Introduction

Top Abstract Introduction Experimental Procedures Results and Discussion References

Alzheimer's disease is associated with a selective loss of cholinergic neurons in the brain (Davies and Maloney, 1976; Whitehouse et al., 1981). The losses are accompanied by decreases in the neurotransmitter acetylcholine as well as in the respective enzymes that synthesize and degrade acetylcholine, choline acetyltransferase and acetylcholinesterase (AChE) (Perry et al., 1977; Bowen et al., 1982). These observations stimulated a cholinergic hypothesis of Alzheimer's disease (Bartus et al., 1982), which postulates that at least some of the cognitive decline experienced by patients with the disease results from a deficiency in acetylcholine and thus in cholinergic neurotransmission. This hypothesis suggested that inhibitors of AChE might elevate levels of acetylcholine in the brains of these patients and reverse the cognitive decline (Muramoto et al., 1979), and experimental evidence has supported this suggestion. The first, and thus far the only, two drugs approved by the U.S. Food and Drug Administration (FDA) for treatment of the cognitive deficit in Alzheimer's disease are both reversible inhibitors of AChE (Fig. 1). Tacrine (Cognex) was approved in 1993 and E2020 (donepezil, Aricept) in 1996. A third potent reversible AChE inhibitor, huperzine A (Fig. 1), is a natural product isolated from the club moss Lycopodium Huperzia serrata used in traditional Chinese herbal medicine (Kozikowski et al., 1992).

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Structures of inhibitors of AChE.

A number of efforts have been undertaken to synthesize even more selective AChE inhibitors as potential therapeutic agents in Alzheimer's disease. Although the highest affinity reversible AChE inhibitors identified to date contain quaternary amine group(s), these efforts have focused on compounds which, like the approved drugs, contain nonquaternary amines and are more likely to pass from the circulation to the brain across the blood-brain barrier. Some new inhibitors have been modeled on tacrine, including bifunctional or bis-tacrine analogs with alkylene linked tacrine moieties (Pang et al., 1996). Other inhibitors have been designed that combine the carbobicyclic substructure of huperzine A with the 4-aminoquinoline substructure of tacrine (Badia et al., 1998). These initial tacrine-huperzine A hybrids, for which we suggest the name huprines, had slightly higher affinity for AChE than tacrine itself. Herein, we introduce a huprine with an ethyl group at position 9 and a chlorine atom at position 3. We designate this compound huprine X (Fig. 1) and show that it inhibits AChE by binding to the enzyme with an affinity that is among the highest yet reported for a reversible inhibitor of AChE.

We also estimate the location of huprine X binding in AChE by reference to the structures of AChE and AChE-inhibitor complexes determined by X-ray crystallography. These structures reveal a deep active site gorge with two sites of ligand interaction: a peripheral site at the surface of the enzyme and an acylation site at the base of the gorge where the substrate acyl group is first transferred to residue S200¹ (Sussman et al., 1991; Harel et al., 1993; Bourne et al., 1995; Harel et al., 1995). Certain ligands can bind selectively to either the acylation site or the peripheral site, and ternary complexes can be formed in which ligands are bound to both sites simultaneously (Taylor and Lappi, 1975; Eastman et al., 1995). X-ray crystallography has shown that edrophonium, huperzine A, and tacrine all are specific for the acylation site, although their contacts with the enzyme surface do not completely overlap (Harel et al., 1993; Raves et al., 1997). Ligands specific for the peripheral site include the small aromatic compound propidium. By examining inhibition of AChE by huprine X in the presence of either propidium or edrophonium, we show that huprine X appears to bind to the acylation site of AChE.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results and Discussion References

Materials

Human erythrocyte AChE (Rosenberry and Scoggin, 1984; Roberts et al., 1987) and recombinant human AChE (Mallender et al., 1999) were purified as outlined previously. Analyses were conducted with human erythrocyte AChE except where noted, and active site AChE concentrations for both enzymes were determined by assuming 410 U/nmol.² Recombinant Drosophila AChE was the SEC1 variant (Incardona and Rosenberry, 1996). Human butyrylcholinesterase was purified from plasma (Lockridge, 1990) and was a gift from Dr. Oksana Lockridge, University of Nebraska Medical Center. Propidium iodide was obtained from Calbiochem (La Jolla, CA), and propidium concentrations were assigned with an extinction coefficient _{493 nm} = 5900 M¹ cm¹ (Taylor and Lappi, 1975). Tacrine was purchased from Sigma Chemical Co. (St. Louis, MO) and E2020 was a gift from Yoshiyuki Kawakami, Ph.D., Eisai Co., Ltd., Ibaraki, Japan.

