Anal. Chem. 2005, 77, 5421-5426
Direct Chromatographic Determination of Dissociation Rate Constants of Ligand-Receptor Complexes: Assessment of the Interaction of Noncompetitive Inhibitors with an Immobilized Nicotinic Acetylcholine Receptor-Based Liquid Chromatography Stationary Phase Ruin Moaddel, Krzysztof Jozwiak,† Rika Yamaguchi,‡ and Irving W. Wainer*
Gerontology Research Center, National institute on Aging, National Institutes of Health, 5600 Nathan Shock Drive, Baltimore, Maryland 21224
A liquid chromatographic stationary phase containing immobilized membranes from a cell line expressing the r3β4 subtype of the neuronal nicotinic acetylcholine receptor (nAChR) has been used to assess dissociation rate constants (kd) of 12 noncompetitive inhibitor-nAChR complexes. The pharmacological effects of the noncompetitive inhibitors, expressed as percent recovery of activity at 7 min and 4 h postexposure to the inhibitor, were also determined. The results demonstrate that the kd values correlated with the pharmacological effect and that this approach can be used to identify molecular structures associated with differences in kd values. The method can be adapted for use with membrane-bound receptors, ion channels, and transporters and represents a direct and facile technique for the assessment of dissociation rate constants (kd) of ligand-receptor complexes. The dissociation of a ligand-receptor complex plays a key role in the pharmacological effect of the ligand. Thus, the ability to rapidly determine dissociation rate constants, kd’s, would be an important addition to drug discovery programs. Currently, stoppedflow analysis is the predominate method used to determine pharmacologically relevant kd’s. While this technique is well-known and accepted, it can also be technically challenging and lengthy. In some instances, the target ligands must be derivatized to enable detection and analysis, For example, the determination of the kinetics associated with the interaction of the 20- and 63-kDa fragments of anthrax protective antigen required the dual fluorescence labeling of the fragments and measurements using fluorescence resonance energy transfer (FRET).1 When membranebased receptors or transporters are studied, these proteins usually * Corresponding author. Phone: (410)-558-8498. Fax: 410-558-8409. E.mail:
[email protected]. † Current address: Medical University of Lublin, Staszica 6, 20-081 Lublin, Poland. ‡ Current address: Shionogi & Co., Ltd., Hyogo 660-0813, Japan. (1) Christensen, K. A.; Krantz, B. A.; Melnyk, R. A.; Collier, R. J. Biochemistry 2005, 44, 1047-1053. 10.1021/ac0504464 CCC: $30.25 Published on Web 07/12/2005
© 2005 American Chemical Society
need to be solubilized and reconstituted in proteoliposomes, as in the recent study of the GLUT1 sugar import and export sites.2 Surface plasmon resonance techniques have also been used to determine kd values. Using this approach, test ligands can be screened using an immobilized target if the target has been isolated and purified.3 Two other approaches have been used with membrane-bound proteins.4,5 In one approach, the ligand, R bungarotoxin, was immobilized, and the target, the nicotinic acetylcholine receptor (nAChR) contained in torpedo membranes, was solubilized and reconstituted in lipids.4 In the second, membranes containing the β-adrenergic receptor were immobilized through avidin-biotin binding using biotinylated membranes or through the initial immobilization of biotinylated antibodies that targeted an amino-terminal FLAG tag on the β-adrenergic receptor.5 Although surface plasmon resonance techniques can be used to determine kd values, in the case of membrane-bound targets, it is difficult to apply the method to the rapid screening of a library of potential ligands. The binding kinetics of agonists to membrane-bound ligandgated ion channels, such as the nAChR, have been studied using electrophysiological approaches.6,7 However, the patch-clamp techniques used in these studies are technically difficult and require multiple experiments. Thus, this approach is also not easily adapted to the rapid screening of compound libraries. An alternative method for the assessment of kd values associated with the binding of compounds to membrane-bound receptors and transporters has been suggested by recent studies involving a column containing membranes obtained from cell lines expressing neuronal nicotinic acetylcholine receptors.8,9 In these studies, noncompetitive inhibitors (NCIs) of the R3β4 nAChR were (2) Sultzman, L. A.; Carruthers, A. Biochemistry 1999, 38, 6640-6650. (3) Cannon, M. J.; Papalia, G. A.; Navratilova, I.; Fisher, R. J.; Roberts, R. L.; Anal. Biochem. 2004, 330, 98-113. (4) Kroger, D.; Hucho, F.; Vogel, H. Anal. Chem. 1999, 71, 3157-3165 (5) Neumann, L.; Wohland, T.; Whelan, R. J.; Zare, R. N.; Kobilka, B. K. ChemBioChem 2002, 3, 993-998. (6) Grosman, C.; Auerbach, A. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 1410214107. (7) Horak, M.; Vlcek, K.; Petrovic, M.; Chodounska, H.; Vyklicky, L. J. Neurosci. 2004, 24, 10318-10325.
