magnesium effect on the acetylcholinesterase inhibition mechanism: a molecular chromatographic...

6
Talanta 79 (2009) 804–809 Contents lists available at ScienceDirect Talanta journal homepage: www.elsevier.com/locate/talanta Magnesium effect on the acetylcholinesterase inhibition mechanism: A molecular chromatographic approach F. Ibrahim, Y.C. Guillaume , M. Thomassin, C. André Equipe des Sciences Séparatives, Biologiques et Pharmaceutiques (2SBP/EA-4267), Laboratoire de Chimie Analytique, Faculté de Médecine-Pharmacie, CHU-Jean Minjoz, Université de Franche-Comté, Place Saint Jacques, 25030 Besanc ¸ on Cedex, France article info Article history: Received 19 January 2009 Received in revised form 30 April 2009 Accepted 5 May 2009 Available online 15 May 2009 Keywords: Acetylcholinesterase Inhibitors Association Magnesium Column liquid chromatography abstract The acetylcholinesterase enzyme (AChE) was immobilized on a chromatographic support to study the effect of magnesium on the binding mechanism of five AChE inhibitors (donepezil, tacrine, galanthamine, physostigmine and huperzine). The determination of the enthalpy and entropy changes of this binding at different magnesium concentration values suggested that van der Waals interactions and hydrogen bonds predominated the donepezil and tacrine association to AChE. As well, hydrophobic and electrostatic forces seemed to be the major interactions controlling the huperzine, galanthamine and physostigmine association with AChE. In addition, it appeared that magnesium cation increased the binding affinity of galanthamine and physostigmine to the active site gorge of AChE. A comparison of the inhibitors hydrophobicity to their relative bound percentage with AChE showed an affinity enhanced with the increase in the molecule hydrophobicity and confirmed that the hydrophobic forces played an important role in the AChEI–AChE binding process. This novel biochromatographic column could be useful to find a specific inhibitor for this enzyme and so open new perspectives to be investigated. © 2009 Elsevier B.V. All rights reserved. 1. Introduction Alzheimer’s disease (AD) is a common age-related neurodegen- erative pathology with neurological and psychiatric manifestations. The greatly reduced presence of acetylcholine in the cerebral cortex is a significant factor in AD [1,2]. The enzyme acetyl- cholinesterase (AChE) catalyzes the hydrolysis of the ester bound of acetylcholine (ACh) to terminate the impulse transmitted action of ACh through cholinergic synapses, therefore the inhibition of acetylcholinesterase (AChE) activity may be one of the most real- istic approaches to the symptomatic treatment of AD [3]. Many medicinal agents, as donepezil, huperzine or rivastigmine, used for treatment of Alzheimer’s disease, belong to the important class of acetylcholinesterase inhibitors (AChEIs) [3,4]. The three- dimensional structure for Torpedo californica AChE was resolved in 1991 [5]. Sussman and co-workers have shown that, structurally, there is a narrow active site gorge about 20Å deep which con- sists of two separated ligand binding sites, the acylation (or active) site which is located at the bottom of the gorge and the periph- eral anionic binding site which is located close to the mouth of the active site gorge. The in vivo and in vitro cholinesterase activ- ity and the effect of inhibitors and cations on this activity have been widely studied in the biomedical literature [6–8], but few Corresponding author. Tel.: +33 3 81 66 55 44; fax: +33 3 81 66 56 55. E-mail address: [email protected] (Y.C. Guillaume). studies have examined the mechanism of inhibitors association with AChE and the effect of biological conditions on this associa- tion [9]. Magnesium, in particular, seems to have a profound effect on dementias of various types. Cilliler et al. [10] suggested that there is a relationship between serum magnesium levels and the degree of Alzheimer’s disease. Glick [11] of the Bionix Corporation reviewed the effects of magnesium in patients with Alzheimer’s disease and other dementias. He reported that magnesium may improve memory and alleviate other symptoms in patients with Alzheimer’s disease. Therefore, recognition of forces during the AChEI–AChE binding process provides us valuable models for fur- ther understanding of acetylcholinesterase inhibitions processes and may open unexplored avenues to other AD treatment. In this study the AChE enzyme was immobilized on amine support by cross-linking reaction with a suitable bifunctional reagent (i.e. glu- taraldehyde). This chromatographic support was used to determine and quantify the forces driving association between a series of acetylcholinesterase inhibitors and this enzyme. The energetic of binding of the inhibitor to the enzyme as both a function of tem- perature and magnesium concentration (x) was studied using a biochromatographic approach. 2. Experimental method 2.1. Reagents and solvents The five drugs AChEIs were depicted in Fig. 1. Galanthamine and donepezil were obtained from Interchim (Montluc ¸ on, France), 0039-9140/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2009.05.005

Upload: f-ibrahim

Post on 28-Oct-2016

221 views

Category:

Documents


2 download

TRANSCRIPT

Page 1: Magnesium effect on the acetylcholinesterase inhibition mechanism: A molecular chromatographic approach

MA

FEU

a

ARRAA

KAIAMC

1

eTccooaimfcd1tssetib

0d

Talanta 79 (2009) 804–809

Contents lists available at ScienceDirect

Talanta

journa l homepage: www.e lsev ier .com/ locate / ta lanta

agnesium effect on the acetylcholinesterase inhibition mechanism:molecular chromatographic approach

