Interaction Kinetics of Reversible Inhibitors and Substrates with Acetylcholinesterase and Its Fasciculin 2 Complex*

Fasciculin 2 (Fas2), a three-fingered peptide of 61 amino acids, binds tightly to the peripheral site of acetylcholinesterases (AChE; EC 3.1.1.7), occluding the entry portal into the active center gorge of the enzyme and inhibiting its catalytic activity. We investigated the mechanism of Fas2 inhibition by studying hydrolysis of cationic and neutral substrates and by determining the kinetics of interaction for fast equilibrating cationic and neutral reversible inhibitors with the AChE·Fas2 complex and free AChE. Catalytic parameters, derived by eliminating residual Fas2-resistant activity, reveal that Fas2 reducesk cat/K m up to 106-fold for cationic substrates and less than 103-fold for neutral substrates. Rate constants for association of reversible inhibitors with the active center of the AChE·Fas2 complex were reduced about 104-fold for both cationic and neutral inhibitors, while dissociation rate constants were reduced 102-to 103-fold, compared with AChE alone. Rates of ligand association with both AChE and AChE·Fas2 complex were dependent on the protonation state of ionizable ligands but were also markedly reduced by protonation of enzyme residue(s) with pK a of 6.1–6.2. Linear free energy relationships between the equilibrium constant and the kinetic constants show that Fas2, presumably through an allosteric influence, markedly alters the position of the transition state in the reaction pathway. Since Fas2 complexation introduces an energetic barrier for hydrolysis of substrates that exceeds that found for association of reversible ligands, Fas2 influences catalytic parameters by a more complex mechanism than simple restriction of diffusional entry and exit from the active center. Conformational flexibility appears critical for facilitating ligand passage in the narrow active center gorge for both AChE and the AChE·Fas2 complex.

The interaction between acetylcholinesterase (EC 3.1.1.7) (AChE) 1 and fasciculin 2 (Fas2) is a well studied example of a tight binding peptide-protein complex. The bound "three-fin-gered" peptide toxin, isolated from venom of Elapidae snakes, resides at the opening of the AChE active center gorge, thus sealing off the entry to the gorge and resulting in virtually complete inhibition of hydrolysis of the neurotransmitter acetylcholine. Static snapshots of the complex structure revealed in crystals of mouse AChE⅐Fas2 (1), Torpedo californica AChE⅐Fas2 (2) and human AChE⅐Fas2 (3) complexes indicate that the "solvent-accessible" surfaces of two molecules overlap in a nearly continuous area of more than 1000 Å 2 around the entry to the AChE active center gorge. The complex, however, retains residual catalytic activity whose magnitude was reported to vary between species, being about 3% in human AChE (4) and about 0.09% in mouse AChE (5), when measured with acetylthiocholine (ATCh). Inhibition of residual catalytic activity of both complexes by conjugation of the active center serine by either organophosphate or trifluoroacetophenone (TFK) inhibitors, showed only a 1-order of magnitude reduction in reaction rates of cationic inhibitors for the AChE⅐Fas2 complex when compared with uncomplexed AChE (5). Rates of association and dissociation of reversible ligand N-methylacridinium were found, however, to be reduced by 3-4 orders of magnitude in the AChE⅐Fas2 complex (6).
Recently, Mallender et al. (7) found that human AChE contained Fas2-resistant esterase activity in purified preparations. Monitoring the phosphorylation rates of fluorescent organophosphates, which exhibit negligible dephosphorylation of the enzyme, enabled these investigators to show that reactions of organophosphates with AChE⅐Fas2 were 2-3 orders of magnitude slower than those for uncomplexed AChE, a reduction greater than observed in earlier studies (4). The predominant residual catalytic activity of the human AChE⅐Fas2 complex appeared to be due to contaminating Fas2-resistant activity (7). Although mouse AChE⅐Fas2 complex appeared to contain only 0.09% activity resistant to Fas2 inhibition (5,8), comparatively high turnover rates of a fasciculin-resistant enzyme may also complicate our analysis of ligand entry into the active center gorge.
By employing an irreversible inhibitor to inhibit selectively the residual Fas2 resistant activity, we demonstrate here that mouse AChE⅐Fas2 complex retains only about 0.004% catalytic activity at 1 mM ATCh instead of the previously reported 0.09% (5). Stopped-flow techniques with fluorescent detection enabled us to determine rates of association and dissociation of several reversible inhibitors that equilibrate with AChE in a subsecond time frame. Furthermore, we have developed a competitive binding assay utilizing a fluorescent carbamate, M7C, to measure binding rates of nonfluorescent inhibitors to both mouse AChE and the AChE⅐Fas2 complex. Utilizing these procedures, we present a quantitative characterization of interaction of * This work was supported by United States Public Health Service Grants GM18360 and DAMD 17-1-8014 (to P. T.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

MATERIALS AND METHODS
Enzyme-Wild-type mouse AChE was expressed in HEK-293 cells stably transfected with cDNA encoding a monomeric form of the enzyme truncated at its carboxyl terminus at position 548. The enzyme secreted into the medium was purified to homogeneity on affinity columns as described previously (9).