Analytical Chemistry, Vol. 77, No. 16, August 15, 2005 5421
identified and characterized using an R3β4 nAChR column and nonlinear chromatography/chemometric techniques.8,9 This method could be used to determine the equilibrium constant for absorption, Ka; the rate constants associated with the formation, ka; and dissociation, kd, of a compound-nAChR complex. In one of these studies, the chromatographically determined kd values for dextromethorphan (1.01 s-1) and levomethorphan (1.55 s-1) were compared to the percent recovery of nicotinestimulated [86Rb+] efflux in a KXR3β4 cell line at 7 min post-NCI exposure, dextromethorphan 38%, levomethorphan 63%.9 This comparison suggested that chromatographically determined kd values could be used to assess the length of the inhibitory effect of NCIs on the pharmacological function of the nAChR. This hypothesis has been confirmed by the results from the current study. In this study, 12 compounds which had been previously identified as NCIs of the R3β4 nAChR10 were chromatographed on a liquid chromatography column containing immobilized membranes obtained from the KXR3β4 cell line. The chromatographically determined kd values were compared to the corresponding percent recovery of nicotine-stimulated [86Rb+] efflux at 7 min and 4 h post-NCI exposure in intact KXR3β4 cells. The results demonstrate that chromatography can be used in drug discovery to rapidly rank compounds by their pharmacological activity, in this case relative length of the inhibitory effect. This approach can also be used to identify molecular structures associated with differences in kd values and can be used in the optimization of lead drug candidates. EXPERIMENTAL SECTION Materials. Benzamidine, NaCl, MgCl2, CaCl2, KCl, MgSO4, cholate, leupeptin, phenyl methyl sulfonyl fluoride (PMSF), EDTA, Trizma, poly(D-lysine), HEPES, and glucose were purchased from Sigma-Aldrich Co. (St. Louis, MO). HPLC grade methanol, ammonium acetate, and 0.1 M ammonium hydroxide solution were purchased from Fisher Scientific (Pittsburgh, PA). Immobilized artificial membrane PC stationary phase (IAM-PC, 12 µm, 300 Å) was purchased from Regis Technologies Inc. (Morton Grove, IL). The NCIs used in this study are presented in Table 1 and Figure 1. Minimum essential medium with Earle’s salts and L-glutamine (MEM), fetal bovine serum (FBS), penicillin/streptomycin (P/S), and geniticin were purchased from Gibco (Carlsbad, CA). [86Rubidium+] chloride ([86Rb+]) was purchased from Perkin-Elmer (Boston, MA). Twenty-four-well plates were purchased from Fisher Scientific. Cell Line. The KXR3β4R2 cell line, which expresses the R3β4 nAChR subtype, has been previously described and was provided by Dr. Kenneth J. Kellar (Georgetown University, Washington, DC).11 The cell line was established and maintained as previously described.11 Preparation of the Immobilized r3β4 nAChR Column. Cellular membranes containing the R3β4 nAChR were (8) Jozwiak, K.; Haginaka, J.; Moaddel, R.; Wainer, I. W. Anal. Chem. 2002, 74, 4618-4624. (9) Jozwiak, K.; Hernandez, S. C.; Kellar, K. J.; Wainer, I. W. J. Chromatogr., B. 2003, 797, 373-379. (10) Jozwiak, K.; Moaddel, R.; Yamaguchi, R.; Ravichandran, S.; Collins, J. R.; Wainer, I. W. J. Chromatogr., B. 2005, 819, 169-174. (11) Xiao, Y.; Meyer, E. L.; Thompson, J. M.; Surin, A.; Wroblewski, J.; Kellar, K. J. Mol. Pharmacol. 1998, 54, 322-333.
