Carmen Nꢀu~nez,^†Marco Catto,^‡Orazio Nicolotti,^‡

Francesco Leonetti,^‡Joaquín M. Campos,^†Miguel A. Gallo,^†Antonio Espinosa,^†and Angelo Carotti*^,‡

^†Departamento de Química Farmacꢀeutica y Orgꢀanica, Facultad de Farmacia, Universidad de Granada, Campus de Cartuja,

18071 Granada, Spain

^‡Dipartimento Farmacochimico, Universitꢁa degli Studi di Bari “Aldo Moro”, Via E. Orabona, 4, 70125 Bari, Italy

S Supporting Information

b

ABSTRACT: A molecular library of quaternary ammonium salts (QASs), mainly

composed of symmetrical bis-quaternary heterocyclic bromides exhibiting choline

kinase (ChoK) inhibitory activity, were evaluated for their ability to inhibit acetyl-

and butyrylcholinesterase (AChE and BChE, respectively). The molecular frame-

work of QASs consisted of two positively charged heteroaromatic (pyridinium or

quinolinium) or sterically hindered aliphatic (quinuclidinium) nitrogen rings kept

at an appropriate distance by lipophilic rigid or semirigid linkers. Many homo-

dimeric QASs showed AChE and BChE inhibitory potency in the nanomolar

range along with a low enzymatic selectivity. Computational studies on AChE,

BChE, and ChoK allowed identiﬁcation of the key molecular determinants for

high aﬃnity and selectivity over either one of the three enzymes and guided

the design of a hybrid bis-QAS (56) exhibiting the highest AChE aﬃnity (IC₅₀

=

15 nM) and selectivity over BChE and ChoK (SI = 50 and 562, respectively) and a

promising pharmacological potential in myasthenia gravis and neuromuscular

blockade.

’ INTRODUCTION

aromatic residues at the bottom where the active site is located.³

Here, two diﬀerent and close subsites contribute to acetylcholine

binding (primary anionic subsite) and hydrolytic activity (esteratic

subsite). One of the most signiﬁcant structural diﬀerences between

the two ChEs resides in their peripheral binding site. In the AChE,

where it gives rise to more speciﬁc interactions with ligands, it is

called the peripheral anionic subsite (PAS) and it is smaller and

narrower than in BChE. In addition, consolidated experimental

evidence supports the involvement of PAS in the acceleration of

amyloid β peptide (Aβ) aggregation in the CNS leading to senile

plaques that represent the major histopathological hallmarks of

Alzheimer’s disease (AD).⁴A strong binding at the PAS accounts

for the high inhibitory potency toward peripheral AChE of many

gorge-spanning, highly polar bis-quaternary ammonium salts

(QASs), e.g., decamethonium andambenonium, whereas less polar

gorge-spanning aminic compounds (e.g., donepezil and homo- and

heterodimers of tacrine), targeting mainly the central AChE, bind

Acetylcholine (ACh) is a neurotransmitter in both the per-

ipheral and central nervous system (PNS and CNS, respectively).

In the PNS, ACh mediates nerve impulse in neuromuscular

junctions and is a major neurotransmitter in the autonomic nervous

system, whereas in the CNS, it chieﬂy controls memory and

learning processes.

Two cholinesterases (ChEs), which are serine hydrolase en-

zymes, contribute to the degradation of the neurotransmitter ACh;

they are acetylcholinesterase (AChE, EC 3.1.1.7) and butyrylcho-

linesterase (BChE, EC 3.1.1.8). Both ChEs hydrolyze ACh, albeit

with slightly diﬀerent kinetics, and coexist ubiquitously in humans.

AChE, mainly expressed in the muscle and cholinergic nervous

system, shows a selectivity toward its endogenous substrate higher

than BChE, which is produced in the liver and represents the

predominant plasmatic ChE. The two ChEs are encoded by

diﬀerent genes and diﬀer in substrate speciﬁcity and sensitivity to

inhibitors mainly because of signiﬁcant structural diﬀerences in

their active site gorge.^1,2AChE enzymatic cavity has the shape of a

narrow gorge nearly 20 Å deep, lined with 14 highly conserved

Received: October 8, 2010

Published: March 21, 2011

r

2011 American Chemical Society

2627

dx.doi.org/10.1021/jm101299d J. Med. Chem. 2011, 54, 2627–2645

|

Journal of Medicinal Chemistry

ARTICLE

Chart 1. Chemical Structures of Diﬀerent AChE Inhibitors Cited in the Text

at the PAS with one of the two molecular heads as revealed by many

cocrystallized enzymeꢀinhibitors complexes.⁵

Chart 2. Chemical Structures of Mono-Quaternary Ammo-

nium Salts 2ꢀ4

The interest in AChE as a biological target in pharmaceutical

research arises from the large number of potential central and

peripheral therapeutic activities of AChE inhibitors. More

recently developed AChE inhibitors (AChE-Is) address the

symptomatic treatment of AD,⁶while older QASs AChE-Is

have found a therapeutic application in the treatment of neuro-

muscular blockade (NMB) in surgical anesthesia,⁷myasthenia

gravis (MG),⁸and to a lesser extent in glaucoma. As for the NMB

reversal agents, at present, all clinically used drugs, such as

neostigmine, pyridostigmine, and edrophonium, are QASs inhi-

bitors of AChE. While for glaucoma the currently used drugs are

β-blockers and prostaglandin analogues,⁹in the case of MG, a

debilitating and rare autoimmune disease involving a progressive

aberration of the neuronal transmission at the neuromuscular

junctions, the ﬁrst-line treatment in the early stage and in mild

disease is still made with oral and low brain-permeable AChE-Is

(e.g., the QASs pyridostigmine and neostigmine). To reduce the

major drawbacks associated with this therapy resulting in the so-

called cholinergic crisis (e.g., hypersalivation, muscular weakness,

diarrhea, and cardiovascular side eﬀects),¹⁰the search for new

more potent and selective peripheral AChE-Is is still ongoing.

Indeed, over the years a plethora of potent and selective

AChE-Is have been reported, showing a wide range of chemical

diversity and biopharmacological actions. Within this broad

cluster of chemical entities, mono- and bis-QASs have kept a

starring role. The bis-QASs are generally spanning along the deep

gorge of AChE binding to both primary and peripheral anionic

subsites which are characterized by the presence of Trp residues

(e.g., Trp84 and Trp279, respectively, in Torpedo californica

AChE (TcAChE)). Therefore, a number of bis-QASs proved

to display outstanding nano- or subnanomolar aﬃnities because

of the highly positive binding stabilization by cationꢀπ interac-

tions with Trp residues in both AChE subsites and, more

occasionally, by polar interactions within the spanning gorge.

heterodimeric edrophoniumꢀcoumarin hybrids¹¹prompted us

to evaluate as ChEs inhibitors, with potential in MG and NMB, a

large series of mono- (Chart 2), bis- (Chart 3), and tris-

quaternary heterocyclic bromides (Chart 4). A number of these

molecules had been already studied as ChoK inhibitors.^12ꢀ20

The present research project stemmed from the observation that

some key molecular features for an eﬀective inhibition of ChEs

were present in many ChoK inhibitors reported in Charts 2ꢀ4.

In particular, the key pharmacophoric elements, which are two

charged quaternary ammonium moieties kept at an appropriate

distance by a linker, characterized many bis-QASs reported in

Chart 3.

After an initial screening of AChE and BChE inhibition of

already reported QASs,^13ꢀ20many new compounds were de-

signed and tested aiming at the detection of the main molecular

determinants of inhibitory potency and enzyme selectivity

through an extension and a critical examination of the structure

ꢀaﬃnity relationships (SAFIRs) and structureꢀselectivity rela-

tionships (SSRs).

Following one of the most exploited design strategy for

enhancing AChE aﬃnity, which is the dimerization of simple

molecular fragments acting on the catalytic or peripheral binding

sites, new and more selective inhibitors were designed by linking

twin cationic heads with a spacer of appropriate length and

ﬂexibility with the purpose of discovering new molecules with

high AChE aﬃnity and potential pharmacological activity in

glaucoma, myasthenia gravis, and neuromuscular blockade. An-

other important goal of our project was the dissociation of the

ChE activity from the ChoK activity. Very satisfactorily, a

As an example, the excellent aﬃnity of ambenonium (IC₅₀

=

0.12 nM) and BW284C51 (IC₅₀= 8 nM), two symmetrical bis-

QASs (Chart 1), might be ascribed to these multiple strong

interactions.

The variety of pharmacological activities displayed by QASs,

their generally facile synthetic accessibility, and our consolidated

expertise in the study of the inhibition of AChE and BChE by

2628

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

Chart 3. Chemical Structures of Bis-Quaternary Ammonium Salts 5ꢀ52

Chart 4. Chemical Structures of Tris-Quaternary Ammo-

nium Salts 53ꢀ55

ChoK selectivity, and with a promising potential in MG and

NMB was successfully designed, synthesized, and tested.

’ CHEMISTRY

The QASs presented in this paper are monocationic (Chart 2),

bis-cationic (Chart 3), and tris-cationic derivatives (Chart 4).

Thirty compounds have been prepared following procedures

already described in the literature.^13ꢀ20

Concerning the synthesis of the new target QASs, the

preparation of monocationic derivatives 2ꢀ4 was carried out

by heating the appropriate nitrogen heterocycle with an equi-

molar ratio of benzyl or p-tolyl bromide in a sealed tube at 100 ꢀC

using 2-butanone as the solvent (Scheme 1).

The synthesis of diﬀerent dibromides was required to insert a

linker between the two cationic fragments of new designed

homodimeric QASs. trans-1,2-Bis(bromomethyl)cyclopropane

was obtained by reduction of the corresponding commercial diester

molecule with excellent inhibitory potency toward AChE, with

good AChE over BChE selectivity, with outstanding AChE over

2629

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

to diol with lithium aluminum hydride²¹and by the subsequent

bromination with phosphorus tribromide.²²The 3,3⁰-bis(bromo-

methyl)biphenyl linker was prepared through the radical bromina-

tion of the commercially available 3,3⁰-dimethylbiphenyl with NBS

and benzoyl peroxide as radical initiator.²³The synthesis of 4,4⁰-

bis(bromomethyl)bibenzyl was carried out by reacting bibenzyl

with formaldehyde and hydrogen bromide in the presence of

o-phosphoric acid.²⁴1,4-Bis[4-(bromomethyl)phenyl]butane was

prepared by the catalytic hydrogenation of trans,trans-1,4-diphenyl-

1,3-butadiene using Pd/C in glacial acetic acid, followed by the bis-

bromomethylation in the previously mentioned conditions.²⁴trans-

4,4⁰-Bis(bromomethyl)stilbene was synthesized from trans-4,4⁰-

bis(hydroxymethyl)stilbene by bromination with phosphorus

tribromide.²⁵trans-4,4⁰-Bis(hydroxymethyl)stilbene was prepared

by the reduction of trans-dimethylstilbene 4,4⁰-dicarboxylate with

lithium aluminum hydride.²⁶

in acetonitrile at 160 ꢀC. 4-Cyclohexylpyridine 1g was prepared

by an iron-mediated coupling reaction between 4-chloropyridine¹³

and freshly prepared cyclohexylmagnesium bromide, using THF

as the solvent and NMP as the ligand. The Grignard reagent was

prepared in situ by classic halogenꢀmetal exchange reaction

starting from cyclohexyl bromide and magnesium turnings in

reﬂuxing THF, in the presence of a catalytic amount of iodine

(Scheme 3). A Suzuki coupling reaction between 4-chloropyr-

idine hydrochloride and phenylboronic acid, in the presence of

Pd-tetrakis(triphenylphosphine) as Pd(0) source and Cs₂CO₃as

the base in a dimethoxyethane/water solvents mixture, allowed

the synthesis of 4-phenylpyridine 1h as described in Scheme 3.

The synthesis of the new target homodimeric bis-QASs was

carried out by heating the suitable dibromides and the appro-

priate nitrogen heterocycles (namely, 4-substituted pyridines or

substituted quinolines and quinuclidines) in a molar ratio 1:2,

thus obtaining the bis-cationic derivatives 5, 6, 9, 10, 12, 15, 16,

20ꢀ22, 27, 28, 36, 41ꢀ44 (Scheme 4), and 47ꢀ50 (Scheme 5).

Scheme 2 describes the preparation of 4,4⁰-bis(bromo-

methyl)biphenyl 1c starting from the commercially available

biphenyl-4,4⁰-dicarboxylic acid that was reacted with SOCl₂to

give the intermediate acyl chloride that aﬀorded the ethyl diester

1a upon reﬂuxing in ethanol. LiAlH₄in reﬂuxing THF was used

to reduce the diester 1a to the corresponding diol 1b. The

subsequent bromination performed with CBr₄and triphenylpho-

sphine aﬀorded dibromide 1c.

Scheme 3. Synthesis of 4-Substituted Pyridines 1dꢀh^a

As shown in Scheme 3, a common synthetic pathway led to

the targeted para-substituted pyridines 1d, 1e, and 1f involving

the microwave-assisted nucleophilic aromatic substitution of 4-

chloropyridine with piperidine, 1,2,3,4-tetrahydroisoquinoline,

and 6,7-dimethoxy-1,2,3,4-tetrahydroisoquinoline, respectively,

Scheme 1. Reaction Pathways for the Synthesis of Mono-

QASs 2ꢀ4^a

^aReagents and conditions: (i) N,N-diisopropylethylamine, dry acetoni-

trile, 160 ꢀC, microwave, 1 h; (ii) magnesium turnings, iodine, dry THF,

reﬂux, 1 h; (iii) 4-chloropyridine, Fe(acac)₃, THF, NMP, room tem-

perature, 15 h; (iv) phenylboronic acid, Cs₂CO₃, Pd-tetrakis(PPh₃),

dimethoxyethane/water 4/1, 80 ꢀC, 4 h.

^aReagents and conditions: (i) benzyl or p-tolyl bromide, 2-butanone,

sealed tube, 100 ꢀC, 48 h; (ii) p-tolyl bromide, 2-butanone, sealed tube,

100 ꢀC, 48 h.

Scheme 2. Synthesis of 4,4⁰-Bis(bromomethyl)biphenyl 1c^a

^aReagents and conditions: (i) thionyl chloride, absolute ethanol, room temperature, overnight then reﬂux, 24 h; (ii) LiAlH₄, dry THF, reﬂux, 7 h; (iii)

triphenylphosphine, CBr₄, dry CH₂Cl₂, 0 ꢀC to room temperature, 6 h.

