Concentration-dependent reversible activation-inhibition of humanbutyrylcholinesterase by tetraethylammonium ion

Jure Stojan1, Marko Golicnik1, Marie-Therese Froment2, Francois Estour2 and Patrick Masson2

1Institute of Biochemistry, Medical Faculty, University of Ljubljana, Slovenia; 2Centre de Recherches du Service de Sante des Armees,

Unite d’Enzymologie, La Tronche, France

Tetraalkylammonium (TAA) salts arewell known reversibleinhibitors of cholinesterases. However, at concentrationsaround 10 mM, they have been found to activate thehydrolysis of positively charged substrates, catalyzed bywild-typehumanbutyrylcholinesterase (EC 3.1.1.8) [Erdoes,E.G., Foldes, F.F., Zsigmond, E.K., Baart, N. & Zwartz,J.A. (1958) Science 128, 92]. The present study was under-taken to determinewhether the peripheral anionic site (PAS)of humanBuChE (Y332,D70) and/or the catalytic substratebinding site (CS) (W82, A328) are involved in this phenom-enon. For this purpose, the kinetics of butyrylthiocholine(BTC) hydrolysis by wild-type human BuChE, by selectedmutants and by horse BuChE was carried out at 25 °C and

pH 7.0 in the presence of tetraethylammonium (TEA). Itappears that human enzymes with more intact structure ofthe PAS showmore prominent activation phenomenon.Thefollowing explanation has been put forward: TEA competeswith the substrate at the peripheral site thus inhibiting thesubstrate hydrolysis at the CS. As the inhibition by TEA isless effective than the substrate inhibition itself, it mimicsactivation.At the concentrations around 40 mM,well withinthe rangeofTEAcompetitionatboth substratebinding sites,it lowers the activity of all tested enzymes.

Keywords: cholinesterases; tetraalkylammonium com-pounds; kinetics; reaction mechanism.

Acetylcholinesterase (AChE; EC 3.1.1.7) and butyrylcholi-nesterase (BuChE; EC 3.1.1.8) are closely related serinehydrolases [1]. No clear physiological function has yet beenassigned to BuChE; it appears to play a role in neurogenesisand neural disorders [2] and it is of pharmacological andtoxicological importance: it hydrolyses numerous estercontaining drugs [3–5] and, like AChE is inhibited bysimilar compounds. Thus, an understanding of BuChEcatalysis and inhibition mechanisms is of paramountimportance, especially for the research of new treatmentsagainst organophosphate and carbamate poisoning [6], i.e.for the design of new reactivators of phosphylated choli-nesterases and of mutated enzymes capable of hydrolyzingorganophosphates or carbamates [7].BuChE catalysis of charged substrates and inhibition by

charged ligands are complex reactions. In particular, theyshow homotropic and heterotropic pseudo-cooperativeeffects. At intermediate substrate concentrations BuChEhydrolyses its optimal substrate BTC with rates exceedingthose expected by simple Michaelis–Menten dependence

and it is slightly inhibited by excess BTC [8,9]. In contrast,AChE shows only negative pseudo-cooperativity at highacetylthiocholine (ATC) concentrations [10,11]. Further-more, some cationic ligands, such as TAA salts, choline, andalso uncharged trialkylammonium compounds act as acti-vators or inhibitors, depending on both the concentration ofthe ligand and the substrate [12], the solvent and thepresence of cosolvent [13,14]. The goal of this work was tolocate the site of interaction between BuChE and tetra-alkylammonium (TAA) salts responsible for activation andto reach a mechanistic explanation of the phenomenon. Inparticular, tetraethylammonium (TEA) at the concentra-tions above 40 mM, reversibly inhibits the wild-type humanBuChE, but at the concentrations around 10 mM itaccelerates BuChE catalyzed hydrolysis of positivelycharged substrates.The active site serine, S198 in humanBuChE, is located at

the bottom of a 20-A deep cleft [15,16]. Ligands can bind ontwo distinct sites: a peripheral anionic site (PAS) located atthe mouth of the active site cleft, regarded as the substrate/ligand recognition site, and the ÔanionicÕ subsite of the CS[1,15]. Residues D70 (D72, Torpedo AChE numbering) andY332 (Y334) are the key elements of the PAS in humanBuChE [9,17]. For positively charged substrates, the CS isW82 (W84) where the binding occurs through p–cationinteractions [7,9,16]. Residue A328 (F330), which is also apart of this hydrophobic subsite, was found to be involvedin substrate/inhibitor binding, too [18]. To determine the siteinvolved in the effect of TAA salts, we carried out thesteady-state and progress curve analysis of BTC hydrolysisby recombinant wild-type human BuChE, by four selectedmutants (Y332A/D70G, Y332D/D70Y, W82A, A328Y)and by commercial horse serum BuChE in the presence ofTEA. Additionally, we tested the hydrolytic activity towardBTC of the mixture between horse enzyme and W82A

Correspondence to J. Stojan, Institute of Biochemistry,

Medical Faculty, Vrazov trg 2, 1000 Ljubljana, Slovenia.