General Chemistry Methods

Melting points were determined in open capillary tubes with a MFB 595010 M Gallenkamp melting point apparatus. ¹H NMR spectrum was recorded at 500 MHz on a Varian VXR 500 spectrometer and ¹³C NMR spectrum was recorded at 75.4 MHz on a Varian Gemini 300 spectrometer. The chemical shifts are reported in parts per million ( scale) relative to internal trimethylsilane and coupling constants are reported in Hertz. Correlation spectroscopy (COSY) ¹H/¹H experiments were performed with standard procedures and COSY ¹H/¹³C experiments with the ¹H-detected heteronuclear multiple-quantum coherence (HMQC) sequence and an indirect detection probe. Infrared (IR) spectra were run on a Perkin Elmer model 1600 Fourier transform/infrared spectrometer. Absorption values are expressed as wavenumbers (cm¹). Optical rotations were measured on a Perkin Elmer model 241 polarimeter. Chiral HPLC analyses were performed on a Waters model 600 liquid chromatograph provided with a Waters model 486 variable  detector, with a CHIRALCEL OD-H column (25 × 0.46 cm) containing the chiral stationary phase cellulose tris(3,5-dimethylphenylcarbamate). Conditions A: mixture of hexane/EtOH in the ratio of 75:25, containing 0.1% diethylamine, as eluent, flow 0.20 ml/min,  = 235 nm. Chiral medium-pressure liquid chromatography (MPLC) separation was carried out on an equipment that consisted of a pump (Büchi 688), a variable UV detector (Büchi) and a column (25 × 2.5 cm) containing microcrystalline cellulose triacetate (15-25 µm) as the chiral stationary phase. Column chromatography was performed on silica gel 60 A C.C. (70-200 mesh; SDS, ref. 2100027). For the thin-layer chromatography, aluminum-backed sheets with silica gel 60 F₂₅₄ (Merck; ref. 1.05554) were used. AlCl₃ and 2-amino-4-chlorobenzonitrile were purchased from Aldrich Chemical (Milwuakee, WI). Analytical grade solvents were used for recrystallizations, whereas pure for synthesis solvents were used in the reaction and column chromatography. Elemental analyses were carried out at the Mycroanalysis Service of the Centro de Investigación y Desarrollo, C.I.D., Barcelona, Spain.

Synthesis of Huprine X and Huprine Y

The synthesis strategy for these compounds is outlined in Fig. 2.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Synthesis strategy for huprine derivatives. Compound rac-3a and its enantiomers (huprine Y is the HCl salt of ()-rac 3a) were prepared as described (Camps et al., 1998, 1999; for the synthesis of related compounds, see Badia et al., 1998). Rac-3b (huprine X is the HCl salt of ()-rac 3b) was prepared in a similar way from the known enone rac-1b (Camps et al., 1996) by Friedländer reaction with 4-chloro-2-aminobenzonitrile, 2 (Camps P, Muñoz-Torrero D, Görbig D, Contreras J, Simon M, Morral J, El Achab R, Badia A, Baños JE, Vivas N. WO patent application 97/13754). Separation of rac-3b was carried out in a similar manner to that described for rac-3a (see Experimental Procedures). Although the absolute configuration of ()-3a was established by X-ray diffraction analysis, that of ()-3b was assigned by analogy with ()-3a (Camps et al., 1998).