2630

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

Scheme 4. Reaction Pathway for the Synthesis of Symmetrical Bis-QASs 5, 6, 9, 10, 12, 15, 16, 20ꢀ22, 27, 28, 36, and 41ꢀ44^a

ARTICLE

^aReagents and conditions: (i) appropriate dibromides, 2-butanone, sealed tube, 100 ꢀC, 15ꢀ192 h.

Scheme 5. Reaction Pathway for the Synthesis of Symme-

assessed in an ex vivo system using recombinant human enzyme

according to a reported method.¹³Inhibition of β-amyloid

aggregation (Table 5) was determined with a spectroﬂuori-

metric, thioﬂavin-T based assay as previously described.²⁸

trical Bis-QASs 47ꢀ50^a

’ MOLECULAR MODELING

The program GOLD²⁹was used in all the docking simulations.

GOLD oﬀers two standard ﬁtness functions, which are Gold-

Score and ChemScore, with generally comparable performances.

However, for any given problem, their predictive capability might

strongly diﬀer, and therefore, the selection of the ﬁtness function

is generally made after assessing its capacity to accurately

reproduce the X-ray binding conformations. In our case, the

two functions were challenged in the prediction of the binding

conformations of donepezil and hemicholinium-3, found in their

complexes with TcAChE (PDB code 1EVE)⁵and hChoK (PDB

code 3F2R),³⁰respectively. As a measure of ﬁt, the rmsd value

was calculated over all the corresponding heavy atoms between

the docking and crystal conformations. As GoldScore aﬀorded

better results for ChEs while ChemScore was better calibrated

toward ChoK, the two diﬀerent ﬁtness functions were accord-

ingly selected for the docking runs on ChEs and ChoK, respec-

tively. The rmsd values for donepezil and hemicholinium-3 were

as small as 0.99 and 1.53 Å, respectively.

^aReagents and conditions: (i) appropriate dibromides, 2-butanone,

sealed tube, 100 ꢀC, 15ꢀ192 h.

The reactions were performed in a sealed tube at 100 ꢀC using

2-butanone as the solvent.

Scheme 6 depicts the synthesis of the heterodimeric com-

pound 56. Intermediate monobromide 1i was obtained through

the nucleophilic substitution of 4-substituted pyridine 1d with an

excess of 1,4-bis(bromomethyl)benzene in 2-butanone in a

sealed tube at 100 ꢀC. Then 1i was reacted in the same conditions

with an equimolar ratio of tetrahydroisoquinoline derivative 1f,

yielding the desired heterodimer 56.

’ BIOLOGICAL ASSAYS

The ChEs in vitro inhibition data of the examined QASs,

determined by using the Ellman’s assay,²⁷are reported in

Tables 1ꢀ5. ChoK inhibitory activity of compound 56 was

Molecular dynamics (MD) simulations were carried out by

using a periodic box of TIP3P water molecules extended for 18 Å

from the protein atoms which were modeled with the parm03

2631

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

Scheme 6. Synthesis of Heterodimer Bis-QAS 56^a

^aReagents and conditions: (i) 1,4-bis(bromomethyl)benzene (excess), 2-butanone, 100 ꢀC, sealed tube, 120 h; (ii) 1f, 2-butanone/acetonitrile 4/1,

100 ꢀC, sealed tube, 48 h.

Table 1. Biological Data of Monomeric Pyridinium

Compounds 2ꢀ4

the bisector, respectively, the distance of their pIC₅₀values from

the bisector being a direct measure of their degree of selectivity.

Two parallel lines traced at one pIC₅₀unit distance below and

above the bisector enable the straightforward location of inhibi-

tors with a SI higher than 10 and lower than 0.1, respectively (that

is, ΔpIC₅₀> 1 and ΔpIC₅₀< ꢀ1, respectively).

At ﬁrst glance, the plotted data indicated that a large number of

potent AChE inhibitors were discovered (i.e., compounds 8, 14,

15, 18, 19, 20, 25, 28, 32, 35, 37, 38, 42, 43, and 49 with an IC₅₀

mean value close to 90 nM) whereas only few BChE inhibitors

(i.e., compounds 32, 35, 45, and 47) exhibited a comparable high

potency. The same considerations hold for the enzymatic

selectivity: a relatively high number of inhibitors showed good

AChE selectivity (SI > 10), whereas only compound 47 exhibited

a good BChE selectivity (SI < 0.1).

Monocationic compounds 2ꢀ4 and bis-quaternary pyridi-

nium derivative 5, bearing a cyclopropyl linker, showed low

binding aﬃnities for both ChEs.

^aCounterions are always bromide ions. ^bThe AChE and BChE inhibi-

tory activities are expressed as IC₅₀(μM) or as percentage of inhibition

( SEM at 10 μM. Values are the mean of two or three independent

experiments. SEM of IC₅₀values are <10%.

SAFIRs of AChE Inhibition. Among derivatives 5ꢀ44, char-

acterized by a common bis-pyridinium cationic head carrying

various, generally basic, R-substituents at position 4, very inter-

esting IC₅₀values (e100 nM) were observed for compounds 14,

18, 19, 25, 28, 32, 35, 37, 38, and 43. A high AChE binding

affinity seemed to be related to the presence of differently

substituted lipophilic aromatic rings in the linker that might

interact with aromatic amino acid side chains spanning the

aromatic gorge of the enzymatic cavity. Actually, moving from

a cyclopropyl to a more lipophilic phenyl linker, a significant

increase of affinity can be appreciated (compare 5, IC₅₀= 8.78

μM, with 6 and 12, IC₅₀= 0.390 and 1.31 μM, respectively).

Regarding compounds bearing xylyl spacers, an increasing size

of the 4-substituent in the cationic heads from N,N-dimethyla-

mino to the piperidinyl group resulted in a signiﬁcant gain of

activity in both p- and m-xylyl linkers (14 > 13 > 12 and 8 > 7 > 6).

When the aliphatic amino substituents were substituted with

larger and more lipophilic 1,2,3,4-tetrahydroisoquinolin-2-yl

(THIQ) moieties, a drop of aﬃnity was observed (8 > 9 > 10

and 14 > 15 > 16), the worst results being obtained with the 6,7-

dimethoxy-THIQ derivatives 10 and 16 (IC₅₀= 2.10 and 1.95

μM, respectively). Interestingly, this decrease of aﬃnity was less

dramatic in the series of inhibitors with a p-xylyl linker. It may be

that more sterically hindered cationic heads were better tolerated

when tethered by a p-xylyl than a m-xylyl spacer because the

former might allow a more extended and stable binding con-

formation. This hypothesis may hold also for sterically hindered

cationic heads tethered to the longer 3,3⁰-bis-methylene- and

4,4⁰-bis-methylenebiphenyl isomeric linkers (compare 28 with

22). Moreover, compounds bearing 3,3⁰-bis-methylenebiphenyl

spacers were more active than the corresponding meta-substi-

tuted xylyl-linked analogues (18 > 7, 19 > 8, and 22 > 10). Within

force ﬁeld. Energy minimization was executed by using AMBER,

version 10,³¹while a 5 ns MD simulation was performed with

NAMD.³²

’ RESULTS AND DISCUSSION

The inﬂuence of the cationic heads and linkers on the AChE

and BChE inhibitory activity was extensively explored with

appropriate structural modiﬁcations that allowed us to shed light

on the most salient aspects of the SAFIRs and SSRs.

Inhibition data are reported in Tables 1ꢀ5 as IC₅₀(μM) or as

percentage of inhibition at 10 μM. The enzyme selectivity was

expressed with the selectivity index SI, which is the IC₅₀(BChE)/

IC₅₀(AChE) aﬃnity ratio.

Inhibitors were grouped according to their shared linkers in

the following order: monocationic (Table 1), bis-cationic

(Tables 2ꢀ4), and tris-cationic (Table 4) QASs. For the sake

of clarity, after some general comments on the SAFIRs and SSRs,

SAFIRs will be discussed separately for AChE and BChE

inhibition.

General Observations on SAFIRs and SSRs. For an immedi-

ate and more efficient analysis of SAFIRs and SSRs, inhibition

data are presented in Figure 1 as a plot of pIC₅₀of AChE (x-axis)

versus pIC₅₀of BChE (y-axis) using the same scale and range for

both axes (square plot). To avoid the loss of informative data

points, compounds 12, 48, and 53, for which only the percentage

of inhibition of one ChE was available, were added to the plot

with an estimated missing pIC₅₀of 5.00. Compounds with close

affinities at both ChEs lie on the bisector (y = x) of the graph,

whereas selective AChE and BChE inhibitors lie below and above

2632

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

Table 2. Biological Data of Dimeric Pyridinium Compounds 5ꢀ44

2633

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

Table 2. Continued

^aCounterions are always bromide ions. ^bDashed lines represent the connection bond of the substituent. ^cThe AChE and BChE inhibitory activities are

expressed as IC₅₀(μM) or as percentage of inhibition ( SEM at 10 μM. Values are the mean of two or three independent experiments. SEM of IC₅₀

values are <10%. ^dSI is the selectivity index, that is, the IC₅₀(BChE)/IC₅₀(AChE) aﬃnity ratio.

this class of compounds the most potent AChE inhibitor 19,

showing an IC₅₀= 26 nM, was found.

steric hindrance at the AChE binding site(s) not seen in the

larger, and more accommodating, BChE binding cavity. An

alternative explanation could be found in a diﬀerent charge

distribution likely resulting from the reduced electron-donor

character of the anilino group compared to the cycloaliphatic

amino groups. Actually, a strong electron-donor eﬀect by

R-substituents at position 4, and the consequent higher

delocalization of the pyridinium positive charge, seemed to

play a key role for good AChE aﬃnity most likely because

πꢀcation interactions with the aromatic residues in the PAS of

AChE might be reinforced. And in fact, piperidine derivative

19 exhibited an aﬃnity signiﬁcantly higher than that of the

comparably sized cyclohexyl (20) and phenyl (21) derivatives

(IC₅₀= 26 vs 134 and 553 nM, respectively), both having

a lower electron-donating character. These ﬁndings were

supported by semiempirical AM1 level calculations of the

charge distribution and the LUMO (lowest unoccupied mo-

lecular orbital) energy of 1-methylpyridinium and of a series

of 4-substituted derivatives, which are the 1-methyl-4-phenyl-

pyridinium, 4-cyclohexyl-1-methylpyridinium, and 4-piperidinyl-

1-methylpyridinium, used as reference model systems. The

obtained results (data not shown) were in good agreement with

An increased ﬂexibility of the spacer was obtained by introdu-

cing methylene bridges between the two phenyl rings of the 4,4⁰-

bis-methylenebiphenyl group (compounds 31ꢀ43 in Table 2).

No clear SAFIRs emerged from the analysis of inhibition data of

this series of more ﬂexible inhibitors, and no optimal length of the

linker could be found. However, in the presence of more ﬂexible

linkers very potent AChE inhibitors, i.e., compounds 32, 35, 37,

38, and 43, all exhibiting IC₅₀e 100 nM, emerged. In addition, it

is worth noting that the reduction of ﬂexibility in the linker and of

the lipophilicity of the R group led to a signiﬁcant decrease of

AChE aﬃnity (compare 44 vs 35 and 35 vs 36, respectively).

The AChE enzymatic region likely binding the R substituents

of the cationic heads appeared not so sterically hindered, as the

ﬁve-, six- and seven-membered cycloaliphatic amines seemed to

be bound with good binding aﬃnities (see compounds 34, 35,

and 37 and compounds 38ꢀ40).

The presence of substituted or unsubstituted N-methylani-

lino R groups in the cationic heads of compounds 11, 17, 23,

24, 29, and 30 promoted BChE selectivity. This result may be

attributed to an increased lipophilic character and/or to some

2634

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

Table 3. Biological Data of Dimeric Quinolinium and Quinuclidinium Derivatives 45ꢀ50

^aCounterions are always bromide ions. ^bDashed lines represent the connection bond of the substituent. ^cThe AChE and BChE inhibitory activities are

expressed as IC₅₀(μM) or as percentage of inhibition ( SEM at 10 μM. Values are the mean of two or three independent experiments; SEM of IC₅₀

values are <10%. ^dSI is the selectivity index, that is, the IC₅₀(BChE)/IC₅₀(AChE) aﬃnity ratio.

Table 4. Biological Data of Bis- and Tris-QASs 51ꢀ55

^aCounterions are always bromide ions. ^bThe AChE and BChE inhibitory activities are expressed as IC₅₀(μM) or as percentage of inhibition ( SEM at

10 μM. Values are the mean of two or three independent experiments. SEM of IC₅₀values are <10%. ^cSI is the selectivity index, that is, the IC₅₀(BChE)/

IC₅₀(AChE) aﬃnity ratio.

those reported in a previous investigation and unequivocally

indicated that electron releasing eﬀects, in particular by the

1-piperidinyl substituent, increased charge delocalization and

the LUMO character which were correlated with the potency of

2635

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

Table 5. Human Cholinesterases and Aβ Aggregation Inhi-

bition Data

Molecular rigidiﬁcation represents a common strategy

adopted by medicinal chemists to increase the activity of a drug

or to reduce its side eﬀects. Incorporating key functional groups

of a ﬂexible ligand into a ring is the usual way of “locking” a

conformation and may provide valuable information on the likely

active binding conformation of that ligand. With this aim, bis-

pyridinium cyclophanes 51 and 52 had been prepared as ChoK

inhibitors and were now tested as ChE inhibitors. All these very

rigid macrocyclic compounds proved to be very poor AChE

inhibitors (Table 4). The same poor inhibition data were

obtained with the tris-cationic compounds 53ꢀ55.

b

compd

hAChE^a

hBChE^a

Aβ_1ꢀ40

14

75 ( 13

60 ( 4

610 ( 25

330 ( 14

36 ( 3

not active

30 ( 2

19

47

620 ( 45

62 ( 4

not active

15 ( 1

56

620 ( 55

3900 ( 360

23 ( 5

donepezil

18 ( 4

not active

tacrine

110 ( 5

not active

^aThe hAChE (human recombinant) and hBChE (from human serum)

inhibitory activities are expressed as IC₅₀(nM). ^bPercentage of Aβ

(30 μM) self-aggregation inhibition measured at 100 μM test com-

pound. Values are the mean ( SEM of at least three independent

experiments.

SAFIRs of BChE Inhibition. SAFIRs at BChE were not so

sharply defined and clear as for AChE. Therefore, only a limited

number, though important, of observations could be made.