Fax: + 386 1 5437641, Tel.: + 386 1 5437649,

E-mail: stojan@ibmi.mf.uni-lj.si

Abbreviations: AChE, acetylcholinesterase; BuChE, butyrylcholine-

sterase; CS, catalytic site; PAS, peripheral anionic site; TAA, tetra-

alkylamonium; TEA, tetraethylammonium; BTC, butyrylthiocholine;

DTNB, dithiobisnitrobenzoic acid; ATC, acetylthiocholine.

Note: a coordinate file of the homology built model of human wild-

type butyrylcholin-esterase with docked TEA can be downloaded from

http://www2.mf.uni-lj.si/�stojan/stojan.html

(Received 27 September 2001, revised 17 December 2001, accepted 19

December 2001)

Eur. J. Biochem. 269, 1154–1161 (2002) Ó FEBS 2002

recombinant human enzyme in order to see whether such alow activity mutant still can tie up substrate by binding itwith high affinity.

M A T E R I A L S A N D M E T H O D S

Chemicals and equipment

Butyrylthiocholine and buffer components of biochemicalgrade were purchased from Sigma Chemical Co. (St Louis,MO, USA). Tetraethylammonium chloride was obtainedfrom Fluka (Buchs, Switzerland), chlorpyrifos-oxon (CPO)was from Dow Chemical Co. (Indianapolis, IN, USA) anddiisopropylfluorophosphate (DFP)was fromAcrosorganicsFrance (Noisy-le-Grand, France).

Classical kinetic experiments were performed on aBeckman DU-7500 diode array spectrophotometer. Rapidkinetic measurements were curried out on a Hi-Tech(Salisbury, UK) PQ/SF-53 stopped-flow apparatus connec-ted to a SU-40 spectrophotometer and Apple E-II micro-computer, equipped with high speed AD converter.

Enzyme sources

Recombinant wild-type and mutant human BuChEs. Twoamino-acid residues (D70 and Y332) in the PAS and two(W82 and A328) in the CS, known to play a role in thebinding of positively charged ligands and in inhibitioncontrol of BuChE, were selected. The BuChE gene wasmutated to make the single mutants W82A and A328Y andthe double mutants Y332A/D70G and Y332D/D70Y.Wild-type and mutant enzymes were expressed in stablytransfected CHO cells as previously described [9].

Horse Serum BuChE. This was purchased from Worth-ington. It was chosen because the major difference betweenhuman and horse BuChEs at the cleft entrance is anadditional negative charge in the loop opposite to the omegaloop. As the two enzymes have 90% identical amino-acidresidues [19], we may see the horse enzyme, in terms ofperipheral site differences, as a human A277V/G283D/P285L triple mutant (W279, D283, I287 homologous, inTorpedo AChE).

Kinetic experiments and data analysis

Hydrolysis of BTC was measured by Ellman’s method in0.1 M potassium phosphate buffer, pH 7.0 at 25 °C [20].The substrate concentration ranges depended on the humanenzyme mutants: 0.6 lM to 90 mM for the wild-type, 0.015–100 mM for double mutants, 3–100 mM for W82A mutantand 0.015–3 mM for A328Y mutant; the substrate concen-trations used with the commercial horse serumBuChEwerebetween 0.05 and 10 mM. The concentration of enzymeactive sites E0, was determined by the method of residualactivity using CPO and/or DFP as the titrating reagents.Inhibition experiments were carried out at TEA concentra-tions from 0 to 100 mM.

Initial rate data in the absence of TEA showed, in mostenzymes, deviations from Michaelis–Menten kinetics: atintermediate substrate concentrations an apparent activa-tion is seen, while inhibition is detectable at the substrateconcentrations approaching maximum solubility. In order

to explain kinetically such observations, we analyzed thedata according to the six parameter model (Scheme 1)introduced previously [21].

S þ SE ÿ!bki SEA þ P1 ÿ!ak3 SE þ P2

"# K1 "# K2

S Sþ þ

S þ E ÿ!ki EA þ P1 ÿ!k3 E þ P2

Scheme 1.

In this scheme, E is the free enzyme, EA the acylatedenzyme, while SE and SEA represent the complexes withthe substrate molecule bound at the modulation site. Theproducts P1 and P2 are thiocholine and butyrate, respect-ively. K1 and K2 are the equilibrium constants for thesubstrate binding to the nonproductive site, while ki andk3 are the rate constants. a and b are the partitioningratios.Mixed equilibrium and steady-state assumptions [22] in

the derivation give the following rate equation:

v0 ¼E0k3½S� 1 þ a ½S�K2

� �½S� 1 þ ½S�

K2

� �þ

k3

ÿ1þ a

½S�K2

�ÿ1þ a

½S�K1

�ki

ÿ1þ b

½S�K1

� ð1Þ

The corresponding kinetic parameters were evaluated byfitting this equation to the initial rate data obtained in theexperiments using recombinant wild-type and the fourmutated human enzymes as well as the horse enzyme.For the analysis of the experiments in the presence of

TEA we made an extension of the model to allow thecompetition between TEA and BTC at both substratebinding sites and consequently also the occupation of thetwo sites by two TEA molecules (Scheme 2).