Enzyme Activity-Hydrolysis of ATCh, PTCh, BTCh, S-PhAc, and pNPAc was measured spectrophotometrically at 412 nm for released thiocholine (10) or at 405 nm for p-nitrophenol. Activities of AChE⅐Fas2 complexes were measured following a 10-min incubation of 640 nM AChE with 1.3 M Fas2 to form the complex and 30-min incubation of the complex with 530 nM S P -3,3-dimethylbutyl methylphosphonothiocholine to inhibit hydrolytic activity not originating from the complex. The very slow reaction of the charged organophosphate with the active center of the Fas2⅐AChE complex gives rise to selective inhibition. Kinetic constants for hydrolysis of the above substrates by AChE and AChE⅐Fas2 complex were determined as described previously (5).
Enzyme Inhibition-The time course of AChE inhibition by trifluoroacetophenones was monitored by measuring ATCh hydrolysis in aliquots from the reaction mixture. Typically, 0.96-ml aliquots were drawn from 10 ml of the reaction mixture and mixed with 30 l of DTNB (10 mM) and 10 l of ATCh (100 mM) at specified time intervals to stop inhibition and determine residual enzyme activity. Both the inhibitory reaction and the assessment of residual activity were conducted in 0.1 M phosphate buffer, pH 7.0. First order rate constants for inhibition (k) were obtained by fitting fractional enzyme inhibition to the following equation, where f, f 0 , and f ϱ are fractions of enzyme inhibited at times t, t ϭ 0, and t ϭ ϱ, respectively. Second order inhibition rate constants were derived from plotting the first order rate constants for inhibition against inhibitor concentration and extrapolating to infinite dilution. Stopped-flow Kinetic Assays-Association and dissociation rate constants for fast equilibrating fluorescent inhibitors were determined using a SX.18 MV stopped-flow instrument equipped with fluorescence detection (Applied Photophysics). Excitation wavelengths were selected to achieve maximum signal amplitude. Acridine, 9-aminoacridine, and N-methyl acridinium were excited at wavelengths between 250 and 260 nm, coumarin at 220 nm, and propidium at 288 nm. For experiments with nonfluorescent inhibitors BW286C51, edrophonium, and decamethonium, but also for selected fluorescent ligands, AChE or AChE⅐Fas2 complex was excited at 276 nm (11). Quenching of ligand or AChE fluorescence upon complex formation was monitored using a 305-nm cut-off filter, and for 9-aminoacridine a 335-nm cut-off filter was used. Tacrine association was detected by excitation at 323 nm, with a 335-nm cut-off filter for emission. Linear plots of observed first order rate constants versus ligand concentrations gave second order associa-tion rate constants from the slope of the line and a first order dissociation rate constant from the ordinate intercept (11,12).
M7C Reaction Assay-The time courses of TFK, (Ϫ)-huperzine, and BW286C51 association with AChE and the AChE⅐Fas2 complex were followed using N,N-dimethyl carbamoyl N-methyl-7-hydroxy-quinolinium (M7C). This nonfluorescent dimethyl carbamoyl ester reacts covalently with the AChE active serine releasing a fluorescent N-methyl-7-quinolinium leaving group (M7H) ( excitation ϭ 410 nm; emission ϭ 510 nm; Ref. 13), thus allowing one to monitor both the rate of M7C reaction with the serine and the number of binding sites. The rate of binding of nonfluorescent covalent inhibitors to AChEs was followed by monitoring a decrease in number of AChE binding sites available for M7C reaction resulting from prior covalent reaction with the nonfluorescent inhibitor. Forty l of 1.0 mM M7C was added to 0.4-ml aliquots drawn at different time intervals from 4 ml of reaction mixture containing buffered solution of an AChE (10 -70 nM) with or without Fas2 (650 nM) and the nonfluorescent inhibitor. Aliquots were quickly transferred into a spectrofluorometric cuvette, and the fluorescence increase was followed at emission ϭ 510 nm ( excitation ϭ 410 nm) for 20 s in the absence of Fas2 or for 100 s in the presence of Fas2. The fluorescence signal exponentially approached a maximum value, which was diminished with increasing times of prior exposure to the nonfluorescent inhibitor, indicating progress of its reaction. The first order rate constant of binding of nonfluorescent inhibitors was determined by fitting the time dependence of relative change in maximal fluorescence to Equation 1.
Equilibrium Binding Measurements-Equilibrium dissociation constants for several fluorescent reversible ligands were determined by titration with increasing concentrations of AChE or AChE⅐Fas2 complex, resulting in quenching of ligand fluorescence. Concentrations of free fluorophores during titrations were determined from corresponding calibration curves constructed for each fluorophore, limiting fluorescence quenching in the bound state. Corrections were made for small changes in reaction volume due to the addition of titrant. Equilibrium dissociation constants (K d ) were determined from the reciprocal of the slopes of Scatchard plots.
Determination of Ligand pK a Values-UV-visible spectra of fast equilibrating reversible AChE inhibitors with ionizable functional groups were recorded in buffers of varying pH. Changes of peak absorbances as a function of buffer pH were analyzed using the following equations (rearranged from Ref. 14), where A min and A max are minimal and maximal peak absorbances and K a is the equilibrium proton dissociation constant. The following buffers were used: 0.10 M citric acid plus 0.20 M sodium phosphate (pH 2.6 -7.8) and 0.050 M sodium phosphate plus sodium pyrophosphate (pH 5.0 -12).