BChE inhibition data in Tables 1ꢀ4 and in Figure 1 indicated

that only compounds 32, 45, and 47, exhibited IC₅₀e 100 nM.

Two of them, compounds 45 and 47, were characterized by bulky

cationic heads, a structural feature previously hypothesized as

responsible for the shift of selectivity toward BChE.

The large sized inhibitors listed in Table 4, with only one

exception that is 53, displayed a poor BChE aﬃnity but higher

than AChE.

Inhibitors with cationic heads carrying substituted or unsub-

stituted N-methylanilino R-groups showed a higher BChE

inhibitory potency than AChE (see compounds 11, 17, 23, 24,

29, and 30). Notably, piperidinyl-substituted inhibitors were

always more potent than the corresponding pyrrolidinyl deriva-

tives, i.e., 8 > 7, 14 > 13, 19 > 18, 26 > 25, 32 > 31, 35 > 34, 39 >

38, and 43 > 42.

As far as the eﬀect of linker ﬂexibility on the BChE activity is

concerned, it may be emphasized that a large number of potent

BChE inhibitors, such as 32, 35, 39, 40, and 43 with IC₅₀of 62,

130, 149, 163, and 206 nM, respectively, carried ﬂexible linkers.

However, this structural feature also determined a potent AChE

inhibition, and as a result, inhibitors with high aﬃnity toward

both enzymes, e.g., compounds 32, 35, 39, and 43, might be

found within this class of molecules.

SSRs. In general, no well-defined SSRs emerged from the

analysis of the SI values, as the data variation was very limited,

Moreover, the selectivity index SI comes from the ratio of two

different experimental values (AChE and BChE IC₅₀values), and

therefore, it may suffer of a stronger error propagation compared

to the single set of AChE and BChE affinity data. Consequently,

only the most relevant aspects of the SSRs will be discussed.

AChE and BChE inhibition data in Tables 1ꢀ4 and in the plot

of Figure 1 indicated that only a limited number of inhibitors

showed a good selectivity toward AChE or BChE (SI > 10 or SI <

0.1, respectively). SI > 10 was exhibited by compounds 6, 7, 18,

25, 28, 49, and 50, whereas only one compound (that is, 47)

displayed SI < 0.1. These compounds are highlighted in Figure 1

and are located below and above the parallel line traced at 1 log

unit from the bisector, respectively.

The inﬂuence of the steric hindrance of the cationic heads and

of the linker length and ﬂexibility on the selectivity was not

clearly deﬁned. In general, a higher steric hindrance on both the

cationic heterocycles and R-substituents increased the BChE

aﬃnity, but the resulting selectivity was generally not very high

(see chloro-substituted N-methylanilino derivatives 17, 24, 29,

and 30, the aminoquinolinium derivatives 45 and 46, and all the

large sized and/or rigid compounds 51, 52, 54, and 55). On the

other hand, for the rigid and bulky quinuclidinium derivatives

47ꢀ50, the length and, to a lesser extent, the ﬂexibility of the

Figure 1. Square plot of AChE/BChE aﬃnity and selectivity. Bordered

by dashed lines, the bottom-right and upper-left corners contain potent

and selective AChE (solid triangles) and BChE (empty circles) inhibi-

tors, respectively. The solid line represents the bisector.

inhibition toward both AChE and BChE as a result of a weaker

solvation eﬀect.¹³

Compounds 45 and 46, carrying the more lipophilic and

bulkier 4-aminoquinolinium cationic heads, showed very low

AChE aﬃnity with both the 3,3⁰-bis-methylenebiphenyl (45) and

the 1,2-bis-p-tolylethane (46) linker.

Also, derivative 47, bearing very lipophilic and bulky quinuclidine

cationic heads bound through a 3,3⁰-bis-methylenebiphenyl linker,

exhibited a very low AChE aﬃnity (IC₅₀= 1.10 μM) that was even

worse when a hydroxyl group was introduced at position 3 in the

quinuclidinium ring (compound 48, 55.8% inhibition at 10 μM).

Conversely, bis-quinuclidine QASs 49 and 50 bearing a 1,2-bi-

sphenylethane and a trans-stilbenyl linker, respectively, proved to be

potent AChE inhibitors (IC₅₀= 97 and 141 nM, respectively).

Therefore, for the series of bis-quinuclidinium derivatives, both the

ﬂexibility and the length of the linker appeared to play a key role for

an eﬃcient interaction at the AChE binding sites.

2636

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

linker clearly appeared as crucial structural determinants for

BChE or AChE selectivity. Compounds 49 and 50, bearing a

1,2-bis-p-tolyl ethane and a 4,4⁰-bis-methylene-trans-stilbenyl

linker, respectively, proved to be highly potent (IC₅₀= 97 and

141 nM, respectively) and selective AChE inhibitors (SI = 22 and

38, respectively), whereas compound 48 with a shorter and less

ﬂexible 3,3⁰-methylenebiphenyl linker presented an inverted

biological proﬁle. In addition to this, derivative 47, characterized

by an unsubstituted quinuclidinium fragment, exhibited a high

BChE aﬃnity (IC₅₀= 89 nM) and the highest BChE selectivity

within the whole series of tested molecules (SI = 0.081).

Figure 2. Heterodimer bis-QAS 56.

identical amino acid residues at both the catalytic and periph-

eral binding sites, apart from the substitution of Tyr337

(hAChE) with Phe330 (TcAChE).³³This single-residue change

was not observed when analyzing primary sequences of either

human (1B41, PDB entry) or bovine (AA123899.1, NCBI

entry) AChEs.

Finally a few words may be spent on the AChE over ChoK

selectivity of the examined QASs. As previously noticed, some of

the QASs reported in this work have been previously studied as

ChoK inhibitors. The inhibition data on this enzyme varied from

submicromolar to the high micromolar range, the very rigid cyclic

bis-QAS 52 being the most potent ChoK inhibitor with an IC₅₀

of 0.30 μM¹⁸and a low selectivity over AChE (SI = 0.063, where

SI is the IC₅₀(ChoK)/IC₅₀(AChE) aﬃnity ratio). With the only

exception of compound 23 (IC₅₀= 0.40 μM ¹⁴and SI = 0.46) all

the other compounds (data not shown) exhibited SI . 1, with

the highest selectivity displayed by compound 7 (SI = 520).

AChE aﬃnity of 7 was, however, relatively low (IC₅₀= 162 nM).

Structure-Based Design of an Inhibitor with Improved

AChE Potency, AChE over BChE Selectivity, and AChE over

ChoK Selectivity. Although a consistent number of homodi-

meric QASs exhibited AChE or BChE affinity in the nanomolar

range, none of them showed a high selectivity toward either one

of the two ChEs. A molecular hybridization approach was

therefore conceived for the design of a potent and selective

AChE inhibitor, possibly endowed with a low affinity toward the

ChoK. This heterodimeric inhibitor was designed by exploring

the combination of two differently 4-substituted pyridinium

cationic heads, chosen among the ones characterizing the most

active AChE inhibitors. Keeping in mind that the presence of an

electron-releasing R-substituent at position 4 is crucial for good

ChE affinity and that highly lipophilic and bulky R-substituents

shifted the affinity toward BChE, unhindered electron-donating

groups at position 4 (i.e., 1-pyrrolidinyl, 1-piperidinyl, and

1-azepinyl) were chosen as the substituent of one of the two

pyridinium heads. Taking into account a previous observation

that a low BChE affinity could have been promoted by the

presence of only one hindered cationic head, the 6,7-dimethoxy-

substituted and unsubstituted THIQ pyridinium moieties were

explored as the second cationic head. Indeed the 6,7-dimethoxy-

THIQ pyridinium cationic head proved to characterize very

active AChE inhibitor (see compound 28), providing that its

potential steric hindrance was partly relieved by an elongated

binding conformation, possibly ensured by para-disubstituted

linkers. As far as ChoK inhibition is concerned, previous SAFIRs

seemed to indicate that this choice could be compatible with a

likely low inhibition. Accordingly, p-xylyl and the 4,4⁰-bis-methy-

lenebiphenyl fragments were explored as the linkers of the two

different cationic heads, selected on the basis of the previously

discussed criteria. The study of the preferred binding conforma-

tions of potentially selective enzyme inhibitors derived from the

combination of the just mentioned molecular fragments was

carried throughout docking simulations.

A series of prospective docking studies were then undertaken

to assist the rational design of a QAS inhibitor conceived to bind

the three enzymes AChE, BChE, and ChoK, with highly diﬀerent

aﬃnities. The task of identifying a novel potent AChE inhibitor

with a remarkable selectivity over BChE and ChoK was very

challenging because of the close sequence similarity between

AChE and BChE (about 66%) and the greater structural diversity

between the two ChEs and ChoK. In view of this, a large series of

virtual compounds deriving from an exhaustive combination of

the molecular features delineated above, i.e., two diﬀerent

cationic heads bound through para-substituted linkers, were

employed to forecast, via molecular docking, AChE, BChE,

and ChoK inhibition. Compounds exhibiting toward the three

enzymes highly diverse aﬃnities, as assessed by the diﬀerent

scoring energies, and binding modes compatible with those of

known inhibitors (e.g., donepezil and decamethonium in the case

of ChEs, and hemicholinium-3 in the case of ChoK) were

selected and further analyzed for the evaluation of their synthetic

feasibility. Among the investigated heterodimeric molecules,

compound 56 was selected and synthesized as the most promis-

ing candidate (Figure 2).

It was in fact observed that the novel designed heterodimer 56

achieved a percentage of scoring vs donepezil equal to 102.9% for

AChE and 103.5% for BChE, whereas vs hemicholinium-3 it was

equal to 113.3% for ChoK. In addition, the visual inspection of its

predicted posing into AChE, BChE, and ChoK (the last two

reported in Figures S1 and S2, respectively, in the Supporting

Information) disclosed to some extent a good overlap with

binding modes of donepezil in TcAChE and hemicholinium-3

in hChoK. The most relevant diﬀerence between the docking

pose of donepezil into hAChE (PDB code 1B41) and its

experimental binding conformation into TcAChE (PDB code

1EVE) was the position of the piperidine ring which resulted

more bent in hAChE and more distant from Tyr337 to prevent

steric clashes. Such evidence was congruent with the observation

that the overlay between the Tyr337 (hAChE numbering) and its

equivalent Phe330 (TcAChE numbering) disclosed an rmsd

equal to 1.99 Å because of the structural ﬁt induced by diﬀerent

cognate ligands within the two orthologous binding sites. How-

ever, a very nice match of the face-to-face interactions occurring

at the catalytic and peripheral binding sites was observed.

As far as the docking of heterodimer 56 is concerned, it was

observed that 56 was likely engaged in favorable interactions

taking place simultaneously at both the primary and peripheral

binding sites of hAChE with the top-scored docking pose

exhibiting a GoldScore ﬁtness energy equal to ꢀ51.69 kJ/mol.

As shown in Figure 3, the prospective use of docking experiments

suggested that the piperidinyl substituent was likely involved in

For this study, the 3D coordinates of the human isoforms of

AChE (PDB code 1B41), BChE (PDB code 1P0I), and ChoK

(PDB code 3F2R) were retrieved from the Protein Data Bank.

hAChE was chosen instead of TcAChE because it has almost

2637

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

Figure 3. Top-scored docking pose of inhibitor 56 into hAChE (PDB code 1B41) is rendered in white capped stick models, while starting MD coordinates

of residues Trp86 and Trp286 are in ball and stick models colored according to the standard atom code. The ﬁtting spheres represent the positions visited

during the 5 ns MD bythephenolic oxygenatoms of the nonconservednarrow gorge residues Tyr337 and Tyr124 and are coloredin magenta and orange as

the corresponding balland stick models oftheirside chains. Black wireframe modelsindicate the aminoacid side chain and inhibitor56conformations atthe

end of the 5 ns MD simulation. Red dashed lines show the occurrence of possible ionꢀdipole contacts established among the side chain oxygen atoms of

Tyr337 and Tyr124 with the pyridinium nitrogen atoms at the distances of 4.82 ((0.42) Å and 4.44 ((0.35) Å, respectively.

hydrophobic and, taking into account the strong charge deloca-

lization, cationꢀπ interactions with Trp86 while, interestingly, a

number of polar residues in the narrow spanning gorge, which are

Tyr124, Tyr337, and to a lesser extent Asp74 and Tyr341, might

engage polar interactions with the pyridinium nitrogens of the

inhibitor. On the other side, the 6,7-dimethoxy-1,2,3,4-tetrahy-

droisoquinoline moiety was likely involved in hydrophobic

interactions with the Trp286 of the PAS. The results acquired

via docking simulations met those found in a previous study

where the interactions of the 5,6-dimethoxy-indan-1-one moiety

of donepezil in the peripheral binding site has been analyzed.³⁴

Incidentally, the docking of 56 in the TcAChE (PDB code 1EVE)

did not return any signiﬁcantly diﬀerent binding mode (see

Figure S3 in the Supporting Information).

Interestingly, the molecular docking of heterodimer 56 into

BChE envisaged a less favorable GoldScore ﬁtness value (ꢀ48

kJ/mol) for the top-scored docking pose presumably because of

the lack of an eﬃcient binding at the peripheral binding site in

BChE. On the other hand, a detrimental eﬀect on the aﬃnity

toward BChE might also be ascribed to the occurrence of some

important residue mutations in the mid-gorge.³⁵And indeed, two

critical residues, Tyr124 and Tyr337 in hAChE were replaced

with Gln119 and Ala328, respectively, in hBChE, and this might

preclude the occurrence of the polar interactions in the mid-

gorge previously pointed out. Molecular docking simulations

nicely suggested a face-to-face interaction in the primary binding

site between the Trp82 (BChE numbering) and inhibitor 56 as

reported in a recent paper for other heterodimeric inhibitors^11a

(see Figure S1 in the Supporting Information). Our prospective

docking study ﬁnally addressed the complex formed between

inhibitor 56 and ChoK. Basically, docking simulations predicted

a prevalent binding mode showing the piperidinyl ring embedded

in a hydrophobic cavity of the catalytic site constituted by

Tyr333, Tyr354, Tyr440, Trp420, Trp423, Phe362, and

Phe435 (ChoK numbering).

For the sake of completeness, a series of molecular dynamic

simulations were undertaken to retrospectively assess the struc-

tural integrity of the binding mode of 56 and, more importantly,

to gain insight on the molecular reasons behind its selectivity

toward AChE. To this end, two parallel 5 ns MD simulations

were run on the complexes of 56 with hAChE and hBChE.