rac-12-Amino-3-chloro-9-ethyl-6,7,10,11-tetrahydro-7,11-methanocycloocta[b]quinoline Hydrochloride (rac-3b.HCl). To a suspension of anhydrous AlCl₃ (3.00 g, 22.5 mmol) and 2-amino-4-chlorobenzonitrile (2.33 g, 15.3 mmol) in 1,2-dichloroethane (20 ml), a solution of 7-ethylbicyclo[3.3.1]non-6-en-3-one rac-1b (1.80 g, 11.0 mmol) in 1,2-dichloroethane (115 ml) was added dropwise. The reaction mixture was stirred under reflux for 7 h, allowed to cool to room temperature, diluted with water (80 ml) and tetrahydrofuran (80 ml), made basic by addition of 5 N NaOH, and stirred at room temperature for 30 min. The organic solvents were removed under reduced pressure, and the residue was filtered. The yellowish solid residue (4.20 g) was submitted to column chromatography [silica gel (125 g), hexane/AcOEt, gradient elution]. On elution with hexane/AcOEt 40:60, rac-3b (1.35 g, 41% yield) was isolated. Subsequent treatment with a solution of HCl (0.37 N solution in MeOH, 3 equiv.), evaporation, and recrystallization of the resulting solid (1.54 g) from MeOH/H₂O 3:10 (26 ml), afforded pure rac-3b.HCl.2/3H₂O (0.96 g, 25% overall) as a white solid: mp 202-206°C (dec.); IR 3500-2000 (max. at 3333, 3177, 2816, 2671) (CH, NH, and NH⁺), 1652, 1634, and 1585 (ar-C-C and ar-C-N) cm¹. ¹H NMR (500 MHz, CD₃OD) : 0.89 (t, J = 7.5 Hz, 3 H, CH₂-CH₃), 1.86 (m, 2H, CH₂-CH₃), 1.95 (dm, J = 12.5 Hz, 1 H, 13-H_syn), 2.00 (broad d, J = 17.5 Hz, 1 H, 10-H_endo), 2.07 (dm, J = 12.5 Hz, 1 H, 13-H_anti), 2.53 (broad dd, J = 17.5 Hz, J' = 4.5 Hz, 1 H, 10-H_exo), 2.80 (m, 1 H, 7-H), 2.87 (broad d, J = 18.0 Hz, 1 H, 6-H_endo), 3.20 (dd, J = 18.0 Hz, J' = 5.5 Hz, 1 H, 6-H_exo), 3.38 (m, 1 H, 11-H), 4.82 (s, NH₂ + NH), 5.56 (broad d, J = 5.5 Hz, 1 H, 8-H), 7.56 (dd, J = 9.0 Hz, J' = 1.5 Hz, 1 H, 2-H), 7.75 (d, J = 1.5 Hz, 1 H, 4-H), 8.34 (d, J = 9.0 Hz, 1 H, 1-H). ¹³C NMR (75.4 MHz, CD₃OD) : 12.6 (CH₃, CH₂-CH₃), 27.6 (CH, C11), 28.1 (CH, C7), 29.4 (CH₂, C13), 30.9 (CH₂, CH₂-CH₃), 34.2 (CH₂, C10), 36.1 (CH₂, C6), 115.4 (C) and 115.5 (C) (C11a and C12a), 119.4 (CH, C4), 123.3 (CH, C8), 126.3 (CH, C1), 127.6 (CH, C2), 139.7 (C, C4a), 140.4 (2 C, C3 and C9), 153.2 (C) and 156.6 (C) (C5a and C12). Anal. Calcd. for C₁₈H₁₉ClN₂.HCl.2/3H₂O: C, 62.25; H, 6.20; N, 8.07; Cl, 20.42. Found: C, 61.95; H, 6.12; N, 8.08; Cl, 20.44.

Preparative Resolution of rac-3b by Chiral MPLC: (+)-(7R,11R)-3b and ()-(7S,11S)-3b (Huprine X). The chromatographic resolution of rac-3b was carried out with MPLC equipment provided with a column containing microcrystalline cellulose triacetate (15-25 µm), pretreated with a 0.1% solution of Et₃N in ethanol, as the chiral stationary phase. The sample of rac-3b (1.59 g) was introduced as free base in three portions (1 × 90 mg + 1 × 500 mg + 1 × 1000 mg) with 96% ethanol (2 ml/min) as the sole eluent and solvent. The chromatographic fractions (5 ml) were analyzed by chiral HPLC under conditions A [()-3b, r.t. 23.87 min, k'₁ = 0.523; (+)-3b, r.t. 27.15 min, k'₂ = 0.733;  = 1.40, Res. = 1.75] and combined conveniently. In this way, ()-3b (650 mg, 98% e.e.) and (+)-3b (390 mg, 94% e.e.) were obtained. The remaining product consisted of mixtures of both enantiomers with lower e.e. values.

