点击显示 收起
【关键词】 Ginsenoside
We have demonstrated previously that the 20(S) but not the 20(R) form of ginsenoside Rg3 inhibited K+ currents flowing through Kv1.4 (hKv1.4) channels expressed in Xenopus laevis oocytes, pointing to the presence of specific interaction site(s) for Rg3 in the hKv1.4 channel. In the current study, we sought to identify this site(s). To this end, we first assessed how point mutations of various amino acid residues of the hKv1.4 channel affected inhibition by 20(S)-ginsenoside Rg3 (Rg3). Lys531 residue is known to be a key site for K+ activation and to be part of the extracellular tetraethylammonium (TEA) binding site; the mutation K531Y abolished the Rg3 effect and made the Kv1.4 channel sensitive to TEA applied to the extracellular side of the membrane. Mutations of many other residues, including the pH sensitive-site (H507Q), were without any significant effect. We next examined whether K+ and TEA could alter the effect of Rg3 and vice versa. We found that 1) raising [K+]o reduced the inhibitory effect of Rg3 on hKv1.4 channel currents, whereas Rg3 shifted the K+ activation curve to the right, and 2) TEA caused a rightward shift of the Rg3 concentration-response curve of wild-type hKv1.4 channel currents, whereas Rg3 caused a rightward shift of the TEA concentration-response curve of K531Y mutant channel currents. The docked modeling revealed that Lys531 plays a key role in forming hydrogen bonds between Rg3 and hKv1.4 channels. These results indicate that Rg3 inhibits the hKv1.4 channel current by interacting with residue Lys531.
Voltage-gated K+ (Kv) channels play critical roles in a wide variety of physiological processes, including the regulation of neurotransmitter release, neuronal excitability, heart rate, muscle contraction, hormone secretion, epithelial electrolyte transport, cell volume, and cell proliferation in neuronal and non-neuronal cells (Hille, 2001). Kv channels consist of tetramers of pore-forming Kv and auxiliary Kvβ subunits (Hille, 2001). The Kv subunit is composed of six -helical transmembrane segments (S1-S6). The S4 segment acts as the voltage-sensing apparatus of the K+ channel (Hille, 2001), whereas the pore-forming S5-S6 segments constitute a selectivity filter and govern voltage-dependent increases in K+ permeability. Site-directed mutagenesis studies using Kv subunits have clarified the detailed action and binding sites of various drugs that regulate Kv channel activity (Hille, 2001). Some Kv channel subunits exhibit transient A-type K+ currents and N-type inactivation, and others exhibit long-lasting, delayed-rectifying C-type K+ currents and C-type inactivation, depending on their channel conductance and gating characteristics (Patel and Campbell, 2005).
Ginseng, the root of Panax ginseng C.A. Meyer, is well known in herbal medicine as a tonic and restorative agent, and it is consumed widely around the world. The molecular bases of ginseng's actions are largely unknown. Numerous reports have suggested that the main molecular ingredients responsible are ginsenosides (also called ginseng saponins), amphiphilic molecules consisting of a hydrophobic aglycone backbone (a hydrophobic four-ring steroid-like structure) linked to monomeric, dimeric, or tetrameric hydrophilic carbohydrate side chains (Fig. 1) (Nah, 1997). However, it is unclear how ginsenosides produce their pharmacological effects. We have reported that ginsenoside Rg3 (20-S-protopanaxadiol-3-[O-β-D-glucopyranosyl (1,2)-β-glucopyranoside]) had an inhibitory effect on voltage-dependent human Kv1.4 (hKv1.4) channel activity expressed in Xenopus laevis oocytes (Jeong et al., 2004). In this report, we present evidences that Rg3 interacts with residue Lys531 to inhibit channel currents. In addition, the docked modeling studies using hKv1.4 channels support the concept that the Lys531 residue plays an important role in the Rg3-mediated regulations of hKv1.4 channel by forming hydrogen bonds between Rg3 and hKv1.4 channels.