M7C Reaction
Assay-Time courses of the M7C reaction with the AChE and the AChE⅐Fas2 complex are shown in Fig. 1. M7C added at a concentration of 100 M binds to AChE at a rate too fast for measurement by conventional mixing and detection by spectrofluorometry. Hence, the carbamoylation step is reflected in a rapid burst of fluorescence. The reaction of M7C with the AChE⅐Fas2 complex was slower but was virtually complete within 100 s. The final fluorescence intensities upon carbamoylation and release of M7H were linearly dependent on AChE and AChE⅐Fas2 concentrations (Fig. 1, B and C).
The rapid and stoichiometric reaction of M7C with AChE or the AChE⅐Fas2 complex enabled us to develop a fluorescent assay for monitoring of binding rates for nonfluorescent ligands. The assay is based on consecutive, rapid M7C titrations of residual catalytically active AChE during the course of its reaction with nonfluorescent ligands. The first order rate constants for reaction were determined by fitting to Equation 1 the fluorescence intensities obtained in subsequent titrations. Good agreement between rate constants for ligand reaction measured from the decrement in the M7C reaction and ligand inhibition of ATCh hydrolysis is evident in Table I.

Hydrolysis of Substrates by AChE and AChE⅐Fas2
Complex-Catalytic hydrolysis of three cationic (ATCh, PTCh, and BTCh) and two neutral (S-PhAc and pNPAc) esters by AChE and AChE⅐Fas2 was compared. Substrate concentration dependences of AChE hydrolysis deviate from Michaelis kinetics at millimolar substrate concentrations, with ATCh, PTCh, and BTCh inhibiting their own hydrolysis. The kinetics of hydrolysis is characterized by substrate inhibition constants (K ss ) and relative activities of enzyme-substrate ternary complexes (b), in addition to Michaelis constants (K m ) and maximum turnover rates (k cat ) (Table II; Refs. 16 and 17). The three cationic substrates have similar Michaelis constants (15)(16)(17)(18), yet the commitment to catalysis (defined as a ratio of the enzyme acylation rate constant and dissociation rate constant of the Michaelis complex (cf. Ref. 19)) and the formation of the tetrahedral transition state, reflected in k cat , appear severely restricted for BTCh and other substrates containing a large acyl group (17,19).
Hydrolysis of neutral substrates by AChE obeyed Michaelis kinetics (Table II), but with K m values up to 2 orders of magnitude larger than for cationic substrates, thus emphasizing the importance of electrostatic interactions in formation of Michaelis complexes. k cat of S-PhAc was similar to k cat values for charged substrates and significantly greater than k cat for pNPAc. A similar relationship between catalytic constants for PhAc and pNPAc has been well described for Electrophorus electricus AChE (20).
Although the AChE⅐Fas2 complex hydrolyzed all substrates significantly more slowly than AChE, cationic substrates with high turnover rates were most affected. Hydrolysis of both cationic and neutral substrates by AChE⅐Fas2 obeyed Michaelis kinetics. Our earlier published curve of ATCh hydrolysis by AChE⅐Fas2 was biphasic, indicating substrate activation (5). The first phase, due to the presence of residual, Fas2-resistant, esterase activity, was inhibited selectively using the organophosphate, S P -3,3-dimethylbutyl methylphosphonothiocholine. Hence with this correction, the ATCh hydrolysis profile for the Fas2⅐AChE complex shows typical Michaelis kinetics. k cat /K m values for ATCh and PTCh were reduced by 5-6 orders of magnitude, whereas k cat /K m for BTCh was reduced by only 3 orders of magnitude. Neutral substrates S-PhAc and pNPAc exhibited much smaller reductions in k cat /K m (680 and 96 times, respectively). The reduction in k cat /K m for the five substrates by AChE⅐Fas2 resulted in distinct clusters of values for charged and neutral substrates. Values of 1.2-4.7 ϫ 10 3 M Ϫ1 s Ϫ1 for neutral substrates approach the k cat /K m value for benzoylthiocholine hydrolysis by AChE (1.1 ϫ 10 4 M Ϫ1 s Ϫ1 ; Ref. 17).
k cat /K m values for neutral substrates with the AChE⅐Fas2 complex were 1-2 orders of magnitude greater than for the charged substrates, suggesting that Fas2 association may impart a more restrictive barrier in the AChE⅐Fas2 complex to diffusional entry and acylation by cationic substrates. The reduction in values of k cat /K m was reflected in both an increase in K m and reduction in k cat . Whereas for neutral substrates and free AChE, the large K m values were limiting, the situation is reversed for the AChE⅐Fas2 complex, where large K m values primarily limit hydrolysis of cationic substrates. k cat values for the AChE⅐Fas2 complex were significantly reduced for all substrates. In addition, k cat values for the AChE⅐Fas2 complex correlate with the size of the acyl moiety in the substrate, suggesting that the steric occlusion by acyl pocket residues also limits substrate hydrolysis for the AChE⅐Fas2 complex. Coulombic repulsion from arginines and lysines of bound Fas2 could also interfere with the orientation of thiocholine in the transition state and its exit from the gorge. Both the reversible binding steps and turnover of two reactive cationic substrates, ATCh and PTCh, appeared to be reduced to the greatest extent in the AChE⅐Fas2 complex, suggesting a similar influence of the acyl pocket on the stability of the Michaelis complexes and tetrahedral transition states. This was not the case for neutral substrates, where k cat alone is affected.
