Investigation of the Phenylalkylamine Binding Site in hKv1.3(H399T), a Mutant with a Reduced C-Type Inactivated State
Department of Applied Physiology, University of Ulm, Ulm, Germany
Abstract
To screen for residues of hKv1.3 important for current block by the phenylalkylamine verapamil, the inactivated-state-reduced H399T mutant was used as a background for mutagenesis studies. This approach was applied mainly to abolish the accumulation in the inactivated blocked state, recovery from which in the wild type is normally slow. Substitution of amino acids in the S6 transmembrane helix indicated a heavy disruption of verapamil block by the A413C mutation, reducing the IC50 from 2.4 to 267 e. Subsequent scanning for verapamil moieties essential for current block was performed by application of derivatives with altered side groups. Neither the removal of the nitrile or the methyl group nor the addition of a methoxy group resulted in major variations of IC50 values for hKv1.3 (H399T) current block. However, disruption of current block by A413C was 4- to 10-fold less pronounced for derivatives lacking the 4-methoxy group of the (3,4-dimethoxyphenyl)ethylmethyl-amino part (devapamil) or all four methoxy groups (emopamil), respectively. Emopamil displayed a Hill coefficient of 2 for hKv1.3 (H399T/A413C) instead of 1 for hKv1.3 (H399T) current block. These results might indicate that the alteration of Ala413 modulates the access of phenylalkylamines to their binding site depending on the occupancy of the phenyl rings with methoxy groups. A computer-based docking model shows a subset of docked PAA conformations, with a spatial proximity between the (4-methoxyphenyl)ethyl-methyl-amino group and Ala413. The PAA binding site might therefore include a binding pocket for the aromatic ring of the ethyl-methyl-amino part in an S6eCS6 interface gap.
Various small molecules have been shown to block ion fluxes through channel proteins by physically occluding the pore. In the case of the voltage-gated potassium channel Kv1.3, such structurally distinct compounds as verapamil (phenylalkylamines), WIN-17317-3 (iminodihydrochinolines), UK-78282 (benzyl piperidines), correolide (triterpenes), 4-phenyl-4-[3-(2-methoxyphenyl)-3-oxo-2-azaprop-1-yl] cyclohexanone (cyclohexyl-substituted benzamides), or Psora-4 (5-phenylalkoxypsoralens) display inhibitory effects (reviewed in Wulff et al., 2003; Vennekamp et al., 2004). An obvious possibility is to use these small molecules as probes to investigate the residues and spatial requisites of the receptor necessary for binding. Enhanced knowledge of these parameters could provide useful information for the design of more specific and potent drugs with potential therapeutic use. The phenylalkylamine (PAA) verapamil was used in this study to explore the binding site of hKv1.3, a channel that plays an important role in T-cell activation and B-cell differentiation (Chandy et al., 1984; DeCoursey et al., 1984; Wulff et al., 2004). Many data have been reported already, determining the basic mode of verapamil block of Kv1.3 currents, thereby providing a good starting point for a detailed investigation of the binding pocket. Verapamil acts as a fast open-channel blocker with an IC50 of approximately 8 e (Rauer and Grissmer, 1996), and block takes place from the intracellular side of the cell. Mutagenesis and competition studies suggest that the binding site is situated in the water-filled cavity below the selectivity filter (Rauer and Grissmer, 1999). The latter results might indicate that side groups of verapamil are in close contact with pore-facing residues of the transmembrane S6 helix. Hanner et al., 2001 reported that various residues in this area are important for the binding of correolides. For L-type Ca2+ channels, it has been shown that the PAA binding site involves residues in the S6 segments of domains III and VI (Hockerman et al., 1995, 1997; Schuster et al., 1996) and a homology model based on the KcsA channel structure has been established presenting the PAA devapamil with the aromatic ring of the phenylethylamine part located inside the S6 interface III/IV crevice (Lipkind and Fozzard, 2003).
A specific property of Kv1.3 that distinguishes it from other Kv subfamily members is the intrinsic C-type inactivated state displayed by the channel. This state is reached slowly by all channel molecules upon depolarization, and recovery from inactivation is slow. Various drugs are known to bind selectively to the open and/or inactivated state; e.g., verapamil application leads to open blocked and inactivated blocked conformations of the channel (Rbe and Grissmer, 2000). A substantial problem of investigating drug effects on Kv1.3 currents is the difficulty to electrically distinguish between the access of the drug to either state. Upon verapamil application, one can observe an acceleration of current decay during the depolarizing steps and an accumulation of peak current block in consecutive voltage steps. The latter effect is caused by an accumulation of the channel in the inactivated blocked state, recovery from which is very slow (Rbe and Grissmer, 2000). To address this problem, we introduced the H399T mutation in the extracellular loop between the pore and the S6 helix of hKv1.3, because it has been described to strongly reduce the C-type inactivated state (Nguyen et al., 1996; Rauer and Grissmer, 1996). This reduction of complexity makes the direct calculation of IC50 values feasible by measurement of steady-state current block. Additional amino acid substitutions in the pore-facing parts of the S6 transmembrane helix (Fig. 1) revealed that the residue at position 413 plays an important role for PAA binding. Thereafter, to detect important side groups of the ligand, block by several verapamil derivatives has been examined and the obtained results pointed out reciprocal interactions between the mutated channel residue and exchanged verapamil side groups. Double-mutant cycle analysis calculations suggested a spatial proximity between both interaction partners, information that was used to evaluate the computer based docking of a PAA molecule into an hKv1.3 homology model.
Materials and Methods
Cells. The COS-7 cell line was obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (Braunschweig, Germany). The cells were maintained in Dulbecco's modified Eagle's medium with high glucose (Invitrogen, Carlsbad, CA) containing 10% fetal calf serum (PAA Laboratories GmbH, Coelbe, Germany) and stored in an incubator at 37°C and 10% p(CO2).
Chemicals and Solutions. All measurements were performed in an external bath solution, high (containing 160 mM NaCl, 2 mM CaCl2, 1 mM MgCl2, 4.5 mM KCl, and 5 mM HEPES) or high (containing 164.5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, and 5 mM HEPES). Osmolarity was 290 to 320 mOsM, and pH was adjusted to 7.4 with NaOH or KOH, respectively. The internal pipette solution contained 155 mM KF, 2 mM MgCl2, 10 mM EGTA, and 10 mM HEPES. Osmolarity was 290 to 320 mOsM, and pH was adjusted to 7.2 with KOH. All verapamil compounds were dissolved in DMSO as stock-solutions and diluted to the final concentration in external bath solution before application. DMSO fraction in the final solution was always <2%. (±)-verapamil, (eC)-verapamil, (+)-verapamil, and (±)-gallopamil were obtained as hydrochloride salts from Sigma-Aldrich (St. Louis, MO). (eC)-devapamil (LU 47093), (±)-norverapamil (LU 41723) and (±)-acyanoverapamil (LU 44935) were provided as hydrochloride salts by Drs. Raschack and Paul of Knoll Pharmaceuticals AG (now merged in Abbott GmbH and Co. KG (Ludwigshafen, Germany). (±)-emopamil (LU 37076) was a generous gift from Dr. Hans-Gether Knaus and Dr. Hartmut Glossmann (Institut fe Biochemische Pharmakologie, Medizinische Universitt Innsbruck, Innsbruck, Austria). KTX (Latoxan, Valence, France; Bachem, Bubendorf, Switzerland) was dissolved in external bath solution containing 0.1% bovine albumin (Sigma-Aldrich). TEA chloride was purchased from Fluka (Buchs, Switzerland) and dissolved in external bath solution to the respective final concentration.
Electrophysiology. The whole-cell recording mode of the patchclamp technique (Hamill et al., 1981) was used throughout all measurements. Experiments were carried out at room temperature (18eC22°C), and external bath solution and diluted drug solutions were applied using a gravity-based permanent perfusion system. Electrodes were pulled from glass capillaries (Science Products, Hofheim, Germany) in three stages and fire-polished to resistances of 2 to 4 M. Data were acquired with an EPC-9 patch-clamp amplifier (HEKA Elektronik, Lambrecht/Pfalz, Germany) connected to a Macintosh computer running Pulse/Pulse Fit ver. 8.40 data acquisition and analysis software (HEKA). All currents were filtered by a 2.9-kHz Bessel filter and recorded with a sampling frequency of 2.00 kHz. Capacitative and leak currents were subtracted, and series resistance compensation (80%) was used for currents exceeding 2 nA. Further data analysis was performed using the Igor Pro 3.1 (Wavemetrics, Lake Oswego, OR) software package. Holding potential was usually eC80 mV (hKv1.3 mutants) or eC120 mV (hKv1.3 wild type). Time constants m and h and steady-state currents with and without drug application were determined by a Hodgkin and Huxley fit to currents elicited by a 200-ms voltage pulse from the holding potential to +40 mV. The deactivation time constant t was obtained by an exponential fit to tail currents at eC60 mV observed after a voltage pulse from the holding potential to +40 mV for 200 ms. Parameters V1/2 and k were deduced from a Boltzmann fit to either normalized maximum peak conductance elicited by 11 consecutive voltage steps from eC50 to +50 mV (V = 10 mV) in high or to maximum tail peak currents at eC80 mV after the voltage steps in high . Drug effects were evaluated by the steady-state current block obtained during 200-ms voltage steps from the holding potential to +40 mV. Time interval between pulses was 30 s, allowing the channel to recover from inactivation completely. IC50 values were calculated by fitting a Hill function (Idrug/I0 = 1/[1+(IC50/Cdrug)nH]) to the normalized steady-state current data points obtained after drug application. I0 represents the steady state current before and Idrug the steady state current after drug application, Cdrug the concentration of the drug in the external bath solution and nH the Hill coefficient.