The top-scored docking poses were thus used as a convenient

starting geometry for determining the structural variation over

time of the AChEꢀ and BChEꢀligand complexes. Both systems

were equilibrated according to standard MD protocols.³⁶Satis-

factorily, the visual inspection of the molecular frames obtained

via MD unequivocally conﬁrmed the correctness of the binding

topology preliminarily predicted by docking in the stage of

rational design. As far as the hAChEꢀligand complex is con-

cerned, the fundamental face-to-face interaction engaged with

Trp86 at the primary binding site was very well conserved during

the simulation, while the Trp286 at the PAS disclosed a weaker

propensity to maintain the πꢀπ stacking interaction with the

6,7-dimethoxy-1,2,3,4-tetrahydroisoquinoline of 56 observed in

the docking study. And in fact, MD simulations showed that the

average distance between the centroid of Trp86 and that of

piperidinyl ring was equal to 4.69 ((0.26) Å, and this was

consistent with the hypothesis of a stable hydrophobic sandwich-

like interaction. On the other hand, the larger average distance of

8.85 ((1.01) Å, measured between the centroid of Trp286 and

that of the 6,7-dimethoxy-1,2,3,4-tetrahydroisoquinoline, indi-

cated only transient hydrophobic contacts at the PAS. Despite

the likely disruption of such interaction at the PAS, the hetero-

dimer 56 never did visit any alternative binding topology along

2638

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

Figure 4. The left-hand side plot (a) shows the variation of the distance between the phenolic oxygen atoms of nonconserved Tyr337 and Tyr124 of

hAChE (equivalent to Ala328 and Gln119 of hBChE) and the closest pyridinium nitrogen of compound 56 (the same color code of Figure 3 has been

adopted). The right-hand side plot (b) shows the variation of the distances between the phenolic oxygen atoms of the conserved residue Tyr332 of

hBChE (equivalent to Tyr 341 of hAChE) and both the pyridinium nitrogen atoms of compound 56. In particular, the red and blue proﬁles represent the

distances to the charged nitrogen atoms of the pyridinium rings joined to the 6,7-dimethoxy-1,2,3,4-tetrahydroisoquinoline and the piperidine moieties,

respectively. All the measures sampled during the 5 ns MD simulation are referenced to the two top-scored poses of 56 resulting from docking

simulations into the binding sites of the two human ChEs.

the entire time of MD simulation. In this regard, nonconserved

Tyr337 and Tyr124 residues of hAChE played an important

role in stabilizing the binding mode of 56 by establishing

ionꢀdipole interactions with both of the charged nitrogen

atoms of the two pyridinium rings of 56. As it can be inferred

from Figure 4a, such interactions were maintained along all the

MD, as the distances between the phenolic oxygen atoms of

Tyr337 and Tyr124 and the corresponding closest charged

nitrogen atom of the pyridinium rings changed within a very

narrow range. More precisely, those distances were equal to

4.82 ((0.42) Å and to 4.44 ((0.35) Å, respectively. For a

better understanding, the Cartesian positions of the phenolic

oxygen atoms of the two tyrosine residues, sampled during the

MD, were conveniently represented as ﬁtting spheres, in

Figure 3. Actually, the signiﬁcant role of nonconserved mid-

gorge residues (i.e., Tyr337 and Tyr124) of hAChE and their

relevance in stabilizing the binding of 56 through a likely

bidentate ionꢀdipole interaction with both the charged nitro-

gen atoms of the pyridinium rings were already observed in the

docking studies. And this was observed despite the well-known

Figure 5. LineweaverꢀBurk plot of inhibition kinetics of 56: reci-

preference of GoldScore scoring function to award hydropho-

bic rather than polar interactions.³⁷On the other hand, the lack

procals of enzyme activity (bovine AChE) vs reciprocals of substrate

(S-acetylthiocholine) concentration in the presence of diﬀerent con-

centrations (0ꢀ30 nM) of inhibitor 56.

of a compact tyrosine-based network in the mid-gorge of BChE

reduced by far the likeness of potential ionꢀdipole interac-

tions. And in fact, MD simulation carried out on BChE

with the recognition of the pivotal role of nonconserved

residues of mid-gorge of hAChE in the binding of inhibitor 56.

Heterodimer 56 was prepared (see Scheme 6) and tested as

AChE, BChE, and ChoK inhibitors. Very gratifyingly, it exhibited

an outstanding AChE aﬃnity (IC₅₀= 15 nM) compared to BChE

(IC₅₀= 752 nM), thus showing a high AChE selectivity (SI = 50)

and being the most potent and selective AChE inhibitor of the

whole series of examined molecules. Moreover, it showed very

poor ChoK inhibitory potency (IC₅₀= 8.43 μΜ), resulting in an

excellent selectivity index (SI = 562).

assigned a secondary role to the conserved residue Tyr332 of

BChE (equivalent to Tyr341 of hAChE). Such a residue acted

like a switch, swapping between the two competing charged

nitrogen atoms of the two pyridinium rings, in an attempt to

ﬁnd an optimal distance with at least one of them. Such a

behavior is described in Figure 4b, which shows how the

chance of ﬁnding an appropriate distance for establishing a

ionꢀdipole interaction with one charged nitrogen atom was

made at the expense of the other one. Retrospectively, MD

simulations reﬁned the preliminary hypothesis emerged in the

step of rational molecular design via docking by furnishing a

more convincing interpretation of the molecular selectivity

The mechanism of AChE inhibition by 56 was assessed by

means of a kinetic study whose results are reported in Figure 5.

2639

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

This study accounted for a mixed-type model of inhibition, in

accordance with a likely dual binding site mode of interaction,³⁴

with an apparent inhibition constant K_iequal to 21 ( 3 nM.

Finally, a further study was undertaken on a selected number of

inhibitors to determine their inhibitory activity on the human

isoforms of both ChEs and on the Aβ_1ꢀ40self-aggregation. Along

with 56, three compounds were chosen as representatives of the

diﬀerent structural classes presented in Tables 1ꢀ4. Taking into

account the diversity of the linkers, we selected the very potent

AChE inhibitors 14 and 19, bearing respectively a p-xylenyl and a

3,3⁰-biphenyl spacer, and the highly BChE-selective compound 47.

IC₅₀on human recombinant acetylcholinesterase and human

serum butyrylcholinesterase and the percentage of inhibition of

the Aβ_1ꢀ40self-aggregation are reported in Table 5, along with the

activities exhibited by two reference compounds, which are done-

pezil (AChE-selective) and tacrine (BChE-selective). Inhibition

data of the two reference compounds are in agreement with

literature data.^38,39The examined QASs exhibited an inhibitory

potency toward the two human ChEs comparable to that observed

for the bovine AChE and serum horse BChE. However, hAChE

aﬃnity and hAChE over hBChE selectivity of QAS 56 slightly

decreased. The Aβ_1ꢀ40antiaggregating activities of compounds in

Table 5 at 100 μM were very low, the most active compound being

the QAS 19 with 30% ((2%) of inhibition.

power to 500 W, and heating ramp times to 5 min. The purity of all the

intermediates, checked by H NMR and HPLC, was always better

1

than 95%. Flash chromatographic separations were performed on

Biotage SP1 purification system using flash cartridges prepacked with

KP-Sil 32ꢀ63 μm, 60 Å silica. Except for those regarding final QASs

compounds, reactions were routinely checked by TLC using Merck

Kieselgel 60 F₂₅₄aluminum plates and visualized by UV light or

iodine. Melting points were determined by the capillary method on a

Stuart Scientific SMP3 electrothermal melting point apparatus and are

uncorrected. Elemental analyses were performed on the EuroEA 3000

analyzer only on the final compounds tested as ChEs inhibitors. The

measured values for C, H, and N agreed to within (0.40% of the

theoretical values. Nuclear magnetic resonance spectra were recorded

at 300 MHz on a Varian Mercury 300 instrument or 400 MHz ¹H and

100 MHz ¹³C NMR Bruker ARX 400 or 300 MHz ¹H and 75 MHz ¹³

C

NMR Bruker AMX-300 spectrometers or 400 MHz ¹H and 100 MHz

¹³C NMR Varian NMR-System-TM 400 or 300 MHz ¹H and 75 MHz

¹³C NMR Varian Inova spectrometers at ambient temperature. Chem-

ical shifts (δ) are quoted in parts per million (ppm) and are referenced

to the residual solvent peak. The following abbreviations were used: s

(singlet), d (doublet), dd (doublet of doublet), t (triplet), pst

(pseudotriplet), psq (pseudoquartet), q (quadruplet), m (multiplet), ddd

(double doublet of doublet). Signals due to OH protons were located by

deuterium exchange with D₂O. High-resolution liquid secondary ion mass

spectra (HR LSIMS) were carried out on a VG AutoSpec Q high-resolution

mass spectrometer (Fisons Instruments). All compounds were dried at

40 ꢀC and 0.1 mmHg for 24 h, but many held on tenaciously to water which

appear to be solvates. Compounds 7,¹⁴8,¹⁴11,¹⁴13,¹⁴14,¹⁴17ꢀ19,¹⁴

23ꢀ26,¹⁴29,¹⁴30,¹⁴31,¹⁶32,¹⁶33ꢀ35,¹³37,¹³38ꢀ40,¹⁶45,¹⁵46,¹⁷51,¹⁸

52,¹⁸53,¹⁹54,¹⁹and 55²⁰were synthesized as previously reported. trans-1,2-

Bis(hydroxymethyl)cyclopropane,²²trans-1,2-bis(bromomethyl)cyclopro-

pane,²¹3,3⁰-bis(bromomethyl)biphenyl,²³4,4⁰-bis[(bromomethyl)biben-

zyl,²⁴trans,trans-1,4-diphenyl-1,3-butadiene,²⁴1,4-bis[4-(bromo-

methyl)phenyl]butane,²⁴trans-4,4⁰-bis(hydroxymethyl)stilbene,²⁶and

trans-4,4⁰-bis(bromomethyl)stilbene²⁵were synthesized according to

literature procedures. 4-(Morpholino)pyridine, 4-(dimethylamino)-

pyridine, 4-(pyrrolidino)pyridine, 4-aminoquinoline, biphenyl-4,4⁰-di-

carboxylic acid, benzyl bromide, 4-methylbenzyl bromide, 1,3-bis(bromo-

methyl)benzene, 1,4-bis(bromomethyl)benzene, quinuclidine, and

3-quinuclidinol were purchased from commercial sources (Sigma-

Aldrich, Europe).

’ CONCLUSION

The work described in the present paper was focused on the

synthesis and biological evaluation as AChE and BChE inhibitors of

a series of mono-, bis-, and tris-QASs. The application of molecular

dimerization and hybridization approaches has laid the groundwork

for the discovery of very active bis-QASs. Following this strategy, by

connecting two cationic heads with a proper substitution pattern

through a spacer of varying lipophilicity, length, and ﬂexibility, we

succeeded in obtaining a large number of compounds active in the

nanomolar range, mainly as AChE inhibitors. A thorough analysis

of the substituent eﬀect in the cationic heads pointed out that the

introduction of powerful electron-releasing groups modulated the

aﬃnity likely through a reinforcement of cationꢀπ interaction

and/or more favorable solvation/desolvation eﬀects.¹³Careful and

stepwise structural modiﬁcations were made aiming at an improve-

ment of the inhibitory potency toward AChE, the AChE over

BChE selectivity, and even more importantly the AChE over ChoK

selectivity. Very satisfactorily, our docking-assisted molecular de-

sign was very successful, leading to the discovery of compound 56,

the most active and selective AChE inhibitor of the whole examined

series of compounds (IC₅₀= 15 nM, SI = 50). Moreover,

compound 56 exhibited a quite poor ChoK inhibitory activity

(IC₅₀= 8.43 ( 0.36 μM) resulting in an excellent selectivity index

(SI = 562; SI is the IC₅₀(ChoK)/IC₅₀(AChE) aﬃnity ratio), and

this was another important goal met in our study. Finally, the

relevant AChE inhibitory potency, the favorable selectivity proﬁle,

and the high aqueous solubility and lipophilicity should guarantee

for heterodimer 56 a prevalent activity at the peripheral level,

strongly required in the therapy of MG and NMB.

Biphenyl-4,4⁰-dicarboxylic Acid Diethyl Ester (1a). Com-

mercially available biphenyl-4,4⁰-dicarboxylic acid (9.7 g, 40.0 mmol)

was suspended in ethanol (100 mL), and thionyl chloride (15 mL, 200

mmol) was slowly added. The reaction mixture was stirred at room

temperature overnight and then refluxed for 24 h. After the mixture was

cooled to room temperature, the solvent was evaporated under reduced

pressure and the crude mixture was diluted with ethyl acetate (500 mL)

and extracted with NaOH 2 N (3 ꢁ 100 mL). The organic phase was

washed with brine (100 mL), dried over Na₂SO₄, concentrated

under vacuum, and finally purified by flash chromatography (gradient

eluent, ethyl acetate in n-hexane 0% f 10%), thus obtaining 1a as a

white solid. Yield: 52%. ¹H NMR (300 MHz, DMSO-d₆) δ: 1.33

(t, J = 7.2 Hz, 6H), 4.33 (q, J = 7.2 Hz, 4H), 7.88 (d, J = 8.5 Hz, 4H),

8.05 (d, J = 8.5 Hz, 4H).

(4⁰-Hydroxymethylbiphenyl-4-yl)methanol (1b). Ester 1a

(6.0 g, 20 mmol) was dissolved in dry THF (80 mL), and LiAlH₄(3.0

g, 80 mmol) was added portionwise in 1 h. The mixture was refluxed for

7 h and then cooled to room temperature. Excess reducing agent was

quenched by careful addition of water while cooling through an external

ice bath. The mixture was kept under stirring for 15 min and extracted

with chloroform (3 ꢁ 200 mL). The organic layers were collected, dried

over Na₂SO₄, and the solvent was removed under rotary evaporation.

Final purification through flash chromatography (gradient eluent,

’ EXPERIMENTAL SECTION

Chemistry. Starting materials, reagents, and analytical grade sol-

vents were purchased from Sigma-Aldrich (Europe). Microwave reac-

tions were performed in a Milestone MicroSynth apparatus, with

setting of temperature and hold times, fixing of maximum irradiation

2640

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

methanol in dichloromethane 0% f 5%) yielded intermediate 1b as an

off-white solid. Yield: 40%. ¹H NMR (300 MHz, DMSO-d₆) δ: 4.50 (d,

J = 5.5 Hz, 4H), 5.18 (t, J = 5.5 Hz, 2H, dis. with D₂O), 7.36 (d, J = 8.3

Hz, 4H), 7.58 (d, J = 8.3 Hz, 4H).

