INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS: ELECTROSTATIC... RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

Cholinesterases (ChEs) are a family of enzymes that fall broadly into two types: acetylcholinesterase (AChE) and butyrylcholinesterase (BChE). They are distinguished primarily by their substrate specificity: AChE hydrolyzes the natural neurotransmitter acetylcholine (ACh) faster than choline esters with bulkier acyl chains; thus it is much less active on the synthetic substrate, butyrylcholine (BCh). In contrast, BChE displays similar activity toward the two substrates (Chatonnet and Lockridge, 1989).

Vertebrates contain both AChE and BChE, which probably originate from the duplication of a single ChE gene (Massoulié et al., 1993). Insects possess a single ChE gene coding for an enzyme with a specificity intermediate between those of AChE and BChE (Massoulié et al., 1993; Taylor and Radic, 1994), while in certain nematode species, up to four ChE genes have been identified (Grauso et al., 1998). The principal physiological function of AChE has long been known to be termination of impulse transmission at cholinergic synapses by rapid hydrolysis of ACh, but the biological role of BChE is still an open question (Chatonnet and Lockridge, 1989).

ChEs are able to catalyze the rapid breakdown of a variety of esters, both cationic and neutral (Quinn, 1987). The highest catalytic rate is displayed by AChE hydrolysis of ACh, at a rate approaching the diffusion-controlled limit (Rosenberry, 1975b; Hasinoff, 1982; Bazelyansky et al., 1986). This efficiency is a common characteristic of ChEs. A tally of the bimolecular rate constants (k_cat/K_M) among ChEs for their best substrates shows a spread of less than an order of magnitude. Values range from 1.6 × 10⁸ M¹ s¹, for hydrolysis of ACh by Electrophorus electricus AChE (EeAChE) (Rosenberry, 1975a), to 4.0 × 10⁷ M¹ s¹ for BChE hydrolysis of both ACh and BCh (Vellom et al., 1993).

Catalytic efficiency of ChEs for neutral substrates is also very high. Values of k_cat/K_M for hydrolysis by AChE of ACh and of its neutral isoster 3,3-dimethylbutylacetate (DBA) do differ by ~40-fold at physiological salt concentration (Hasan et al., 1981), but there is evidence that the hydrolysis of the thio analog of DBA is also diffusion-controlled (Bazelyansky et al., 1986). Moreover, it has been shown that BChE turns over o-nitrophenyl butyrate (o-NPB) faster than it breaks down butyrylthiocholine (Masson et al., 1997).

Studies of the pH and charge dependence of catalytic hydrolysis of substrates and of binding of reversible inhibitors suggested that the active site of ChEs contains two major subsites, the "esteratic" and the "anionic" (Wilson and Bergmann, 1950), corresponding, respectively, to the catalytic machinery and the choline-binding pocket (Froede and Wilson, 1971). The high bimolecular association constants for cationic ligands and their ionic strength dependence suggested a high charge density in the active site. This led to the prediction that up to nine negative charges were present in the "anionic" site (Rosenberry and Neumann, 1977; Nolte et al., 1980). A second "anionic" site, which became known as the "peripheral" anionic site (PAS), was proposed based on binding of bis-quaternary ammonium compounds (Bergmann et al., 1950). Binding of ligands to the peripheral anionic site causes inactivation of the enzyme, though the mechanism of inhibition is not clear. It has been speculated that the PAS is involved in the phenomenon of substrate inhibition and activation through binding of a second substrate molecule. Its function may involve either allosteric modification of the active site (Radic et al., 1991; Shafferman et al., 1992; Barak et al., 1995) or alteration of the traffic of substrate and products by blocking access to the catalytic machinery (Berman and Nowak, 1992; Haas et al., 1992; Schalk et al., 1992). In addition, there is evidence for an involvement of the PAS in functions distinct from catalysis. Recent studies have presented evidence for a role of the PAS of AChE in neurite regeneration and outgrowth (Layer et al., 1993; Willbold and Layer, 1994; Jones et al., 1995; Srivatsan and Peretz, 1997) and in the growth and differentiation of spinal motor neurons (Bataillé et al., 1998).