. Ibrahim, Y.C. Guillaume ∗, M. Thomassin, C. Andréquipe des Sciences Séparatives, Biologiques et Pharmaceutiques (2SBP/EA-4267), Laboratoire de Chimie Analytique, Faculté de Médecine-Pharmacie, CHU-Jean Minjoz,niversité de Franche-Comté, Place Saint Jacques, 25030 Besancon Cedex, France

r t i c l e i n f o

rticle history:eceived 19 January 2009eceived in revised form 30 April 2009ccepted 5 May 2009vailable online 15 May 2009

a b s t r a c t

The acetylcholinesterase enzyme (AChE) was immobilized on a chromatographic support to study theeffect of magnesium on the binding mechanism of five AChE inhibitors (donepezil, tacrine, galanthamine,physostigmine and huperzine). The determination of the enthalpy and entropy changes of this bindingat different magnesium concentration values suggested that van der Waals interactions and hydrogenbonds predominated the donepezil and tacrine association to AChE. As well, hydrophobic and electrostatic

eywords:cetylcholinesterase

nhibitorsssociationagnesium

forces seemed to be the major interactions controlling the huperzine, galanthamine and physostigmineassociation with AChE. In addition, it appeared that magnesium cation increased the binding affinityof galanthamine and physostigmine to the active site gorge of AChE. A comparison of the inhibitorshydrophobicity to their relative bound percentage with AChE showed an affinity enhanced with theincrease in the molecule hydrophobicity and confirmed that the hydrophobic forces played an important

indinenzym

olumn liquid chromatography role in the AChEI–AChE bspecific inhibitor for this

. Introduction

Alzheimer’s disease (AD) is a common age-related neurodegen-rative pathology with neurological and psychiatric manifestations.he greatly reduced presence of acetylcholine in the cerebralortex is a significant factor in AD [1,2]. The enzyme acetyl-holinesterase (AChE) catalyzes the hydrolysis of the ester boundf acetylcholine (ACh) to terminate the impulse transmitted actionf ACh through cholinergic synapses, therefore the inhibition ofcetylcholinesterase (AChE) activity may be one of the most real-stic approaches to the symptomatic treatment of AD [3]. Many

edicinal agents, as donepezil, huperzine or rivastigmine, usedor treatment of Alzheimer’s disease, belong to the importantlass of acetylcholinesterase inhibitors (AChEIs) [3,4]. The three-imensional structure for Torpedo californica AChE was resolved in991 [5]. Sussman and co-workers have shown that, structurally,here is a narrow active site gorge about 20 Å deep which con-ists of two separated ligand binding sites, the acylation (or active)ite which is located at the bottom of the gorge and the periph-

ral anionic binding site which is located close to the mouth ofhe active site gorge. The in vivo and in vitro cholinesterase activ-ty and the effect of inhibitors and cations on this activity haveeen widely studied in the biomedical literature [6–8], but few

∗ Corresponding author. Tel.: +33 3 81 66 55 44; fax: +33 3 81 66 56 55.E-mail address: [email protected] (Y.C. Guillaume).

039-9140/$ – see front matter © 2009 Elsevier B.V. All rights reserved.oi:10.1016/j.talanta.2009.05.005

g process. This novel biochromatographic column could be useful to find ae and so open new perspectives to be investigated.

© 2009 Elsevier B.V. All rights reserved.

studies have examined the mechanism of inhibitors associationwith AChE and the effect of biological conditions on this associa-tion [9]. Magnesium, in particular, seems to have a profound effecton dementias of various types. Cilliler et al. [10] suggested thatthere is a relationship between serum magnesium levels and thedegree of Alzheimer’s disease. Glick [11] of the Bionix Corporationreviewed the effects of magnesium in patients with Alzheimer’sdisease and other dementias. He reported that magnesium mayimprove memory and alleviate other symptoms in patients withAlzheimer’s disease. Therefore, recognition of forces during theAChEI–AChE binding process provides us valuable models for fur-ther understanding of acetylcholinesterase inhibitions processesand may open unexplored avenues to other AD treatment. In thisstudy the AChE enzyme was immobilized on amine support bycross-linking reaction with a suitable bifunctional reagent (i.e. glu-taraldehyde). This chromatographic support was used to determineand quantify the forces driving association between a series ofacetylcholinesterase inhibitors and this enzyme. The energetic ofbinding of the inhibitor to the enzyme as both a function of tem-perature and magnesium concentration (x) was studied using abiochromatographic approach.

2. Experimental method

2.1. Reagents and solvents

The five drugs AChEIs were depicted in Fig. 1. Galanthamineand donepezil were obtained from Interchim (Montlucon, France),

Page 2: Magnesium effect on the acetylcholinesterase inhibition mechanism: A molecular chromatographic approach

F. Ibrahim et al. / Talanta 7

pFF

eE(IfEfipPcma0pThflmsA1si

2

c

of a single component equilibrium of a compound dissolved in a

Fig. 1. Chemical structures of AChEIs.

hysostigmine and huperzine were purchased from Sigma (Paris,rance) and tacrine from Cayman Chemical-Interchim (Montlucon,rance).