Molecular Biology. The hKv1.3 wild-type plasmid was a generous gift from Prof. Dr. O. Pongs (Institut fe Neurale Signalverarbeitung, Zentrum fe Molekulare Neurobiologie, Hamburg, Germany). It contains the human Kv1.3 K+ channel gene in a pRc/CMV vector (Invitrogen) with a CMV promoter for protein expression in mammalian cells. Channel mutants were generated by introducing the corresponding point mutations in the cloned hKv1.3 gene with the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). Transfection of COS-7 cells was done using the Fugene 6 transfection reagent (Roche Molecular Biochemicals, Mannheim, Germany). Cells were grown to 80% confluence and cotransfected with 1 e of hKv1.3 DNA and 0.5 e of eGFP-N1 DNA (BD Biosciences Clontech, Palo Alto, CA). One day after transfection sufficient protein for electrophysiological measurements was expressed.
Double Mutant Cycle Analysis. The coupling coefficient was calculated by the equation = [IC50(Rwt, Lwt) x IC50(Rmt, Lmt)]/[IC50(Rwt, Lmt) x IC50(Rmt, Lwt)], according to Hidalgo and MacKinnon (1995). Rwt was the receptor "wild type" hKv1.3 (H399T) and Rmt was the H399T/A413C double mutant. Lwt represents the "wild-type" drug [(±)-verapamil or (eC)-verapamil] and Lmt represents the "mutant" drug (a verapamil derivative with an exchanged side group). If coupling coefficients of less than 0 were obtained, the reciprocal value was taken. The change of free binding energy (G) was calculated by the equation G = RTln, where R is the gas constant and T is absolute temperature.
Results
hKv1.3 (H399T) As a Tool for Determination of Steady State PAA Block. To minimize the time the channel spends in the C-type inactivated state of hKv1.3, the H399T mutation, which is analogous to the H404T mutation in mKv1.3, was introduced (Nguyen et al., 1996; Rauer and Grissmer, 1996). Figure 2 shows the K+ currents through hKv1.3 wild type (A) and H399T mutant (B) channels, without drug application (pulse 1). The inactivation time constant (h) of hKv1.3 (H399T) is approximately 2.3-fold slower compared with wild type (Fig. 2; Table 1), and investigation of the kinetics for recovery from inactivation (Fig. 3, A and B) revealed that the corresponding time constant rec of hKv1.3 H399T is approximately 10 times faster (Table 1) than in case of the wild type. The intrinsic activation parameters (V1/2, k, and m) of hKv1.3 (H399T) are similar to the corresponding wild-type values, whereas the deactivation time constant t at eC60 mV is increased by a factor of approximately 3.7 (Table 1). Application of 20 e (±)-verapamil resulted in a steady-state block of hKv1.3 (H399T) currents at the end of a 200-ms pulse and no further accumulation of peak current block in consecutive pulses (Fig. 2B, pulse 2eC4). The normalized steady-state current after PAA application should therefore directly reflect the ratio of channels in the open blocked and open states. The remaining inactivation of the mutant hKv1.3 channels after 200 ms was ignored for the evaluation of steady-state block.
All values are given as mean ± S.D.; numbers of independent experiments are given in parentheses.
Identification of hKv1.3 (H399T) Residues Important for PAA Block. Several amino acids in the S6 transmembrane helix of hKv1.3 (H399T) have been substituted, because effects of these residues on current block by verapamil or similar drugs to Kv1.3 or other ion channels have been described or assumed. Substitutions were performed mainly with cysteines to allow a chemical modification in future experiments. Exceptions were the H399T/S410G and H399T/G416A double mutants. In the first case, the corresponding glycine of Kv2.1 was introduced; in the second case, an alanine was introduced because of the critical function of Gly416 as the potential gating hinge (Jiang et al., 2002b). The obtained double mutants, H399T/S410G, H399T/C412A, H399T/L418C, and H399T/I420C, had only minor or moderate effects on (±)-verapamil block (data not shown). The H399T/G416A mutation resulted in no functional expression of channel protein, possibly because of the disruption of the gating hinge. Another double mutant, H399T/A413C, displayed an affinity for (±)-verapamil that was reduced by a factor of approximately 100 (Table 2) compared with hKv1.3 (H399T). Steady-state current block with no further accumulation of block could still be observed in this channel mutant (Fig. 2C, pulses 2eC4). Furthermore, hKv1.3 (H399T/A413C) exhibits an even slower intrinsic inactivation rate (2-fold), a similar recovery from inactivation (Fig. 3C; Table 1), and an accelerated deactivation rate (10-fold) compared with hKv1.3 (H399T) (Fig. 2C, pulse 1; Table 1). The activation parameters (V1/2, k, m) also remain similar.
All values were obtained from at least four independent experiments and are given as mean ± S.D.
Current Block by Externally Applied KTX and TEA Is Affected by Introduction of A413C in hKv1.3 (H399T). Effects on the affinity for externally applied KTX and TEA have been examined to investigate whether the introduced cysteine decreases the affinity for (±)-verapamil merely by a direct interaction with the drug or by allosteric changes of the three dimensional protein structure. Both inhibitors are known to bind to the extracellular vestibule of Kv1.3 (Kavanaugh et al., 1991; Wrisch and Grissmer, 2000) and could serve as probes for structural differences in the extracellular region of the double mutant. In the case of hKv1.3 (H399T), IC50 values of 25.9 ± 2.5 mM for TEA and 0.25 ± 0.02 nM for KTX (Fig. 4, A and B, filled bars) were obtained. IC50 values for KTX and TEA block differed for hKv1.3 (H399T/A413C), with 5.8 ± 0.4 mM for TEA and 0.66 ± 0.06 nM for KTX (Fig. 4, A and B, open bars).
Importance of A413 for the Stereoselectivity of Verapamil Block. Verapamil is a chiral compound and current block by both enantiomers was investigated. These experiments were necessary for comparison with current block by (eC)-devapamil and to select the appropriate enantiomer for subsequent docking computations. (eC)-Verapamil blocked hKv1.3 (H399T) currents with an approximately 3.6-fold higher potency compared with the (+)-enantiomer (Table 2). It is surprising that the introduced A413C mutation reduced the affinity for (eC)-verapamil by a factor of 200, whereas the affinity for (+)-verapamil was reduced by a factor of 80 only (Table 2). Thus, the ability of the channel to discriminate between both enantiomers seems to be almost lost in the double mutant, leaving a preference of 1.4-fold for (eC)-verapamil over (+)-verapamil.
Inhibition Properties among Distinct Verapamil Derivatives Are Different for hKv1.3 (H399T) and hKv1.3 (H399T/A413C). Verapamil contains various functional groups, each of which might or might not play an important role in blocking Kv1.3 currents. For this reason, block of currents through hKv1.3 (H399T) and the double mutant H399T/A413C by several verapamil derivatives (Fig. 5) was investigated. Compared with (±)-verapamil, the compounds (±)-emopamil, (±)-gallopamil, (±)-norverapamil, and (±)-acyanoverapamil displayed only subtle variations in the IC50 values for hKv1.3 (H399T) current block. The same result was obtained for the (eC)-verapamil compared with the (eC)-devapamil IC50 value. However, in case of the hKv1.3 (H399T/A413C) double mutant, the results were partially different. Whereas (±)-gallopamil, (±)-norverapamil, and (±)-acyanoverapamil IC50 values are still similar to that of (±)-verapamil, (±)-emopamil block is approximately 10 times less affected by the A413C mutation. Another remarkable finding was that (±)-emopamil block of hKv1.3 (H399T/A413C) currents was the only case in which the dose-response fits resulted in a Hill coefficient of 2 instead of 1 (Fig. 6). Comparison of (eC)-devapamil and (eC)-verapamil IC50 values for the double mutant also suggests a smaller influence of A413C on (eC)-devapamil current block, which is 4-fold more potent compared with (eC)-verapamil.