5422 Analytical Chemistry, Vol. 77, No. 16, August 15, 2005
Table 1. The Dissociation Rate Constants (kd) Determined for the Interaction of the Compounds Used in This Study with Immobilized r3β4 NAChRs Using Nonlinear Chromatography and the Percent Recovery of Nicotine-induced [86Rb+] Efflux in KXr3β4 Cells after Exposure to the Compounds Used in This Study % recovery (n ) 4) compound Nifedipineb,c (M + H+) ) 347 Nicardapineb,c (M + H+) ) 480 Dizocilpine(MK-801)b,c (M + H+) ) 226 Mecamylamineb,c (M + H+) ) 271 Laudanosineb,c (M + H+) ) 372 Verapamilb,c (M + H+) ) 456 Methoxyverapamilb,c (M + H+) ) 486 Dextromethorphanb,c (M + H+) ) 272 3-Hydroxymorphinanb,d (M + H+) ) 258 Diltiazemb,e (M + H+) ) 416 Desacetyl Diltiazemb,d (M + H+) ) 374 N-demethyl Diltiazemb,e (M + H+) ) 402
kd
(s-1)
r2 a
Fa
7-min
4-h
4.4
0.996 3.3 × 105 78.1 ( 2.4
98.5 ( 1.5
0.5
0.995 4.1 × 105 28.8 ( 2.9
96.8 ( 2.2
3.5
0.974 0.7 × 105 92.4 ( 5.3
98.4 ( 6.9
5.0
0.980 1.0 × 105 92.9 ( 4.1 101.7 ( 3.2
2.2
0.978 0.9 × 105 88.4 ( 2.7
96.9 ( 4.8
0.7
0.992 2.5 × 105 37.9 ( 7.3
98.8 ( 4.9
0.7
0.995 2.3 × 105 48.9 ( 19.9 96.5 ( 5.7
1.0
0.998 1.6 × 105 50.9 ( 6.6
96.4 ( 2.0
2.0
0.987 0.9 × 105 100.5 ( 3.8
99.3 ( 5.1
1.7
0.979 0.9 × 105 52.1 ( 2.4
97.9 ( 3.0
1.9
0.971 0.8 × 105 67.4 ( 3.4
98.5 ( 2.9
1.8
0.970 0.8 × 105 68.1 ( 3.3
97.1 ( 4.2
a r2 and F are the measurement of the fit of the chromatographic data with the calculated kd values. b (M + H+) represents the molecular ions used in the mass spectrometric detection of the compounds. c Purchased from Sigma Chemical Company (St.Louis, MO). d Kindly donated from Hoffmann La Roche (Nutley, NJ). e Kindly donated from Marion Merril Dow, Inc. (Kansas City, KS).