10 min. Then a volume of 2.0 mL of dry THF was introduced, followed

by the dropwise addition of a solution of iodine (catalytic amount) in dry

THF (4.0 mL) and finally of cyclohexyl bromide (0.99 mL, 8.0 mmol).

When the dark brown color of the solution disappeared and turned to a

gray suspension, the mixture was refluxed for 1 h. After the mixture was

cooled to room temperature, the Grignard reagent was added dropwise

through a syringe under an argon atmosphere to a solution of

4-chloropyridine¹³(0.60 g, 5.3 mmol), Fe(acac)₃(0.19 g, 0.53 mmol),

and NMP (0.85 mL) in dry THF (11.0 mL) while the mixture cooled to

0 ꢀC. The reaction mixture turned red and then was stirred at room

temperature. After 15 h, water (150 mL) was added and the mixture was

extracted with diethyl ether (3 ꢁ 60 mL). The organic layers were

collected, dried over Na₂SO₄, and then the mixture was kept under an

argon stream until evaporation of the solvent and disappearance of the

spot of low-boiling 4-chloropyridine from the TLC control. In this way,

4-cyclohexylpyridine 1g was obtained as a dark yellow oil as confirmed

by ¹H NMR analysis and with a purity higher than 95% as checked by

HPLC. Yield: 41%. ¹H NMR (300 MHz, DMSO-d₆) δ: 0.73ꢀ0.83 (m,

1H), 1.15ꢀ1.27 (m, 2H), 1.32ꢀ1.40 (m, 4H), 1.66ꢀ1.71 (m, 1H),

1.75ꢀ1.78 (m, 2H), 2.42ꢀ2.51 (m, 1H), 7.22 (d, J = 6.1 Hz, 2H), 8.42

(d, J = 6.1 Hz, 2H).

4-Phenylpyridine (1h). 4-Bromopyridine hydrochloride (0.39 g,

2.0 mmol) and phenylboronic acid (0.29 g, 2.4 mmol) were suspended

in a dimethoxyethane/water 4/1 (v/v) mixture (15 mL). Cs₂CO₃was

added (1.3 g, 4.0 mmol) followed by Pd(PPh₃)₄(0.23 g, 0.2 mmol). The

reaction mixture was heated to 80 ꢀC for 4 h, and then DME was

evaporated under vacuum. The mixture was partitioned between water

(50 mL) and ether (200 mL), and then the organic phase was washed

with a saturated aqueous solution of Na₂CO₃(3 ꢁ 80 mL), dried over

Na₂SO₄, and concentrated to dryness by rotary evaporation. Purification

by flash chromatography (gradient eluent, ethyl acetate in n-hexane 0%

f 40%) yielded 1h as a white solid. Yield: 93%. ¹H NMR (300 MHz,

DMSO-d₆) δ: 7.46ꢀ7.54 (m, 3H), 7.70 (d, J = 6.3 Hz, 2H), 7.77ꢀ7.81

(m, 2H), 8.63 (d, J = 6.3 Hz, 2H).

4,4⁰-Bis(bromomethyl)biphenyl (1c). Carbon tetrabromide

(2.7 g, 8.3 mmol) was added to a solution of 1b (1.6 g, 7.5 mmol) in

dry CH₂Cl₂(50 mL). Triphenylphosphine (2.4 g, 9.0 mmol), previously

dissolved in dry CH₂Cl₂(20 mL), was added dropwise at 0 ꢀC. The

reaction mixture was slowly warmed to room temperature and then

stirred for 6 h. After evaporation of the solvent, purification through flash

chromatography (gradient eluent, ethyl acetate in n-hexane 0% f 10%)

furnished derivative 1c. Yield: 67%. ¹H NMR (300 MHz, DMSO-d₆) δ:

4.75 (s, 4H), 7.52 (d, J = 8.3 Hz, 4H), 7.66 (d, J = 8.3 Hz, 4H).

4-Piperidin-1-ylpyridine (1d). A Pyrex vessel was charged with a

magnetic stirring bar, and then 4-chloropyridine hydrochloride (0.60 g,

4.0 mmol) was added and suspended in dry acetonitrile (10 mL). N,

N-Diisopropylethylamine (2.1 mL, 12 mmol) was added followed by

piperidine (0.40 mL, 4.0 mmol). The reactor was placed in a microwave

apparatus and irradiated at 160 ꢀC for 1 h. After cooling to room

temperature, the mixture was partitioned between ethyl acetate (60 mL)

and water (100 mL) and then extracted with ethyl acetate (3 ꢁ 60 mL).

The organic layers were collected, washed with brine (50 mL), dried

over Na₂SO₄, and then the solvent was removed under reduced

pressure. Ether (5 mL) was added to the resulting crude oil, and the

inorganic precipitate was filtered off. Evaporation of the solvent gave 1d

as a dark yellow solid of satisfactory purity, as confirmed by HPLC

analysis. Yield: 82%. ¹H NMR (300 MHz, DMSO-d₆þ D₂O)

δ: 1.48ꢀ1.61 (m, 6H), 3.29ꢀ3.32 (m, 4H), 6.74ꢀ6.76 (m, 2H),

8.07ꢀ8.10 (m, 2H).

2-Pyridin-4-yl-1,2,3,4-tetrahydroisoquinoline (1e). In a Pyr-

ex vessel, charged with a magnetic stirring bar, a suspension of 1,2,3,4-

tetrahydroisoquinoline hydrochloride (0.51 g, 3.0 mmol) and 4-chlor-

opyridine hydrochloride (0.50 g, 3.3 mmol) in dry acetonitrile (10 mL)

was prepared. N,N-Diisopropylethylamine (1.7 mL, 10 mmol) was

added, and the reactor was introduced in a microwave apparatus and

irradiated for 1 h at 160 ꢀC. After cooling to room temperature, the

mixture was poured onto ice (100 mL) and then extracted with ethyl

acetate (3 ꢁ 60 mL). The organic layers were collected, washed with

brine (50 mL), and dried over Na₂SO₄. The organic solution was then

concentrated under reduced pressure and purified through flash chro-

matography (gradient eluent, methanol in acetone 0% f 5%), yielding

1e as a yellow solid. Yield: 74%. ¹H NMR (300 MHz, DMSO-d₆) δ: 2.89

(t, J = 6.1 Hz, 2H), 3.59 (t, J = 6.1 Hz, 2H), 4.48 (s, 2H), 6.81ꢀ6.84 (m,

2H), 7.17ꢀ7.25 (m, 4H), 8.13ꢀ8.15 (m, 2H).

General Procedure for the Preparation of Mono- and Bis-

pyridinium Bromides 2ꢀ6, 9, 10, 12, 15, 16, 20ꢀ22, 27, 28,

36, 41ꢀ44, and 47ꢀ50. A solution of the appropriate bromomethyl

or bis(bromomethyl) linker and the appropriate 4-substituted pyridine

(in a 1:1 or a 1:2 molar ratios) was heated to 100 ꢀC in a sealed tube in

2-butanone (4.0 mL) from 15 to 192 h on a 0.2 mmol scale. After

filtration and thorough washing with 2-butanone, ethyl acetate, and

diethyl ether, the solid was purified by recrystallization from ethanol or

methanol after adding diethyl ether to turbidity.

1-Benzyl-4-(pyrrolidin-1-yl)pyridinium Bromide (2). White

solid. Yield: 27%. Mp: 227ꢀ229 ꢀC. ¹H NMR (300 MHz, CD₃OD) δ:

2.11 (m, 4H), 3.55 (m, 4H), 5.37 (s, 2H), 6.87 (d, J = 7.8 Hz, 2H), 7.41

(m, 5H), 8.21 (d, J = 7.8 Hz, 2H). HR LSIMS (m/z) calcd for C₁₆H₁₉N₂

6,7-Dimethoxy-2-pyridin-4-yl-1,2,3,4-tetrahydroisoqui-

noline (1f). In a Pyrex vessel, charged with a magnetic stirring bar, a

suspension of 6,7-dimethoxy-1,2,3,4-tetrahydroisoquinoline hydro-

chloride (0.69 g, 3.0 mmol) and 4-chloropyridine hydrochloride (0.50

g, 3.3 mmol) in dry acetonitrile (10 mL) was prepared. N,N-Diisopro-

pylethylamine (1.7 mL, 10 mmol) was added, and the reactor was

introduced in a microwave apparatus and irradiated for 1 h at 160 ꢀC.

After cooling to room temperature, the mixture was poured onto

crushed ice (150 g) and the resulting solid was filtered and washed with

water, thus obtaining 1f as an off-white solid. Yield: 78%. ¹H NMR (300

MHz, DMSO-d₆) δ: 2.79 (t, J = 5.8 Hz, 2H), 3.57 (t, J = 5.8 Hz, 2H),

3.70 (s, 3H), 3.71 (s, 3H), 4.39 (s, 2H), 6.76ꢀ6.82 (m, 4H), 8.13 (d, J =

5.5 Hz, 2H).

4-Cyclohexylpyridine (1g). Magnesium turnings (0.21 g, 8.8

mmol) were suspended in dry diethyl ether (3.0 mL) in a flame-dried

round-bottom flask equipped with a reflux condenser. While the mixture

was being stirred, the solvent was evaporated under reduced pressure.

The system was purged under argon stream for 10 min. This procedure

was repeated twice, and then the flask was gently heated under vacuum

for 10 min. The system was again purged through an argon stream for

(M-Br)^þ239.1543; found 239.1545. Anal. (C₁₆H₁₉BrN₂0.5H₂O)

3

C, H, N.

1-(4-Methylbenzyl)-4-(pyrrolidin-1-yl)pyridinium Bro-

mide (3). White solid. Yield: 87%. Mp: 188ꢀ190 ꢀC. ¹H NMR

(400 MHz, CD₃OD) δ: 2.11 (psq, J = 6.7 Hz, 4H), 2.33 (s, 3H), 3.54

(pst, J = 6.7 Hz, 4H), 5.31 (s, 2H), 6.85 (d, J = 7.5 Hz, 2H), 7.23 (d, J =

8.1 Hz, 2H), 7.28 (d, J = 8.1 Hz, 2H), 8.20 (d, J = 7.5 Hz, 2H). HR

LSIMS (m/z) calcd for C₁₇H₂₁N₂(M ꢀ Br)^þ253.1705; found

253.1698. Anal. (C₁₇H₂₁BrN₂0.7H₂O) C, H, N.

3

4-Amino-1-(4-methylbenzyl)quinolinium Bromide (4).

White solid. Yield: 71%. Mp: 261ꢀ263 ꢀC. ¹H NMR (400 MHz,

CD₃OD) δ: 2.29 (s, 3H), 5.75 (s, 2H), 6.92 (d, J = 7.3 Hz, 1H), 7.13

(d, J = 8.2 Hz, 2H), 7.17 (d, J = 8.2 Hz, 2H), 7.68 (pst, J = 7.7 Hz, 1H),

7.90 (pst, J₁= 1.2 Hz, J₂= 6.9 Hz, J₃= 8.6 Hz, 1H), 7.98 (d, J = 8.8 Hz,

1H), 8.40 (d, J = 8.4 Hz, 1H), 8.58 (d, J = 7.3 Hz, 1H). HR LSIMS (m/z)

calcd for C₁₇H₁₇N₂(M ꢀ Br)^þ249.1392; found 249.1398. Anal.

(C₁₇H₁₇BrN₂0.2H₂O) C, H, N.

3

2641

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

1,1⁰-[trans-Cyclopropane-1,2-diylbis(methylene)]bis(4-

(dimethylamino)pyridinium Dibromide (5). White solid. Yield:

62%. Mp: 284ꢀ5 ꢀC. ¹H NMR (300 MHz, CD₃OD) δ: 0.98 (t, J = 6.7

Hz, 2H), 1.66ꢀ1.68 (m, 2H), 3.26 (s, 12H), 3.93 (dd, J_vic= 8.1 Hz,

J_gem= 14.1 Hz, 2H), 4.29 (dd, J_vic= 5.8 Hz, J_gem= 14.1 Hz, 2H), 6.65 (d,

J = 7.7 Hz, 4H), 8.19 (d, J = 7.7 Hz, 4H). HR LSIMS (m/z) calcd for

C₁₉H₂₈Br₂N₄Na (M þ Na)^þ493.0578; found 493.0580. Anal.

(s, 2H), 7.22 (d, J = 6.7 Hz, 4H), 7.38ꢀ7.41 (m, 2H), 7.49ꢀ7.55 (m, 2H),

7.67ꢀ7.69 (m, 2H), 7.78 (s, 2H), 8.36 (d, J = 6.7 Hz, 4H). Anal.

(C₄₆H₄₈Br₂N₄O₄2.5H₂O) C, H, N.

3

1,1⁰-[Biphenyl-4,4⁰-diylbis(methylene)]bis[4-(3,4-dihy-

droisoquinolin-2(1H)-yl)pyridinium] Dibromide (27). Yellow

solid. Yield: 62%. Mp: 88ꢀ91 ꢀC (dec). ¹H NMR (300 MHz, CD₃OD)

δ: 3.06 (t, J = 6.1 Hz, 4H), 3.83 (t, J = 6.1 Hz, 4H), 4.77 (s, 4H), 5.44 (s,

4H), 7.26ꢀ7.29 (m, 12H), 7.47ꢀ7.49 (m, 4H), 7.68ꢀ7.71 (m, 4H),

(C₁₉H₂₈Br₂N₄1.5H₂O) C, H, N.

3

1,1⁰-[1,3-Phenylenebis(methylene)]bis[4-(dimethylamino)

pyridinium] Dibromide (6). Off-white solid. Yield: 64%. Mp >

300 ꢀC. ¹H NMR (300 MHz, CD₃OD) δ: 3.26 (s, 12H), 5.43 (s, 4H),

7.04 (d, J = 7.7 Hz, 4H), 7.52ꢀ7.39 (m, 3H), 7.57 (s, 1H), 8.28 (d, J = 7.7

Hz, 4H). HR LSIMS (m/z) calcd for C₂₂H₂₈Br₂N₄Na (M þ Na)^þ

8.31 (d, J = 7.2 Hz, 4H). Anal. (C₄₂H₄₀Br₂N₄2.5H₂O) C, H, N.