Catalysis by ChEs occurs by a mechanism similar to that of the serine proteases, via an acyl-enzyme intermediate. It is assumed to involve a tetrahedral transition state produced by nucleophilic attack on the substrate by a serine, followed by general-base catalysis assisted by a histidine. The transition state subsequently collapses to the acyl-enzyme by general-acid-catalyzed expulsion of choline (Quinn, 1987). Solution of the three-dimensional (3D) structure of AChE from Torpedo californica (TcAChE) (Sussman et al., 1991) was followed by determination of the crystal structures of several complexes of AChE with specific inhibitors (Harel et al., 1993, 1995, 1996). Analysis of these structures, taken together with systematic site-directed mutagenesis studies (Radic et al., 1991, 1993; Shafferman et al., 1992; Barak et al., 1995), has permitted identification of the key functional residues in the active site and contributed to clarification of their role in recognition of substrates and inhibitors. (A comprehensive and updated list of references for cholinesterase mutants can be obtained through the ESTHER server, http://meleze.ensam.inra.fr/cholinesterase (Cousin et al., 1997).) It has thus been possible to construct a picture of the structural factors governing both the mechanism and specificity of ChEs. At the same time the data obtained have given rise to a new set of still unanswered questions.

The structure/function relationships that have emerged over the past seven years paint a picture of a family of rapid enzymes specific for cationic substrates that have evolved to very high catalytic efficiency by adopting some rather counterintuitive solutions. The active-site serine of TcAChE, S200, was unexpectedly found to be located near the bottom of a 24-Å-deep gorge, ~4.4 Å wide at its narrowest point and 8.0 Å wide at its mouth. This serine forms a catalytic triad with H440 and E327. [Residue numbers follow the guidelines established at the 1992 OHOLO meeting, Eilat (Massoulié et al., 1993). The numbering used is that of the sequence of TcAChE. When species-specific numbers are employed, the homologous position in TcAChE will follow, printed in italics and appearing in parentheses.] Contrary to the assumption of a concentration of negative charges within the active site (Nolte et al., 1980), no more than two formal negative charges were found near the active site serine (E199 and E443). In fact, the walls of the gorge were found to be lined by the side-chains of 14 conserved aromatic residues (Sussman et al., 1991). In analogy to studies of ACh binding to model hosts (Dougherty and Stauffer, 1990), it was suggested that the binding of ACh to the sites within the gorge involved in substrate recognition is stabilized by interactions between the quaternary ammonium group of ACh and the  electrons of some of the conserved aromatic residues of AChE. This hypothesis was confirmed by inspection of the structures of complexes of AChE with various quaternary inhibitors (Harel et al., 1993). Analysis of the 3D structure of the complex of AChE with the transition-state analog, m-trimethylammonium trifluoroacetophenone (TFK⁺), has highlighted the specific intermolecular interactions between the substrate and the active site that are responsible for the catalytic efficacy of AChE. These structural data suggest that the particular active-site configuration of AChE allows the enzyme to efficiently sequester the acylation transition state in a preorganized polar environment formed by the oxyanion hole, consisting of the main-chain N-H dipoles, G118, G119, and A201, and the side chains of key aromatic residues such as W84 and F330 (Harel et al., 1996). This environment is able to stabilize the catalytic transition state by providing it with larger electrostatic stabilization than in the solvent and is the basis of the catalytic power of the ChEs (Fuxreiter and Warshel, 1998).

Inspection of the overall 3D structure of TcAChE revealed a marked asymmetrical surface distribution of charged residues. These residues were shown to segregate roughly into a "northern" negative hemisphere, considering the mouth of the gorge as the north pole, and a "southern" positive one, giving rise to a large "dipole moment" roughly oriented along the axis of the active-site gorge (Ripoll et al., 1993; Tan et al., 1993; Antosiewicz et al., 1994; Felder et al., 1997). The magnitude of this "dipole moment" was estimated, by electrooptical measurements on Bungarus fasciatius AChE (BfAChE), to be ~1000 Debyes (Porschke et al., 1996). The biological significance of these unusual electrostatic properties has been the subject of much controversy. It has been suggested that the "macrodipole" might act to steer cationic ligands to the mouth of the gorge (Sussman et al., 1991; Tan et al., 1993), where they would bind to the aromatic residues lining it and subsequently be committed to moving down the gorge, toward the active site, in a fashion similar to that of an affinity column (Sussman et al., 1991). Calculations of the rates of encounter of charged ligands and substrates based on Brownian dynamics (BD) simulations predicted that the electrostatic properties of AChE would be responsible for a rate enhancement of up to 240-fold (Zhou et al., 1996). The possibility of a large electrostatic steering effect on positively charged substrates has raised the question of the route of clearance of choline (Ch), the cationic product of enzymatic hydrolysis, which would seem to be trapped at the bottom of the gorge by a strong electric field.