Double Mutant Cycle Analysis of the Obtained IC50 Values. To detect coupled interactions caused by the substitution of the alanine at position 413 and the alteration of the different side groups of verapamil, the coupling coefficients () and the change of free binding energies (G) were calculated for all reactions (Table 3). In this approach, interdependency between two interaction partners is indicated by values deviating from 1. Whereas the coupling coefficients calculated for gallopamil, norverapamil, and acyanoverapamil stayed near 1, calculated for the (eC)-verapamil/(eC)-devapamil pair was 2.5, and the corresponding change of free binding energy was 0.52 kcal/mol. These data suggest a mutual influence of the deletion of the methoxy group of verapamil and the substitution of A413 of hKv1.3 (H399T). Furthermore, an interpretation of the correlations made by Schreiber and Fersht, 1995 suggests that a change in free binding energy of >0.5 kcal/mol might indicate a spatial distance between both interaction partners of less than 5 .
Building an hKv1.3 Homology Model and Docking of a PAA. To accommodate the inability of verapamil to bind to the closed conformation of Kv1.3 (DeCoursey, 1995), a suitable model for a K+ channel in the open state had to be selected. The X-ray structure of MthK, a Ca2+ gated K+ channel in the open conformation (Jiang et al., 2002a), was chosen as a scaffold for the homology model of hKv1.3. The hKv1.3 primary sequence from positions 352 to 431 was aligned to MthK, covering the S5 and S6 transmembrane helices and the pore region (Deep View software package; Guex and Peitsch, 1997). Thereafter, the hKv1.3 homology model was calculated using the SWISS-MODEL server (Peitsch, 1995; Guex and Peitsch, 1997; Schwede et al., 2003). Finally, docking was performed with the AutoDock 3.0 software package (Goodsell and Olson, 1990; Morris et al., 1996, 1998) using a modified (eC)-verapamil molecule as a ligand. This molecule lacks the R4 nitrile and R5 methyl groups, because these side groups did not show an important effect on current block in either the H399T or the A413C/H399T mutant. Of 256 docked conformations obtained, only those 31 were selected with the C atom of the methoxy group situated less than 5 away from the C atom of one of the four A413 side chains. No strength of interaction was implicated. Despite the symmetrical nature of the channel model, 24 methoxy groups were in proximity to Ala413 from a single subunit (Fig. 7, cyan). Within this group, two populations of similar conformations could be identified: cluster I (Fig. 7, magenta), which consisted of 12 PAAs situated in a stretched conformation in the inner vestibule of the channel with their ,-dimethoxyphenyl rings inserted slightly between two S6 helices, and cluster II (Fig. 7, gray), which contained four molecules in a similar stretched configuration but with their ,-dimethoxyphenyl rings extended deeper into the S6 interface crevice.
Discussion
The purpose of this study was to characterize the binding site of verapamil in the hKv1.3 mutant H399T, a channel mutant with a reduced C-type inactivated state. Mutations of residues in the S6 domain and application of different verapamil derivatives revealed interdependencies between position 413 of the channel and functional groups of verapamil. This information was used to deduce a more detailed description of the binding pocket.
As anticipated, the introduced H399T mutation reduced the channel's C-type inactivation but not to the extent described for an analogous mKv1.3 (H404T) mutant (Rauer and Grissmer, 1996). Nonetheless, hKv1.3 (H399T) was selected as a Kv1.3 model channel with a strongly reduced C-type inactivated state because the time constant for recovery from inactivation was approximately 10-fold faster compared with the wild-type channel. Steady-state currents were clearly observable with or without drug application after 1 s (data not shown) voltage pulses to +40 mV and no accumulation of steady-state current block in consecutive pulses was present.
Subsequent substitutions of amino acids in the S6 domain of hKv1.3 (H399T) pointed out that the residue at position 413 plays an important role in PAA binding. By exchanging the alanine at position 413 for a cysteine, (±)-verapamil affinity was decreased by approximately 100-fold. Similar results were obtained by Hanner et al. (2001) for an hKv1.3 (A413C) mutant that displayed a 40-fold lower affinity for diTC binding compared with the wild type. Furthermore, the corresponding residue in IVS6 of L-type Ca2+ channels had a strong influence on PAA affinity (Hockerman et al., 1995).
The additional A413C mutation slowed the C-type inactivation rate even further, an effect that had also been observed for the hKv1.3 (A413C) single mutant (data not shown). More evidence for A413 playing a role in the inactivation process was found by Panyi et al. (1995), where a strong increase of the inactivation rate in an hKv1.3 (A413V) mutant was described.
It has been reported that binding of external TEA slowed C-type inactivation (Grissmer and Cahalan, 1989; Choi et al., 1991) and that the transition to the inactivated state induced structural changes in the outer pore region (Yellen et al., 1994; Liu et al., 1996; Kiss et al., 1999). Further disruption of the inactivation process by introduction of A413C might therefore explain the affinity shift observed for TEA. However, no correlations between KTX binding and C-type inactivation have been described yet. The 3-fold decrease of KTX and 5-fold increase of TEA affinity in hKv1.3 (H399T/A413C) might be due to a structural rearrangement of the channel protein in the double mutant that affects the extracellular vestibule. This rearrangement might change the intracellular vestibule as well, which could explain the dramatic decrease in verapamil affinity.
Another explanation for the disruption of PAA block might be the loss of direct effects of the residue at position 413 on binding of distinct verapamil side groups. Verapamil derivatives lacking the methyl group at R5 (norverapamil) or the nitrile group at R4 (acyanoverapamil) did not show major variations in hKv1.3 (H399T) and hKv1.3 (H399T/A413C) current block compared with (±)-verapamil. Thus, neither the R5 methyl group nor the R4 nitrile group seems to play an important role for PAA binding in both channel mutants. The importance of the four methoxy groups at the R1, and R2, positions of both phenyl rings was assessed by investigation of (±)-emopamil block, a verapamil derivative that lacks all four moieties. In case of hKv1.3 (H399T) only little difference could be observed between the affinities for (±)-verapamil and (±)-emopamil, suggesting no major relevance of the methoxy groups. However, as an inhibitor of hKv1.3 (H399T/A413C) currents, (±)-emopamil is an approximately 10-fold more potent than (±)-verapamil. In addition, fits of the corresponding (±)-emopamil dose-response curve indicated a 2:1 ligand/receptor stoichiometry. So far, the binding of two small molecule inhibitors to hKv1.3 has been demonstrated only for disubstituted cyclohexyl compounds (Schmalhofer et al., 2002, 2003) and 5-phenylalkoxypsoralens (Vennekamp et al., 2004). To investigate the importance of the methoxy groups in more detail, we examined the effects of (±)-gallopamil, which has an additional methoxy group at R3, and (eC)-devapamil, that lacks the methoxy group at . Although the R3 methoxy group seems to have little influence on block of hKv1.3 (H399T) and hKv1.3 (H399T/A413C) currents, the missing methoxy group increases apparent inhibition 4-fold in the double mutant only. This effect is weaker but similar to the one obtained by removal of all four methoxy groups. However, the Hill coefficient for (eC)-devapamil block of hKv1.3 (H399T/A413C) currents remains at 1. Johnson et al. (1996) demonstrated that in wild-type Ca2+ channels, (eC)-devapamil inhibits currents much more strongly than verapamil, whereas in the Y1463F mutant, this preference is lost. This mutation seems to have the opposite effect on PAA binding compared with A413C in hKv1.3 (H399T). It is striking that the removal of a single hydroxyl group in Y1463F was responsible for the affinity changes, whereas a single thiol group was added by the A413C mutation.
The missing methoxy group of (eC)-devapamil accounts solely for the altered affinity to hKv1.3 (H399T/A413C) compared with (eC)-verapamil. This might indicate coupled effects of the removal of the OCH3 group and the substitution at position 413. The double mutant cycle analysis results support this line of argument, suggesting a distance of less than 5 between both interaction partners.
The assignment of this spatial constriction to the hKv1.3/PAA docking results yielded two clusters of PAA molecules with the , -dimethoxyphenyl ring located at the entrance (cluster I) or the core (cluster II) of an S6eCS6 interface gap. These results are similar to those obtained by Lipkind and Fozzard (2003) for a detailed PAA docking model for an L-type Ca2+ channel.
Figure 8 represents a possible explanation for the observed effects. If the permeation of the ethyl-methylaminophenyl ring into the S6 interface crevice is an essential mechanism of PAA binding (Fig. 8A), it might be sterically constricted by introduction of the A413C mutation (Fig. 8B). However, the degree of constriction seems to be less pronounced with a decreasing number of methoxy groups present at the ethyl-methyl-aminophenyl ring. In addition, the introduction of A413C enables emopamil, which lacks all four methoxy groups at both phenyl rings, to bind with a 2:1 stoichiometry (Fig. 8B, right). The altered PAA binding pocket might therefore implicate a constricted access and a spatial dilation that allows a second PAA to bind to the channel, if the first is able to enter deeply into the S6 interface gap.