SEI ¢K7

Iþ SEþ S ÿ!bki SEAþP1 ÿ!ak3

SEþP2

"# "#K1 "#K2

S S Sþ þ þEI ¢

K5

IþEþ S ÿ!ki EAþP1 ÿ!k3

EþP2

þ þ þI I I"# "#K3 "#K4

IEI ¢K6

Iþ IEþ S ÿ!dki IEAþ P1 ÿ!ck3

IEþ P2

Scheme 2.

In this scheme, I stands for TEA and c and d are again thecorresponding partitioning ratios.An analogous derivation as described for Scheme 1 leads

to the following rate equation:

v0 ¼E0k3½S� 1þ a ½S�K2

þ c ½I�K4

� �½S� 1þ ½S�K2

þ ½I�K4

� �þ

k3

ÿ1þa ½S�K2

þc ½I�K4�ÿ

1þ½S�K1þ ½I�K3þ

½I�K5þ ½S�½I�K1K7

þ ½I�2K3K6

�ki

ÿ1þb ½S�K1

þd ½I�K3�

ð2Þ

Final evaluation of kinetic constants relevant for eachindividual enzyme was carried out by fitting this equation

Ó FEBS 2002 Activation inhibition of human butyrylcholinesterase (Eur. J. Biochem. 269) 1155

simultaneously to the data in the absence and presence ofTEA. We started with fixed values of parameters obtainedfrom the analysis without the inhibitor, to determine roughestimates of TEA binding parameters. Eventually, allparameters were released to achieve the best accordancebetween theoretical curves and the data. It should bestressed that some parameters in the reaction Scheme areclosely related to certain parts of data. For instance, theparameter a set to zero, would denote complete blocking ofdeacylation. Solubility maximum of the substrate, however,only allows to statistically anticipate the real value unless theclear plateau is reached [23].

The initial rate data for the W82A mutant differedsubstantially from the data for other enzymes. It appearedthat the hydrolysis of BTC by this enzyme obeyedMichaelis–Menten kinetics. In order to investigate thekinetics of this mutant more closely, we measured thehydrolysis of BTC catalyzed by the W82A mutant, bythe horse enzyme and by the mixture of the two enzymes ona stopped-flow apparatus. Aliquots of two solutions, onecontaining the enzyme and the other the substrate andDTNB were mixed together in the mixing chamber of theapparatus. The absorbance of the reaction mixture wasrecorded spectrophotometrically [20] at various concentra-tions of the substrate in the presence of 0.66 mM DTNB. Inorder to avoid possible product modulation, we stopped themeasurement when approximately 60 lM concentration ofthe product was formed. The stock solutions of the twoenzymes were prepared by dilution of the aliquots with the

same amount of buffer. The mixture was prepared bymixing together the aliquots without adding buffer. In thisway the mixture contained the same concentrations of thetwo enzymes as the solution of each individual enzyme. Theactivities of the three solutions were now tested at varioussubstrate concentrations in the range from 5 lM to 75 mM.The concentration of W82A was 16 lM and that of thehorse enzyme was 10 nM. The experimental conditions werethe same as in classical initial rate measurements (pH 7.0and 25 °C).We analyzed the data forW82Aby fitting a system of stiff

differential equations, that described the six-parametermodel in Fig. 1 under combined steady-state and equilib-rium assumptions (cf. [24]) to the data of all experimentalprogress curves simultaneously. Initial rates were obtainedas numerical derivatives at zero time of each progress curve.The same procedure was used to evaluate data obtainedwith commercial horse BuChE. The initial rates at varioussubstrate concentrations for the mixture of the two enzymes(W82A mutant and horse serum BuChE) were determinedanalytically by fitting the equation for single exponentialcurve to each individual progress curve and than takingderivatives at time zero.

Model building

Modelling was performed with WHATIF [25], starting withthe homology built model of human BuChE (CODEP06276) from Swiss-Model, an automated protein modeling

Fig. 1. pS curves for the hydrolysis of butyryl-

thiocholine catalyzed by the wild-type, by var-

ious human butyrylcholinesterase mutants and

by horse butyrylcholinesterase in 0.1 M phos-

phate buffer at pH 7.0 and 25 °C.

1156 J. Stojan et al. (Eur. J. Biochem. 269) Ó FEBS 2002

server [26], on an IBM compatible PC running under LINUX.Further refinement and the molecular dynamics werecarried out using the macromolecular simulation programCHARMM [27] on a cluster of four PCs. Topology and forcefield parameters for TEA from CHARMM distribution c27n1were used. Energy minimizations were performed with aconstant dielectric constant (e ¼ 1). Electrostatic force wastreated without cutoffs and van der Waals forces werecalculated with the shift method with a cutoff of 10 A. Alllysines and arginines were protonated and aspartic andglutamic acids were deprotonated. Histidines were neutralwith a hydrogen on N d1.