Molecular dynamics (MD) simulations have suggested that an alternative route of access to the active site might open via the concerted movement of residues W84, V129, and G441 (Gilson et al., 1994), while a more recent MD study has identified a number of "side entrances" to the gorge, all involving concerted movements of a number of side chains in the  loop (C67-C92) of AChE (Wlodek et al., 1997b). As yet, no simple and predictive model for the clearance of the products of hydrolysis of ACh or BCh has been introduced into the treatment of the molecular traffic of substrates and ligands through the active site of ChEs.

The importance of electrostatic interactions in the steering of cationic substrates to the active site of ChEs was challenged by a study of a series of mutants of human recombinant AChE (hAChE), in which up to seven negative residues around the outer rim of the gorge were neutralized, thus substantially reducing the magnitude of the "macrodipole," without producing large changes in the second-order hydrolysis rate constant (Shafferman et al., 1994). The contradiction between data documenting the electrostatic properties of ChEs and the apparent lack of correlation between their experimental modification and a major catalytic effect has been the topic of various studies. Antosiewicz and co-workers (Antosiewicz et al., 1994, 1995a,b, 1996; Antosiewicz and McCammon, 1995) have attempted to correlate the electrostatic properties of AChE with the on rates for charged ligands and the second-order hydrolysis rate constants. The picture emerging from their work supports the existence of an electrostatic steering effect in ChEs. Moreover, calculations performed by the same group, on structural models of the mutants analyzed kinetically by Shafferman et al. (1994), account for the small changes in catalytic rates observed. This reconciles the hypothesis of a role for electrostatics in accelerating the encounter between the enzyme and reactive species with experimental data that seemed to argue against it (Antosiewicz et al., 1995b). Nevertheless, these studies were unable to strongly correlate experimental results with any one particular aspect of the electrostatic properties of the ChEs, since neither the total charge nor the dipole moment fully accounted for the electrostatic steering of ligand to the active site.

Calculations of the potential gradient along the axis of the gorge and of its dependence on salt concentration have been performed both for wild-type (WT) AChE of different species and for a series of surface and active-site mutants (Antosiewicz et al., 1995b; Wlodek et al., 1997a; Felder et al., 1997). These calculations have revealed that AChEs display a similar negative potential gradient, beginning several angstroms outside the gorge entrance, and continuing down the gorge toward the active site. This potential is not correlated with ChE surface charge distribution, but is due to a combined effect of the overall charge distribution in the ChE molecule, including the effect of several -helix dipoles. The results of these studies suggest the existence of a long-range electrostatic interaction, attributable to this potential gradient, that contributes to both the enhancement of encounter rates between cationic ligands and the catalytic machinery buried at the bottom of the active-site gorge of ChEs. Radic et al. (1997) have performed a detailed analysis of the influence of electrostatics on the kinetics of ligand binding to AChE. This study focused on the ionic-strength dependence of the binding of reversible inhibitors to AChE after neutralization, by site-directed mutagenesis, of anionic side chains in the surface area around the entrance to the active-site gorge, in the PAS and in the active center. Comparison of the experimental data to BD simulations of the on rates of the ligands revealed good agreement for surface mutants, while predictions were less accurate when some key residues in the PAS were neutralized. The results were interpreted in the framework of two distinct types of electrostatic interactions: the first, dependent on salt concentration, causing acceleration of the initial encounter rates of cationic ligands with the enzyme, and the second, independent of salt concentration, resulting in trapping of these ligands by specific residues in the PAS or within the active site. The presence of a trap for cationic ligands in AChE and BChE has been proposed in several other studies (Rosenberry and Neumann, 1977; Hasinoff, 1982; Hosea et al., 1996; Masson et al., 1996, 1997), but no analytical and quantitative treatment of the effect of a trapping surface on the encounter rate of cationic ligands with the ChEs has been advanced until now.

In the following sections we will present a model for the molecular traffic of neutral and cationic ligands, substrates, and products through the active site of ChEs, and we will evaluate the contributions of electrostatic interactions to the traffic of cationic species.

First, we will illustrate a common treatment for the diffusion of both neutral and cationic ligands toward an enzyme characterized by a buried active site. Subsequently, we will show how the specificity of ChEs for cationic ligands and substrates can be treated by introducing two additional modules to this common treatment:

1.