Fig. 1. Chemical structures of ginsenosides and effects of various ginsenosides on hKv1.4 channel currents. A, ginsenosides differ in the three side chains attached to a common steroid-like ring. Abbreviations for carbohydrates are as follows: Glc, glucopyranoside; Ara (pyr), arabinopyranoside; Rha, rhamnopyranoside. Superscripts indicate the glucose ring carbon that links the two carbohydrates. B, summary histograms showing the effects of various ginsenosides (100 µM each) on the wild-type hKv1.4 channel current. Inset, representative hKv1.4 channel currents elicited by 500-ms voltage steps to +50 mV from a holding potential of -80 mV, in the absence and presence of 100 µM concentrations of ginsenosides Rh2, Rg3, and compound K (CK).
Materials. Ginsenosides were kindly provided by the Korean Ginseng Cooperation (Taejon, Korea). The cDNA for human K+ channel Kv1.4 (GenBank accession number NM_002233) was kindly provided by Dr. O. Pongs (University of Hamburg, Germany). Other agents were purchased from Sigma-Aldrich (St. Louis, MO).
Preparation of Xenopus laevis Oocytes and Microinjection. X. laevis frogs were purchased from Xenopus I (Ann Arbor, MI). Their care and handling were in accordance with the highest standards of institutional guidelines. For isolation of oocytes, frogs were anesthetized with an aerated solution of 3-amino benzoic acid ethyl ester, and the ovarian follicles were removed. The oocytes were separated with collagenase followed by agitation for 2 h in Ca2+-free medium containing 82.5 mM NaCl, 2 mM KCl, 1 mM MgCl2, 5 mM HEPES, 2.5 mM sodium pyruvate, 100 units/ml penicillin, and 100 µg/ml streptomycin. Stage V-VI oocytes were collected and stored in ND96 medium (96 mM NaCl, 2 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2, and 5 mM HEPES, pH 7.5) supplemented with 0.5 mM theophylline and 50 µg/ml gentamicin. The oocyte-containing solution was maintained at 18°C with continuous gentle shaking and renewed every day. Electrophysiological experiments were performed within 5 to 6 days of oocyte isolation, with chemicals applied to the bath. For K+ channel experiments, Kv channel-encoding cRNAs (40 nl) were injected into the animal or vegetal pole of each oocyte 1 day after isolation, using a 10-µl microdispenser (VWR Scientific, West Chester, PA) fitted with a tapered glass pipette tip (15-20 µm in diameter) (Lee et al., 2005).
Site-Directed Mutagenesis of the Kv1.4 Subunit and in Vitro Transcription of Kv1.4 Channel cDNAs. Single or double amino acid substitutions were made using a QuikChange XL site-directed mutagenesis kit (Stratagene, La Jolla, CA), along with Pyrococcus furiosus DNA polymerase and sense and antisense primers encoding the desired mutations. Overlap extension of the target domain by sequential polymerase chain reaction (PCR) was carried out according to the manufacturer's protocol. The final PCR products were transformed into Escherichia coli strain DH5, screened by PCR, and confirmed by sequencing of the target regions. The mutant DNA constructs were linearized at their 3' ends by digestion with XhoI, and run-off transcripts were prepared using the methylated cap analog m7G(5')ppp(5')G. The cRNAs were prepared using a mMessage mMachine transcription kit (Ambion, Austin, TX) with T7 RNA polymerase. The absence of degraded RNA was confirmed by denaturing agarose gel electrophoresis followed by ethidium bromide staining. Likewise, recombinant plasmids containing Kv channel cDNA inserts were linearized by digestion with the appropriate restriction enzymes, and cRNAs were obtained using the mMessage mMachine in vitro transcription kit with SP6 RNA polymerase or T7 polymerase. The final cRNA products were resuspended at a concentration of 1 µg/µl in RNase-free water, and stored at -80°C (Lee et al., 2005).