The corrections of the starting structure were performedin iterative steps as follows: the protein molecule was put inthe cube of water molecules (9091), subjected to 150relaxation steps (50 steps of steepest descent optimization,50 steps of optimization by adopted basisNewton–Raphsonmethod, 50 steps of steepest descent lattice optimization)and followed by 10 picoseconds constant pressure andtemperature (CPT) dynamic simulation (300 K, 1 bar, timestep of 1 fs) invoking the Ewald summation for calculatingthe electrostatic interactions. The last framewas devoided ofall watermolecules but those in the coat of 2.9 A around theprotein, relaxed with 100 optimization steps and checked bythe CHECK module in WHATIF. The unrealistic proteinportions were exchanged by the DGFIX command or byusing the SCAN LOOP command in SPDBVIEWER [26]. Aftersome 20 steps the check score improved substantially, so acontinuous simulation run was performed for 300 ps. Thefinal frame was used in a subsequent simulation involvingTEA. Docking was performed by superimposing TEA tothe trimethylamino group of docked acetycholine fromProtein Data Bank entry 2ACE [15]. In order to removeoverlapping between existing water molecules and newlyintroduced TEA we performed 150 relaxation steps (seepreviously) with fixed protein and TEA, followed by further150 steps without any constrains. Finally, the dynamicssimulation as described was run for 180 ps.

R E S U L T S

Initial rate data for the hydrolysis of BTC by five (wild-typeand four mutants) recombinant human BuChEs and horseBuChE are presented in Fig. 1. The pS diagrams show thatactivation at intermediate substrate concentrations is present

in all selected enzymes except in the W82A mutant, thatapparently obeys Michaelis–Menten kinetics. Additionally,to obtain comparable activities, the concentration of thismutanthad tobe raised almost hundred times in comparisonto the wild-type enzyme and theA328Ymutant andwas still10 times higher than the concentrations of the doublemutants. Experimental data in all diagrams cannot predictthe extentof inhibitionat saturating substrate concentrationsand in thecaseofW82Aenzymeeven theplateau/optimumisnot reached. On the other hand, the theoretical curves forother enzymes, that were obtained by putting kineticparameters from Table 1 into the Eqn 1, are in very goodagreement with the data and they stipulate completesubstrate inhibition. In other words, the fitting convergedwith the parameter a set to zero. It should be recalled that thedata in the absence and presence of TEA were used for thedetermination of the kinetic parameters listed in Table 1.Figure 2 shows the dependence of activity on the TEA

concentration of all enzymes at different substrate concen-trations. From the panels in this figure we can see someimportant characteristics: (a) activation at low TEAconcentrations is clearly visible in the wild-type enzyme, inthe A328Y mutant and in the ÔcompensatoryÕ mutant(Y332D/D70Y). It can only be perceived in Y332A/D70Gmutant but is absent in the W82A and in horse enzyme. (b)In the wild-type enzyme the activation is the most prom-inent at intermediate substrate concentrations. (c) Increas-ing inhibition at higher TEA concentrations is seen in allenzymes and the curves in the presence of the lowestsubstrate concentration approach to zero. This is the mostevident in the wild-type and A328Y enzymes. The lineardecrease in double mutants also indicates such a tendency.(d) Interestingly, TEA shows no activation of commercialwild-type horse BuChE at any concentration. Moreover,inhibition by TEA is very effective and the rate of hydrolysisclearly approaches zero even at the highest substrateconcentration. (e) Inhibition by TEA is the most prominentin the A328Y mutant of human BuChE. It occurs at muchlower TEA concentrations as in other enzymes, butactivation is also present. Unlike in the wild-type enzyme,activation in the A328Y mutant appears stronger at highersubstrate concentrations. (f) Regarding TEA inhibition, theW82A mutant is a special case: the inhibition emerges onlyat higher substrate concentrations indicating that either theinteraction of TEA with the free enzyme is very weak or an

Table 1. Characteristic constants for the interactions of various human butyrylcholinesterases and horse butyrylcholinesterse with butyrylthiocholine

and tetraethylammonium according to Scheme 2. Values in parenthesis are for the Michaelis–Menten mechanism (see Discussion).

Wild-type

(39.5 nM)

Y332D/D70Y

(220 nM)

Y332A/D70G

(245 nM)

W82A

(2.4 lM)

A328Y

(39 nM)

Horse BuChE

(10 nM)

ki (M)1Æs)1) 3.45 ± 0.46 · 106 6.88 ± 0.97 · 105 7.82 ± 0.57 · 105 (1.44) 88.2 ± 1.3 2.08 ± 0.12 · 107 8.51 ± 0.2 · 106

k3 (s)1) 467 ± 26 113 ± 2 132 ± 7 (0079) 18.0 ± 0.6 3800 ± 1700 1282 ± 87

K1 (lM) 46.9 ± 7.7 292 ± 136 60.9 ± 9.9 17.0 ± 3.4 25.5 ± 2.9 100 ± 9.9

K2 (mM) 77.2 ± 12.5 88.6 ± 4.2 85.3 ± 9.3 0.27 ± 0.05 1.03 ± 0.92 38.8 ± 9.3

a 0 0 0 0.0242 ± 0.0032 0.117 ± 0.059 0.376 ± 0.053

b 0.028 ± 0.005 0.440 ± 0.056 0.166 ± 0.015 0.0119 ± 0.0022 0.0114 ± 0.0017 0.134 ± 0.032