Acetylcholinesterase from Electrophorus Electricus (electricel) was purchased from Sigma (Paris, France), glutaraldehyde,llman’s reagent and sodium cyanoborohydride from InterchimMontlucon, France) and S-acetylthiocholine from ALFA AESAR-nterchim (Montlucon, France). Monoethanolamine was obtainedrom Aldrich–Sigma (Paris, France). Water was obtained from anlgastat option water purification system (Odil, Talant, France)tted with a reverse osmosis cartridge. Sodium dihydrogenophos-hate and di-natriumhydrogenophosphate were obtained fromrolabo and Merck (Paris, France) respectively. The mobile phaseonsisted of 0.1 M sodium phosphate buffer adjusted at pH 7.4 withagnesium concentrations (x) equal to 0.0, 0.5, 0.75, 1.0, 1.5, 2.0

nd 2.5 mmol L−1 (including its biological concentration range i.e..75–1.0 mmol L−1). Experiments were carried out over the tem-erature range 278–308 K (278, 283, 288, 293, 298, 303 and 308 K).he detection wavelength was at 280 nm for donepezil, 210 nm foruperzine and 254 nm for the other molecules. The mobile phaseow-rate was 0.5 mL/min. AChEIs solutions were prepared in theobile phase, and 20 �L was injected into the chromatographic

ystem. For the determination of the adsorption isotherms, for eachChEI studied, the equilibration of the column was carried out with5 concentrations of AChEI (0–7 �M) in the mobile phase to obtain atable detection. 20 �L of the most concentrated AChEI sample wasnjected at least three times and the retention time was measured.

.2. Apparatus

The high performance liquid chromatography (HPLC) systemonsisted of a Hewlett-Packard quaternary pump (1050), an Agi-

9 (2009) 804–809 805

lent (G1365B, serie 1100) UV-Visible detector (Paris, France) anda Rheodyne 7725i injection valve (Cotati, CA, USA) fitted witha 20-�L sample loop. The Modulo-Cart HS Uptisphere3 NH2(50 mm × 4.6 mm) was purchased from Interchim (Montlucon,France), where the propylamino groups were bound onto 3 �m sil-ica particles of 120 Å pore size. The acetylcholinesterase stationaryphase (i.e., the AChE column) prepared via the in situ technique isgiven below.

2.3. Covalent immobilization technique of AChE

The in situ immobilization technique was considered in thisstudy [12–14]. The immobilization of AChE via the amino groupsof the enzyme on aminopropyl silica pre-packed column activatedwith glutaraldehyde was carried out as follows. Briefly, the columnwas first washed with phosphate buffer (50 mM, pH = 7.0) for 30 minat 0.5 mL/min. Then the stationary phase was activated by recyclingglutaraldehyde 10% in phosphate buffer (50 mM, pH 6.0) for 12 h inthe dark, followed by washing with phosphate buffer (50 mM, pH6.0) for 30 min at 0.5 mL/min. A solution of 250 U AChE in 80 mLphosphate buffer (50 mM, pH 6.0) was recirculated through the col-umn at a flow-rate of 0.5 mL/min for 24 h. After immobilization,the enzyme solution was analysed with the Ellman’s assay in orderto determine the unreacted enzyme units. The Schiff bases werereduced by recycling the derivatized of 0.1 M cyanoborohydridesolution in phosphate buffer (50 mM, pH 6.0) for 2 h at 25 ◦C. Thematrix was then washed with phosphate buffer (50 mM, pH 6.0) andthen with 0.2 M monoethanolamine solution in phosphate buffer(50 mM, pH 6.0) for 3 h at room temperature. Then the column wasrinsed for 1 h with phosphate buffer (pH 7.4, 50 mM) at a flow-rateof 0.5 mL/min and was stored at 4 ◦C after rinsing with phosphatebuffer (50 mM, pH 7.4) containing 0.1% sodium azide. By analysingthe enzymatic activity of the enzyme solutions before and after theimmobilization using Ellman’s assay, the activity decreased about70%, this means that about 70% of the enzyme in the AChE solutionwere immobilized on the column stationary phase.

2.4. Langmuir distribution isotherms

Single and multi-component isotherms are now measured bydynamic methods. The most widespread of this is frontal analysis,but this technique is time consuming and requires large amountsof pure compounds [15]. Another popular method, elution by char-acteristic point (ECP), derives the isotherm from the profile of thediffuse front of the band obtained in response to a single injec-tion of a highly concentrated sample [16]. This method is fastand needs only small amounts of sample, but it requires accuratecalibration of the detector and an efficient column. Distributionisotherms can also be apprehended using the perturbation tech-nique originally developed for measuring gas-adsorbent equilibria.The perturbation technique makes possible the determination ofadsorption isotherms by measuring the retention times of smallsample sizes injected onto a column equilibrated with sample solu-tions at different concentration levels. The column used for thedetermination of the isotherm is first equilibrated with a solutioncontaining the compound dissolved in a non-adsorbable solvent.Then a small sample volume containing higher concentration of thecompound is injected onto the column. After the injection, the equi-librium condition is disturbed and the perturbation waves reachesthe column outlet, a peak is registered by the detector. In the case

non-adsorbable solvent, one peak is observed and the distributionisotherms depends only on the concentration of a single solute[17,18]. The well-known Langmuir theoretical approach relates thetotal concentration of the sample in the stationary phase (CS) and

Page 3: Magnesium effect on the acetylcholinesterase inhibition mechanism: A molecular chromatographic approach

806 F. Ibrahim et al. / Talanta 79 (2009) 804–809

Table 1Values of the retention contribution of the two kind of sites k′

A and k′B, the retention factor k′(k′ = k′

A+ k′

B) (extrapolated at Cm = 0), the relative bound percentage b%, the

log P and the non-linear regression coefficients r2 and F (Langmuir model; Lang and bi-Langmuir model; bi-Lang), for the five AChEIs at a bulk solvent pH = 7.4 and T = 298 K,standard deviations are in parentheses.