Summary. Screening of the inactivated state reduced hKv1.3 (H399T) channel for residues participating in verapamil block revealed an important influence of the residue at position 413 in the S6 domain. Further inhibition experiments with verapamil derivatives suggest interdependence between the number of methoxy substituents at the PAA phenyl rings and alterations of Ala413. These data, taken together with information obtained by a computer-based docking model, might indicate that residue 413 is important for access of the PAA ethyl-methyl-aminophenyl ring to a binding pocket in an S6 interface gap.
This work was presented in part as a poster [Dreker T, Lehmann-Horn F, and Grissmer S. (2005) Utilization of Kv1.3 double mutants to explore the binding site of verapamil. Biophys J 88:278a].
doi:10.1124/mol.105.012401.
References
Chandy KG, DeCoursey TE, Cahalan MD, McLaughlin C, and Gupta S (1984) Voltage-gated potassium channels are required for human T lymphocyte activation. J Exp Med 160: 369eC385.
Choi KL, Aldrich RW, and Yellen G (1991) Tetraethylammonium blockade distinguishes two inactivation mechanisms in voltage-activated K+ channels. Proc Natl Acad Sci USA 88: 5092eC5095.
DeCoursey TE (1995) Mechanism of K+ channel block by verapamil and related compounds in rat alveolar epithelial cells. J Gen Physiol 106: 745eC779.
DeCoursey TE, Chandy KG, Gupta S, and Cahalan MD (1984) Voltage-gated K+ channels in human T lymphocytes: a role in mitogenesis Nature (Lond) 307: 465eC468.
Goodsell DS and Olson AJ (1990) Automated docking of substrates to proteins by simulated annealing. Proteins 8: 195eC202.
Grissmer S and Cahalan M (1989) TEA prevents inactivation while blocking open K+ channels in human T lymphocytes. Biophys J 55: 203eC206.
Guex N and Peitsch MC (1997) SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis 18: 2714eC2723.
Hamill OP, Marty A, Neher E, Sakmann B, and Sigworth FJ (1981) Improved patch-clamp techniques for high-resolution current recording from cells and cellfree membrane patches. Pflueg Arch Eur J Physiol 391: 85eC100.
Hanner M, Green B, Gao YD, Schmalhofer WA, Matyskiela M, Durand DJ, Felix JP, Linde AR, Bordallo C, Kaczorowski GJ, et al. (2001) Binding of correolide to the K(v)1.3 potassium channel: characterization of the binding domain by site-directed mutagenesis. Biochemistry 40: 11687eC11697.
Hidalgo P and MacKinnon R (1995) Revealing the architecture of a K+ channel pore through mutant cycles with a peptide inhibitor. Science (Wash DC) 268: 307eC310.
Hockerman GH, Johnson BD, Abbott MR, Scheuer T, and Catterall WA (1997) Molecular determinants of high affinity phenylalkylamine block of L-type calcium channels in transmembrane segment IIIS6 and the pore region of the alpha1 subunit. J Biol Chem 272: 18759eC18765.
Hockerman GH, Johnson BD, Scheuer T, and Catterall WA (1995) Molecular determinants of high-affinity phenylalkylamine block of L-type calcium channels. J Biol Chem 270: 22119eC22122.
Jiang Y, Lee A, Chen J, Cadene M, Chait BT, and MacKinnon R (2002a) Crystal structure and mechanism of a calcium-gated potassium channel. Nature (Lond) 417: 515eC522.
Jiang Y, Lee A, Chen J, Cadene M, Chait BT, and MacKinnon R (2002b) The open pore conformation of potassium channels. Nature (Lond) 417: 523eC526.
Johnson BD, Hockerman GH, Scheuer T, and Catterall WA (1996) Distinct effects of mutations in transmembrane segment IVS6 on block of L-type calcium channels by structurally similar phenylalkylamines. Mol Pharmacol 50: 1388eC1400.
Kavanaugh MP, Varnum MD, Osborne PB, Christie MJ, Busch AE, Adelman JP, and North RA (1991) Interaction between tetraethylammonium and amino acid residues in the pore of cloned voltage-dependent potassium channels. J Biol Chem 266: 7583eC7587.
Kiss L, LoTurco J, and Korn SJ (1999) Contribution of the selectivity filter to inactivation in potassium channels. Biophys J 76: 253eC263.
Lipkind GM and Fozzard HA (2003) Molecular modeling of interactions of dihydropyridines and phenylalkylamines with the inner pore of the L-type Ca2+ channel. Mol Pharmacol 63: 499eC511.
Liu Y, Jurman ME, and Yellen G (1996) Dynamic rearrangement of the outer mouth of a K+ channel during gating. Neuron 16: 859eC867.
Morris GM, Goodsell DS, Halliday RS, Huey R, Hart WE, Belew RK, and Olson AJ (1998) Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function. J Comp Chem 19: 1639eC1662.
Morris GM, Goodsell DS, Huey R, and Olson AJ (1996) Distributed automated docking of flexible ligands to proteins: parallel applications of AutoDock 2.4. J Comput-Aided Mol Des 10: 293eC304.
Nguyen A, Kath JC, Hanson DC, Biggers MS, Canniff PC, Donovan CB, Mather RJ, Bruns MJ, Rauer H, Aiyar J, et al. (1996) Novel nonpeptide agents potently block the C-type inactivated conformation of Kv1.3 and suppress T cell activation. Mol Pharmacol 50: 1672eC1679.
Panyi G, Sheng Z, and Deutsch C (1995) C-type inactivation of a voltage-gated K+ channel occurs by a cooperative mechanism. Biophys J 69: 896eC903.
Peitsch MC (1995) Protein modelling by E-mail. Bio/Technology 13: 658eC660.
Rauer H and Grissmer S (1996) Evidence for an internal phenylalkylamine action on the voltage-gated potassium channel Kv1.3. Mol Pharmacol 50: 1625eC1634.
Rauer H and Grissmer S (1999) The effect of deep pore mutations on the action of phenylalkylamines on the Kv1.3 potassium channel. Br J Pharmacol 127: 1065eC1074.
Rbe RJ and Grissmer S (2000) Block of the lymphocyte K+ channel mKv1.3 by the phenylalkylamine verapamil: kinetic aspects of block and disruption of accumulation of block by a single point mutation. Br J Pharmacol 131: 1275eC1284.
Schmalhofer WA, Bao J, McManus OB, Green B, Matyskiela M, Wunderler D, Bugianesi RM, Felix JP, Hanner M, Linde-Arias AR, et al. (2002) Identification of a new class of inhibitors of the voltage-gated potassium channel, Kv1.3, with immunosuppressant properties. Biochemistry 41: 7781eC7794.
Schmalhofer WA, Slaughter RS, Matyskiela M, Felix JP, Tang YS, Rupprecht K, Kaczorowski GJ, and Garcia ML (2003) Di-substituted cyclohexyl derivatives bind to two identical sites with positive cooperativity on the voltage-gated potassium channel, K(v)1.3. Biochemistry 42: 4733eC4743.
Schreiber G and Fersht AR (1995) Energetics of protein-protein interactions: analysis of the barnase-barstar interface by single mutations and double mutant cycles. J Mol Biol 248: 478eC486.
Schuster A, Lacinova L, Klugbauer N, Ito H, Birnbaumer L, and Hofmann F (1996) The IVS6 segment of the L-type calcium channel is critical for the action of dihydropyridines and phenylalkylamines. EMBO (Eur Mol Biol Organ) J 15: 2365eC2370.
Schwede T, Kopp J, Guex N, and Peitsch MC (2003) SWISS-MODEL: An automated protein homology-modeling server. Nucleic Acids Res 31: 3381eC3385.
Vennekamp J, Wulff H, Beeton C, Calabresi PA, Grissmer S, Hansel W, and Chandy KG (2004) Kv1.3-blocking 5-phenylalkoxypsoralens: a new class of immunomodulators. Mol Pharmacol 65: 1364eC1374.
Wrisch A and Grissmer S (2000) Structural differences of bacterial and mammalian K+ channels. J Biol Chem 275: 39345eC39353.
Wulff H, Beeton C, and Chandy KG (2003) Potassium channels as therapeutic targets for autoimmune disorders. Curr Opin Drug Discov Dev 6: 640eC647.
Wulff H, Knaus HG, Pennington M, and Chandy KG (2004) K+ channel expression during B cell differentiation: implications for immunomodulation and autoimmunity. J Immunol 173: 776eC786.