Kinetics of Binding of Reversible Ligands to AChE-The second order association and the first order dissociation rate constants for two substituted trifluoroacetophenones and 10 noncovalent reversible ligands to AChE and AChE⅐Fas2 complex are listed in Table III. The constants for slow equilibrating   inhibitors such as the trifluoroacetophenones and (Ϫ)-huperzine A were determined using the M7C reaction assay, whereas for the faster equilibrating inhibitors, stopped-flow techniques with fluorescence monitoring were used (Figs. 2 and 3). Rates of interaction for BW286C51 were analyzed by both techniques. The association rate constants of the 10 ligands with AChE covered a narrow range of less than 2 orders of magnitude (4.5 ϫ 10 7 to 2.8 ϫ 10 9 M Ϫ1 s Ϫ1 ), and only (Ϫ)-huperzine A had a significantly slower association rate constant. Since several fast equilibrating ligands from Table III bear ionizable groups, their protonation states at pH 7.0 were analyzed by determining their pK a values (Table IV). These pK a values indicate that, , where s is substrate concentration, K ss is the substrate inhibition or activation constant, and b is a measure of relative activity of ternary enzyme-substrate complexes. ATCh, K ss ϭ 15 mM, b ϭ 0.23 (data of Radić et al. (16)); PTCh, K ss ϭ 1.3 mM, b ϭ 0.21; BTCh, K ss ϭ 7.1 mM, b ϭ 0.48 (data of Hosea et al. (17). Shown are means from 2-4 experiments. The typical S.E. of determination was 30 -40% of the mean value.  Fig. 3. The typical S.E. was 30 -40% of the mean value of constants. The equilibrium dissociation constant (K d ) was derived as the ratio of k Ϫ1 /k 1 . n.d., not determined due to slow rates of dissociation (TFK ϩ and TFK 0 ) or dissociation rates too rapid for stopped-flow measurements (propidium).
in addition to quaternary ligands, tacrine and 9-aminoacridine will be positively charged at pH 7.0; acridine will be neutral and coumarin exists as approximately a 3:1 ratio of neutral to anionic species. The fastest association rate constants were observed for quaternary ammonium ligands bearing a permanent cation (k 1 ϭ 1.3 ϫ 10 8 to 2.8 ϫ 10 9 M Ϫ1 s Ϫ1 ). Protonated cationic ligands, tacrine and 9-aminoacridine (k 1 ϭ 1.3 ϫ 10 8 to 1.5 ϫ 10 8 M Ϫ1 s Ϫ1 ), approached the lower end of the range of rates for quaternary compounds, while neutral ligands (acridine and coumarin) had the slowest association rate constants (k 1 ϭ 4.5 ϫ 10 7 to 1.2 ϫ 10 8 M Ϫ1 s Ϫ1 ). Association rate constants for peripheral site ligands, propidium and coumarin, exhibited the same rates as those ligands associating with the active center, despite the greater surface accessibility of the peripheral site.
The rate constants for dissociation of the 12 ligands extending over 7 orders of magnitude (1.8 ϫ 10 Ϫ5 to 4.2 ϫ 10 2 s Ϫ1 ), spanned a greater range than the association rate constants. Trifluoroacetophenones were the slowest dissociating ligands. These slowly reversible ligands interact covalently at the esteratic site and noncovalently in the choline binding site (12,21). Comparison of quaternary, protonated, and neutral ligands binding at the active center reveals that apart from trifluoroacetophenones and (Ϫ)-huperzine A, cationic bisquaternary BW286C51 and the cationic tertiary amines, tacrine and 9-aminoacridine, had the slowest dissociation rate constants.
Kinetics of Binding of Reversible Ligands to the AChE⅐Fas2 Complex-Ligand interactions with the AChE⅐Fas2 complex reveal marked reductions in both association and dissociation rate constants, compared with free AChE (Table III). The most affected should be the peripheral site ligands, propidium and coumarin, whose binding is competitive with Fas2 and the bisquaternary ligands, decamethonium and BW286C51, which share a portion of the peripheral binding site with Fas2. The 8-order of magnitude slower association rate constants of BW286C51 and decamethonium accompanied with only 1-3order of magnitude slower dissociation constants indicate that these bisquaternary ligands will form ternary complex with mAChE⅐Fas2.
By contrast, the interaction of coumarin with the AChE⅐Fas2 complex was altered in a different manner than for propidium, BW286C51, and decamethonium. The small reduction in the coumarin association rate constant is characteristic of active center ligands. Coumarin's dissociation rate constant was, however, reduced significantly more than those for the above ligands. Taken together, the comparative kinetics indicate the existence of a secondary binding site for coumarin in the active center gorge of the AChE⅐Fas2 complex. Unlike other peripheral site ligands coumarin appears small enough to enter the active center gorge in AChE. Previous mutagenesis and molecular modeling studies also suggest coumarin binding to the active center when the peripheral site is occupied (22)(23)(24).