AchEI k′A

k′B

k′ b (%) log P r2; F-Lang r2; F-bi-Lang

Huperzine 0.54 (0.01) 0.01 (0.01) 0.55 (0.01) 35.40 (0.09) 0.20 0.9993; 2990 0.9996; 6281PGTD

t

C

wcptb

k

wmn(vit([iitpAttcwa(g

k

Ersuacotr

l

w�cvsic

hysostigmine 0.73 (0.01) 0.02 (0.01) 0.75 (0.01)alanthamine 0.73 (0.01) 0.03 (0.01) 0.76 (0.01)acrine 2.28 (0.01) 0.02 (0.01) 2.30 (0.01)onepezil 2.98 (0.02) 0.03 (0.01) 3.01 (0.01)

hat in the mobile phase (Cm) [17–19]:

S = ˛KCm

1 + KCm(1)

here ˛ is the column saturation capacity and K is the associationonstant between AChEI in the sample and the AChE stationaryhase. The sample AChEI retention factor k′ was directly propor-ional to the slope of its adsorption isotherm and can be thus giveny the following equation [17–19]:

′ = t − t0

t0= �

dCS

dCm= �˛K

(1 + KCm)2(2)

here t is the retention time of the solute determined from the peakaximum, t0 is the column hold up time, i.e., the elution time of a

on-retained compound, and � is the column phase ratio (VS/Vm)VS is the volume of the stationary phase in the column and Vm theoid volume). By plotting the k′ value vs. the sample concentrationn the bulk solvent, the parameters �˛K corresponding to the reten-ion contribution of the AChE binding site under linear conditioni.e., the k′ value extrapolated at Cm = 0) were calculated using Eq. (2)17–19]. The main advantage of the perturbation technique consistsn using a simpler instrumentation for the acquisition of the exper-mental data than frontal analysis method: the determination ofhe concentration of the individual compounds at the intermediatelateaus of the frontal analysis curves is no longer needed [17,18].s well, using the AChE stationary phase, AChEI could tightly bind

o the matrix of the column. Then if AChEI bound on two sites onhe stationary phase, i.e. a specific site (site A with an adsorptiononstant KA) and a column saturation capacity ˛A and second sitehich is non-specific (sites B with an adsorption constants KB andcolumn saturation capacity ˛B), then the AChEI retention factor

k′) directly proportional to the slope of its adsorption isotherm isiven by the following equation [17–19]:

′ = t − t0

t0= �

dCS

dCm= �

(˛AKA

(1 + KACm)2+ ˛BKB

(1 + KBCm)2

)(3)

q. (3) was fitted to the solute retention factor k′ by a non-linearegression and the parameters k′

A = �KA˛A and k′B = �KB˛B corre-

ponding to the retention contributions of the two kinds of sitesnder linear conditions were calculated. Valuable informationsbout the processes driving AChEI–AChE association mechanisman be further gained by examining the temperature dependencef AChEI retention [20,21]. Under linear conditions, the tempera-ure dependence of the retention factor is given by the followingelationship:

n k′ =(

−�H◦

RT

)+

(�S◦

R

)+ ln � (4)

here R is the gas constant, T is the column temperature in Kelvin,H◦ and �S◦ are, respectively, the solute enthalpy and entropy

hanges accompanying the transfer of the AChEIs from the bulk sol-ent to the AChE surface. If the AChE stationary phase, AChEI andolvent properties are temperature invariant, a linear van’t Hoff plots obtained and from the slope and intercept �H◦ and �S◦ can bealculated.

43.00 (0.10) 1.70 0.9998; 8980 0.9999; 60031043.30 (0.11) 1.80 0.9998; 7710 0.9998; 882069.74 (0.12) 2.20 0.9998; 9040 0.9999; 42393075.10 (0.12) 3.70 0.9998; 9980 0.9999; 530896

3. Results and discussion

3.1. Langmuir distribution isotherms

The Langmuir distribution isotherms were calculated at pH = 7.4and 298 K. For each AChEI and for each AChEI concentration in thebulk solvent, the most concentrated AChEI sample was injectedinto the chromatographic system and its retention factor was deter-mined (see Section 2.1). The variation coefficients of the k′ valueswere <0.3%, indicating a high reproducibility and a good stabil-ity for the chromatographic system. Using a weighted non-linearregression (WNLIN) procedure, the constants of Eq. (2) were usedto estimate the retention factors. The slope of the curve represent-ing the variation of the estimated retention factors (k′) (Eq. (2)) vs.the experimental values (0.998; ideal is 1.000) and r2 (0.996) indi-cate that there is an excellent correlation between the predictedand experimental retention factors. The non-linear regression coef-ficient r2 and the F value (from the Fisher test with the confidencelevel at 95%) were determined. These are shown in Table 1. The Fvalue constitutes a more discriminating parameter than the r2 valuewhen assessing the significance of the model equation. From the fullregression model, a Student’s t-test was used to provide the basisfor the decision as to whether or not the model coefficients weresignificant. Results of the Student’s t-test show that no variable canbe excluded from the model. These results showed that the Lang-muir model describes accurately the association behaviour of AChEIwith AChE. However, the immobilization of AChE on silica supportcould lead to non-specific interactions. Using the non-linear regres-sion, the retention contributions of the two kind of sites k′