Yellen G, Sodickson D, Chen TY, and Jurman ME (1994) An engineered cysteine in the external mouth of a K+ channel allows inactivation to be modulated by metal binding. Biophys J 66: 1068eC1075., http://www.100md.com(Tobias Dreker, and Stepha)
Abstract
To screen for residues of hKv1.3 important for current block by the phenylalkylamine verapamil, the inactivated-state-reduced H399T mutant was used as a background for mutagenesis studies. This approach was applied mainly to abolish the accumulation in the inactivated blocked state, recovery from which in the wild type is normally slow. Substitution of amino acids in the S6 transmembrane helix indicated a heavy disruption of verapamil block by the A413C mutation, reducing the IC50 from 2.4 to 267 e. Subsequent scanning for verapamil moieties essential for current block was performed by application of derivatives with altered side groups. Neither the removal of the nitrile or the methyl group nor the addition of a methoxy group resulted in major variations of IC50 values for hKv1.3 (H399T) current block. However, disruption of current block by A413C was 4- to 10-fold less pronounced for derivatives lacking the 4-methoxy group of the (3,4-dimethoxyphenyl)ethylmethyl-amino part (devapamil) or all four methoxy groups (emopamil), respectively. Emopamil displayed a Hill coefficient of 2 for hKv1.3 (H399T/A413C) instead of 1 for hKv1.3 (H399T) current block. These results might indicate that the alteration of Ala413 modulates the access of phenylalkylamines to their binding site depending on the occupancy of the phenyl rings with methoxy groups. A computer-based docking model shows a subset of docked PAA conformations, with a spatial proximity between the (4-methoxyphenyl)ethyl-methyl-amino group and Ala413. The PAA binding site might therefore include a binding pocket for the aromatic ring of the ethyl-methyl-amino part in an S6eCS6 interface gap.
Various small molecules have been shown to block ion fluxes through channel proteins by physically occluding the pore. In the case of the voltage-gated potassium channel Kv1.3, such structurally distinct compounds as verapamil (phenylalkylamines), WIN-17317-3 (iminodihydrochinolines), UK-78282 (benzyl piperidines), correolide (triterpenes), 4-phenyl-4-[3-(2-methoxyphenyl)-3-oxo-2-azaprop-1-yl] cyclohexanone (cyclohexyl-substituted benzamides), or Psora-4 (5-phenylalkoxypsoralens) display inhibitory effects (reviewed in Wulff et al., 2003; Vennekamp et al., 2004). An obvious possibility is to use these small molecules as probes to investigate the residues and spatial requisites of the receptor necessary for binding. Enhanced knowledge of these parameters could provide useful information for the design of more specific and potent drugs with potential therapeutic use. The phenylalkylamine (PAA) verapamil was used in this study to explore the binding site of hKv1.3, a channel that plays an important role in T-cell activation and B-cell differentiation (Chandy et al., 1984; DeCoursey et al., 1984; Wulff et al., 2004). Many data have been reported already, determining the basic mode of verapamil block of Kv1.3 currents, thereby providing a good starting point for a detailed investigation of the binding pocket. Verapamil acts as a fast open-channel blocker with an IC50 of approximately 8 e (Rauer and Grissmer, 1996), and block takes place from the intracellular side of the cell. Mutagenesis and competition studies suggest that the binding site is situated in the water-filled cavity below the selectivity filter (Rauer and Grissmer, 1999). The latter results might indicate that side groups of verapamil are in close contact with pore-facing residues of the transmembrane S6 helix. Hanner et al., 2001 reported that various residues in this area are important for the binding of correolides. For L-type Ca2+ channels, it has been shown that the PAA binding site involves residues in the S6 segments of domains III and VI (Hockerman et al., 1995, 1997; Schuster et al., 1996) and a homology model based on the KcsA channel structure has been established presenting the PAA devapamil with the aromatic ring of the phenylethylamine part located inside the S6 interface III/IV crevice (Lipkind and Fozzard, 2003).
A specific property of Kv1.3 that distinguishes it from other Kv subfamily members is the intrinsic C-type inactivated state displayed by the channel. This state is reached slowly by all channel molecules upon depolarization, and recovery from inactivation is slow. Various drugs are known to bind selectively to the open and/or inactivated state; e.g., verapamil application leads to open blocked and inactivated blocked conformations of the channel (Rbe and Grissmer, 2000). A substantial problem of investigating drug effects on Kv1.3 currents is the difficulty to electrically distinguish between the access of the drug to either state. Upon verapamil application, one can observe an acceleration of current decay during the depolarizing steps and an accumulation of peak current block in consecutive voltage steps. The latter effect is caused by an accumulation of the channel in the inactivated blocked state, recovery from which is very slow (Rbe and Grissmer, 2000). To address this problem, we introduced the H399T mutation in the extracellular loop between the pore and the S6 helix of hKv1.3, because it has been described to strongly reduce the C-type inactivated state (Nguyen et al., 1996; Rauer and Grissmer, 1996). This reduction of complexity makes the direct calculation of IC50 values feasible by measurement of steady-state current block. Additional amino acid substitutions in the pore-facing parts of the S6 transmembrane helix (Fig. 1) revealed that the residue at position 413 plays an important role for PAA binding. Thereafter, to detect important side groups of the ligand, block by several verapamil derivatives has been examined and the obtained results pointed out reciprocal interactions between the mutated channel residue and exchanged verapamil side groups. Double-mutant cycle analysis calculations suggested a spatial proximity between both interaction partners, information that was used to evaluate the computer based docking of a PAA molecule into an hKv1.3 homology model.
Materials and Methods
Cells. The COS-7 cell line was obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (Braunschweig, Germany). The cells were maintained in Dulbecco's modified Eagle's medium with high glucose (Invitrogen, Carlsbad, CA) containing 10% fetal calf serum (PAA Laboratories GmbH, Coelbe, Germany) and stored in an incubator at 37°C and 10% p(CO2).
Chemicals and Solutions. All measurements were performed in an external bath solution, high (containing 160 mM NaCl, 2 mM CaCl2, 1 mM MgCl2, 4.5 mM KCl, and 5 mM HEPES) or high (containing 164.5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, and 5 mM HEPES). Osmolarity was 290 to 320 mOsM, and pH was adjusted to 7.4 with NaOH or KOH, respectively. The internal pipette solution contained 155 mM KF, 2 mM MgCl2, 10 mM EGTA, and 10 mM HEPES. Osmolarity was 290 to 320 mOsM, and pH was adjusted to 7.2 with KOH. All verapamil compounds were dissolved in DMSO as stock-solutions and diluted to the final concentration in external bath solution before application. DMSO fraction in the final solution was always <2%. (±)-verapamil, (eC)-verapamil, (+)-verapamil, and (±)-gallopamil were obtained as hydrochloride salts from Sigma-Aldrich (St. Louis, MO). (eC)-devapamil (LU 47093), (±)-norverapamil (LU 41723) and (±)-acyanoverapamil (LU 44935) were provided as hydrochloride salts by Drs. Raschack and Paul of Knoll Pharmaceuticals AG (now merged in Abbott GmbH and Co. KG (Ludwigshafen, Germany). (±)-emopamil (LU 37076) was a generous gift from Dr. Hans-Gether Knaus and Dr. Hartmut Glossmann (Institut fe Biochemische Pharmakologie, Medizinische Universitt Innsbruck, Innsbruck, Austria). KTX (Latoxan, Valence, France; Bachem, Bubendorf, Switzerland) was dissolved in external bath solution containing 0.1% bovine albumin (Sigma-Aldrich). TEA chloride was purchased from Fluka (Buchs, Switzerland) and dissolved in external bath solution to the respective final concentration.
Electrophysiology. The whole-cell recording mode of the patchclamp technique (Hamill et al., 1981) was used throughout all measurements. Experiments were carried out at room temperature (18eC22°C), and external bath solution and diluted drug solutions were applied using a gravity-based permanent perfusion system. Electrodes were pulled from glass capillaries (Science Products, Hofheim, Germany) in three stages and fire-polished to resistances of 2 to 4 M. Data were acquired with an EPC-9 patch-clamp amplifier (HEKA Elektronik, Lambrecht/Pfalz, Germany) connected to a Macintosh computer running Pulse/Pulse Fit ver. 8.40 data acquisition and analysis software (HEKA). All currents were filtered by a 2.9-kHz Bessel filter and recorded with a sampling frequency of 2.00 kHz. Capacitative and leak currents were subtracted, and series resistance compensation (80%) was used for currents exceeding 2 nA. Further data analysis was performed using the Igor Pro 3.1 (Wavemetrics, Lake Oswego, OR) software package. Holding potential was usually eC80 mV (hKv1.3 mutants) or eC120 mV (hKv1.3 wild type). Time constants m and h and steady-state currents with and without drug application were determined by a Hodgkin and Huxley fit to currents elicited by a 200-ms voltage pulse from the holding potential to +40 mV. The deactivation time constant t was obtained by an exponential fit to tail currents at eC60 mV observed after a voltage pulse from the holding potential to +40 mV for 200 ms. Parameters V1/2 and k were deduced from a Boltzmann fit to either normalized maximum peak conductance elicited by 11 consecutive voltage steps from eC50 to +50 mV (V = 10 mV) in high or to maximum tail peak currents at eC80 mV after the voltage steps in high . Drug effects were evaluated by the steady-state current block obtained during 200-ms voltage steps from the holding potential to +40 mV. Time interval between pulses was 30 s, allowing the channel to recover from inactivation completely. IC50 values were calculated by fitting a Hill function (Idrug/I0 = 1/[1+(IC50/Cdrug)nH]) to the normalized steady-state current data points obtained after drug application. I0 represents the steady state current before and Idrug the steady state current after drug application, Cdrug the concentration of the drug in the external bath solution and nH the Hill coefficient.