1

1

Calibration of Huprine Concentrations by Titration with AChE

AChE active site concentrations can be estimated to an accuracy of better than 10% from the AChE activity assay described below, and stock concentrations of inhibitors can be determined by stoichiometric titration with the enzyme (Eastman et al., 1995; Mallender et al., 1999). Titration of AChE with huprine X gave a linear decrease in AChE activity with increasing concentrations of huprine X (Fig. 3), consistent with the formation of an inactive huprine X-AChE complex. However, the stock concentration of huprine X calculated by assuming 1:1 stoichiometry in this complex was about half of that calculated directly from the dry weight. Because there is no crystallographic or kinetic evidence that a high-affinity inhibitor binds with >1:1 stoichiometry to the AChE active site, we calculated an apparent extinction coefficient of 18 ± 1 mM¹ cm¹ for huprine X from the absorbance at 326 nm and the concentration determined by the titration in duplicate experiments. This extinction coefficient also was used to estimate concentrations of huprine Y. 

View larger version (11K): [in this window] [in a new window]   Fig. 3.   Titration of recombinant human AChE with huprine X. Each point represents one mixture with a final concentration of 1.21 µM AChE and the indicated concentration of huprine X relative to the maximal concentration [huprine X]o incubated for 3 h at 23°C. Aliquots (10 µl) were then diluted 300- to 15,000-fold into assay solution containing 0.5 mM acetylthiocholine. Points represent observed v (mM/min) normalized to a 300-fold dilution into assay solution. The line was fitted to the points in a nonequilibrium analysis with the program SCoP, assuming that huprine X interacted with AChE according to Scheme 1 with rate constants kI = kAI and kI = kAI given in Fig. 5. The fitted value of [huprine X]o = 1.87 µM was used to determine an apparent extinction coefficient for huprine X.

Steady-State Inhibition of Enzyme-Catalyzed Substrate Hydrolysis

Inhibitor was preincubated with AChE or butyrylcholinesterase in buffer (20 mM sodium phosphate and 0.02% Triton X-100 at pH 7.0) for 30 min, except for 1 to 6 h preincubations of huprine X and Y with human AChE. Substrate hydrolysis rates v were measured at 25°C after addition of small aliquots of acetylthiocholine and DTNB (to a final concentration of 0.33 mM) in a total volume of 1.0-3.0 ml. Hydrolysis was monitored in an Ellman assay by formation of the thiolate dianion of 5,5'-dithiobis-(2-nitrobenzoic acid) (DTNB) at 412 nm (_{412 nm} = 14.15 mM¹ cm¹; Riddles et al., 1979) for 1 to 3 min on a Varian Cary 3A spectrophotometer (Ellman et al., 1961), and substrate concentrations were corrected for substrate depletion resulting from hydrolysis during this interval. It was assumed that acetylthiocholine concentrations were maintained at low enough values (<1.0 mM) to ignore substrate inhibition in the absence of inhibitors (Szegletes et al., 1999) and that inhibitors that bind exclusively to the acylation site in AChE are unable to form ternary complexes with substrate. These assumptions allow analysis of the inhibition data to be based on Scheme 1.

View larger version (7K): [in this window] [in a new window]   Scheme 1.

In Scheme 1, the initial enzyme-substrate complex ES can form an acyl enzyme intermediate EA with the rate constant k₂. The inhibitor (I) can bind to the free enzyme E or to EA with respective equilibrium constants K_I (=k_I/k_I) and K_AI (=k_AI/k_AI), and the deacylation rate constant k₃ is altered by a factor b in the EAI complex. According to Scheme 1, v is given by eq. 1.