Data Recording. A custom-made Plexiglas net chamber was used for two-electrode voltage-clamp recordings as reported previously (Lee et al., 2005). The oocytes were impaled with two micro-electrodes filled with 3 M KCl (0.2-0.7 MÙ), and electrophysiological experiments were carried out at room temperature using an Oocyte Clamp (OC-725C; Warner Instruments, Hamden, CT). Stimulation and data acquisition were controlled with a pClamp 8 (Molecular Devices, Sunnyvale, CA). For most electrophysiological experiments, oocytes were perfused initially with ND96 solution (96 mM NaCl, 3 mM KCl, 2 mM CaCl2, and 5 mM HEPES, pH adjusted to 7.4 with NaOH), and control current recordings were obtained. To measure K+ activation of the Kv1.4 channel, a solution was applied in which the NaCl was replaced with various concentrations of KCl. In all cases, the solution was perfused at a flow rate of 3 ml/min, and the system was allowed 30 to 60 s to reach steady state before current recording. The oocytes were then clamped at a holding potential of -80 mV, membrane potential was depolarized to +50 mV for 500 ms every 10 s, and currents were recorded.
Fig. 2. Rg3 effects on wild-type and mutant hKv1.4 channel currents. A, concentration-response curves for Rg3-induced inhibition of wild-type and mutant hKv1.4 channel currents. The same voltage protocol as in A was used to elicit the K+ currents. Amino acids in outer pore residues near selectivity filter were mutated into alanine. B, representative wild-type and K531Y mutant hKv1.4 channel currents elicited in the absence and presence of various concentrations of Rg3. C, concentration-response curves for Rg3-induced inhibition of wild-type and mutant hKv1.4 channel currents. D, I-V relationships of wild-type and mutant hKv1.4 channels in the absence or presence of Rg3. Voltage pulses of 500-ms duration were applied in 10-mV increments and at 10-s intervals from a holding potential of -80 mV. The peaks of the evoked currents, normalized to the peak current evoked by the voltage step to +60 mV in the absence of Rg3, were used in the I-V plot. *, P < 0.01 compared with wild-type hKv1.4 channel.
Homology Modeling. A homology model of the hKv1.4 was built on the basis of the 2.9-Å crystal structure of the rat Kv1.2 channel (Protein Data Bank code 2A79), using the homology modeling program MODELLER 8v2 (Sali and Blundell, 1993). Sequence alignment was carried out using the AlignX module of the Vector NTI (Lu and Moriyama, 2004). A stretch of amino acids in the Kv1.2 channel corresponding to the region from Met438 to Thr571 in hKv1.4 was chosen as the template for homology modeling, because the other regions had structures missing in the Protein Data Bank file. The chosen region contained the pore and its neighboring areas, and the sequence identity between hKv1.4 with rat Kv1.2 through this region was 90%. The homotetrameric structure was restrained to maintain symmetry during homology modeling. A total of 200 structures was generated, and the one with the lowest DOPE score from MODELLER was chosen for further minimization. Hydrogen atoms were added to the homology model using Sybyl v7.0 (Tripos Inc., St. Louis, MO). The homology model structure was energy-minimized using the Tripos forcefield protocol in Sybyl. The same strategies were used to generate a homology model of K531Y.
Virtual Docking. The structure of Rg3 was constructed using Chemdraw ultra 8.0 (Cambridgesoft, Cambridge, MA) and converted to a 3-dimensional model and energy-minimized using Chem3D ultra 8.0 (Cambridgesoft, Cambridge, MA), followed by a second round of energy minimization using SYBYL forcefield (Tripos, St. Louis, MO). The virtual dockings of Rg3 to the homology model of hKv1.4 wild-type and K531Y mutant channels were performed using GOLD v3.0 (The Cambridge Crystallographic Data Centre, Cambridge, UK), a program that uses stochastic genetic algorithms for conformational searching (Verdonk et al., 2003). The Lys531 residues in each of the four subunits were designated as the active site residues, and the active radium was set as 10 Å from the active site residues. The docked models with the best GOLD scores were selected for final complex structural analysis. The interactions between the ligand and each homology model were examined using the SILVER tool of the GOLD software package. All structural figures were prepared using PyMol v0.98 (DeLano Scientific LLC, San Francisco, CA).