immobilized on the IAM stationary phase to produce the R3β4 nAChR-IAM stationary phase as previously described.12 Chromatographic System and Conditions. The LC/MS system consisted of a LC10AD pump (Shimadzu, Columbia, MD), an ESA 540 autoinjector (Spark-Holland), and a Micromass Q-TOF mass spectrometer (Micromass, Beverly, MA). The data were recorded and processed using MassLynx v.3.5. (Micromass). The mobile phase was composed of ammonium acetate (10 mM, pH 7.4) modified with methanol in a ratio of 85:15 (v/v) delivered at a flow rate of 0.2 mL/min. Aqueous solutions (10 µM) of each compound were prepared, and after equilibration of the column (∼2 h), 20 µL of each solution was sequentially injected into column. Positive ionization mass spectrometry (ESI+) was used to monitor the parent molecular ions (M + H+) (see Table 1). Determination of the Dissociation Rate Constant (kd). The details of the nonlinear chromatographic approach and this application of this technique to the determination of kd on an immobilized R3β4 nAChR stationary phase have been presented elsewhere.8 Briefly, the Impulse Input Solution program implemented in commercially available software for chromatographic signal deconvolution was used to fit the experimentally obtained peak profiles. Nicotine-Stimulated [86Rb+] Efflux Studies in KXr3β4R2 Cells. Nicotine-stimulated [86Rb+] efflux studies were carried out (12) Moaddel, R.; Jozwiak, K.; Whittington, K.; Wainer, I. W. Anal. Chem. 2005, 77, 895-901.
Figure 2. The nonlinear chromatographic traces for 3-hydroxymorphinan (1) and dextromethorphan (2) obtained on an immobilized R3β4 nAChR column.
RESULTS AND DISCUSSION
Figure 1. Chemical structures of the compounds used in this study.
with the KXR3β4R2 cells using a modified version of a previously described procedure.9 In this study, cells were grown in selection growth medium (500 mL MEM, 10% FBS, 1% p/s, 350 mg of geniticin). Once the cells reached >90% confluence, they were plated (1 mL/well) on 24-well plates coated with poly(D-lysine) (>300 000 MW). These plates were then incubated for 48 h at 37 °C to reach >90% confluence. A 1-µL aliquot of the NCIs (10 mM in methanol) was added to each well for a final concentration of 10 µM NCI and the plates were incubated for 24 h. The medium was removed, and the cells were incubated for 4 h at 37 °C with 0.5 mL of [86Rb+] (2 µCi/well) in growth medium containing a 10 µM concentration of the NCIs. After incubation, the medium was aspirated, and the wells were washed for 2 min with 1 mL of Hepes buffer (15 mM, pH 7.4) containing 140 mM NaCl, 2 mM KCl, 1 mM MgSO4, 1.8 mM CaCl2, and 11 mM glucose (buffer A). The process was repeated two additional times, first with a wash of 2 min and then with a final wash of 7 min. One-milliliter solutions of either buffer A (control) or buffer A with 100 µM nicotine (experimental) were then added to the wells for 7 min or 4 h, after which they were transferred to scintillation vials and counted for [86Rb+] concentration using liquid scintillation counting. Subsequently, 500 µL of 0.1 N NaOH was added to each well, the plates were rotated at room temperature for 30 min, and the cell lysates were collected and counted. The amount of [86Rb+] efflux was calculated as a percentage of the total [86Rb+] loaded. The experiments were run in quadruplicate.
Nonlinear Chromatography and Calculation of kd Values. The compounds used in this study are presented in Figure 1. The chromatographic traces produced by each of these compounds contained an asymmetric peak, as illustrated by the chromatographic traces produced by dextromethorphan and 3-hydroxymorphinan, Figure 2. The shape of a chromatographic peak is a function of the specific and nonspecific interactions between the solute and the stationary phase, in particular, the kinetics involved in the formation and dissolution of the solute-stationary phase complex, that is, the association and dissociation rate constants, ka and kd, respectively. When the stationary phase contains an immobilized protein, the dissociation of a ligand-protein complex is usually slower than the rate of complex formation producing non-Gaussian peaks. The position and width of non-Gaussian chromatographic peaks have been used to determine the affinity, Kd, and kinetics, kd and ka, of the binding of D- and l-tryptophan to immobilized human serum albumin.13,14 This method provides pharmacologically relevant parameters but has not been utilized with membranebound proteins. An alternative approach to the analysis of non-Gaussian peak shapes was developed by Carr et al. and defined in the context of nonlinear chromatography.15 The nonlinear chromatography model assumes that (1) the stationary phase contains a limited number of immobilized active sites, (2) kinetic rates of adsorption and desorption are the primary source of band-broadening and peak skew, and (3) dispersive and extracolumn effects are negligible. Using the analysis developed by Carr et al.,15 the dissymmetries of the chromatographic peaks were used to calculate the kd values (Table 1). In this approach, the Impulse Input Solution program, implemented in commercially available software for chromatographic signal deconvolution, was used to fit the experimentally obtained peak profiles. The kd values were obtained from the (13) Yang, J.; Hage, D. S. J. Chromatogr., A 1997, 766, 15-25. (14) Talbert, A. M.; Tranter, G. E.; Holmes, E.; Francis, P. L. Anal. Chem. 2002, 74, 446-452. (15) Wade, J. L.; Bergold, A. F.; Carr, P. W. Anal. Chem. 1987, 59, 1286-1295.