(1)

In this equation, V_max and K_app have been defined previously (Szegletes et al., 1998) and K_U is defined below. Consistent with eq. 1, reciprocal plots of v¹ versus [S]¹ at all inhibitor (I) concentrations were linear with a slope of K_app/V_max in the absence of inhibitor. Slopes and intercepts of these plots were the respective reciprocals of the second- and first-order rate constants for substrate hydrolysis at a given concentration of I and were calculated by a weighted linear regression analysis (Fig. P, version 6.0; Biosoft, Milltown, NJ) that assumed that v has a constant percentage of error. The slopes were normalized by K_app/V_max by K_app/V_max as in eq. 2, and values of K_I were determined by linear regression analyses of these normalized slopes versus [I] according to eq. 2, with slope values weighted by the reciprocal of their variance (Eastman et al., 1995).

(2)

A similar analysis of the intercepts of the reciprocal plots gave K_U in eq. 1, where K_U = k₃(K_R + b[I])/(k_cat (1  b + (bk₃/k_I))), K_R = (k_AI + bk₃)/k_AI, and k_cat = k₂k₃/(k₂ + k₃) (Rosenberry, 1975). When the intercepts increase linearly with [I], K_U is a constant that corresponds to k₃K_AI/k_cat (when bk₃ k_AI) or to k₃k_I/k_catk_AI (when k_I bk₃). For acetylthiocholine hydrolysis by human AChE, k₃ is ~10⁴ s¹ and k_cat is about one-half of k₃ (Szegletes et al., 1999).

When an inhibitor (I1) specific to the acylation site is incubated with AChE together with varying concentrations of a second inhibitor (I2) specific for the AChE peripheral site, ternary complexes with both inhibitors bound simultaneously to the enzyme can occur (Taylor and Lappi, 1975; Szegletes et al., 1998). At equilibrium, the residual concentration of free enzyme [E] in the presence of both inhibitors relative to the concentration of free enzyme [E]_[I2] = 0 when only I1 is present is given by eq. 3.

(3)

In this equation, K_I1 is the equilibrium dissociation constant for I1 with E, K_I2 is the equilibrium dissociation constant for I2 with E, and K_I12 is the equilibrium dissociation constant for I1 with the EI2 complex. The concentrations [E] and [E]_[I2] = 0 are proportional to the second-order rate constants for substrate hydrolysis noted under eq. 1, and their ratio herein was estimated from the relative v for acetylthiocholine hydrolysis.

Slow Equilibration of an Inhibitor with AChE

Association rate constants were determined from the progressive inhibition of AChE with time as the inhibitor approached equilibrium binding (Eastman et al., 1995; Szegletes et al., 1998). Binding was initiated at 23°C in buffer, in most cases under pseudo first-order conditions in which the concentration of I in the incubation mixture was adjusted to at least four times the concentration of AChE. At various times t, a 1.0-ml aliquot was removed to a cuvette, and 40 µl of acetylthiocholine and DTNB was added to a final concentration of 0.5 and 0.33 mM, respectively. The initial hydrolysis rate v was immediately recorded at 412 nm, and the series of v at various t was fitted to eq. 4 by a weighted nonlinear regression analyses (Fig. P), which assumed that v has a constant percentage of error.

(4)

In eq. 4, k is the observed pseudo first-order rate constant for the approach to equilibrium, and v_initial and v_final are the calculated values of v at time zero and at equilibrium, respectively. The dependence of k on [I] was analyzed by weighted linear regression analysis according to eq. 5, with k values weighted by the reciprocal of their variance.

(5)

Equation 5 assumes a simple bimolecular reaction of I with AChE, where k_I is the association rate constant and k_I is the dissociation rate constant (Eastman et al., 1995).

In some cases, hydrolysis rates v were fitted directly to a general steady-state solution of Scheme 1 with the program SCoP (version 3.51; Simulation Resources, Inc., Redlands CA) (Szegletes et al., 1998, 1999). This program solves rate equations numerically and allows fitting of v at times before inhibitor equilibrium.

AChE Peripheral Site Binding Affinity Determined by Fluorescence Titration

Binding of propidium to the AChE peripheral site was monitored by enhancement of the propidium fluorescence on a Perkin-Elmer LS-50B luminescence spectrometer in 1 mM sodium phosphate buffer and 0.02% Triton X-100 at 23°C (Taylor and Lappi, 1975). Measurements were performed in a low ionic strength buffer to increase the affinity of propidium (Taylor and Lappi, 1975). Propidium fluorescence was monitored with excitation at 500 nm and emission from 590 to 630 nm. Fluorescence contributions from scatter in the buffer and enzyme were subtracted. Total areas under the fluorescence emission curves (f) were fitted by nonlinear regression analysis (Fig. P) to eq. 6.