For the eight remaining active center ligands, the association rate constants with AChE⅐Fas2 were about 4 orders of magnitude smaller than the corresponding constants for AChE, irrespective of ligand charge or size. Their dissociation rate constants were reduced by only 1-3 orders of magnitude. The greater reduction of association than dissociation rate constants leads to an increase in K d for active center ligands in the AChE⅐Fas2 complex typically by 1 order of magnitude but in one case by 2 orders. The K d for coumarin, with its presumed occupation of the active center in the AChE⅐Fas2 complex is actually slightly decreased.
Equilibrium Binding Measurements-Equilibrium dissociation constants for several reversible ligands, determined also by fluorescence titrations (Fig. 4), are listed in Table V and show good agreement with corresponding constants derived from the ratios of dissociation to association reaction rate constants (cf . Table III).
pH Dependence for Interaction of Reversible Ligands with AChE and the AChE⅐Fas2 Complex-AChE and several reversible ligands from Table III bear ionizable groups, and their mutual interactions may depend on pH of the reaction medium. Interaction of a reversible ligand and enzyme, both bearing ionizable groups, can be described as follows.
It can be shown that for association between ligand and enzyme in the above scheme, the following equation holds, where k 1 and k 1 max are second order association rate constants for ligand and enzyme at a given pH and a maximal rate constant of the reactive species, respectively. Factor b is a measure of the difference in the second order rate constants for reaction between enzyme and protonated versus nonprotonated forms of ligand. The species EH ϩ is assumed not to bind ligand. K a E and K a L are equilibrium dissociation constants for a proton from ionizable groups in the enzyme and ligand, respectively.
The pH dependence of reaction rate constants of four ligands with AChE was studied in the range of pH 4.5 to 11.0 (Table  VI). Association rate constants (k 1 ) for all ligands decreased at  pH values below 6.0 by 1-3 orders of magnitude; the greatest reduction was observed for quaternary cation edrophonium. Analysis of dependences of association rate constants for edrophonium and 9-aminoacridine on pH, using the model from Scheme 1 and the corresponding Equation 4, is shown in Fig.  5A. Association of both ligands in the low pH range depends on ionization of the same group, presumably in AChE, with a pK a E value of 6.1-6.2. The AChE group with pK a of 6.2 was implicated as critical for carbamoylation and phosphorylation of AChE by inhibitors containing quaternary ammonium in the leaving group (25). The reduction of association constants for edrophonium and coumarin in the pH range above pH 7.0 is consistent with their deprotonation (cf. Table IV). The asymmetric shape of the pH dependence curve for edrophonium, reflected in a value of b larger than 0 (b ϭ 0.14), may suggest that zwitterionic edrophonium species formed by deprotonation of the hydroxyl group associates with enzyme with k 1 max of 1.7 ϫ 10 7 M Ϫ1 s Ϫ1 , which is about 7-fold slower than k 1 max for cationic edrophonium.
On the other hand reduction of association constants of 9-aminoacridine is reflected in a pK a L value of 10.3. This value is close to both the pK a value of 9-aminoacridine (10.1) and the pK a value found for carbamoylation and phosphorylation of human AChE (10.25;Ref. 25) and therefore could be either ligand-or enzyme-related. Minimal variations of dissociation rate constants of cationic ligands (Table VI) in the pH interval studied suggests an absence of a pH related conformational change in the active center gorge. Protonation of acridine slowed the dissociation rate 10-fold, emphasizing the role of electrostatic interactions in stabilization of cationic ligands in the AChE active center gorge.
The pH dependence of 9-aminoacridine association with AChE⅐Fas2 complex at low pH is very similar to the dependence observed for association with AChE alone, and reflected in pK a E value of 6.1 (Fig. 5B). At the high pH, the association rate constants were only slightly reduced, much less than observed for association with AChE alone. Thus, both protonated and neutral 9-aminoacridine appear to associate with AChE⅐Fas2 at similar association rates. In addition, protonation of a basic AChE residue(s), with a presumed pK a of ϳ10.3 (Fig. 5A), may be affected by bound Fas2 or is no longer a rate-limiting factor in ligand association. The deviations of experimental points from the calculated curve observed in Fig. 5B at intermediate and high pH appeared reproducible and are possibly an indication of the interacting species being more complex than described in Scheme 1. Rates of 9-aminoacridine dissociation from AChE⅐Fas2 complex were largely independent on pH. DISCUSSION Our comparison of the interaction kinetics of five substrates and 12 reversible inhibitors between the AChE⅐Fas2 complex and free AChE (Fig. 6) showed second order reaction rate constants for substrates and association rate constants for reversible ligands to be markedly reduced for the AChE⅐Fas2 complex. When the contribution of the fasciculin-resistant enzyme species is accounted for, the magnitude of the rate reductions approaches 4 -6 orders of magnitude.