A module that incorporates the effects of short-range and ionic strength independent interactions between key residues in the area of the entrance to the active-site gorge and the quaternary ammonium moiety of cationic substrates and ligands. This local module will be shown to describe the effects of a putative trapping mechanism for cationic species. The effect of this surface trap on the enhancement of encounter rates between cationic ligands and ChEs will be analyzed quantitatively by correlating the electrostatic potentials in the area surrounding the entrance to the active-site gorge with the experimentally measured encounter rates of cationic ligands with ChEs.

2.

A module that incorporates the effects of long-range and ionic-strength dependent interactions. This global module, which is shown to describe the steering of cationic substrates and ligands to the gorge floor, can be analyzed quantitatively by evaluating the overall electric potential around the entrance and within the active-site gorge and its dependence on ionic strength.

The clearance of positively charged products and ligands from the active-site gorge will be considered as analogous to the Brownian migration of a charged particle out of a one-dimensional box against an electrostatic potential linear in distance along the length of the box. For neutral molecules, the value of this potential will be set to zero, while for charged species we will introduce values of the gorge potential either derived from experimental data or calculated with the Poisson-Boltzmann equation (PB).

This modular approach will enable us to rationalize the apparent paradox of electric fields, which act to steer cationic species toward the active site while at the same time not hindering the clearance of positively charged products. The results obtained will be used to discuss the mechanisms involved in the emergence of the specificity of ChEs for cationic substrates from the framework of an already very efficient catalytic machinery for the hydrolysis of small esters.

	METHODS: ELECTROSTATIC CALCULATIONS AND HOMOLOGY MODELS

TOP ABSTRACT INTRODUCTION METHODS: ELECTROSTATIC... RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

We calculated the electrostatic potential along the active-site gorge of ChEs, by solving the PB equation, to study the influence on gorge electrostatic potentials of salt concentration and of the neutralization of a number of key residues in the area of the surface cationic trap and in the active-site gorge. The electrostatic calculations were performed on the following enzymes (where crystallographic coordinates are available, the PDB ID code will follow in parentheses): AChE from the following species: Torpedo californica (TcAChE-PDB ID: 2ACE), mouse (mAChE-PDB ID 1MAH), Bungarus fasciatus (BfAChE), Drosophila melanogaster (DmAChE), and human (hAChE); and human BChE (hBChE). Where the crystallographic coordinates were not available, homology models were constructed. The residues to be mutated were chosen on the basis of their presumed involvement in the recognition of cationic substrates and in the contribution to the steering of cationic ligands toward the active site of ChEs (Shafferman et al., 1992, 1994; Radic et al., 1993, 1995; Barak et al., 1995; Masson et al., 1996; Ordentlich et al., 1996). Three classes of mAChE mutants, as shown in Table 1, were generated by homology modeling from the respective WT structures. The positions of these mutations on the 3D fold of ChE are illustrated in Fig. 1. In the first class, based on the studies of Shafferman et al. (1994) and Radic et al. (1997), we generated five mutant structures in which up to seven acidic residues on the enzyme surface near the gorge entrance were neutralized. In the second and third classes, the mutations focused on the modification of residues in the area of the cationic trap and at the bottom of the gorge, respectively, to analyze the correlation between local and overall electrostatic potentials and the encounter rates of cationic ligands and substrates.

View larger version (58K): [in this window] [in a new window]   FIGURE 1   Ribbon diagram of AChE showing the relative positions of the mutated residues. The axis of the gorge is indicated by the yellow bar; the side chains of the surface mutants are colored in red; E199, at the bottom of the gorge, is in green; and D72, the main component of the cationic trap, is in purple.

Construction of homology models

In brief, models were constructed by use of the automated knowledge-based model-building tool resident on the Swiss-Model Server (Peitsch, 1996) (http://www.expasy.ch/swissmod/SWISS_MODEL.html), employing 3D structures of TcAChE (PDB ID 2ACE, 1ACJ, 1FSS) and mAChE (PDB ID 1MAH) as templates. The automatic procedure involved alignment of the sequence to be modeled with the template sequences, by application of the BLAST algorithm (Altschul et al., 1990). Regions of sequence similarity were automatically selected and employed to build a framework for the model structure. Missing loops were automatically constructed by searching the PDB database, employing either the best fitting fragment corresponding to the sequence or a framework constructed by the average of the five best fragments. The last step of the procedure involved automated rebuilding of side chains, verification of the quality of the model, and refinement of the final structure by energy minimization and molecular dynamics. The models were then checked again manually, and any missing part that was not successfully built automatically was added manually as described previously (Felder et al., 1997).