Data Analysis. To obtain the concentration-response curve of the effect of Rg3 on the K+ current from the hKv1.4 channel, the peak amplitudes at different concentrations of Rg3 were plotted, and Origin software (OriginLab Corp., Northampton, MA) was used to fit the plot to the Hill equation: y/ymax = [A]nH/([A]nH + [IC50]nH), where y is the peak current at a given concentration of Rg3, ymax is the maximal peak current, IC50 is the concentration of Rg3 producing a half-maximal effect, [A] is the concentration of Rg3, and nH is the Hill coefficient. All values are presented as means ± S.E.M. The significance of differences between mean control and treatment values was determined using Student's t test. P < 0.05 was considered statistically significant.
Rg3 Inhibited hKv1.4 Channel Currents More Potently than Other Ginsenosides. Using the two-electrode voltage-clamp technique, we recorded hKv1.4 channel currents from X. laevis oocytes injected with cRNA encoding the hKv1.4 channel protein. To elicit the currents, we applied a voltage step (500-ms duration) to +50 mV at 10-s intervals from a holding potential of -80 mV. The currents evoked by this voltage-clamp protocol were transient A-type K+ currents that decayed rapidly (Fig. 1B, inset) (Gòmez-Hernandez et al., 1997). Rg3 at 100 µM inhibited the hKv1.4 channel currents by an average of 65%, and other ginsenosides (Rb1, Rc, Rd, Rf, Rg1, Rh2, and compound K) were much less effective (Fig. 1B). The Rg3 effect was concentration-dependent (Fig. 2A) and reversible (data not shown). The IC50 value and Hill coefficient were 32.6 ± 2.2 µM and 1.59 ± 0.13, respectively (Table 1).
TABLE 1 Effects of Rg3 on wild-type and mutant hKv1.4 channels expressed in X. laevis oocytes
Currents were elicited by single-step voltage pulses from –80 to +50 mV.
To assess the effect of Rg3 on the current-voltage (I-V) relationship, we constructed I-V curves with and without Rg3 in the bath. The current responses evoked by voltage steps (i.e., a series of voltage pulses of 500-ms duration given in 10-mV increments and 10-s intervals from the holding potential of -80 mV) were used to construct the I-V curve. In the absence of Rg3, hKv1.4 currents were elicited by voltage pulses more positive than -40 mV, and current amplitude increased linearly with further depolarization (Fig. 2D). The presence of Rg3 reduced current amplitude over the entire voltage range in which the current was activated (Fig. 2D).
Fig. 3. Mutual antagonistic actions of extracellular [K+] and Rg3 on hKv1.4 channel currents. A, hKv1.4 channel currents evoked in 3 and 99 mM K+ by 500-ms voltage pulses to +50 mV from a holding potential of 0 mV, in the absence and presence of Rg3.B, plot showing the effect of extracellular [K+]onRg3 (100 µM) inhibition of the current. C, K+ current activation curves constructed from the currents evoked in the absence and presence of Rg3. D, I-V plots for hKv1.4 channel currents obtained in 3 and 99 mM K+, in the absence and presence of Rg3 (100 µM). Currents are normalized to that evoked in 3 mM K+ by the voltage step to +60 mV in the absence of Rg3.
The K531Y Substitution Affected Rg3 Inhibition of Channel Current. Previous works showed that Rg3 inhibits hvKv1.4 channel current in a stereospecific manner (Jeong., 2004) and that Rg3 regulates 5-HT3A receptor channel activity through interaction with amino acid residues in the channel pore region (Lee et al., 2007). We therefore hypothesized that Rg3 might have a specific interaction site(s) on the hKv1.4 channel and that Rg3 interaction site(s) with the hKv1.4 channel might be related with channel pore region. To identify that site(s), we first made the following changes by site-directed mutations in channel pore regions: channel pore sites (S510K, D513Q, V525L, and V535Q), outer pore sites (K531A, P532A, I533A, T534A, V535A). Next, we also constructed mutant channels as follows: N-glycosylation site (N353Q) (Watanabe et al., 2004), voltage sensor site (R447C and R450C) (Fedida and Hesketh, 2001), voltage shift sites (L478F and G548P) (Judge et al., 1999; Magidovich and Yifrach, 2004), pH sensitive site (H507Q) (Claydon et al., 2004) and C-type inactivation site (V560A) (Bett and Rasmusson, 2004). We found that in one of outer pores, K531A substitution significantly attenuated Rg3 inhibition of the hKv1.4 channel currents (Fig. 2A), whereas the other changes had no significant effects (Table 1). These results showed a possibility that Rg3 regulates hKv1.4 channel activity by interacting with Lys531, which is also known as one of K+ activation sites. Therefore, we constructed mutant channels at K+ activation sites (K531Y, I533M, and K531Y-I533M) (Pardo et al., 1992; Claydon et al., 2004). We found that the K531Y substitution and the K531Y-I533M double substitution almost abolished Rg3 inhibition of the hKv1.4 channel currents (Fig. 2, B-D). These results indicate that Rg3-induced regulation of hKv1.4 channel activity is closely related to the Lys531 residue.