A andk′

B were determined from Eq. (3). The corresponding non-linearregression coefficients r2 and F values of this bi-Langmuir modelwere determined and given in Table 1. The non-linear coefficientresults (r2 > 0.99) and the F values proved that the two Langmuirmodel described accurately the binding mechanism of AChEI withthe AChE stationary phase. As well, the results showed that theinteractions between ACEI and the matrix of the stationary phasewere neglected (the k′

A and k′B values were given in Table 1 and

k′B � k′

A).

3.2. Thermodynamic origins of the AChEI binding to AChE

The k′ values were determined for a sample concentration in themobile phase equal to zero; i.e. Cm = 0. The van’t Hoff plots (ln k′ vs.1/T) of Eq. (4) were drawn for all the AChEI molecules. Linear van’tHoff plots were obtained with correlation coefficients r2 higher than0.91 for donepezil, tacrine and huperzine. An example of plot wasgiven for tacrine in Fig. 2. For galanthamine the variations of ln k′

vs. 1/T were negligible (Fig. 2). These van’t Hoff plot behaviourswere thermodynamically expected when the solute–AChE associa-tion mechanism was independent of temperature. According to Eq.

(4) these van’t Hoff plots provided a conventional way of calculat-ing the thermodynamic parameters (�H◦, �S◦). For donepezil andtacrine which are dual acetylcholinesterase inhibitors that couldbind simultaneously to the peripheral and catalytic sites of theenzyme [22], both �H◦, �S◦ were negative values (Table 2). Nega-
Page 4: Magnesium effect on the acetylcholinesterase inhibition mechanism: A molecular chromatographic approach

F. Ibrahim et al. / Talanta 79 (2009) 804–809 807

Fc

tfvcAsblTdohsbtddiswtwMNo[pappghcaia

t[e

TVm

A

DTGH

ig. 2. Plot of ln k′ vs. 1/T for tacrine and galanthamine at pH = 7.4 and a magnesiumoncentration equal to 0.75 mM.

ive �H◦ values indicated that it was energetically more favourableor these drugs to be linked to AChE rather to be in the bulk sol-ent. Negative �S◦ values showed an increase in the order of thehromatographic system when these drugs were included in theChE binding gorge. The negative �H◦ and �S◦ values demon-trated that the binding was controlled enthalpically, i.e., hydrogenonds and van der Waals forces were the major interactions stabi-

izing the AChE–drug (i.e., donepezil, tacrine) association [23–25].he �S◦ values obtained for tacrine were more negative than foronepezil (Table 2), showing that the tacrine molecule was morerganized in the binding gorge of AChE than donepezil due to theydrophobic interactions. These results in accordance with othertudies [26,27] showed that donepezil was fixed inside the gorgey direct hydrogen bonds, water bridges, and hydrophobic interac-ions. Niu et al. [26] showed that all oxygen and nitrogen atoms ofonepezil take part in the formation of hydrogen bonds, but mostly,onepezil forms hydrogen bonds with the enzyme residues through

ts carbonyl oxygen atom of the dimethoxyindanone group. At theame time, the aromatic residues lining the enzyme gorge wallere the major components contributing to hydrophobic interac-

ions with donepezil [26] and tacrine [28,29]. Tacrine can be bindith the aromatic amino acid lining the gorge by �–� stacking.oreover the aromatic nitrogen of tacrine and the tacrine aminoH group can be hydrogen-bonded to the main-chain carbonylxygen of some amino acids (His, Trp, . . .) of the binding gorge28,29]. For huperzine, the positive �H◦ and �S◦ values indicatedredominant hydrophobic forces between AChE and this molecule,nd draw attention to the role that solvent reorganization must belaying in determining the strength of the huperzine–AChE com-lex [30–32]. In addition to the hydrophobic interactions whichovern the huperzine–AChE association, other interactions as theydrogen bonds due to the huperzine electronegative atoms (O, N)an get involved in this association. This result in accordance withprevious study [33] showed that huperzine engaged favourable

nteractions in the hydrophobic core and electrostatic region of thective site gorge of the enzyme.

Enthalpy–entropy compensation (EEC) temperature is a usefulhermodynamic approach to the analysis of physico-chemical data

34]. Mathematically the entropy-enthalpy compensation can bexpressed by the following equation:

H◦ = ˇ �S◦ + �G◦ˇ (5)

able 2alues of �H◦ (kJ/mol) and �S◦ (J/mol K) for the five AChEIs binding on AChE at aagnesium concentration equal to 0.75 mM. Standard deviations are in parentheses.