Calculation of electrostatic potentials along the active-site gorge

The electrostatic potential along the active-site gorge was calculated by generating a string of dummy atoms at 1-Å intervals along the gorge axis (Fig. 1) and evaluating the electric potential for the position of each dummy atom. The axis of the active-site gorge in the structures examined was defined as extending from atom I444-CD (gorge bottom) to the center of mass of atoms E73-CA, N280-CB, D285-CG, and L333-O (gorge entrance) (Antosiewicz and McCammon, 1995). The portion of the gorge extending from the bottom to S200-OG was defined as the binding region (~4 Å long), and the remainder as the transit region (~20 Å long). The width of the gorge mouth was measured by taking the average of the values of gorge radii originating from the center of mass of atoms E73-CA, N280-CB, D285-CG, and L333-O and intersecting the gorge rim at the CB atoms of residues D72, W279, D273, and D365. To study the local electrostatic potentials generated in the gorge by the residues involved in the cationic trap, we defined a region of the gorge extending from its mouth (as defined above) to a depth of 6 Å. The value of the electrostatic potential along this portion of the axis was then averaged and taken as a measure of the local potential in the trap region.

The PB equation was solved by the finite-difference method (Warwicker and Watson, 1982), using the QDIFFXS algorithm of version 3.0 of DelPhi (Gilson and Honig, 1988; Honig and Nicholls, 1995). A grid of up to 90 Å³ was used. Calculations were performed, using an initial coarser grid with a 35-Å border and 1.45-Å grid spacing and subsequently focusing onto a second grid with a 10-Å border and 0.89-Å spacing. The internal and external dielectric constants were fixed at values of 2 and 80, respectively. Calculations were performed for salt concentrations of 5, 145, and 670 mM. The Stern ion exclusion layer was set at 2 Å, and the dielectric boundary between protein and solvent was constructed using a probe radius of 1.4 Å. Calculations were performed at 298.15 K and pH 7.0, using the Parse partial atomic charge and radius set (Sitkoff et al., 1994). The protonation states of the ionizable amino acids were assigned by examination of the solvent accessibility of their side chains in the 3D structure. On the basis of this analysis, all Glu, Asp, Lys, and Arg residues were set to be fully ionized, and the average charge on the active-site H440 and on all other histidines was set to zero (M. K. Gilson, personal communication). Potential values are expressed in kT/e units (1 kT/e = 25.6 mV = 0.593 kcal/mol/e).

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS: ELECTROSTATIC... RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

Because our study focuses on providing a model for the diffusive and binding events occurring before and after the bond rearrangement and cleavage steps of catalysis, the experimental parameters best suited for comparison with our model are the on rates and binding constants (k_on and K_i) of transition-state analogs, whose magnitudes are dependent only on the processes of diffusion and binding to the active site of ChEs. The contributions of both long- and short-range electrostatic interactions to the stability of the complex formed between a charged transition-state analog and AChE (or BChE) and to the on rate of the ligand can be evaluated by comparing the values of K_i and k_on for the charged species with those for a neutral isosteric ligand. These results can then be used to segregate the contributions of electrostatic interactions to molecular traffic from their effect on the chemical steps of catalysis.

Peptidyl trifluoromethyl ketones have been used extensively as transition-state analogs of various serine hydrolases (Imperiali and Abeles, 1986; Takahashi et al., 1988; Allen and Abeles, 1989b; Brady et al., 1989). In particular, they have been employed to assess the role of global electrostatic interactions in the stabilization of the catalytic transition state of subtilisin BPN' (Jackson and Fersht, 1993). A large body of kinetic evidence demonstrates that another series of trifluoromethyl ketones serves as transition-state analogs of ChEs (Allen and Abeles, 1989a; Nair et al., 1993, 1994). The structure of the complex of the phenyl trifluoromethyl ketone, TFK⁺, with TcAChE has been solved by x-ray crystallography, illustrating in detail the structural interactions responsible for the tight binding of this particular transition-state analog (Harel et al., 1996). The effect of salt concentration and of the mutation of residues D72, E199, and several negatively charged surface residues on the K_i of both TFK⁺ and its isosteric neutral analog, m-tertbutyltrifluoroacetophenone (TFK⁰) (Quinn et al., 1995; Radic et al., 1995, 1997; Hosea et al., 1996), have been measured.

Accordingly, the results of our calculations, presented in the following sections, will be compared to experimental data collected for TFK⁺, TFK⁰, and another ligand, N-methylacridinium (NMA), which has been employed to study the role of electrostatic properties in ChE catalysis (Nolte et al., 1980).