Fas2-resistant Esterase-A Fas2-resistant esterase in quantities consisting of 0.09% of the total activity is retained and eluted from the affinity columns. Characterization of this species is problematic due to its comparatively small quantities. Its specificity for cationic organophosphates differs from mouse butyrylcholinesterase and suggests that it may be a proteolysis product or a spontaneous modification of AChE. Allowing for the contaminant, we now estimate the catalytic efficiency for ATCh with the AChE⅐Fas2 complex to be 0.004% of uncomplexed AChE activity instead of the previously reported 0.09% (5). Mallender et al. (7) found that a large fraction of catalytic activity of the human AChE⅐Fas2 complex, that represented 2-31% of uncomplexed human AChE activity, was in fact contaminating, Fas2-resistant esterase activity. Their findings prompted us to develop multiple approaches to examine whether mouse AChE, despite its low residual activity in the presence of Fas2, also contained a Fas2-resistant component.
Development of a novel competitive binding assay based on

FIG. 4. Scatchard plots for fluorescence titrations of 9-aminoacridine (circles) and tacrine (triangles) by AChE (closed symbols) and the AChE⅐Fas2 complex (open symbols).
Concentration of both ligands was 10 nM in titration with 5-500 nM AChE and was 50 nM (tacrine) or 100 nM (9-aminoacridine) in titration with 25-3000 nM AChE⅐Fas2 complex. Inset, decreases in fluorescent emission spectrum of 9-aminoacridine during titration with increasing concentrations of AChE and AChE⅐Fas2 complex. Shown are 10 nM 9-aminoacridine titrated with concentrations between 0 and 500 nM AChE and 100 nM 9-aminoacridine titrated with concentrations between 0 and 3000 nM AChE⅐Fas2 complex.  Aminoacridine  38  760  Tacrine  44  800  Coumarin 14,000 ND fluorescent product of the carbamoyl ester, M7C, enabled us to determine rate constants for binding of nonfluorescent ligands to mouse AChE⅐Fas2 complex and to unliganded mouse AChE. The carbamoylation reaction weights each active center equivalently and avoids amplification from the high turnover by residual enzyme not inhibited by Fas2. In addition, incubation of a mixture of the AChE⅐Fas2 complex and Fas2-resistant esterase with a bulky organophosphate selectively inhibits the latter, enabling one to account for the disparity of findings with previous studies. Finally, employing stopped-flow kinetics with direct fluorescent detection of the complexes allowed us to separate association and dissociation reaction rate constants for binding of several rapidly equilibrating reversible inhibitors to both free AChE and the AChE⅐Fas2 complex.
Influence of Fasciculin on Hydrolysis of Substrates-Analysis of the kinetics indicates that inhibition of AChE activity by Fas2 is not due solely to a diminished rate of substrate entry into the active center in the AChE⅐Fas2 complex. The commonly considered reaction steps in substrate hydrolysis by AChE (14,28,29) are shown in Scheme 2, where enzyme (E) and substrate (S) associate to form a Michaelis complex (ES). In turn, the complex is converted to an un-stable acyl enzyme (EA) with loss of the substrate leaving group (P 1 ). EA is hydrolyzed to free enzyme and acyl group of substrate (P 2 ). The corresponding catalytic parameters, K m , k cat , and k cat /K m are described as follows.
Second order rate constants for reaction of cationic substrates with AChE⅐Fas2, reflected in k cat /K m values between 60 and 100 M Ϫ1 s Ϫ1 , are slower than rate constants for entry of reversible ligands into the AChE⅐Fas2 complex. The exceptionally slow association rates of BW286C51 and decamethonium are most easily explained by partial overlap of their binding site with the site of Fas2 residence due to their extended conformation (cf. Fig. 6). Thus, additional conformational constraints apply to their association rates. Furthermore, the reduction of nearly 6 orders of magnitude in k cat /K m for ATCh and PTCh exceeds the reduction in association rate constants for the reversible cationic inhibitors with the AChE⅐Fas2 complex. Catalyzed hydrolysis by AChE⅐Fas2 for various substrates yielded sets of limiting values for k cat and for k cat /K m . The similarity of k cat values for charged (ATCh) and two neutral acetate esters (S-PhAc and pNPAc) suggests that the rate of deacylation of EA (k 3 ) may be involved in rate limitation for the chemical part of the catalytic reaction under the influence of bound Fas2. The second order rate constants of association were fitted to Equation 4 with the following parameters. A, , pK a E ϭ 6.2, pK a L ϭ 7.6, b ϭ 0.19 Ϯ 0.03 (for edrophonium) and k 1 max ϭ (1.4 Ϯ 0.1) ϫ 10 8 M Ϫ1 s Ϫ1 , pK a E ϭ 6.1, pK a L ϭ 10.3, b ϭ 0 (for 9-aminoacridine). B, k 1 max ϭ (5.2 Ϯ 0.2) ϫ 10 4 M Ϫ1 s Ϫ1 , pK a E ϭ 6.1, pK a L ϭ 11.9, b ϭ 0 (for 9-aminoacridine).