Molecular Biology. The hKv1.3 wild-type plasmid was a generous gift from Prof. Dr. O. Pongs (Institut fe Neurale Signalverarbeitung, Zentrum fe Molekulare Neurobiologie, Hamburg, Germany). It contains the human Kv1.3 K+ channel gene in a pRc/CMV vector (Invitrogen) with a CMV promoter for protein expression in mammalian cells. Channel mutants were generated by introducing the corresponding point mutations in the cloned hKv1.3 gene with the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). Transfection of COS-7 cells was done using the Fugene 6 transfection reagent (Roche Molecular Biochemicals, Mannheim, Germany). Cells were grown to 80% confluence and cotransfected with 1 e of hKv1.3 DNA and 0.5 e of eGFP-N1 DNA (BD Biosciences Clontech, Palo Alto, CA). One day after transfection sufficient protein for electrophysiological measurements was expressed.
Double Mutant Cycle Analysis. The coupling coefficient was calculated by the equation = [IC50(Rwt, Lwt) x IC50(Rmt, Lmt)]/[IC50(Rwt, Lmt) x IC50(Rmt, Lwt)], according to Hidalgo and MacKinnon (1995). Rwt was the receptor "wild type" hKv1.3 (H399T) and Rmt was the H399T/A413C double mutant. Lwt represents the "wild-type" drug [(±)-verapamil or (eC)-verapamil] and Lmt represents the "mutant" drug (a verapamil derivative with an exchanged side group). If coupling coefficients of less than 0 were obtained, the reciprocal value was taken. The change of free binding energy (G) was calculated by the equation G = RTln, where R is the gas constant and T is absolute temperature.
Results
hKv1.3 (H399T) As a Tool for Determination of Steady State PAA Block. To minimize the time the channel spends in the C-type inactivated state of hKv1.3, the H399T mutation, which is analogous to the H404T mutation in mKv1.3, was introduced (Nguyen et al., 1996; Rauer and Grissmer, 1996). Figure 2 shows the K+ currents through hKv1.3 wild type (A) and H399T mutant (B) channels, without drug application (pulse 1). The inactivation time constant (h) of hKv1.3 (H399T) is approximately 2.3-fold slower compared with wild type (Fig. 2; Table 1), and investigation of the kinetics for recovery from inactivation (Fig. 3, A and B) revealed that the corresponding time constant rec of hKv1.3 H399T is approximately 10 times faster (Table 1) than in case of the wild type. The intrinsic activation parameters (V1/2, k, and m) of hKv1.3 (H399T) are similar to the corresponding wild-type values, whereas the deactivation time constant t at eC60 mV is increased by a factor of approximately 3.7 (Table 1). Application of 20 e (±)-verapamil resulted in a steady-state block of hKv1.3 (H399T) currents at the end of a 200-ms pulse and no further accumulation of peak current block in consecutive pulses (Fig. 2B, pulse 2eC4). The normalized steady-state current after PAA application should therefore directly reflect the ratio of channels in the open blocked and open states. The remaining inactivation of the mutant hKv1.3 channels after 200 ms was ignored for the evaluation of steady-state block.
All values are given as mean ± S.D.; numbers of independent experiments are given in parentheses.
Identification of hKv1.3 (H399T) Residues Important for PAA Block. Several amino acids in the S6 transmembrane helix of hKv1.3 (H399T) have been substituted, because effects of these residues on current block by verapamil or similar drugs to Kv1.3 or other ion channels have been described or assumed. Substitutions were performed mainly with cysteines to allow a chemical modification in future experiments. Exceptions were the H399T/S410G and H399T/G416A double mutants. In the first case, the corresponding glycine of Kv2.1 was introduced; in the second case, an alanine was introduced because of the critical function of Gly416 as the potential gating hinge (Jiang et al., 2002b). The obtained double mutants, H399T/S410G, H399T/C412A, H399T/L418C, and H399T/I420C, had only minor or moderate effects on (±)-verapamil block (data not shown). The H399T/G416A mutation resulted in no functional expression of channel protein, possibly because of the disruption of the gating hinge. Another double mutant, H399T/A413C, displayed an affinity for (±)-verapamil that was reduced by a factor of approximately 100 (Table 2) compared with hKv1.3 (H399T). Steady-state current block with no further accumulation of block could still be observed in this channel mutant (Fig. 2C, pulses 2eC4). Furthermore, hKv1.3 (H399T/A413C) exhibits an even slower intrinsic inactivation rate (2-fold), a similar recovery from inactivation (Fig. 3C; Table 1), and an accelerated deactivation rate (10-fold) compared with hKv1.3 (H399T) (Fig. 2C, pulse 1; Table 1). The activation parameters (V1/2, k, m) also remain similar.
All values were obtained from at least four independent experiments and are given as mean ± S.D.
Current Block by Externally Applied KTX and TEA Is Affected by Introduction of A413C in hKv1.3 (H399T). Effects on the affinity for externally applied KTX and TEA have been examined to investigate whether the introduced cysteine decreases the affinity for (±)-verapamil merely by a direct interaction with the drug or by allosteric changes of the three dimensional protein structure. Both inhibitors are known to bind to the extracellular vestibule of Kv1.3 (Kavanaugh et al., 1991; Wrisch and Grissmer, 2000) and could serve as probes for structural differences in the extracellular region of the double mutant. In the case of hKv1.3 (H399T), IC50 values of 25.9 ± 2.5 mM for TEA and 0.25 ± 0.02 nM for KTX (Fig. 4, A and B, filled bars) were obtained. IC50 values for KTX and TEA block differed for hKv1.3 (H399T/A413C), with 5.8 ± 0.4 mM for TEA and 0.66 ± 0.06 nM for KTX (Fig. 4, A and B, open bars).
Importance of A413 for the Stereoselectivity of Verapamil Block. Verapamil is a chiral compound and current block by both enantiomers was investigated. These experiments were necessary for comparison with current block by (eC)-devapamil and to select the appropriate enantiomer for subsequent docking computations. (eC)-Verapamil blocked hKv1.3 (H399T) currents with an approximately 3.6-fold higher potency compared with the (+)-enantiomer (Table 2). It is surprising that the introduced A413C mutation reduced the affinity for (eC)-verapamil by a factor of 200, whereas the affinity for (+)-verapamil was reduced by a factor of 80 only (Table 2). Thus, the ability of the channel to discriminate between both enantiomers seems to be almost lost in the double mutant, leaving a preference of 1.4-fold for (eC)-verapamil over (+)-verapamil.
Inhibition Properties among Distinct Verapamil Derivatives Are Different for hKv1.3 (H399T) and hKv1.3 (H399T/A413C). Verapamil contains various functional groups, each of which might or might not play an important role in blocking Kv1.3 currents. For this reason, block of currents through hKv1.3 (H399T) and the double mutant H399T/A413C by several verapamil derivatives (Fig. 5) was investigated. Compared with (±)-verapamil, the compounds (±)-emopamil, (±)-gallopamil, (±)-norverapamil, and (±)-acyanoverapamil displayed only subtle variations in the IC50 values for hKv1.3 (H399T) current block. The same result was obtained for the (eC)-verapamil compared with the (eC)-devapamil IC50 value. However, in case of the hKv1.3 (H399T/A413C) double mutant, the results were partially different. Whereas (±)-gallopamil, (±)-norverapamil, and (±)-acyanoverapamil IC50 values are still similar to that of (±)-verapamil, (±)-emopamil block is approximately 10 times less affected by the A413C mutation. Another remarkable finding was that (±)-emopamil block of hKv1.3 (H399T/A413C) currents was the only case in which the dose-response fits resulted in a Hill coefficient of 2 instead of 1 (Fig. 6). Comparison of (eC)-devapamil and (eC)-verapamil IC50 values for the double mutant also suggests a smaller influence of A413C on (eC)-devapamil current block, which is 4-fold more potent compared with (eC)-verapamil.