Part I: Molecular trafficdiffusion

A common treatment for the diffusion of cationic and neutral isosteric ligands to a buried active site points toward two different limits for diffusion to the active site of ChEs

View larger version (56K): [in this window] [in a new window]   FIGURE 2   Schematic diagram of the model for diffusion of neutral and cationic ligands and substrates to the active site of ChE. The spherical model is superimposed on a slab view of TcAChE showing the active-site gorge and the positions of D72 and of the catalytic serine, S200. The inner and outer spherical caps are colored in red. The inner cap includes the catalytic machinery of AChE, as can be seen by the position of S200.

(1)

(2)

View larger version (13K): [in this window] [in a new window]   FIGURE 3   (A) Molecular structure of the transition-state analogs, TFK+ and TFK0, and of NMA. (B) Molecular structures of ACh, BCh, and some neutral substrates of the ChEs.

7

1

1

1

(3)

1

1

Module I: a surface trap for cationic species operating via short-range local interactions

(4)

View larger version (29K): [in this window] [in a new window]   FIGURE 4   (A) Correlation between average potentials in the trap region and encounter rates for TFK+ at high (600 mM), physiological (145 mM), and low (2 mM) salt concentrations. The potential values are calculated for WT mAChE and for the following mutants (the abbreviations used in the list on the right of the figure are given in parentheses): D72N (D72N); D72N/D275V/D278N (72/275/278); E82Q/E89Q/D275V/D278N (4m); E82Q/E89Q/D275V/D278N/D365N (5m); E82Q/E89Q/D275V/D278N/E285Q/D365N (6m). (B) Correlation between overall gorge potential and encounter rates for TFK+ at high (600 mM), physiological (145 mM), and low (2 mM) salt concentrations. The potential values are calculated for WT mAChE and for the following mutants (the abbreviations used in the list on the right of the figure are given in parentheses): D275V (D275V); D275V/D278N (275/278); E82Q/E89Q/D275V/D278N (4m); E82Q/E89Q/D275V/D278N/D365N (5m); E82Q/E89Q/D275V/D278N/E285Q/D365N (6m); E199Q (E199Q).

1.	Cation- interactions are predominantly electrostatic, involving the interaction of the cation with the large, permanent quadrupole moment of the aromatic ring (Dougherty, 1996).
2.	Gas-phase measurements of binding energy of cations to benzene and to toluene have shown that in this phase a cation would preferentially bind to an aromatic compound rather than to water (Sunner et al., 1981).
3.	In an aqueous environment, a pocket lined with the side chains of amino acids such as Trp, Phe, and Tyr can efficiently compete with full water solvation for the stabilization of a positive charge, because of the sizable quadrupole moment of the rings of aromatic residues (Luhmer et al., 1994). The importance of cation- interactions in the catalytic function of ChEs has been confirmed by a large body of structural (Sussman et al., 1991; Harel et al., 1993, 1996) and kinetic (Ordentlich et al., 1993; Radic et al., 1993; Nair et al., 1994; Barak et al., 1995) data.

Module II: the potential gradient along the active-site gorge is responsible for steering cationic species to the active site of ChE

View larger version (22K): [in this window] [in a new window]   FIGURE 5   Correlation between overall gorge potentials of WT AChEs and encounter rates of TFK+ and NMA as a function of salt concentration. The potential values were calculated for TcAChE for salt concentration values varying from 2 to 600 mM.

View larger version (38K): [in this window] [in a new window]   FIGURE 6   Electrostatic potentials along the gorge axis for ChEs. The potentials were calculated for TcAChE, mAChE, hAChE, DmAChE, BfAChE, and hBChE.

Part II: quantitative treatment of the electrostatic forces underlying ChE-ligand binding interactions

Contribution of electrostatic interactions to the binding of transition-state analogs to ChEs

1.	A global term arising from long-range electrostatic contributions that are screened at high salt concentrations.
2.	A local term that can be identified with cation- and local, short-range electrostatic interactions between the charged substrates and key binding residues, such as W84, in the active-site gorge.

(5)

(6)

(7)

(8)

(9)

(10)

(11)

View this table: [in this window] [in a new window]   TABLE 5   Experimental values of Ki for TFK+ and TFK0 employed in the calculation of Gcharge, GL, and Gg

Calculation of the potential inside the active-site gorge of AChE from G values for the activated complex of AChE with TFK⁺ and TFK⁰

(12.1)

(12.2)

(13.1)

(13.2)