ChEIs �H◦ (kJ/mol) �S◦ (J/mol K)

onepezil −6.59 (0.05) −9.73 (0.08)acrine −10.73 (0.09) −27.73 (0.07)alanthamine – +0.99 (0.01)uperzine +14.76 (0.10) +45.74 (1.01)

Fig. 3. Plot of �H◦ (J/mol) vs. �S◦ (J/mol K) for donepezil, tacrine and huperzinebinding on AChE at all the magnesium concentrations.

�G◦ˇ

is the corresponding Gibbs free energy variation at the

compensation temperature ˇ. According to this last equation, whenenthalpy–entropy compensation is observed with a group of com-pounds in a particular chemical interaction, all the compounds havethe same free energy �G

◦ˇ

at temperature ˇ [35,36]. The plot of�H◦ vs. �S◦ obtained for donepezil, tacrine and huperzine waslinear at all the magnesium concentration values of the bulk sol-vent (r2 > 0.99) (Fig. 3). The correlation coefficient of this plot washigher than 0.99, and this value can be considered adequate to ver-ify enthalpy–entropy compensation [37]. It can be deduced that therelative contributions of enthalpy and entropy to the overall freeenergy are the same for this three AChEIs (donepezil, tacrine andhuperzine) and the magnesium concentrations in the bulk solvent.But, since different mechanisms could result in the same propor-tion of enthalpy and entropy relative to the overall free energy, itcannot be deduced rigorously that the association mechanism ofthese molecules with the AChE was independent of their struc-tures and the magnesium concentration in the bulk solvent [38].However, donepezil, tacrine and huperzine have similar inhibitoractivity for AChE. These two conditions (EEC and similar biologi-cal effects) seem to imply a similarity of properties of these threeAChEIs.

For galanthamine the ln k′ vs. 1/T variations were negligible(Fig. 2). Thus, for this molecule, hydrophobic and electrostatic inter-actions were the major interactions governing its association withAChE (entropically driven mechanism). Greenblatt et al. [39] havesolved the X-ray crystal structure of galanthamine bound in theactive site of Torpedo californica acetylcholinesterase (TcAChE) at2.3 Å resolution. They have shown that galanthamine binds at thebase of the gorge interacting with both the acyl-binding pocketand the principal quaternary ammonium-binding site, the indolering of Trp-84. The relatively tight binding of galanthamine to AChEappears to arise from a number of moderate to weak interactionswith the protein, coupled to a low entropy cost for binding due tothe rigid nature of the inhibitor [39,40]. These interactions camefrom polar and non-polar interactions of the tertiary amino group,the terminal hydroxyl group and the aromatic rings of galanthaminewith the active binding gorge of the enzyme. In addition, two prin-cipal hydrogen bonds could be formed with some amino acid of theactive site gorge (Trp, Asp, . . .) thanks to the hydroxyl group andthe O-methyl group of galanthamine.

Physostigmine (AB) combined also with both the peripheraland catalytic sites of the enzyme (EH) active gorge, this associa-tion involves a reversible complex (EH–AB) formation. Barak et al.[41] have shown that the accommodation of physostigmine and

its analogues by AChE is dominated by hydrophobic interactions[41,42]. The reversible complex is followed by carbamoylation of theenzyme, and production of a covalent adduct (EA). The carbamoy-lated enzyme is then hydrolyzed to regenerate the free enzyme as
Page 5: Magnesium effect on the acetylcholinesterase inhibition mechanism: A molecular chromatographic approach

808 F. Ibrahim et al. / Talanta 7

Fa

f

E

dtadp

(

b

gmlmph(ficaTtf[

3m

Aicm

wb

ptppdi

ig. 4. Plot of ln k′ vs. PMg2+ for donepezil and physostigmine at a column temper-ture equal to 308 K.

ollowing [40,41]:

H + ABk1�k−1

EH· · ·ABk2−→EA + BH

k3−→EH + AOH

ki represents the rate constant of the different equilibrium. Theecarbamoylation phase of the reaction is considerably slow thanhe other phases and is therefore the most clarified step of the inter-ction [40]. Therefore, concerning the physostigmine molecule, theetermination of the retention parameters with the AChE stationaryhase, give only an initial evaluation of its association mechanism.

The relative bound percentages of the AChEIs with the enzymeb%) were calculated at 298 K using the following equation [43]:

% = k′

1 + k′ % (6)

These values were given in Table 1. Eq. (6) has been shown toive a good correlation vs. reference methods for compounds withedium-to-strong binding to protein [43,44]. The corresponding

og P (partition coefficient octanol/water) values were exposed byany scientific sites (Pubchem, Drugbank, Chemspider, . . .). Table 1

resents the log P values (http://pubchem.ncbi.nlm.nih.gov/) whichave been derived from an atomic fragment database using X log Pa program for the prediction of the octanol/water partition coef-cients of organic compounds). A significant linear relationshiporrelating the AChE relative bound percentage to log P takens a measure of lipophilicity was obtained (r upper than 0.87).his result showed an affinity enhanced with the increase inhe molecule hydrophobicity and confirmed that the hydrophobicorces play an important role in the AChEI–AChE binding processes44].

.3. Magnesium cation effect on the AChEI–AChE associationechanism

The magnesium effects on association mechanism between theChEIs and the enzyme can be modelled at a thermodynamic level

n terms of the direct stoichiometric participation of the magnesiumations in the association reaction. The dependence of k′ on theagnesium concentration (x) can be formulated as follows [45,46]:

∂ log k′

∂PMg2+ = nMg2+ (7)

here nMg2+ was the number of magnesium cations displaced oround upon AChEI–AChE complex formation.