K3 (mM) 8.58 ± 3.91 23.8 ± 2.2 13.1 ± 3.4 – 0.325 ± 0.029 7.19 ± 2.8

K4 (mM) 177 ± 82.9 129.4 ± 38.6 397 ± 7.5 (340) 5.7 ± 0.4 39.0 ± 8.9 –

c 0.393 ± 0.096 0.257 ± 0.118 0.407 ± 0.030 – 0 –

d 0.926 ± 0.395 0.915 ± 0.236 0.511 ± 0.049 – 0.093 ± 0.013 –

K6 (mM) 2.97 ± 1.15 59.8 ± 27.0 296 ± 93 – 1.75 ± 0.22 93.7 ± 12.7

Ó FEBS 2002 Activation inhibition of human butyrylcholinesterase (Eur. J. Biochem. 269) 1157

independent binding of TEA and substrate at low concen-trations occurs on different sites.In order to find out the reason for the very low activity of

the W82A mutant, we tested the activity of the enzymemixture: W82A human BuChE and horse BuChE. Figure 3shows the progress curves obtained in this experiment andthe pS diagram of calculated initial rates. It can be clearly

seen that at low substrate concentrations the mixture of theenzymes is less active than the horse enzyme alone.Additionally, the theoretical curves for W82Amutant agreevery good with the experimental progress curves. It shouldbe stressed, however, that we could only achieve such anagreement with six-parameter model according to Scheme 1andnotwith simpleMichaelis–Menten reactionmechanism.

Fig. 2. Dependence of the activity of various

human BuChEs and horse BuChE on the con-

centration of TEA at various butyrylthiocholine

concentrations. BTC concentrations for

human enzymes are 15 lM, 25 lM, 50 lM,100 lM, 1 mM, 2 mM, 3 mM, from the lowest

to the highest curve. For horse enzyme BTC

concentrations are: 50 lM, 200 lM, 1 mM and

2 mM.

Fig. 3. Progress curves for the hydrolysis of

butyrylthiocholine catalyzed by the horse

butyrylcholinesterase, by the W82A mutant of

human butyrylcholinesterase and by the enzyme

mixture. Measurements were performed at

substrate concentrations ranging from 5 lM to

75 mM. Lower right panel shows the depend-

ence of the initial rates in the form of pS

curves.

1158 J. Stojan et al. (Eur. J. Biochem. 269) Ó FEBS 2002

Molecular dynamics calculations on the wild-type humanBuChE in water reveal after 180 ps very interesting TEApositioning. From the starting site in the vicinity of W82indole ring it moved upward the cleft and accommodatedjust belowY332 andD70, themajor constituents of the PASin human BuChE (Fig. 4). It seems that the A328 plays arole in this rapid movement (compareK3 values). In A328Ymutant and in vertebrate AChEs, the homologous F330 orY330 would prevent such positioning of TEA.

D I S C U S S I O N

The kinetic behavior of ChEs shows deviations from theMichaelis–Menten model. Although it has long beenbelieved that the only deviation in vertebrate AChEs isinhibition by excess substrate and that BuChEs are analo-gically activated, a more detailed investigations on insectAChEs, nematode enzymes and also BuChEs from varioussources reveal both phenomena [28].

Recent studies on various mutated enzymes showed, thatan appropriate mutation can mask one or the otherdeviation, but can also introduce it, if missing (cf.[10,23,29]). Exactly this can be seen from our experimentsin Fig. 1. While the pS curve for the wild-type enzymeshows clearly deviations at intermediate and very highsubstrate concentrations none of them is evident in theW82A mutant. However, the progress curves for W82A inFig. 3, which include the information at very low substrateconcentrations (the plateaus confirm complete hydrolysis),can only be explained by introducing an additionaldeviation from Michaelian kinetics into the reaction mech-anism (see parameters in Table 1). Moreover, we could alsospeculate that unless prevented by the solubility maximum,inhibition too might occur. Consequently, in W82Amutanta plateau/optimum shift towards higher substrate concen-trations appeared to take place. The explanation for such ashift may be the very low turnover of this mutant(kcat ¼ 100 min)1, [30]). The probability for the substrateto encounter the acylation site correctly is so low, that the

possible perturbations at the PAS are kinetically invisible.Slow acylation, again, should be the consequence ofchanged architecture in the catalytic site which firstly,cannot help to accommodate the substrate in formingMichaelis–Menten complex and secondly, enhance thestability of the acylated enzyme.In order to find out whether the extremely poor activity of