3

1,1⁰-[Biphenyl-4,4⁰-diylbis(methylene)]bis[4-(6,7-di-

methoxy-3,4-dihydroisoquinolin-2(1H)-yl)pyridinium] Di-

bromide (28). Yellow solid. Yield: 43%. Mp: 224ꢀ225 ꢀC (dec). ¹H

NMR (300 MHz, CD₃OD) δ: 2.98 (t, J = 5.8 Hz, 4H), 3.81ꢀ3.84 (m,

16H), 4.70 (s, 4H), 5.43 (s, 4H), 6.85 (s, 2H), 6.87 (s, 2H), 7.23 (d, J =

6.9 Hz, 4H), 7.48 (d, J = 8.0 Hz, 4H), 7.70 (d, J = 8.0 Hz, 4H), 8.30 (d, J =

529.0578; found 529.0576. Anal. (C₂₂H₂₈Br₂N₄0.5H₂O) C, H, N.

3

1,1⁰-[1,3-Phenylenebis(methylene)]bis[4-(3,4-dihydroiso-

quinolin-2(1H)-yl)pyridinium] Dibromide (9). Yellow solid.

6.9 Hz, 4H). Anal. (C₄₆H₄₈Br₂N₄O₄2H₂O) C, H, N.

3

1

1,1⁰-[4,4⁰-(Ethane-1,2-diyl)bis(4,1-phenylene)](methylene)bis-

(4-morpholinopyridinium) Dibromide (36). Off-white solid. Yield:

100%. Mp: 180ꢀ182 ꢀC. ¹H NMR (300 MHz, CD₃OD) δ: 2.90 (s, 4H),

3.70 (t, J = 4.5 Hz, 8H), 3.80 (t, J = 4.5 Hz, 8H), 5.36 (s, 4H), 7.19 (d, J = 7.9

Hz, 4H), 7.23 (d, J= 8.2 Hz, 4H), 7.32 (d, J= 8.2 Hz, 4H), 8.29 (d, J=7.9Hz,

4H). HR LSIMS (m/z) calcd for C₃₄H₄₀BrN₄O₂(M ꢀ Br)^þ615.2335;

Yield: 69%. Mp: 128ꢀ130 ꢀC (dec). H NMR (300 MHz, CD₃OD)

δ: 3.06 (t, J = 6.1 Hz, 4H), 3.82 (t, J = 6.1 Hz, 4H), 4.77 (s, 4H), 5.42

(s, 4H), 7.20ꢀ7.26 (m, 12H), 7.37ꢀ7.52 (m, 4H), 8.30 (d, J = 7.2 Hz,

4H). Anal. (C₃₆H₃₆Br₂N₄2H₂O) C, H, N.

3

1,1⁰-[1,3-Phenylenebis(methylene)]bis[4-(6,7-dimethoxy-

3,4-dihydroisoquinolin-2(1H)-yl)pyridinium] Dibromide (10).

Yellow solid. Yield: 51%. Mp: 261ꢀ263 ꢀC. ¹HNMR(300 MHz, CD₃OD)

δ: 2.95 (t, J = 6.1 Hz, 4H), 3.78 (t, J = 6.1 Hz, 4H), 3.80 (s, 12H), 4.66 (s,

4H), 5.42 (s, 4H), 6.83 (s, 2H), 6.87 (s, 2H), 7.21 (d, J = 6.8 Hz, 4H),

found 615.2335. Anal. (C₃₄H₄₀Br₂N₄O₂1.5H₂O) C, H, N.

3

1,1⁰-[4,4⁰-(Butane-1,4-diyl)bis(1,4-phenylene)]bis(methy-

lene)bis(4-(dimethylamino)pyridinium) Dibromide (41). Off-

white solid. Yield: 89%. Mp: 161ꢀ163 ꢀC. ¹H NMR (300 MHz,

CD₃OD) δ: 1.70 (m, 4H), 2.72 (t, 4H), 3.33 (s, 12H), 5.42 (s, 4H),

7.09 (d, J = 7.9 Hz, 4H), 7.31 (d, J = 8.2 Hz, 4H), 7.39 (d, J = 8.2 Hz, 4H),

8.32 (d, J = 7.9 Hz, 4H). HR LSIMS (m/z) calcd for C₃₂H₄₀BrN₄

7.42ꢀ7.50 (m, 4H), 8.29 (d, J = 6.8 Hz, 4H). Anal. (C₄₀H₄₄Br₂N₄O₄

3

1.5H₂O) C, H, N.

1,1⁰-[1,4-Phenylenebis(methylene)]bis[4-(dimethylamino)

pyridinium] Dibromide (12). White solid. Yield: 51%. Mp >

300 ꢀC.¹H NMR (400 MHz, CD₃OD) δ: 3.25 (s, 12H), 5.40 (s, 4H),

7.01 (d, J = 7.9 Hz, 4H), 7.46 (s, 4H), 8.23 (d, J = 7.9 Hz, 4H). HR

LSIMS (m/z) calcd for C₂₂H₂₈BrN₄(M ꢀ Br)^þ427.1497; found

427.1494. Anal. (C₂₂H₂₈Br₂N₄) C, H, N.

(M ꢀ Br)^þ559.2431; found 559.2429. Anal. (C₃₂H₄₀Br₂N₄1.2H₂O)

3

C, H, N.

1,1⁰-[4,4⁰-(Butane-1,4-diyl)bis(1,4-phenylene)]bis(methy-

lene)bis[4-(pyrrolidin-1-yl)pyridinium] Dibromide (42). Off-

white solid. Yield: 67%. Mp: 107ꢀ109 ꢀC. ¹H NMR (300 MHz,

CD₃OD) δ: 1.61 (m, 4H), 2.11 (m, 8H), 2.63 (m, 4H), 3.55 (m,

8H), 5.32 (s, 4H), 6.86 (d, J = 7.8 Hz, 4H), 7.22 (d, J = 8.2 Hz, 4H), 7.30

(d, J = 8.2 Hz, 4H), 8.20 (d, J = 7.8 Hz, 4H). HR LSIMS (m/z) calcd for

1,1⁰-[1,4-Phenylenebis(methylene)]bis[4-(3,4-dihydroiso-

quinolin-2(1H)-yl)pyridinium] Dibromide (15). Off-white solid.

Yield: 55%. Mp: 78ꢀ80 ꢀC (dec). ¹H NMR (300 MHz, D₂O) δ: 2.34 (t,

J = 5.8 Hz, 4H), 3.03 (t, J = 5.8 Hz, 4H), 3.76 (s, 4H), 5.17 (s, 4H), 6.28

(d, J = 7.4 Hz, 4H), 6.48ꢀ6.53 (m, 4H), 6.64ꢀ6.67 (m, 2H), 6.71ꢀ6.76

C₃₆H₄₄BrN₄(M

ꢀ

Br)^þ611.2744; found 611.2747. Anal.

(m, 2H), 7.52 (s, 4H), 7.96 (d,

J

=

7.4 Hz, 4H). Anal.

(C₃₆H₄₄Br₂N₄2.5H₂O) C, H, N.

3

1,1⁰-[4,4⁰-(Butane-1,4-diyl)bis(1,4-phenylene)]bis(methy-

lene)bis[4-(piperidin-1-yl)pyridinium] Dibromide (43). White

solid. Yield: 60%. Mp: 259ꢀ261 ꢀC. ¹H NMR (300 MHz, CD₃OD) δ:

1.65ꢀ1.81 (m, 16H), 2.67 (m, 4H), 3.71ꢀ3.75 (m, 8H), 5.33 (s, 4H),

7.15 (d, J = 7.9 Hz, 4H), 7.25 (d, J = 8.4 Hz, 4H), 7.33 (d, J = 8.4 Hz, 4H),

8.20 (d, J = 7.5 Hz, 4H). HR LSIMS (m/z) calcd for C₃₈H₄₇N₄(M ꢀ

(C₃₆H₃₆Br₂N₄2H₂O) C, H, N.

3

1,1⁰-[1,4-Phenylenebis(methylene)]bis[4-(6,7-dimethoxy-

3,4-dihydroisoquinolin-2(1H)-yl)pyridinium] Dibromide (16).

1

Off-white solid. Yield: 68%. Mp: 157ꢀ159 ꢀC (dec). H NMR (300

MHz, CD₃OD) δ: 2.97 (t, J = 5.8 Hz, 4H), 3.81 (t, J = 5.8 Hz, 4H), 3.82 (s,

6H), 3.83 (s, 6H), 4.69 (s, 4H), 5.41 (s, 4H), 6.85 (s, 2H), 6.87 (s, 2H), 7.21

(d, J = 6.6 Hz, 4H), 7.46 (s, 4H), 8.27 (d, J = 6.6 Hz, 4H). Anal.

HBr ꢀ Br)^þ559.3801; found 559.3803. Anal. C₃₈H₄₈Br₂N₄2H₂O)

3

(C₄₀H₄₄Br₂N₄O₄2H₂O) C, H, N.

C, H, N.

3

1,1⁰-[Biphenyl-3,3⁰-diylbis(methylene)]bis(4-cyclohexyl-

(E)-1,1⁰-[4,4⁰-(Ethene-1,2-diyl)bis(1,4-phenylene)]bis-

(methylene)bis[4-(piperidin-1-yl)pyridinium] Dibromide (44).

White solid. Yield: 84%. Mp > 300 ꢀC.¹H NMR (400 MHz, CD₃OD) δ:

1.70ꢀ1.77 (m, 12H), 3.68ꢀ3.71 (m, 8H), 5.34 (s, 4H), 7.13 (d, J = 7.5 Hz,

4H), 7.23 (s, 2H), 7.39 (d, J = 7.8 Hz, 4H), 7.63 (d, J = 7.8 Hz, 4H), 8.19 (d,

J = 7.8 Hz, 4H). HR LSIMS (m/z) calcd for C₃₆H₄₂Br₂N₄Na (M þ Na)^þ

pyridinium) Dibromide (20). Off-white solid. Yield: 48%. Mp:

1

244ꢀ246 ꢀC. H NMR (300 MHz, CD₃OD) δ: 1.31ꢀ1.42 (m, 2H),

1.46ꢀ1.55 (m, 8H), 1.78ꢀ1.82 (m, 2H), 1.88ꢀ2.01 (m, 8H),

2.86ꢀ2.93 (m, 2H), 5.85 (s, 4H), 7.49ꢀ7.59 (m, 4H), 7.74ꢀ7.78

(m, 2H), 7.95ꢀ7.96 (m, 2H), 7.99 (d, J = 6.9 Hz, 4H), 9.49 (d,

J = 6.9 Hz, 4H). Anal. (C₃₆H₄₂Br₂N₂2H₂O) C, H, N.

711.1674; found 711.1665. Anal. (C₃₆H₄₂Br₂N₄1.8H₂O) C, H, N.

3

1,1⁰-[Biphenyl-3,3⁰-diylbis(methylene)]bis(4-phenylpyri-

dinium) Dibromide (21). Off-white solid. Yield: 73%. Mp:

156ꢀ158 ꢀC. ¹H NMR (300 MHz, CD₃OD) δ: 5.92 (s, 4H),

7.57ꢀ7.64 (m, 10H), 7.78ꢀ7.81 (m, 2H), 7.98ꢀ8.00 (m, 6H), 8.42 (d,

1,1⁰-[Biphenyl-3,3⁰-diylbis(methylene)]bis(quinuclidinium)

Dibromide (47). White solid. Yield: 84%. Mp > 300 ꢀC.¹H NMR (300

MHz, CD₃OD) δ: 1.98ꢀ2.10 (m, 12H), 2.17ꢀ2.20 (m, 2H),

3.59ꢀ3.64 (m, 12H), 4.59 (s, 4H), 7.58 (d, J = 7.9 Hz, 2H), 7.66 (t,

J = 7.7 Hz, 2H), 7.48 (d, J = 7.5 Hz, 2H), 8.01 (s, 2H). HR LSIMS (m/z)

calcd for C₂₈H₃₈BrN₂(M ꢀ Br)^þ481.2218; found 481.2219. Anal.

J = 6.9 Hz, 4H), 9.12 (d, J = 6.9 Hz, 4H). Anal. (C₃₆H₃₀Br₂N₂H₂O).

3

1,1⁰-[Biphenyl-3,3⁰-diylbis(methylene)]bis[4-(6,7-dimethoxy-

3,4-dihydroisoquinolin-2(1H)-yl)pyridinium] Dibromide (22).

(C₂₈H₃₈Br₂N₂2H₂O) C, H, N.

3

1

Yellow solid. Yield: 63%. Mp: 164 ꢀC (dec), 250ꢀ251 ꢀC. H NMR

1,1⁰-[Biphenyl-3,3⁰-diylbis(methylene)]bis(3-hydroxyqui-

nuclidinium) Dibromide (48). White solid. Yield:76%. Mp > 300 ꢀC.

¹H NMR (300 MHz, CD₃OD) δ: 1.85ꢀ1.94 (m, 4H), 2.08ꢀ2.20

(300 MHz, CD₃OD) δ: 2.96 (t, J = 6.0 Hz, 4H), 3.80 (s, 6H), 3.81 (s, 6H),

3.82 (t, J = 6.0 Hz, 4H), 4.69 (s, 4H), 5.47 (s, 4H), 6.84 (s, 2H), 6.87

2642

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

(m, 4H), 2.31ꢀ2.36 (m, 2H), 3.25ꢀ3.30 (m, 2H), 3.44ꢀ3.67 (m, 8H),

3.85ꢀ3.92 (m, 2H), 4.23ꢀ4.26 (m, 2H), 4.63 (m, 4H), 7.48 (d, J = 7.5

Hz, 2H), 7.66 (t, J = 7.7 Hz, 2H), 7.90 (d, J = 7.5 Hz, 2H), 8.02 (s, 2H).

HR LSIMS (m/z) calcd for C₂₈H₃₇N₂O₂(M ꢀ HBr ꢀ Br)^þ433.2855;

bond directionalities) were taken from the GOLD parameter file. In the

present study, the 3D coordinates of hAChE (PDB code 1B41), hBChE

(PDB code 1P0I), TcAChE (PDB code 1EVE), and hChoK (PDB code

3F2R) were retrieved from the Protein Data Bank. The target proteins were

prepared by adding hydrogen atoms, completing and optimizing missing

residues, and removing water and cocrystallized molecules. As is well-

known,⁴⁰the histidine side chains cannot normally be placed into the

electron density map unambiguously. As a result, the protonation state of

such residues was adjusted according to the formation of HB networks that

was further confirmed upon visual inspection. The basic amino functional

groups were protonated. Aromatic amino functional groups were left

uncharged, and carboxylic groups were considered to be deprotonated.