TABLE VI
Effect of pH on association and dissociation rate constants for selected reversible ligands and mouse AChE Reactions were monitored in 0.1 M citrate-0.2 M phosphate buffers (pH 4.5-7.0), 0.1 M phosphate buffers (pH 5.5-8.0), and 0.050 M phosphatepyrophosphate buffers (pH 6.0 -9.0). Values of constants measured in different buffers of the same pH were similar and were therefore averaged. Experiments were performed in duplicate. pH 9-Aminoacridine Acridine Edrophonium Coumarin Parallel reductions in k 2 by Fas2 are also likely, since the inhibition rates of several organophosphates were shown to be reduced in the AChE⅐Fas2 complex by several orders of magnitude more than the rates of association of reversible inhibitors (6). 2 Furthermore, a second clustering of k cat /K m values for hydrolysis of charged acyl thiocholines by AChE⅐Fas2 (ϳ83 M Ϫ1 s Ϫ1 ) as opposed to neutral acetates (ϳ3000 M Ϫ1 s Ϫ1 ) indicates that the corresponding k Ϫ1 and k 2 constants are of the same magnitude and, consequently, together with k 1 both influence values of k cat /K m . This follows because the reduction of k cat /K m for ATCh, imposed by Fas2, is larger than reductions observed in rates of association of reversible inhibitors. Charge on the leaving group influences both reactions, described by k 1 and k 2 , which both participate in limiting the rates of hydrolysis by the AChE⅐Fas2 complex. The absence of the selective steric constraints of the acyl pocket in the AChE⅐Fas2 complex (Table II) further suggests that the leaving group of the substrate is limiting for catalytic throughput (k cat /K m ) in the AChE⅐Fas2 complex. Thus, in addition to the pronounced effects of bound Fas2 on rates of ligand entry into AChE, reflected in k 1 , an allosteric interaction between the Fas2 binding site and the active center appears to affect the efficiency of AChE catalysis primarily through reduction of both k 3 and k 2 .
Influence of Fasciculin on the Binding of Reversible Ligands-Our compilation of reaction rate constants for diverse reversible inhibitors of AChE permits one to dissect the factors controlling ligand entry to the gorge and steric constraints within the gorge. The affinities of these inhibitors are governed by two factors. First, the association rate constants for all but one ligand studied approach the diffusion-limited range, ensuring rapid formation of inhibitory complexes. Similar rates of association for ligands binding at the peripheral site and active center reflect minimal steric constraints for entry into the narrow, 18 -20 Å in depth active center gorge (cf. Fig. 6) for those ligands not exceeding a critical volume (propidium) or shape ((Ϫ)-huperzine A).
Second, the residence times of ligands in the stabilized complexes with AChE varied over 7 orders of magnitude. Previ-ously reported rate constants for ligand association with AChE, limited to slow dissociating fasciculins, huprines, trifluoroacetophenones, huperzine A, ambenonium, and bisquaternary o-CIB-BQ (Table VII), and only three fast equilibrating ligands, N-methylacridinium, 7-hydroxyquinolinium, and propidium, were in good agreement with the constants reported here (cf .  Tables III and VII The binding of several ligands, such as edrophonium and decamethonium, that lack a spectral overlap with tryptophan fluorescence emission nonetheless results in diminished AChE fluorescence in the ligand-enzyme complex. Since fluorescence resonance energy transfer is not possible in this situation, the quantum yields of several tryptophans have been affected by ligand occupation of the enzyme active center. This observation not only increases the range of ligands whose association with the enzyme can be measured directly but also indicates that the microenvironments of several aromatic residues are perturbed in the complex. The implications of this finding will be addressed in subsequent studies. Linear Free Energy Relationships-Reduced association and dissociation rate constants for ligands interacting with the AChE⅐Fas2 complex reveal significant steric hindrance for entry of ligands into the gorge. The different magnitude of the steric intervention on association and dissociation steps, however, suggests that Fas2 alters the position of the transition state relative to the free and bound species. This is evident from the strikingly different linear free energy relationships between association or dissociation rate constants of the ligands and their equilibrium dissociation constants for AChE and AChE⅐Fas2 complex (Fig. 7).
Relating free energy changes in the transition state to those in the ground states for the various ligands enables one to better describe the energetic pathway for ligand entry and exit from the gorge for AChE and the AChE⅐Fas2 complex (37-39). Thus, for a series of ligands, L, the following linear free energy 2 Radić and Taylor, manuscript in preparation. relationships can be developed.
⌬G ‡ A,L and ⌬G ‡ D,L are the activation free energies for association and dissociation, ⌬G 0 L is the respective equilibrium free energy for each ligand, and ⌬G ‡ 0 is a constant, intrinsic activation barrier. The slope of the relationships is related to ⌽, an TABLE VII Literature values of kinetic constants for interaction of reversible inhibitors with acetylcholinesterase Second order association rate constants (k 1 ) and first order dissociation rate constants (k Ϫ1 ) were determined under the following experimental conditions: 50 mM sodium phosphate buffer, pH 7.2, with ϳ100 mM NaCl and 25 M Triton X-100 (12), 100 mM phosphate buffer, pH 7.0, with 0.01% bovine serum albumin (5, 31), 10 mM Tris/HCl buffer, pH 7.5, with 100 mM NaCl and 0.1 mg/ml bovine serum albumin (30), 20 mM phosphate buffer, pH 7.0, with 0.02% Triton X-100 (4,32), 100 mM phosphate buffer, pH 7.0, with 1.0% Triton X-100 (33) estimate of the position of the transition state for ligand association, which can vary between 0 and 1. Analysis of the data in Fig. 7 shows a ⌽ value of 0.093 for ligand association with AChE and 0.93 for ligand association with AChE⅐Fas2. For dissociation, 1 Ϫ ⌽ is 0.91 for AChE and 0.07 for the AChE⅐Fas2 complex. In each case, huperzine was not included in the analysis, since it was an obvious exception to the correlations, which may relate to its slow association rate. The slow rate of trifluoroacetophenone dissociation and the mutually exclusive binding between propidium and Fas2 preclude the analysis of their interactions with the AChE⅐Fas2 complex.