Double Mutant Cycle Analysis of the Obtained IC50 Values. To detect coupled interactions caused by the substitution of the alanine at position 413 and the alteration of the different side groups of verapamil, the coupling coefficients () and the change of free binding energies (G) were calculated for all reactions (Table 3). In this approach, interdependency between two interaction partners is indicated by values deviating from 1. Whereas the coupling coefficients calculated for gallopamil, norverapamil, and acyanoverapamil stayed near 1, calculated for the (eC)-verapamil/(eC)-devapamil pair was 2.5, and the corresponding change of free binding energy was 0.52 kcal/mol. These data suggest a mutual influence of the deletion of the methoxy group of verapamil and the substitution of A413 of hKv1.3 (H399T). Furthermore, an interpretation of the correlations made by Schreiber and Fersht, 1995 suggests that a change in free binding energy of >0.5 kcal/mol might indicate a spatial distance between both interaction partners of less than 5 .
Building an hKv1.3 Homology Model and Docking of a PAA. To accommodate the inability of verapamil to bind to the closed conformation of Kv1.3 (DeCoursey, 1995), a suitable model for a K+ channel in the open state had to be selected. The X-ray structure of MthK, a Ca2+ gated K+ channel in the open conformation (Jiang et al., 2002a), was chosen as a scaffold for the homology model of hKv1.3. The hKv1.3 primary sequence from positions 352 to 431 was aligned to MthK, covering the S5 and S6 transmembrane helices and the pore region (Deep View software package; Guex and Peitsch, 1997). Thereafter, the hKv1.3 homology model was calculated using the SWISS-MODEL server (Peitsch, 1995; Guex and Peitsch, 1997; Schwede et al., 2003). Finally, docking was performed with the AutoDock 3.0 software package (Goodsell and Olson, 1990; Morris et al., 1996, 1998) using a modified (eC)-verapamil molecule as a ligand. This molecule lacks the R4 nitrile and R5 methyl groups, because these side groups did not show an important effect on current block in either the H399T or the A413C/H399T mutant. Of 256 docked conformations obtained, only those 31 were selected with the C atom of the methoxy group situated less than 5 away from the C atom of one of the four A413 side chains. No strength of interaction was implicated. Despite the symmetrical nature of the channel model, 24 methoxy groups were in proximity to Ala413 from a single subunit (Fig. 7, cyan). Within this group, two populations of similar conformations could be identified: cluster I (Fig. 7, magenta), which consisted of 12 PAAs situated in a stretched conformation in the inner vestibule of the channel with their ,-dimethoxyphenyl rings inserted slightly between two S6 helices, and cluster II (Fig. 7, gray), which contained four molecules in a similar stretched configuration but with their ,-dimethoxyphenyl rings extended deeper into the S6 interface crevice.
Discussion
The purpose of this study was to characterize the binding site of verapamil in the hKv1.3 mutant H399T, a channel mutant with a reduced C-type inactivated state. Mutations of residues in the S6 domain and application of different verapamil derivatives revealed interdependencies between position 413 of the channel and functional groups of verapamil. This information was used to deduce a more detailed description of the binding pocket.
As anticipated, the introduced H399T mutation reduced the channel's C-type inactivation but not to the extent described for an analogous mKv1.3 (H404T) mutant (Rauer and Grissmer, 1996). Nonetheless, hKv1.3 (H399T) was selected as a Kv1.3 model channel with a strongly reduced C-type inactivated state because the time constant for recovery from inactivation was approximately 10-fold faster compared with the wild-type channel. Steady-state currents were clearly observable with or without drug application after 1 s (data not shown) voltage pulses to +40 mV and no accumulation of steady-state current block in consecutive pulses was present.
Subsequent substitutions of amino acids in the S6 domain of hKv1.3 (H399T) pointed out that the residue at position 413 plays an important role in PAA binding. By exchanging the alanine at position 413 for a cysteine, (±)-verapamil affinity was decreased by approximately 100-fold. Similar results were obtained by Hanner et al. (2001) for an hKv1.3 (A413C) mutant that displayed a 40-fold lower affinity for diTC binding compared with the wild type. Furthermore, the corresponding residue in IVS6 of L-type Ca2+ channels had a strong influence on PAA affinity (Hockerman et al., 1995).
The additional A413C mutation slowed the C-type inactivation rate even further, an effect that had also been observed for the hKv1.3 (A413C) single mutant (data not shown). More evidence for A413 playing a role in the inactivation process was found by Panyi et al. (1995), where a strong increase of the inactivation rate in an hKv1.3 (A413V) mutant was described.
It has been reported that binding of external TEA slowed C-type inactivation (Grissmer and Cahalan, 1989; Choi et al., 1991) and that the transition to the inactivated state induced structural changes in the outer pore region (Yellen et al., 1994; Liu et al., 1996; Kiss et al., 1999). Further disruption of the inactivation process by introduction of A413C might therefore explain the affinity shift observed for TEA. However, no correlations between KTX binding and C-type inactivation have been described yet. The 3-fold decrease of KTX and 5-fold increase of TEA affinity in hKv1.3 (H399T/A413C) might be due to a structural rearrangement of the channel protein in the double mutant that affects the extracellular vestibule. This rearrangement might change the intracellular vestibule as well, which could explain the dramatic decrease in verapamil affinity.
Another explanation for the disruption of PAA block might be the loss of direct effects of the residue at position 413 on binding of distinct verapamil side groups. Verapamil derivatives lacking the methyl group at R5 (norverapamil) or the nitrile group at R4 (acyanoverapamil) did not show major variations in hKv1.3 (H399T) and hKv1.3 (H399T/A413C) current block compared with (±)-verapamil. Thus, neither the R5 methyl group nor the R4 nitrile group seems to play an important role for PAA binding in both channel mutants. The importance of the four methoxy groups at the R1, and R2, positions of both phenyl rings was assessed by investigation of (±)-emopamil block, a verapamil derivative that lacks all four moieties. In case of hKv1.3 (H399T) only little difference could be observed between the affinities for (±)-verapamil and (±)-emopamil, suggesting no major relevance of the methoxy groups. However, as an inhibitor of hKv1.3 (H399T/A413C) currents, (±)-emopamil is an approximately 10-fold more potent than (±)-verapamil. In addition, fits of the corresponding (±)-emopamil dose-response curve indicated a 2:1 ligand/receptor stoichiometry. So far, the binding of two small molecule inhibitors to hKv1.3 has been demonstrated only for disubstituted cyclohexyl compounds (Schmalhofer et al., 2002, 2003) and 5-phenylalkoxypsoralens (Vennekamp et al., 2004). To investigate the importance of the methoxy groups in more detail, we examined the effects of (±)-gallopamil, which has an additional methoxy group at R3, and (eC)-devapamil, that lacks the methoxy group at . Although the R3 methoxy group seems to have little influence on block of hKv1.3 (H399T) and hKv1.3 (H399T/A413C) currents, the missing methoxy group increases apparent inhibition 4-fold in the double mutant only. This effect is weaker but similar to the one obtained by removal of all four methoxy groups. However, the Hill coefficient for (eC)-devapamil block of hKv1.3 (H399T/A413C) currents remains at 1. Johnson et al. (1996) demonstrated that in wild-type Ca2+ channels, (eC)-devapamil inhibits currents much more strongly than verapamil, whereas in the Y1463F mutant, this preference is lost. This mutation seems to have the opposite effect on PAA binding compared with A413C in hKv1.3 (H399T). It is striking that the removal of a single hydroxyl group in Y1463F was responsible for the affinity changes, whereas a single thiol group was added by the A413C mutation.
The missing methoxy group of (eC)-devapamil accounts solely for the altered affinity to hKv1.3 (H399T/A413C) compared with (eC)-verapamil. This might indicate coupled effects of the removal of the OCH3 group and the substitution at position 413. The double mutant cycle analysis results support this line of argument, suggesting a distance of less than 5 between both interaction partners.
The assignment of this spatial constriction to the hKv1.3/PAA docking results yielded two clusters of PAA molecules with the , -dimethoxyphenyl ring located at the entrance (cluster I) or the core (cluster II) of an S6eCS6 interface gap. These results are similar to those obtained by Lipkind and Fozzard (2003) for a detailed PAA docking model for an L-type Ca2+ channel.
Figure 8 represents a possible explanation for the observed effects. If the permeation of the ethyl-methylaminophenyl ring into the S6 interface crevice is an essential mechanism of PAA binding (Fig. 8A), it might be sterically constricted by introduction of the A413C mutation (Fig. 8B). However, the degree of constriction seems to be less pronounced with a decreasing number of methoxy groups present at the ethyl-methyl-aminophenyl ring. In addition, the introduction of A413C enables emopamil, which lacks all four methoxy groups at both phenyl rings, to bind with a 2:1 stoichiometry (Fig. 8B, right). The altered PAA binding pocket might therefore implicate a constricted access and a spatial dilation that allows a second PAA to bind to the channel, if the first is able to enter deeply into the S6 interface gap.