Molecular docking on human AChE and ChoK resulted in 10 poses per

inhibitor in a sphere of 10 and 15 Å radius centered on the centroid atom of

donepezil (cocrystallized with TcAChE, PDB code 1EVE, previously

aligned to hAChE) and of hemicholinium-3, respectively.

Molecular Dynamics. The top-scored solution from molecular

docking of compound 56 was used as the initial enzymeꢀinhibitor

complex for MD simulations. hAChEꢀ and hBChEꢀ56 complexes

were immersed in a cubic TIP3P water box that extended 18 Å from the

protein surface atoms. The system was neutralized by addition of Na^þ

counterions using the AMBER Leap module. This led to a total of

108 129 atoms and of 107 934 for the hAChEꢀ and hBChEꢀ56

complexes, respectively. The parm03 version of the all-atom AMBER

force field was used to model the system. Solvent molecules were initially

relaxed by means of energy minimizations and 30 ps of MD. Subse-

quently, the full system was minimized to remove bad contacts in the

initial geometry and heated gradually to 310 K during 600 ps of MD. The

SHAKE algorithm was employed to constrain all RꢀH bonds, and

periodic boundary conditions were applied to simulate a continuous

system. A nonbonded cutoff of 12 Å was used, whereas the particle mesh

Ewald (PME) method was employed to include the contributions of

long-range interactions. The pressure (1 atm) and the temperature (310

K) of the system were controlled by the Langevin thermostatꢀbarostat

via NAMD, the software used for the 5 ns MD simulations. The time-

step was 1.5 fs, and the system was simulated for a total of 5 ns, with

coordinates saved every 2 ps.

Cholinesterase Inhibition. The in vitro inhibition assays of

AChE from bovine erythrocytes (0.36 U/mg) and BChE from equine

serum (13 U/mg) were run in phosphate buffer 0.1 M, at pH 8.0. Acetyl-

and butyrylthiocholine iodides were used, respectively, as substrates, and

5,5⁰-dithiobis(2-nitrobenzoic acid) (DTNB) was used as the chromo-

phoric reagent.²⁷Inhibition assays were carried out on an Agilent 8453E

UVꢀvisible spectrophotometer equipped with a cell changer. Solutions

of tested compounds were prepared starting from 10 mM stock

solutions in DMSO, which were diluted with aqueous assay medium

to a final content of organic solvent always lower than 1%. AChE

inhibitory activity was determined in a reaction mixture containing 200

μL of a solution of AChE (0.415 U/mL in 0.1 M phosphate buffer, pH

8.0), 100 μL of a 3.3 mM solution of DTNB in 0.1 M phosphate buffer

(pH 7.0) containing 6 mM NaHCO₃, 100 μL of a solution of the

inhibitor (six to seven concentrations ranging from 1 ꢁ 10^ꢀ10to 1 ꢁ

10^ꢀ4M), and 500 μL of phosphate buffer, pH 8.0. After incubation for

20 min at 25 ꢀC, acetylthiocholine iodide (100 μL of 5 mM aqueous

solution) was added as the substrate, and AChE-catalyzed hydrolysis was

followed by measuring the increase of absorbance at 412 nm for 5.0 min

at 25 ꢀC. The concentration of compound which determined 50%

inhibition of the AChE activity (IC₅₀) was calculated by nonlinear

regression of the responseꢀlog(concentration) curve, using GraphPad

Prism, version 5. BChE inhibitory activity was assessed similarly using

butyrylthiocholine iodide as the substrate.

found 433.2852. Anal. (C₂₈H₃₈Br₂N₂O₂1.6H₂O) C, H, N.

3

1,1⁰-[4,4⁰-(Ethane-1,2-diyl)bis(1,4-phenylene)]bis(methy-

lene)bis(3-hydroxyquinuclidinium) Dibromide (49). White

1

solid. Yield: 53%. Mp: 274ꢀ276 ꢀC. H NMR (300 MHz, CD₃OD)

δ: 1.78ꢀ1.97 (m, 4H), 2.05 (m, 2H), 2.22ꢀ2.35 (m, 2H), 2.29 (m, 2H),

3.00 (s, 4H), 3.11 (ddd, J₁= 2.5 Hz, J₂= 3.0 Hz, J₃= 13.0 Hz, 2H),

3.31ꢀ4.49 (m, 8H), 3.71 (ddd, J₁= 3.0 Hz, J₂= 8.3 Hz, J₃= 13.0 Hz,

2H), 4.18ꢀ4.20 (m, 2H), 4.38 (d, J = 13.0 Hz, 2H), 4.44 (d, J = 13.0 Hz,

2H), 7.33 (d, J = 8.3 Hz, 4H), 7.42 (d, J = 8.3 Hz, 4H). HR LSIMS (m/z)

calcd for C₃₀H₄₂BrN₂O₂(M ꢀ Br)^þ541.2424; found 541.2419. Anal.

(C₃₀H₄₂Br₂N₂O₂1H₂O) C, H, N.

₀3

(E)-1,1⁰-[4,4 -(Ethene-1,2-diyl)bis(1,4-phenylene)]bis-

(methylene)bis(3-hydroxyquinuclidinium) Dibromide (50).

1

Off-white solid. Yield: 39%. Mp 153ꢀ155 ꢀC. H NMR (400 MHz,

CD₃OD) δ: 1.81ꢀ1.96 (m, 4H), 2.07 (m, 2H), 2.16ꢀ2.18 (m, 2H),

2.25ꢀ2.35 (m, 2H), 3.13 (d, J = 12.9 Hz, 2H), 3.33ꢀ3.52 (m, 8H), 3.73

(m, 2H), 4.19ꢀ4.21 (m, 2H), 4.40 (d, J = 12.9 Hz, 2H), 4.45 (d, J = 12.9

Hz, 2H), 6.77 (s, 2H), 7.36 (d, J = 8.2 Hz, 4H), 7.43 (d, J = 8.2 Hz, 4H).

HR LSIMS (m/z) calcd for C₃₀H₄₀BrN₂O₂(M ꢀ Br)^þ539.2268; found

539.2271. Anal. (C₃₀H₄₀Br₂N₂O₂2.2H₂O) C, H, N.

3

1-[4-(Bromomethyl)benzyl]-4-piperidin-1-ylpyridinium

Bromide (1i). 1,4-Bis(bromomethyl)benzene (0.11 g, 0.40 mmol) was

stirred in 2-butanone (3.5 mL) until complete dissolution, and 1d (0.026

g, 0.16 mmol), previously dissolved in butanone (0.5 mL), was added in

5 portions of 0.10 mL each, within 96 h during which the mixture was

heated to 100 ꢀC. Then stirring and heating were continued for an

additional 24 h, and the mixture was cooled to room temperature and

washed with butanone and diethyl ether, yielding monobromide 1i as an

off-white solid. Yield: 89%. ¹H NMR (300 MHz, CD₃OD) δ: 1.70ꢀ1.77

(m, 6H), 3.68ꢀ3.72 (m, 4H), 4.57 (s, 2H), 5.32 (s, 2H), 7.12

(d, J = 7.7 Hz, 2H), 7.35 (d, J = 8.3 Hz, 2H), 7.48 (d, J = 8.3 Hz,

2H), 8.16 (d, J = 7.7 Hz, 2H).

4-(6,7-Dimethoxy-3,4-dihydroisoquinolin-2(1H)-yl)-1-{4-

[(4-piperidin-1-ylpyridinium-1-yl)methyl]benzyl}pyridinium

Dibromide (56). Intermediate 1i (0.043 g, 0.10 mmol) was suspended

in a 2-butanone/acetonitrile 4/1 (v/v) mixture (4.0 mL), and 1f (0.27 g,

0.10 mmol) was added. The mixture was refluxed for 48 h. After the

mixture was cooled, the solid was washed with 2-butanone, acetonitrile,

and ethanol, giving an off-white solid that was recrystallized from

ethanol/diethyl ether, yielding an off-white solid. Yield: 75%. Mp:

134ꢀ136 ꢀC (dec), 194ꢀ196 ꢀC. ¹H NMR (300 MHz, CD₃OD)

δ: 1.70ꢀ1.78 (m, 6H), 2.98 (t, J = 5.8 Hz, 2H), 3.68ꢀ3.72 (m, 4H),

3.80ꢀ3.84 (m, 8H), 4.70 (s, 2H), 5.35 (s, 2H), 5.41 (s, 2H), 6.85

(s, 1H), 6.87 (s, 1H), 7.12 (d, J = 7.4 Hz, 2H), 7.22 (d, J = 7.4 Hz, 2H),

7.46 (s, 4H), 8.16 (d, J = 7.4 Hz, 2H), 8.27 (d, J = 7.4 Hz, 2H). Anal.

(C₃₄H₄₀Br₂N₄O₂2H₂O) C, H, N.

3

Molecular Modeling. GOLD (version 4.1.2), a genetic algorithm

based software, was used for the docking study. GoldScore and Chem-

Score were used as fitness functions to rank the docking simulations on

ChEs and ChoK, respectively. GoldScore is made up of four compo-

nents that account for proteinꢀligand binding energy: proteinꢀligand

hydrogen bond energy, proteinꢀligand van der Waals energy, ligand

internal van der Waals energy, and ligand torsional strain energy.

ChemScore was instead derived empirically from a set of 82 pro-

teinꢀligand complexes for which measured binding affinities were

available. ChemScore estimates the total free energy change that occurs

on ligand binding by accounting for a number of physical contributions

to free energy (e.g., hydrogen bonding) whose weight is determined by

regression. Parameters used in the fitness function (e.g., hydrogen bond

energies, atom radii and polarizabilities, torsion potentials, and hydrogen

Kinetic studies on bovine AChE reported in Figure 5 were performed

in the same incubation conditions, using six concentrations of substrate

(from 0.033 to 0.2 mM) and four concentrations of compound 56 (from

2643

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

0 to 30 nM). Apparent inhibition constants and kinetic parameters were

calculated within the “enzyme kinetics” module of Prism.

Inhibition tests on human recombinant AChE (2770 U/mg) and

human serum BChE (7.5 U/mg) were run under the same experimental

conditions used for bovine AChE and horse BChE, respectively.

Choline Kinase Inhibition. The ex vivo ChoK inhibition assays

for compound 56 were performed according to a previously reported

protocol.⁴¹

Inhibition of Aβ_1ꢀ40Aggregation. Aβ_1ꢀ40(purity of >95%;

EzBiolab, Carmel, CA, U.S.) was treated as previously reported²⁸in

order to obtain batches free from preaggregates. Inhibitory activities

were determined for test compounds at 100 μM in 30 μM Aβ

aggregation samples, using a previously described spectrofluorimetric

method.²⁸

(3) Sussman, J. L.; Harel, M.; Frolow, F.; Oefner, C.; Goldman, A.;

Toker, L.; Silman, I. Atomic Structure of Acetylcholinesterase from

Torpedo californica: A Prototypic Acetylcholine-Binding Protein. Science

1991, 253, 872–879.

(4) Inestrosa, N. C.; Alvarez, A.; Perez, C. A.; Moreno, R. D.;

Vicente, M.; Linker, C.; Casanueva, O. I.; Soto, C.; Garrido, J. Acet-

ylcholinesterase Accelerates Assembly of Amyloid-beta-peptides into

Alzheimer’s Fibrils: Possible Role of the Peripheral Site of the Enzyme.

Neuron 1996, 16, 881–891.

(5) Kryger, G.; Silman, I.; Sussman, J. L. Structure of Acetylcholi-

nesterase Complexed with E2020 (Aricept): Implications for the Design

of New Anti-Alzheimer Drugs. Structure 1999, 7, 297–307.

(6) Rama Sastry, B. V. Anticholinergic Drugs. In Burger’s Medicinal

Chemistry and Drug Discovery, 6th ed.; Abraham, D. J., Ed.; Wiley: New

York, 2003; pp 109ꢀ166.

(7) Bevan, D. R.; Donati, F.; Kopman, A. F. Reversal of Neuromus-

cular Blockade. Anesthesiology 1992, 77, 785–805.

’ ASSOCIATED CONTENT

(8) Meriggioli, M. N.; Sanders, D. B. Autoimmune Myasthenia

Gravis: Emerging Clinical and Biological Heterogeneity. Lancet 2009,

8, 475–490.

(9) McKinnon, S. J.; Goldberg, L. D.; Peeples, P.; Walt, J. G.;

Bramley, T. J. Current Management of Glaucoma and the Need for

Complete Therapy. Am. J. Manag. Care 2008, 14, S20–S27.

(10) García-Carrasco, M.; Escꢀarcega, R. O.; Fuentes-Alexandro, S.;

Riebeling, C.; Cervera, R. Therapeutic Options in Autoimmune

Myasthenia Gravis. Autoimmun. Rev. 2007, 6, 373–378.

(11) (a) Pisani, L.; Catto, M.; Giangreco, I.; Leonetti, F.; Nicolotti,

O.; Stefanachi, A.; Cellamare, S.; Carotti, A. Design, Synthesis

and Biological Evaluation of Coumarin Derivatives Tethered to an

Edrophonium-like Fragment as Highly Potent and Selective Dual

Binding Site Acetylcholinesterase Inhibitors. ChemMedChem 2010,

5, 1616–1630. (b) Leonetti, F.; Catto, M.; Nicolotti, O.; Pisani, L.;

Cappa, A.; Stefanachi, A.; Carotti, A. Homo- and Hetero-Bivalent

Edrophonium-like Ammonium Salts as Highly Potent, Dual Binding

Site AChE Inhibitors. Bioorg. Med. Chem. 2008, 16, 7450–7456.

(12) (a) Campos, J.; Nꢀu~nez, M. C.; Conejo-García, A.; Sꢀanchez-

Martín, R. M.; Hernꢀandez-Alcoceba, R.; Rodríguez-Gonzꢀalez, A.;

Lacal, J. C.; Gallo, M. A.; Espinosa, A. QSAR-Derived Choline Kinase

Inhibitors: How Rational Can Antiproliferative Drug Design Be?

Curr. Med. Chem. 2003, 10, 1095–1112. (b) Campos, J. M.; Sꢀanchez-

Martín, R. M.; Conejo-García, A.; Entrena, A.; Gallo, M. A.; Espinosa,

A. (Q)SAR Studies To Design New Human Choline Kinase In-

hibitors as Antiproliferative Drugs. Curr. Med. Chem. 2006, 13,

1231–1248.