Analytical Chemistry, Vol. 77, No. 16, August 15, 2005
5423
Figure 3. Experimental peak profiles overlaid with fitted nonlinear functions (1) and fitted nonlinear functions alone (2) for verapamil and dilitiazem.
broadness of the chromatographic traces according to the equation
kd ) 5424
1 a2t0
Analytical Chemistry, Vol. 77, No. 16, August 15, 2005
(1)
where a2 is the peak’s broadness (width parameter) and t0 is the void time of the chromatographic system. The relationship between the broadness of the chromatographic peak and the calculated kd values is illustrated by dextromethorphan and 3-hydroxymorphinan (Figure 2). The
Figure 4. The correlation observed between the dissociation rate constants (kd) determined for the interaction of the compounds used in this study with immobilized R3β4 nAChRs using nonlinear chromatography and the percent recovery of nicotine-induced [86Rb+] efflux in KXR3β4 cells at 7 min after exposure to the compounds used in this study.
pronounced asymmetry of the dextromethorphan peak relative to the 3-hydroxymorphinan peak results in a 2-fold lower calculated kd value, 1.0 s-1 and 2.0 s-1, respectively (Table 1). The experimentally obtained peaks and the peaks obtained from the fitted curves generated using the nonlinear chromatography functions were compared using the “goodness of fit” parameters r2 and F (Table 1). The peak-tailing observed in the experimentally obtained chromatographic traces was not completely described by the nonlinear chromatography modeling (Figure 3). This “extra tailing” was displayed by each compound studied during this project and was treated as a systematic error. This approach does not appear to interfere with the relative comparison of chromatographically determined kd values and functional receptor recovery (see below). Percent Recovery of Pharmacological Activity. The percent recovery of the nicotine-induced [86Rb+] efflux in KXR3β4R2 cells was determined at 7 min and 4 h post-NCI exposure. Previous studies carried out at 7 min and at 4 and 24 h had determined that the first two time periods were enough to describe recovery.9 In this study, significant differences in the effect of the NCIs were observed at the 7-min time point, with the percent recovery ranging from a low of ∼29% (nicardapine) to essentially complete recovery, ∼100% (3-hydroxymorphinan) (Table 1). At the 4-h time point, all of the efflux activity had been fully recovered for each of the NCIs tested (Table 1). Comparison of kd Values and Percent Recovery. It was not possible to compare chromatographically obtained kd values to similar rate constants obtained using an alternative experimental approach, since these data do not exist for NCIs of the nAChR and are not easily attainable. An alternative is to compare the chromatographic data with an indirect measure of the dissociation rate, the time course of the recovery of functional activity after exposure to the NCI.16 The chromatographically determined kd values were compared to the corresponding percent recovery using Microsoft Excel, and a nonlinear correlation was observed between these two sets of data with an r2 value of 0.7458 (p ) 0.0048) (Figure 3). The data (16) Papke, R. L.; Sanberg, P. R.; Shytle, R. D. J. Pharmacol. Exp. Ther. 2001, 297, 646-656.
were also tested for statistical significance using a Spearman correlation, and the two-tailed P value was