The log k′ vs. PMg2+ (=−log x) were plotted for all AChEIs. Theselots showed that for donepezil, tacrine and huperzine the reten-

ion was magnesium concentration independent. An example oflot for donepezil was given in Fig. 4. For galanthamine andhysostigmine the retention was magnesium concentration depen-ent and increased significantly with salt concentration increase

n the bulk solvent (Fig. 4). This increase of the binding intensity

[[[[

9 (2009) 804–809

with magnesium concentration can be explained by a change in thewater activity (i.e. the hydrophobic effect) classically attributed tosalt effect [45,46]. This trend of the hydrophobic interaction inten-sity of galanthamine and physostigmine with the active site gorgewith magnesium concentration is affected by the binding of magne-sium cations to the anionic groups of the peripheral anionic bindingsite. From Eq. (7), the slope of the curve log k′ vs. PMg2+ (Fig. 4) givesthe number of magnesium at the AChEIs–AChE interface impliedin the binding process. Negatives values of nMg2+ were obtainedfor galanthamine and physostigmine at all temperatures of thebulk solvent. For example, at 35 ◦C the corresponding nMg2+ valuefor galanthamine was −0.10, and for physostigmine was −0,16.The negative values of nMg2+ reflected that the galanthamine orphysostigmine displacement to their AChE active binding gorge wasaccompanied with magnesium exclusion. Many studies showedthat AChE can bind divalent inorganic cations (Mg2+, Ca2+, . . .) [9,47]and this binding was showed to have a competitive effect againstcationic substrates and inhibitors binding on AChE [9]. Our resultsshowed that the ionic competitive effect of magnesium for galan-thamine and physostigmine binding on AChE was less importantthan the hydrophobic interactions. In addition, these hydropho-bic interactions were increased as the magnesium concentrationin the bulk solvent increased. These results demonstrated that inthe biological magnesium concentration range (0.75–1 mmol L−1),an increase in the magnesium concentration led an enhancementof the galanthamine and physostigmine association with AChE andconsequently an increase of the AChE inhibition. Thus, a magnesiumsupplementation for patients who suffer from Alzheimer diseasecan be useful during the treatment with these AChEIs.

4. Conclusion

The mechanism of donepezil, tacrine, galanthamine, huperzineand physostigmine binding to AChE was analyzed. The determina-tion of the thermodynamic data of this association showed that fordonepezil and tacrine the inhibitor binding with AChE was gov-erned principally by hydrogen bonding and van der Waals forces.Whereas for huperzine, galanthamine and physostigmine, the asso-ciation was dominated by hydrophobic interactions. The effect ofmagnesium on the association with AChE was studied. No impor-tant change in the donepezil, tacrine and huperzine associationwith AChE stationary phase was observed. On the other hand anincrease of the magnesium concentration led an enhancement ofthe galanthamine and physostigmine association with the enzymeand consequently an increase of the AChE inhibition. This studydemonstrated that it seems to be interesting to test in vivo, the mag-nesium supplementation during galanthamine and other AChEIstreatment for patients who suffer from Alzheimer disease.

References

[1] B.A. Yankner, A. Caceres, L.K. Duffy, Proc. Natl. Acad. Sci. U.S.A. 87 (1990) 9020.[2] D.C. German, U. Yazdani, S.G. Speciale, P. Pasbakhsh, D. Games, C.L. Liang, J.

Comp. Neurol. 462 (2003) 371.[3] J. Kao, G. Grossberg, Alzheimer’s Disease, vol. 2, Springer, Berlin, Heidelberg,

2008, p. 25.[4] J. Kaur, M.Q. Zhang, Curr. Med. Chem. 7 (2000) 273.[5] J.L. Sussman, M. Harel, F. Frolow, C. Oefner, A. Goldman, L. Toker, I. Silman,

Science 253 (1991) 872.[6] M. Alcalá Mdel, N.M. Vivas, S. Hospital, P. Camps, D. Munoz-Torrero, A. Badia,

Neuropharmacology 44 (2003) 749.[7] I. Silman, J.L. Sussman, Chem.-Biol. Interact. 175 (2008) 3.[8] F. Stock, J. Hoffmann, J. Ranke, R. Störmann, B. Ondruschka, B. Jastorff, Green

Chem. 6 (2004) 286.[9] V. Tõugu, T. Kesvatera, Biochim. Biophys. Acta 1298 (1996) 12.10] A.E. Cilliler, S. Ozturk, S. Ozbakir, Gerontology 53 (2007) 419.11] J.L. Glick, Med. Hypotheses 31 (1990) 211.12] M. Bartolini, V. Cavrini, V. Andrisano, J. Chromatogr. A 1031 (2004) 27.13] A.M. Girelli, E. Mattei, J. Chromatogr. B 819 (2005) 3.

Page 6: Magnesium effect on the acetylcholinesterase inhibition mechanism: A molecular chromatographic approach

lanta 7

[

[[[[

[[[[

[

[

[

[

[[

[[

[[

[

[[[

[[[

[

[

[

[[

Matoga, Y.C. Guillaume, J. Pharm. Biomed. Anal. 41 (2006) 228.