(6)

In eq. 6, f_a is the fluorescence intensity coefficient for free propidium, f_b is the fluorescence intensity coefficient for bound propidium, E_tot is the total enzyme concentration, L_tot is the total propidium concentration, K_D is the equilibrium dissociation constant, and D = E_tot + L_tot + K_D. Estimates of K_D are obtained with greatest accuracy from eq. 6 when L_tot < K_D. Data were fitted to eq. 6 with the calculated E_tot as the independent variable and L_tot at a fixed propidium concentration (0.2 and 2 µM for the binding to AChE and the huprine X-AChE complex, respectively) to give f_a, f_b, and K_D.

Molecular Modeling of Ternary Complex of AChE with Huprine X and Propidium

Construction and analyses of three-dimensional models were performed on a Silicon Graphics workstation Indigo2 IMPACT with QUANTA98 (Molecular Simulations, Inc., Waltham, MA). Modeling of the ternary complex of Torpedo californica AChE (TcAChE) with propidium and huprine X began with the crystal structure coordinates for the TcAChE-tacrine complex (Protein Data Bank file: 1ACJ) (Harel et al., 1993). Propidium was manually docked into the peripheral site with the same procedure as described previously (Barak et al., 1994; Szegletes et al., 1998). Huprine X was built from atomic coordinates of 9-methyl huprine (provided by Dr. Javier Luque, Departament de Fisicoquímica, Universitat de Barcelona) and manually docked into the acylation site to maximize the molecular overlap with the aminoquinoline portion of tacrine (Fig. 1). The resulting structure was optimized by energy minimization with CHARMm module of QUANTA98 (conjugate gradient) starting with refinement of the propidium, huprine X and gorge solvent molecules. Subsequent molecular optimization added all amino acid side residues in proximity to the propidium and huprine X ligands.

	Results and Discussion

Top Abstract Introduction Experimental Procedures Results and Discussion References

Comparison of Inhibition Constants for Various Inhibitors with AChE. Scheme 1 provides a framework for the analysis of inhibitors that bind to the acylation site of AChE. Steady-state inhibition data for huprine X and two other inhibitors in Fig. 1 is shown in Fig. 4. The reciprocal plots for all three inhibitors (Fig. 4, A-C) show both increasing slopes and increasing intercepts with higher inhibitor concentration. This pattern is often termed "mixed inhibition," and in Scheme 1, it is consistent with significant inhibitor interaction with both the free enzyme E and the acetyl enzyme EA. Replots of the normalized slopes versus the inhibitor concentrations (Fig. 4, D and F) gave estimates of K_I, the dissociation constant for inhibitor binding to E, as reported in Table 1. The K_I of 26 pM for huprine X indicates that it is one of the highest affinity reversible inhibitors for AChE yet reported. The peptide neurotoxin fasciculin, with a K_I for human AChE of ~10 pM under these buffer conditions (Eastman et al., 1995), has a slightly higher affinity. This K_I for huprine X was comparable with that for the bisquaternary inhibitor ambenonium (Hodge et al., 1992) and the transition state analog m-(N, N,N-trimethylammonio)trifluoroacetophenone (TMTFA) (Szegletes et al., 1998).³ As shown in Table 1, the affinity of huprine X was 180 times higher than that of huperzine A and 1200 times higher than that of tacrine, the two well known inhibitors of AChE upon which the hybrid structure of huprine X was based. Furthermore, the huprine X affinity was 40 times higher than that of E2020. Huprine Y, the analog of huprine X in which a methyl group replaces the ethyl group at position 9, gave a K_I similar to that of huprine X (Table 1). The high affinity of huprine X also was very selective for vertebrate AChE relative to other cholinesterases. K_I values for the inhibition of human butyrylcholinesterase and Drosophila AChE were 120 ± 12 and 55 ± 10 nM, respectively, more than three orders of magnitude higher than that for human AChE.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Steady-state inhibition of AChE hydrolysis of acetylthiocholine. A-C, reciprocal plots of initial velocities (A412 nm/min) and substrate concentrations were analyzed according to eq. 1. Huprine X concentrations in A were 0 (), 240 (), and 400 pM (). Tacrine concentrations in B were 0 (), 60 (), and 300 nM (). E2020 concentrations in C were 0 (), 10 (), and 40 nM (). D-F, slopes of plots in A-C were normalized by dividing by the slope in the absence of inhibitor and plotted against the inhibitor concentration according to eq. 2 to derive KI values (Table 1). Additional points from data not shown in A and C were included in D and F, respectively.