the W82A mutant is due to low affinity for the substrateand/or to the slow acylation-deacylation, we mixed W82Amutant with horse serum BuChE and tested the hydrolyticactivity of the mixture at low BTC concentrations. The aimwas to perform the experiments where the concentrations ofthe substrate and the W82A mutant were similar, while theconcentration of the horse enzyme was at least thousandtimes lower. Under such conditions it might be incorrectlyassumed that the low activity enzyme in such largeconcentration must tie up substrate by binding it to anumber of sites with varying affinity. Of course, only a singlespecific interaction is possible when the enzyme is a reactionpartner in stoichiometric amount to the substrate. We canconclude therefore that the lower activity of the mixture,compared to horse enzyme alone, is a consequence of goodaffinity of W82A for BTC (17 lM) and rather ineffectivecatalysis.Our experiment is the first kinetic evidence that in spite of

high substrate affinity the activity of a cholinesterasemay bevery low. It is well known that transition state analogues areextremely good inhibitors. As BTC is the substrate, asubstantional shift of the pS curve towards high concentra-tions suggests the inability to reach transition state rather tostabilize in it. It is well founded therefore to corroborate thisfinding with apparent activating deviation from Michaelis–Menten kinetics, which has also been reported for severalother cholinesterases. It was suggested that deviations fromMichaelis–Menten kinetics reflect the binding of thesubstrate molecule to the PAS [9,10,31,32]. In enzymes,showing apparent activation, the substrate affinity for thePAS appears to be relatively high, but overall catalyticpower of such enzymes is low [23]. It seems that inhibition at

Fig. 4. Stereo view of important active site residues in superimposed structures of Torpedo acetylcholinesterase (2ACE) and human butyryl-

cholinesterase.Docked as a tetrahedral adduct is butyrylthiocholine. The starting position of tetraethylammonium is position superimposed on the

substrate trimethyl group. An intermediate position of tetraethylammonium and final position (uppermost) after 180 picoseconds molecular

dynamics are also seen. Note the overlaping of TorpedoAChE residue F330 and tetraethylammonium in the final position. Corresponding A328 in

butyrylcholinesterase does not prevent the final orientation of tetraethylammonium. Labelling and numbering are according to human

butyrylcholinesterase.

Ó FEBS 2002 Activation inhibition of human butyrylcholinesterase (Eur. J. Biochem. 269) 1159

substrate concentrations in the range of high affinitybinding constant, mimics apparent activation.All this is supported by the inhibitory pattern of TEA on

various enzymes. The activation of some enzymes by TEAin low concentrations might be the consequence of thecompetition between the substrate and TEA at the PAS. Incomparison to the substrate, TEA inhibits the substratehydrolysis less effectively (d > b, see Table 1). This mightbe true for all enzymes showing activation by TEA,especially because in different enzymes it ÔappearsÕ atdifferent substrate concentrations (compare the wild-typeand the A328Y). At the highest TEA concentrations itcompetes with the substrate at both sites, thus, inhibiting theenzyme. The question rises, why some enzymes do not showactivation by TEA, but show substrate affinity at the PAS.Two explanations come tomind. The first one would be thatthe same substrate orientation at the PAS, in variousenzymes, cannot affect the events at the acylation site. As inthe W82A mutant, the missing bulky indole ring at thebottom of the cleft allows multiple substrate orientations atthe acylation site, thus preventing the influence of the ligandfrom the PAS. In this enzyme the weak inhibition at thehighest TEA concentrations corroborates the explanation.The second plausible possibility would be different orien-tation of the substrate at the PAS. It might be the case in theY332A/D70G double mutant and in the horse enzyme. Inaddition to an extra negative charge at themouth of the cleftin horse enzyme (D283, identical in Torpedo), the substrateaffinity at the PAS of both these enzymes appears lower asin other tested enzymes.The important role of PAS residues is further supported

by docking and dynamics simulations of TEA in the cleft ofthe wild-type human BuChE. Similar to the simulations onhuman AChE [33], a gradual movement is seen of the TEAmolecule from its starting position at p-electron interactionswith W82 indole ring, upwards to the vicinity of the twoPAS constituting residues, Y332 and D70. Although longersimulation run might reveal yet another position, such amovement indicates that at low concentrations TEA mightpreferentially occupy PAS, thus, preventing substrate tobind and to inhibit its own metabolization.Finally, we would like to discuss the significance of

kinetic parameters that we evaluated with our six parametermodel. One could argue at this point that the model canvery exactly reproduce the experimental data but does notreflect the realistic events during the catalytic process andthus the constants and their values are meaningless. Threepoints should be emphasized in this connection. Firstly, thekinetic model is one possible reduction of the traditionalreaction scheme generally valid for all cholinesterases. Itassumes that Michaelis–Menten complex is not accumula-ting, but it does not deny it. Such an approach is welljustified in the kinetic analysis and the simplification isintroduced according to well known principles [34]. More-over, some parameters can easily be interpreted with theclassical terms of Michalis–Menten kinetics. For instance,k3 in the model represents kcat and ki is in fact kcat/Km.Secondly, the six parameters are sufficient but also neces-sary to reproduce two deviations from Michaelian kinetics,for which more and more evidence exists, that they are arule rather an exception with cholinesterases. It should bevery clear, that many realistic models with more than sixparameters can equally precise reproduce the data, but only