F. Ibrahim et al. / Ta

14] T. Bagnost, Y.C. Guillaume, M. Thomassin, J.F. Robert, A. Berthelot, A. Xicluna, C.André, J. Chromatogr. B 856 (2007) 113.

15] G. Guiochon, S. Golshan-Shirazi, A.M. Katti, Academic Press, Boston, 1994.16] J.F.K. Huber, R.G. Gerriste, J. Chromatogr. 58 (1971) 137.17] C. Blumel, P. Hugo, A. Seidel Morgenstern, J. Chromatogr. A 865 (1999) 51.18] P. Jandera, S. Berncekova, K. Mihlbachler, G. Guiochon, V. Backovska, J. Planeta,

J. Chromatogr. A 925 (2001) 19.19] C. André, Y.C. Guillaume, Chromatographia 58 (2003) 193.20] Y.C. Guillaume, C. Guinchard, Anal. Chem. 68 (1996) 2869.21] C. André, Y.C. Guillaume, J. Chromatogr. A 1029 (2004) 21.22] D. Alonso, I. Dorronsoro, L. Rubio, P. Munoz, E. García-Palomero, M. Del Monte,

A. Bidon-Chanal, M. Orozco, F.J. Luque, A. Castro, M. Medina, A. Martínez, Bioorg.Med. Chem. 13 (2005) 6588.

23] C. Andre, M. Thomassin, C. Guyon, T.T. Truong, Y.C. Guillaume, J. Pharm. Biomed.Anal. 32 (2003) 217.

24] S. Urien, P. Nguyen, S. Berlioz, F. Brée, F. Vacherot, J.P. Tillement, J. Biochem. 302(1994) 69.

25] E. Peyrin, Y.C. Guillaume, N. Morin, C. Guinchard, J. Chromatogr. A 808 (1998)113.

26] C. Niu, Y. Xu, Y. Xu, X. Luo, W. Duan, I. Silman, J.L. Sussman, W. Zhu, K. Chen, J.Shen, H. Jiang, J. Phys. Chem. B 109 (2005) 23730.

27] G. Kryger, I. Silman, J.L. Sussman, Structure 7 (1999) 297.

28] P. Munoz-Ruiz, L. Rubio, E. García-Palomero, I. Dorronsoro, M. del Monte-Millán,

R. Valenzuela, P. Usán, C. de Austria, M. Bartolini, V. Andrisano, A. Bidon-Chanal,M. Orozco, F.J. Luque, M. Medina, A. Martínez, J. Med. Chem. 48 (2005) 7223.

29] D. Han, P. Yang, J. Mol. Struct. (Theochem.) 668 (2004) 25.30] D.M. Wong, H.M. Greenblatt, H. Dvir, P.R. Carlier, Y.F. Han, Y.P. Pang, I. Silman,

J.L. Sussman, J. Am. Chem. Soc. 125 (2003) 363.

[[

[

9 (2009) 804–809 809

31] D. Zhong, A. Douhal, A.H. Zewail, Proc. Natl. Acad. Sci. U.S.A. 97 (2000) 14056.32] F. Darrouzain, C. André, L. Ismaili, M. Matoga, Y.C. Guillaume, J. Chromatogr. B

820 (2005) 283.33] A. Saxena, N. Qian, I.M. Kovach, A.P. Kozikowski, Y.P. Pang, D.C. Vellom, Z. Radic,

D. Quinn, P. Taylor, B.P. Doctor, Protein Sci. 3 (1994) 1770.34] W. Melander, C.S. Horvath, Academic Press, New York. 23, 1986.35] R.R. Krug, Ind. Eng. Chem. Fundam. 19 (1980) 50.36] C. André, L. Ping, M. Thomassin, J.F. Robert, Y.C. Guillaume, Anal. Chem. Acta

542 (2005) 199.37] L.A. Cole, J.G. Dorsey, K.A. Dill, Anal. Chem. 64 (1992) 1324.38] R. Ranatunga, M.F. Vitha, P.W. Carr, J. Chromatogr. A 946 (2002) 47.39] H.M. Greenblatt, G. Krygera, T. Lewisb, I. Silmanc, J.L. Sussman, FEBS Lett. 463

(1999) 321.40] J. Stojan, M. Zorko, Biochim. Biophys. Acta (BBA)—Protein Struct. and Mol. Enzy-

mol. 1337 (1997) 75.41] D. Barak, A. Ordentlich, D. Stein, Q.S. Yu, N.H. Greig, A. Shafferman, Biochem J.

417 (2009) 213.42] A. Shafferman, D. Barak, D. Stein, C. Kronman, B. Velan, N.H. Greig, A. Ordentlich,

Chem.-Biol. Interact. 175 (2008) 166.43] D.S. Hage, J. Chromatogr. B 768 (2002) 3.44] F. Darrouzain, P. Dallet, J.P. Dubost, L. Ismaili, F. Pehourcq, B. Bannwarth, M.

45] F.S. Sarr, Y.C. Guillaume, C. André, J. Pharm. Biomed. Anal. 47 (2008) 651.46] C.J. Van Oss, Interfacial Forces in Aqueous Media, Marcel dekker, New York,

1994.47] A.C.D. Bainy, M.H.G. de Medeiros, P. Di Mascio, E.A. de Almeida, Biotemas 19

(2006) 35.