The capacity of Fas2 to alter the relationship between the transition and ground states points to the role of the Fas2 in influencing the energetics of the pathway for small ligand entry into the gorge. In the absence of Fas2, the activation barrier for formation of ligand-AChE complexes correlates with the stabilization energy of the complex as ligand structure is varied. In the presence of Fas2, the transition barrier energetics correlate with the energies of the free species. Hence, diffusional entry of the ligand to the gorge requires bypassing the cationic bound Fas2, and this becomes the new energetic barrier limiting for the interaction. In the ligand-bound AChE⅐Fas2 complex, escape of the ligand from its residence site deep in the gorge no longer is the rate limitation for dissociation. This altered energetic requirement arises from not simply the charge and steric barriers imparted by the Fas2 molecule per se (6) but also the capacity of Fas2 to restrict conformational flexibility of the ligand binding site, deep in the AChE gorge.
Factors Controlling Kinetics of Ligand Interaction-By using stopped-flow instrumentation, we examined interaction kinetics for rapid and slow equilibrating ligands with selectivity for the active and the peripheral sites. It has been suggested (7, 40 -42) that binding of cationic substrates to peripheral site of mammalian, fish, or insect AChE effectively increases the local substrate concentration available for binding to the active center and thus enhances catalytic efficiency. Such a mechanism seems plausible where substrates would have to traverse an energy barrier imposed by a narrow gorge to reach to the site of catalysis in the active center. As would be expected, the kinetic barriers imposed by the pre-Michaelis complex are minimal, since active center ligands and peripheral site ligands associate with AChE at comparable rates. The modular analysis of molecular traffic through the AChE active center (41) suggested distinct limits for diffusion of neutral and cationic ligands, differing by about 1 order of magnitude. Our determined average second order association rate constants of 8.9 ϫ 10 8 M Ϫ1 s Ϫ1 (for cationic ligands) and 8.2 ϫ 10 7 M Ϫ1 s Ϫ1 (for neutral ligands) are in accord with the prediction.
Protonation of neutral ligands (9-aminoacridine, tacrine) only modestly affects association rates with AChE and the AChE⅐Fas2 complex, perhaps reflecting a balance between rate acceleration by the addition of a charge and deceleration from the increased hydrated molecular volume. The hydrated cation structure, however, decreases the dissociation rates, as evidenced by a comparison of N-methylacridinium and acridine with 9-aminoacridine and tacrine as well as by pH-dependent decrease of the dissociation rate of acridine. Net neutralization of charged edrophonium by turning it into zwitterion, as well as the addition of negative charge to neutral coumarin markedly reduced their association rate constants, as reported for 1-methyl 7-hydroxyquinolinium (36). We find that protonation of enzyme group(s) of pK a 6.1-6.2 critically decreases association rates of positively charged ligands with both AChE and the AChE⅐Fas2 complex in addition to reduction of phosphorylation and carbamoylation rates of the AChE active serine (25).
Rapid conformational isomerizations may control the opening times of the active center gorge to admit ligands (43). In addition to pathway restrictions imposed by charge and steric constraints of Fas2, these isomerizations could be far less frequent in AChE⅐Fas2 due to restrictions in the breathing motions of domains of AChE immobilized by Fas2. Fas2 at the gorge entry can also be expected to influence allosterically the conformation of residues controlling catalysis and ligand dissociation at the base of the gorge.
Comparative analysis of the kinetics of ligand entry and exit from the active center gorge for free AChE and AChE⅐Fas2 complex reveals that the steric influences of Fas2 at the gorge entry are insufficient to explain the differences in linear free energy relationships between the kinetic and equilibrium constants for free AChE and the Fas2⅐AChE complex (Fig. 7).  Table III. Quaternary ammonium, cationic inhibitors are noted by circles, protonated cationic inhibitors by diamonds, and uncharged inhibitors by squares. Lines were generated by linear regression of Equations 8 and 9. (Ϫ)-Huperzine A (inhibitor number 4) was not included in the calculation. A, linear free energy relationship for dissociation from AChE and AChE⅐Fas2. Calculated slopes were (1.0 Ϯ 0.1) for AChE, and (0.074 Ϯ 0.13) for AChE⅐Fas2. B, linear free energy relationship for association with AChE and AChE⅐Fas2. Calculated slopes were Ϫ(0.093 Ϯ 0.056) for AChE, and Ϫ(0.93 Ϯ 0.13) for AChE⅐Fas2.