Summary. Screening of the inactivated state reduced hKv1.3 (H399T) channel for residues participating in verapamil block revealed an important influence of the residue at position 413 in the S6 domain. Further inhibition experiments with verapamil derivatives suggest interdependence between the number of methoxy substituents at the PAA phenyl rings and alterations of Ala413. These data, taken together with information obtained by a computer-based docking model, might indicate that residue 413 is important for access of the PAA ethyl-methyl-aminophenyl ring to a binding pocket in an S6 interface gap.
This work was presented in part as a poster [Dreker T, Lehmann-Horn F, and Grissmer S. (2005) Utilization of Kv1.3 double mutants to explore the binding site of verapamil. Biophys J 88:278a].
doi:10.1124/mol.105.012401.
References
Chandy KG, DeCoursey TE, Cahalan MD, McLaughlin C, and Gupta S (1984) Voltage-gated potassium channels are required for human T lymphocyte activation. J Exp Med 160: 369eC385.
Choi KL, Aldrich RW, and Yellen G (1991) Tetraethylammonium blockade distinguishes two inactivation mechanisms in voltage-activated K+ channels. Proc Natl Acad Sci USA 88: 5092eC5095.
DeCoursey TE (1995) Mechanism of K+ channel block by verapamil and related compounds in rat alveolar epithelial cells. J Gen Physiol 106: 745eC779.
DeCoursey TE, Chandy KG, Gupta S, and Cahalan MD (1984) Voltage-gated K+ channels in human T lymphocytes: a role in mitogenesis Nature (Lond) 307: 465eC468.
Goodsell DS and Olson AJ (1990) Automated docking of substrates to proteins by simulated annealing. Proteins 8: 195eC202.
Grissmer S and Cahalan M (1989) TEA prevents inactivation while blocking open K+ channels in human T lymphocytes. Biophys J 55: 203eC206.
Guex N and Peitsch MC (1997) SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis 18: 2714eC2723.
Hamill OP, Marty A, Neher E, Sakmann B, and Sigworth FJ (1981) Improved patch-clamp techniques for high-resolution current recording from cells and cellfree membrane patches. Pflueg Arch Eur J Physiol 391: 85eC100.
Hanner M, Green B, Gao YD, Schmalhofer WA, Matyskiela M, Durand DJ, Felix JP, Linde AR, Bordallo C, Kaczorowski GJ, et al. (2001) Binding of correolide to the K(v)1.3 potassium channel: characterization of the binding domain by site-directed mutagenesis. Biochemistry 40: 11687eC11697.
Hidalgo P and MacKinnon R (1995) Revealing the architecture of a K+ channel pore through mutant cycles with a peptide inhibitor. Science (Wash DC) 268: 307eC310.
Hockerman GH, Johnson BD, Abbott MR, Scheuer T, and Catterall WA (1997) Molecular determinants of high affinity phenylalkylamine block of L-type calcium channels in transmembrane segment IIIS6 and the pore region of the alpha1 subunit. J Biol Chem 272: 18759eC18765.
Hockerman GH, Johnson BD, Scheuer T, and Catterall WA (1995) Molecular determinants of high-affinity phenylalkylamine block of L-type calcium channels. J Biol Chem 270: 22119eC22122.
Jiang Y, Lee A, Chen J, Cadene M, Chait BT, and MacKinnon R (2002a) Crystal structure and mechanism of a calcium-gated potassium channel. Nature (Lond) 417: 515eC522.
Jiang Y, Lee A, Chen J, Cadene M, Chait BT, and MacKinnon R (2002b) The open pore conformation of potassium channels. Nature (Lond) 417: 523eC526.
Johnson BD, Hockerman GH, Scheuer T, and Catterall WA (1996) Distinct effects of mutations in transmembrane segment IVS6 on block of L-type calcium channels by structurally similar phenylalkylamines. Mol Pharmacol 50: 1388eC1400.
Kavanaugh MP, Varnum MD, Osborne PB, Christie MJ, Busch AE, Adelman JP, and North RA (1991) Interaction between tetraethylammonium and amino acid residues in the pore of cloned voltage-dependent potassium channels. J Biol Chem 266: 7583eC7587.
Kiss L, LoTurco J, and Korn SJ (1999) Contribution of the selectivity filter to inactivation in potassium channels. Biophys J 76: 253eC263.
Lipkind GM and Fozzard HA (2003) Molecular modeling of interactions of dihydropyridines and phenylalkylamines with the inner pore of the L-type Ca2+ channel. Mol Pharmacol 63: 499eC511.
Liu Y, Jurman ME, and Yellen G (1996) Dynamic rearrangement of the outer mouth of a K+ channel during gating. Neuron 16: 859eC867.
Morris GM, Goodsell DS, Halliday RS, Huey R, Hart WE, Belew RK, and Olson AJ (1998) Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function. J Comp Chem 19: 1639eC1662.
Morris GM, Goodsell DS, Huey R, and Olson AJ (1996) Distributed automated docking of flexible ligands to proteins: parallel applications of AutoDock 2.4. J Comput-Aided Mol Des 10: 293eC304.
Nguyen A, Kath JC, Hanson DC, Biggers MS, Canniff PC, Donovan CB, Mather RJ, Bruns MJ, Rauer H, Aiyar J, et al. (1996) Novel nonpeptide agents potently block the C-type inactivated conformation of Kv1.3 and suppress T cell activation. Mol Pharmacol 50: 1672eC1679.
Panyi G, Sheng Z, and Deutsch C (1995) C-type inactivation of a voltage-gated K+ channel occurs by a cooperative mechanism. Biophys J 69: 896eC903.
Peitsch MC (1995) Protein modelling by E-mail. Bio/Technology 13: 658eC660.
Rauer H and Grissmer S (1996) Evidence for an internal phenylalkylamine action on the voltage-gated potassium channel Kv1.3. Mol Pharmacol 50: 1625eC1634.
Rauer H and Grissmer S (1999) The effect of deep pore mutations on the action of phenylalkylamines on the Kv1.3 potassium channel. Br J Pharmacol 127: 1065eC1074.
Rbe RJ and Grissmer S (2000) Block of the lymphocyte K+ channel mKv1.3 by the phenylalkylamine verapamil: kinetic aspects of block and disruption of accumulation of block by a single point mutation. Br J Pharmacol 131: 1275eC1284.
Schmalhofer WA, Bao J, McManus OB, Green B, Matyskiela M, Wunderler D, Bugianesi RM, Felix JP, Hanner M, Linde-Arias AR, et al. (2002) Identification of a new class of inhibitors of the voltage-gated potassium channel, Kv1.3, with immunosuppressant properties. Biochemistry 41: 7781eC7794.
Schmalhofer WA, Slaughter RS, Matyskiela M, Felix JP, Tang YS, Rupprecht K, Kaczorowski GJ, and Garcia ML (2003) Di-substituted cyclohexyl derivatives bind to two identical sites with positive cooperativity on the voltage-gated potassium channel, K(v)1.3. Biochemistry 42: 4733eC4743.
Schreiber G and Fersht AR (1995) Energetics of protein-protein interactions: analysis of the barnase-barstar interface by single mutations and double mutant cycles. J Mol Biol 248: 478eC486.
Schuster A, Lacinova L, Klugbauer N, Ito H, Birnbaumer L, and Hofmann F (1996) The IVS6 segment of the L-type calcium channel is critical for the action of dihydropyridines and phenylalkylamines. EMBO (Eur Mol Biol Organ) J 15: 2365eC2370.
Schwede T, Kopp J, Guex N, and Peitsch MC (2003) SWISS-MODEL: An automated protein homology-modeling server. Nucleic Acids Res 31: 3381eC3385.
Vennekamp J, Wulff H, Beeton C, Calabresi PA, Grissmer S, Hansel W, and Chandy KG (2004) Kv1.3-blocking 5-phenylalkoxypsoralens: a new class of immunomodulators. Mol Pharmacol 65: 1364eC1374.
Wrisch A and Grissmer S (2000) Structural differences of bacterial and mammalian K+ channels. J Biol Chem 275: 39345eC39353.
Wulff H, Beeton C, and Chandy KG (2003) Potassium channels as therapeutic targets for autoimmune disorders. Curr Opin Drug Discov Dev 6: 640eC647.
Wulff H, Knaus HG, Pennington M, and Chandy KG (2004) K+ channel expression during B cell differentiation: implications for immunomodulation and autoimmunity. J Immunol 173: 776eC786.
Yellen G, Sodickson D, Chen TY, and Jurman ME (1994) An engineered cysteine in the external mouth of a K+ channel allows inactivation to be modulated by metal binding. Biophys J 66: 1068eC1075., http://www.100md.com(Tobias Dreker, and Stepha)