I

I

Determination of Association and Dissociation Rate Constants for Huprine X and AChE. AChE inhibitors with K_I values in the subnanomolar range often have very low dissociation rate constants that permit measurement of the kinetics of inhibitor association. This was the case with huprine X and huprine Y, as demonstrated in Fig. 5 and Table 1. Rate constants for the approach to equilibrium binding at a fixed concentration of huprine X were estimated as in Fig. 5A, and from the time course of these reactions values of k_I = 4.4 ± 0.3 × 10⁸ M¹min¹ and k_I = 0.009 ± 0.003 min¹ were obtained (Fig. 5B). The value of k_I we previously reported for huperzine A in the same buffer (Szegletes et al., 1998) is also shown in Table 1, and this value is only ~1% of the k_I value for either huprine X or huprine Y. In fact, the difference in k_I accounts for almost all of the difference in K_I between huperzine A and the two hybrid compounds, as the k_-I values were nearly comparable. However, the k_I value reported herein for huprine X still remained <20% of that previously reported for the bisquaternary AChE inhibitor ambenonium (Hodge et al., 1992) and <1% of that for the monoquaternary inhibitor N-methylacridinium (Rosenberry et al., 1996). We measured a pK_a of 8.9 for huprine X, indicating that the ring N should be largely protonated at pH 7, so huprine X should be cationic. Thus, the lower k_I for huprine X relative to ambenonium or N-methylacridinium appears to indicate some remaining steric or conformational restraints on its rapid association with the AChE active site.

View larger version (10K): [in this window] [in a new window]   Fig. 5.   Association and dissociation rate constants for huprine X and AChE. A, progressive inhibition of AChE (6 pM) with time as huprine X binding approached equilibrium was measured by removing 1-ml aliquots of the reaction mixture and immediately adding DTNB and the substrate acetylthiocholine for assay as outlined in Experimental Procedures. Hydrolysis rates v obtained with 25 () and 75 () pM huprine X were fitted to eq. 4 to obtain a value of k for each curve. B, values of k obtained at several concentrations of huprine X as in A were plotted against [huprine X]. Although the plot was nearly linear, the concentration of huprine X in some cases was not sufficiently in excess of the concentration of AChE to justify analysis with eq. 5. The SCoP fitting program was applied directly to all the association rate data (Szegletes et al., 1998) to obtain kI = 440 ± 30 µM1min1 and k-I = 0.009 ± 0.003 min1.

Huprine X Binds to Acylation Site but Slightly Interferes with Binding of Propidium to Peripheral Site. Because crystal structures of AChE show that both tacrine and huperzine A bind to the acylation site (Harel et al., 1993; Raves et al., 1997), we anticipated that huprine X also would bind to the acylation site. To explore this point, we used propidium, a ligand specific for the AChE peripheral site (Taylor and Lappi, 1975). If huprine X binding is confined to the acylation site, a ternary complex of AChE with huprine X and propidium should be able to form. In principle, such a complex can be detected by equilibrium titration or by kinetic analyses. For example, kinetic studies have shown that propidium and huperzine A form a ternary complex with AChE (Szegletes et al., 1998).