with additional, more or less realistic assumptions. Ourkinetic model needs no additional assumptions and aunique set of six parameters can be evaluated if the twodeviations can be inspected.In conclusion, the major arguing point in the interpret-

ation of the results obtained by this model is a greatdifference between the binding of the substrate to themodifier site of free and acylated enzyme (K1 vs. K2). Wehave designed the experiment with the mixture of a normaland a low activity enzyme to confirm at least one highaffinity substrate binding site in the enzyme showing twodeviations fromMichaelian kinetics (see K1 for W82A). Weagree, that the model does not predict the exact spot andorientation in this binding but clearly explains the observeddeviation at that substrate concentration as homotropicinhibition rather as activation. The concentration dependentactivation-inhibition pattern by TMA and other quaternaryand tertiary ammonium compounds [12] strongly supportsthis interpretation.

A C K N O W L E D G E M E N T S

We thank Dr Oksana Lockridge (Eppley Institute, University of

Nebraska, Omaha, USA) for generously providing us with human

butyrylcholinesterase mutants. This work was partially supported by

the Ministry of Science and Technology of the Republic of Slovenia,

Grant No. P3-8720-0381 to J. S. and by DGA/DSP/STTC, grant no.

97/08 and 99 CO 029 to P. M.

R E F E R E N C E S

1. Massoulie, J., Pezzementi, L., Bon, S., Krejci, E. & Vallette, F.M.

(1993) Molecular and cellular biology of cholinesterases. Prog.

Neurobiol. 41, 31–91.

2. Mack, A.&Robitzki, A. (2000) The key role butyrylcholinesterase

during neurogenesis and neural disorders. Prog. Neurobiol. 60,

607–628.

3. Lockridge, O. (1992) In Pharmacogenetics of Drug Metabolism

(Kalow, W., ed.), pp. 15–50. Pergamon Press, New York.

4. Cashman, J.R., Perotti, B.Y.T., Berkman, C.E. & Lin, J. (1996)

Pharmacokinetics and molecular detoxication. Health Persp. 104,

23–40.

5. Mattes, C.E., Belendiuk, G.W., Lynch, T.J., Brady, R.O. &

Dretchen, K.L. (1998) Butyrylcholinesterase: an enzyme antidote

for cocaine intoxication. Addict. Biol. 3, 171–178.

6. Ballantyne, B. & Marss, T.C. (1992) Clinical and Experimental

Toxicology of Organophosphates and Carbamates. Butterworth-

Heinemann, Oxford.

7. Broomfield, C.A., Lockridge, O. & Millard, C.B. (1999) Protein

engineering of a human enzyme that hydrolyses V and G nerve

agents: design, construction and characterization. Chem. Biol.

Interact. 119/120, 413–418.

8. Masson, P., Froment, M.Th, Bartels, C.F. & Lockridge, O. (1996)

Asp 70 in the peripheral anionic site of human butyryl-

cholinesterase. Eur. J. Biochem. 235, 36–48.

9. Masson, P., Legrand, P., Bartels, C.F., Froment, M.Th, Schopfer,

C.M. & Lockridge, O. (1997) Role of aspartate 70 and trypto-

phane 82 in binding of succinyldithiocholine to human butyryl-

cholinesterase. Biochemistry 36, 2266–2277.

10. Radi\¢C.Z., Pickering, N.A., Vellom, D.C., Champ, S. & Taylor,

P. (1993) Three distinct domains in the cholinesterase molecule

confer selectivity for acetyl- and butyrylcholinesterase inhibitors.

Biochemistry 32, 12074–12084.

11. Ordentlich, A., Barak, D., Kronman, C., Ariel, N., Segal, B.,

Velan, B. & Shafferman, A. (1995) Contribution of aromatic

1160 J. Stojan et al. (Eur. J. Biochem. 269) Ó FEBS 2002

moities of tyrosine 133 and of the anionic subsite tryptophane 86

to catalytic efficiency and allosteric modulation of acetyl-

cholinesterase. J. Biol. Chem. 270, 2082–2091.

12. Erdoes, E.G., Foldes, F.F., Zsigmond, E.K., Baart, N. & Zwartz,

J.A. (1958) Acceleration of plasma colinesterase activity by qua-

ternary ammonium salts. Science 128, 92.

13. Clery, C., Heiber-Langen, I., Channac, L., David, L., Balny, C.

& Masson, P. (1995) Substrate dependence of amiloride- and

soman-induced conformation changes of butyrylcholinesterase

as evidence by high-pressure perturbation. Biochim. Biophys. Acta

1250, 19–28.

14. Levitsky,V.,Xie,W., Froment,M.Th,Lockridge,O.&Masson,P.

(1999) Polyol-induced activation by excess substrate of the D70G

butyrylcholinesterase mutant. Biochim. Biophys. Acta 1429, 422–

430.

15. Sussman, J.L., Harel, M., Frolow, F., Oefner, C., Goldman, A.,

Toker, L. & Silman, I. (1991) Atomic structure of acetylcholine-

sterase from Torpedo californica: a prototypic acetylcholine-

binding protein. Science 253, 872–878.