S

Supporting Information. Docking models of 56 into

b

BChE, ChoK, and TcAChE (Figures S1ꢀS3). This material is

available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION

Corresponding Author

*Phone: þþ390805442782. Fax: þþ39-0805442230. E-mail:

carotti@farmchim.uniba.it.

Author Contributions

^{^}A.C.G. and L.P. contributed equally to this work.

’ ACKNOWLEDGMENT

The Spanish authors thank the Consejería de Innovaciꢀon,

Ciencia y Empresa of the Junta de Andalucía (Excellence

Research Project No. P07-CTS-03210) for ﬁnancial support

and M. Paz Carrasco-Jimꢀenez and Pablo Rios from the Depart-

ment of Biochemistry and Molecular Biology, University of

Granada (Spain), for the determination of ChoK IC₅₀for

compound 56. Financial support to the Italian authors from

MIUR (Rome, Italy; Grant PRIN 20085HR5JK) is gratefully

acknowledged.

(13) Campos, J.; Nꢀu~nez, M. C.; Rodríguez, V.; Entrena, A.; Hernꢀandez-

Alcoceba, R.; Fernꢀandez, F.; Lacal, J. C.; Gallo, M. A.; Espinosa, A.

LUMO Energy of Model Compounds of Bispyridinium Compounds as

an Index for the Inhibition of Choline Kinase. Eur. J. Med. Chem. 2001,

36, 215–225.

(14) Conejo-García, A.; Bꢀa~nez-Coronel, M.; Sꢀanchez-Martín, R.;

Rodríguez-Gonzꢀalez, A.; Ramos, A.; Ramírez de Molina, A.; Espinosa,

A.; Gallo, M. A.; Campos, J.; Lacal, J. C. Inﬂuence of the Linker in

Bispyridium Compounds on the Inhibition of Human Choline Kinase.

J. Med. Chem. 2004, 47, 5433–5440.

(15) Sꢀanchez-Martín, R.; Campos, J. M.; Conejo-García, A.; Cruz-

Lꢀopez, O.; Bꢀa~nez-Coronel, M.; Rodríguez-Gonzꢀalez, A.; Gallo, M. A.;

Lacal, J. C.; Espinosa, A. Symmetrical Bis-Quinolinium Compounds:

New Human Choline Kinase Inhibitors with Antiproliferative Activity

against the HT-29 Cell Line. J. Med. Chem. 2005, 48, 3354–3363.

(16) Campos, J.; Nꢀu~nez, M. C.; Sꢀanchez, R. M.; Gꢀomez-Vidal, J. A.;

Rodríguez-Gonzꢀalez, A.; Bꢀa~nez, M.; Gallo, M. A.; Lacal, J. C.; Espinosa,

A. Quantitative StructureꢀActivity Relationships for a Series of Sym-

metrical Bisquaternary Anticancer Compounds. Bioorg. Med. Chem.

2002, 10, 2215–2231.

’ ABBREVIATIONS USED

Aβ, β-amyloid; ACh, acetylcholine; AChE, acetylcholinesterase;

AD, Alzheimer’s disease; BChE, butyrylcholinesterase; ChE, cho-

linesterase; ChoK, choline kinase; CNS, central nervous system;

DTNB, 5,5⁰-dithiobis(2-nitrobenzoic acid); HB, hydrogen bond;

MG, myasthenia gravis; NMB, neuromuscular blockade; PNS, per-

ipheral nervous system; QAS, quaternary ammonium salt; SAFIR,

structureꢀaﬃnity relationship; SI, selectivity index; SSR, structureꢀ

selectivity relationship; THIQ, 1,2,3,4-tetrahydroisoquinolin-2-yl

’ REFERENCES

(1) Gentry, M. K.; Doctor, B. P. Alignment of Amino Acid

Sequences of Acetylcholinesterases and Butyrylcholinesterases. In Cho-

linesterases: Structure, Function, Mechanism, Genetics and Cell Biology;

Massouliꢀe, J., Bacou, F., Barnard, E. A., Chatonnet, A., Doctor, B. P.,

Quinn, D. M., Eds.; American Chemical Society: Washington, DC,

1991; pp 394ꢀ398.

(2) Saxena, A.; Redman, A. M.; Jiang, X.; Lockridge, B. P.; Doctor,

B. P. Diﬀerences in Active Site Gorge Dimensions of Cholinesterases

Revealed by Binding of Inhibitors to Human Butyrylcholinesterase.

Biochemistry 1997, 36, 14642–14651.

(17) Rosa, J. C.; Galanakis, D.; Ganellin, C. R.; Dunn, P. M.

Synthesis, Molecular Modeling, and K^þChannel-Blocking Activity of

Dequalinium Analogues Having Semirigid Linkers. J. Med. Chem. 1996,

39, 4247–4254.

2644

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Journal of Medicinal Chemistry

ARTICLE

(18) Conejo-García, A.; Campos, J.; Sꢀanchez-Martín, R. M.; Gallo,

M. A.; Espinosa, A. Bispyridinium Cyclophanes: Novel Templates for

Human Choline Kinase Inhibitors. J. Med. Chem. 2003, 46, 3754–3757.

(19) Conejo-García, A.; Campos, J.; Sꢀanchez, R. M.; Rodríguez-

Gonzꢀalez, A.; Lacal, J. C.; Gallo, M. A.; Espinosa, A. Choline Kinase

Inhibitory Eﬀect and Antiproliferative Activity of New 1,1⁰,1⁰⁰-(Benzene-

1,3,5-triylmethylene)tris{4-[(disubstituted)amino]pyridinium} Tribro-

mides. Eur. J. Med. Chem. 2003, 38, 109–116.

(20) Conejo-García, A.; Campos, J.; Sꢀanchez, R. M.; Rodríguez-

Gonzꢀalez, A.; Lacal, J. C.; Gallo, M. A.; Espinosa, A. Synthesis and NMR

Studies on a C3-Symmetrical Triquinoline Triscationic Bicyclophane.

J. Org. Chem. 2005, 70, 5748–5751.

(21) Tao, B.; Huang, T. L.; Zhang, Q.; Jackson, L.; Queener, S. F.;

Donkor, I. O. Synthesis and Anti-Pneumocystis carinii Activity of Con-

formationally Restricted Analogues of Pentamidine. Eur. J. Med. Chem.

1999, 34, 531–538.

(22) Neset, S.; Hope, H.; Undheim, K. Stereoselective Synthesis

of Cyclopropane-1,2-bis(glycine) Derivatives. Tetrahedron 1997, 53,

10459–10470.

J. M.; Saxena, A. Development of Molecular Probes for the Identiﬁcation

of Extra Interaction Sites in the Mid-Gorge and Peripheral Sites of

Butyrylcholinesterase (BuChE). Rational Design of Novel, Selective,

and Highly Potent BuChE Inhibitors. J. Med. Chem. 2005, 48,

1919–1929.

(36) Cheng, Y.; Cheng, X.; Radiꢀc, Z.; McCammon, J. A. Acetylcho-

linesterase: Mechanisms of Covalent Inhibition of Wild-Type and

H447I Mutant Determined by Computational Analysis. J. Am. Chem.

Soc. 2007, 129, 6562–6570.

(37) Jones, G.; Willett, P.; Glen, R. C.; Leach, A. R.; Taylor, R.

Further Development of a Genetic Algorithm for Ligand Docking and

Its Application to Screening Combinatorial Libraries. In Rational Drug

Design, Novel Methodology and Practical Applications; Parrill, A. L., Red-

dy, M. R., Eds.; American Chemical Society: Washington, DC, 1999; pp

271ꢀ291.

(38) Camps, P.; Formosa, X.; Galdeano, C.; Gꢀomez, T.; Mu~noz

Torrero, D.; Ramírez, L.; Viayna, E.; Gꢀomez, E.; Isambert, N.; Lavilla, R.;

Badia, A.; Clos, M. V.; Bartolini, M.; Mancini, F.; Andrisano, V.; Bidon-

ꢀ

Chanal, A.; Huertas, O.; Dafni, T.; Luque, F. J. Tacrine-Based Dual

Binding Site Acetylcholinesterase Inhibitors as Potential Disease-Mod-

ifying Anti-Alzheimer Drug Candidates. Chem.-Biol. Interact. 2010,

187, 411–415.

(39) Mu~noz-Ruiz, P.; Rubio, L.; Garcia-Palomero, E.; Dorronsoro,

I.; del Monte-Millan, M.; Valenzuela, R.; Usan, P.; de Austria, C.;

Bartolini, M.; Andrisano, V.; Bidon-Chanal, A.; Orozco, M.; Luque,

F. J.; Medina, M.; Martinez, A. Design, Synthesis, and Biological

Evaluation of Dual Binding Site Acetylcholinesterase Inhibitors: New

Disease-Modifying Agents for Alzheimer’s Disease. J. Med. Chem. 2005,

48, 7223–7233.

(40) Hartshorn, M. J.; Verdonk, M. L.; Chessari, G.; Brewerton,

S. C.; Mooij, W. T. M.; Mortenson, P. N.; Murray, C. W. Diverse, High-

Quality Test Set for the Validation of ProteinꢀLigand Docking Perfor-

mance. J. Med. Chem. 2007, 50, 726–741.

(41) (a) Hernandez-Alcoceba, R.; Fernandez, F.; Lacal, J. C. In Vivo

Antitumor Activity of Choline Kinase Inhibitors: A Novel Target for

Anticancer Drug Discovery. Cancer Res. 1999, 59, 3112–3118. (b)

Ramirez de Molina, A.; Penalva, V.; Lucas, L.; Lacal, J. C. Regulation

of Choline Kinase Activity by Ras Proteins Involves Ral-GDS and PI3K.

Oncogene 2002, 21, 937–994.

(23) Wenner, W. Bis(bromomethyl) Compounds. J. Org. Chem.

1952, 17, 523–528.

(24) Cram, D. J.; Steinberg, H. Macro Rings. I. Preparation and

Spectra of the Paracyclophanes. J. Am. Chem. Soc. 1951, 73, 5691–5704.

(25) Raston, I.; Wennerstr€om, O. [2(6)]Paracyclophanes—Large

Ring Compounds with Extended pi-Systems. Acta Chem. Scand. 1982,

B36, 655–660.

(26) Drefahl, G.; Ploetner, G. Untersuchungen €uber Stilbene, XVIII.

Polyphenyl-polyene. Chem. Ber. 1958, 91, 1274–1280.

(27) Ellman, G. L.; Courtney, K. D.; Andres, V., Jr.; Feartherstone, R.

M. A New and Rapid Colorimetric Determination of Acetylcholinester-

ase Activity. Biochem. Pharmacol. 1961, 7, 88–95.

(28) Cellamare, S.; Stefanachi, A.; Stolfa, D. A.; Basile, T.; Catto, M.;

Campagna, F.; Sotelo, E.; Acquafredda, P.; Carotti, A. Design, Synthesis,

and Biological Evaluation of Glycine-Based Molecular Tongs as Inhibi-

tors of Aβ_1-40Aggregation in Vitro. Bioorg. Med. Chem. 2008,

16, 4810–4822.

(29) Verdonk, M. L.; Cole, J. C.; Hartshorn, M. J.; Murray, C. W.;

Taylor, R. D. Improved ProteinꢀLigand Docking Using GOLD.

Proteins 2003, 52, 609–623.

(30) Hong, B. S.; Allali-Hassani, A.; Tempel, W.; Finerty, P. J., Jr.;

MacKenzie, F.; Dimov, S.; Vedadi, M.; Park, H. Crystal Structures of

Human Choline Kinase Isoforms in Complex with Hemicholinium-3:

Single Amino Acid near the Active Site Inﬂuences Inhibitor Sensitivity.

J. Biol. Chem. 2010, 285, 16330–16340.

(31) Case, D. A.; Darden, T. A.; Cheatham, T. E., III; Simmerling,

C. L.; Wang, J.; Duke, R. E.; Luo, R.; Crowley, M.; Walker, R. C.; Zhang,

W.; Merz, K. M.; Wang, B.; Hayik, S.; Roitberg, A.; Seabra, G.;

Kolossvꢀary, I.; Wong, K. F.; Paesani, F.; Vanicek, J.; Wu, X.; Brozell,

S. R.; Steinbrecher, T.; Gohlke, H.; Yang, L.; Tan, C.; Mongan, J.;

Hornak, V.; Cui, G.; Mathews, D. H.; Seetin, M. G.; Sagui, C.; Babin, V.;

Kollman, P. A. AMBER, version 10; University of California: San

Francisco, CA, 2008.

(32) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid,

E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K. Scalable

Molecular Dynamics with NAMD. J. Comput. Chem. 2005,

26, 1781–1802.

(33) Barril, X.; Orzoco, M.; Luque, F. J. Predicting Relative Binding

Free Energies of TacrineꢀHuperzine A Hybrids as Inhibitors of

Acetylcholinesterase. J. Med. Chem. 1999, 42, 5110–5119.

(34) Piazzi, L.; Rampa, A.; Bisi, A.; Gobbi, S.; Belluti, F.; Cavalli, A.;

Bartolini, M.; Andrisano, V.; Valenti, P.; Recanatini, M. 3-(4-{[Benzyl-

(methyl)amino]methyl}-phenyl)-6,7-dimethoxy-2H-2-chromenone

(AP2238) Inhibits Both Acetylcholinesterase and Acetylcholinesterase-

Induced β-Amyloid Aggregation: A Dual Function Lead for Alzheimer’s

Disease Therapy. J. Med. Chem. 2003, 46, 2279–2282.

(35) Campiani, G.; Fattorusso, C.; Butini, S.; Gaeta, A.; Agnusdei,

M.; Gemma, S.; Persico, M.; Catalanotti, B.; Savini, L.; Nacci, V.;

Novellino, E.; Holloway, H. W.; Greig, N. H.; Belinskaya, T.; Fedorko,

2645

dx.doi.org/10.1021/jm101299d |J. Med. Chem. 2011, 54, 2627–2645

Article Doi

Homodimeric bis-quaternary heterocyclic ammonium salts as potent acetyl- and butyrylcholinesterase inhibitors: A systematic investigation of the influence of linker and cationic heads over affinity and selectivity

DOI: 10.1021/jm101299d

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jm101299d

Full text of DOI:10.1021/jm101299d

Products guided by the article

R&D Labs maybe for 203334-73-0

Relevant to this article

Hot Product