View larger version (9K): [in this window] [in a new window]   Fig. 6.   Fluorescence titration of recombinant human AChE with propidium in the absence (A) and presence (B) of huprine X. A total of three fluorescence spectra was collected and averaged for each AChE titration point as outlined in Experimental Procedures, and total areas under the fluorescence emission curves (F) were plotted against the final AChE concentration in the assay sample. A, propidium was added to a total concentration of 0.2 µM to AChE and then successively diluted with a solution of 0.2 µM propidium. The data were fitted to eq. 6 to give fa = 39 ± 4, fb = 200 ± 4, and KD = 0.63 ± 0.10 µM. B, because of the shift in the binding curve for propidium in the presence of huprine X, propidium was added to a higher total concentration of 2 µM to AChE preincubated with 100 µM huprine X and then successively diluted with a solution of 2 µM propidium and 100 µM huprine X. In the presence of huprine X, a plateau in the fluorescence enhancement indicating propidium saturation in the ternary complex was not obtained. However, if the same degree of fluorescence enhancement was assumed for propidium in the ternary complex as in the propidium-AChE binary complex (fb/fa = 5.1), the data fitted to eq. 6 gave fa = 26 ± 1 and KD = 190 ± 25 µM. This assumption is not necessarily valid, because some disruption of the propidium environment occurs in the presence of huprine X (Fig. 8).

View larger version (10K): [in this window] [in a new window]   Fig. 7.   Determination of the relative affinity of huprine X in the propidium-AChE complex (A) or the edrophonium-AChE complex (B) by inhibition of substrate hydrolysis. AChE (, 25 pM; , 1200 pM) was incubated without () or with () huprine X (4 nM) and the indicated concentration of propidium or edrophonium for 6 h at 23°C, and a 2.9-ml aliquot was mixed with 100 µl of DTNB and 12 µl of acetylthiocholine to give final concentrations of 0.33 mM DTNB and 20 µM acetylthiocholine. Points represent v observed over a 2- to 5-min interval normalized by the corresponding v obtained at the same [huprine X] in the absence of propidium (v[prop] = 0) or edrophonium (v[edro] = 0). Lines correspond to normalized v obtained by fitting the experimental data to a complete nonequilibrium reaction scheme with the program SCoP, with kinetic rate constants previously determined for acetylthiocholine and propidium or edrophonium (Eastman et al., 1995; Szegletes et al., 1999) and those in Table 1 for huprine X. The dashed lines correspond to the absence of huprine X and the solid lines, to 4 nM huprine X. The dotted lines correspond to the curves calculated when no ternary complex forms (1/KI12 = 0). Based on data with 4 nM huprine X as well as with 1 and 2 nM huprine X (data not shown) with both propidium and edrophonium, the procedure in A gave KI2 = 1.4 µM for propidium, KI1 = 20 pM for huprine X, and KI12/KI1 = 17 ± 3. The procedure in B gave KI2 = 0.2 µM for edrophonium, KI1 = 29 pM for huprine X, and KI12/KI1 > 1000. Lines also were fitted with the much simpler analysis scheme in eq. 3, which required the approximations that v reflects the true second-order rate constant, that the free huprine X concentration remains unchanged through the entire range of propidium or edrophonium concentrations, and that no slow reequilibration occurs in the presence of substrate. Although these approximations held only approximately at best, ratios of KI12/KI1 of 12 ± 2 with propidium and >200 with edrophonium were in reasonable agreement with those from the SCoP program.

View larger version (20K): [in this window] [in a new window]   Fig. 8.   Molecular modeling of the ternary complex of AChE with huprine X and propidium (see Experimental Procedures). AChE ligands and amino acids in the vicinity of the ligands are displayed with all other residues, solvent molecules and water omitted. The thin line indicates the orientations of propidium and amino acids in the propidium-TcAChE complex. The thick line indicates the effect of huprine X in the acylation site on the orientation of propidium, F330, and D72.

Huprine X May Be Useful in the Treatment of Alzheimer's Disease. The selectivity of an inhibitor of AChE in treating patients with Alzheimer's disease appears to parallel the affinity of the inhibitor for the AChE active site. Thus, E2020, with an affinity some 30 times higher than that of tacrine, is administered at a daily dose of only one-tenth that of tacrine.⁴ The lower dosage regimen results in fewer undesirable side effects with E2020 than with tacrine. If this trend is extended by huprine X, which has an AChE affinity yet 40 times higher than that of E2020, huprine X may prove to be an even more attractive drug for the treatment of Alzheimer's disease. Such optimization of the AChE inhibitor of therapeutic choice may provide the most effective treatment until other classes of drugs become available that target the basic cause of Alzheimer's disease.