16. Harel, M., Sussman, J.L., Krejci, E., Bon, S., Chanal, P.,

Massoulie, J. & Silman, I. (1992) Conversion of acetyl-

cholinesterase to butyrylcholinesterase: modeling and mutagen-

esis. Proc. Natl Acad. Sci. USA 89, 10827–10831.

17. Nachon, F., Ehret-Sabatier, L., Loew, D., Colas, C. & van

Dorsselaer and Goldner, M. (1998) Trp82 and Tyr332 are

involved in two quaternary ammonium binding domains of

human butyrylcholinesterase as revealed by photoaffinity labeling

with [3H]DDF. Biochemistry 37, 10507–10513.

18. Saxena, A., Redman, A.M.G., Quian, N., Lockridge, O. & Doc-

tor, B.P. (1997) Differences in active site gorge dimensions of

cholinesterases revealed by binding of inhibitors to human

butyrylcholinesterase. Biochemistry 36, 14642–14651.

19. Wierdl, M., Morton, C.L., Dunks, M.K. & Potter, P.M. (2000)

Isolation and characterization of a cDNA encoding a horse liver

butyrylcholinesterase: evidence for CPT-11 drug activation. Bio-

chem. Pharmacol. 59, 773–781.

20. Ellman, G.L., Courtney, K.D., Andres, V. & Feathersone, R.M.

(1961) A new and rapid colorimetric determination of acet-

ylcholinesterase activity. Biochem. Pharmacol. 7, 88–95.

21. Stojan,J.,Marcel,V.,Estrada-Mondaca,S.,Klaebe,A.,Masson,P.

& Fournier, D. (1998) A putative kinetic model for substrate

metabolization by Drosophila acetylcholinesterase. FEBS Lett.

440, 85–88.

22. Cha, S. (1968) A simple method for derivation of rate equation for

enzyme-catalyzed reactions under the rapid equlibrium assump-

tion or combined assumptions of equlibrium and steady state.

J. Biol. Chem. 243, 820–825.

23. Golinik, M., Fournier, D. & Stojan, J. (2001) Interaction of

Drosophila acetylcholinesterases with D-tubocurarine: an explan-

ation of the activation by an inhibitor. Biochemistry 40, 1214–

1219.

24. Stojan, J. (1997) Analysis of progress curves in an acetyl-

cholinesterase reaction: a numerical integration treatment.

J. Chem. Inf. Comput. Sci. 37, 1025–1027.

25. Vriend, G. (1990) WHAT IF: a molecular modeling and drug

design program. J. Mol. Graph. 8, 52–56.

26. Guex, N. & Peitsch, M.C. (1997) SWISS-MODEL and the Swiss-

PdbViewer: an environment for comparative protein modeling.

Electrophoresis 18, 2714–2723.

27. Brooks, B.R., Bruccoleri, R.E., Olafson, B.D., States, D.J.,

Swaminathan, S. & Karplus, M. (1983) CHARMM: a program

for macromolecular energy minimization and dynamic calcula-

tions. Comput. Chem. 4, 187–217.

28. Marcel, V., Palacios, L.G., Pertuy, C., Masson. P. & Fournier, D.

(1998) Two invertabrate acetylcholinesterases show activation

followed by inhibition with substrate concentration. Biochem. J.

329, 329–334.

29. Marcel, V., Estrada-Mondaca, S., Magne, F., Stojan, J., Klaebe,

A. & Fournier, D. (2000) Exploration of the Drosophila acet-

ylcholinesterase substrate activation site using a reversible

inhibitor (Triton X-100) andmutated enzymes. J. Biol. Chem. 275,

11603–11609.

30. Masson, P., Xie, W., Froment, M.-T. & Lockridge, O. (2000)

Effects of mutations of active site residues and amino acids

interacting with omega loop on substrate activation of butyr-

ylcholinesterase. Biochim. Biophys. Acta 1564, 166–176.

31. Szegletes, T., Mallender, W.D. & Rosenberry, T.L. (1998)

Nonequilibrium analysis alters the mechanistic interpretation of

inhibition of acetylcholinesterase by peripheral site ligand. Bio-

chemistry 37, 4206–4216.

32. Mallender, W.D., Szegletes, T. & Rosenberry, T.L. (2000) Acet-

ylcholine binds to Asp74 at the peripheral site of human acet-

ylcholinesterase as the first step in the catalytic pathway.

Biochemistry 39, 7753–7763.

33. Van Belle, D., De Maria, L., Iurcu, G. & Wodak, S.J. (2000)

Pathways of ligand clearance in acetylcholinesterase by multiple

copy sampling. J. Mol. Biol. 298, 705–726.

34. Cleland, W.W. (1977) Determining the chemical mechanisms of

enzyme-catalyzed reactions by kinetic studies. Adv. Enzymol. 45,

273–387.

Ó FEBS 2002 Activation inhibition of human butyrylcholinesterase (Eur. J. Biochem. 269) 1161