Extracellular K+ and Rg3 Each Antagonized the Other's Effects on hKv1.4 Channel Currents. If indeed Rg3 produces its effect by interacting with Lys531, the K+ activation site, an increase in extracellular [K+]o would compete with Rg3 for Lys531 and thus inhibit the action of the ginsenoside. Conversely, Rg3 would inhibit K+ activation by competing with K+ for Lys531. We found that extracellular K+ and Rg3 each indeed antagonized the other's effect. Figure 3, A-C, shows that raising extracellular [K+]o inhibited the effect of Rg3 (IC50 of extracellular K+ for the Rg3 effect: 6.4 ± 2.9 mM), whereas Fig. 3D shows that Rg3 (100 µM) inhibited the effect of K+, thus shifting the K+ activation curve to the right (EC50 values of K+ before and after Rg3 treatment: 4.2 ± 0.9 and 9.2 ± 1.5 mM, respectively; P < 0.01). These findings confirm that Rg3 competes with extracellular [K+] for the Lys531 residue.
Fig. 4. Mutually antagonistic action of TEA and Rg3 on wild-type and K531Y mutant hKv1.4 channel currents. A, concentration-dependent effects of TEA on wild-type and K531Y hKv1.4 channel currents. B, concentration-response curves for Rg3 inhibition of the wild-type hKv1.4 channel current in the absence and presence of TEA (10 µM). Inset, representative K+ current traces from control oocytes () and oocytes treated with 100 µM Rg3 + 10-µM TEA (), or 100-µM Rg3 alone (). C, concentration-response curves for TEA inhibition of K531Y channel currents evoked in the absence and presence of Rg3 (100 µM). Inset, representative K+ current traces from controls () and oocytes treated with 100-µM Rg3 + 30-µM TEA (), or 30 µM TEA alone ().
TEA and Rg3 Each Inhibited the Other's Effect on hKv1.4 Channel Currents. Residue Lys531 is also known to form a part of the extracellular binding site for tetraethylammonium (TEA) (Heginbotham and Mackinnon, 1992; Gòmez-Hernandez et al., 1997). We therefore reasoned that if Rg3 produced its effect by interacting with Lys531, extracellular application of TEA should antagonize the effect of Rg3 on channel currents, and conversely, Rg3 should antagonize the TEA effect. To test these possibilities, we used oocytes expressing wild-type or K531Y mutant hKv1.4 channels, because extracellularly applied TEA binds to both types of channel but inhibits only the mutant channels (Fig. 4A; IC50, 26.5 ± 2.6 µM). Figure 4B shows that TEA had no effect by itself on wild-type hKv1.4 channel currents but that it inhibited the action of Rg3, thus causing the Rg3 concentration-response curve to shift to the right (IC50 values of Rg3 before and during 10-µM TEA treatment: 35.1 ± 3.6 and 93.1 ± 6.7 µM, respectively; P < 0.001). Conversely, Rg3 antagonized the effect of TEA on K531Y channel currents, causing a rightward shift of the TEA concentration-response curve (IC50 values of TEA before and during 100-µM Rg3 treatment: 23.3 ± 2.7 and 40.1 ± 7.1 µM, respectively; P < 0.01) (Fig. 4C). These results lend further support to the hypothesis that Rg3 interacts with Lys531 to inhibit hKv1.4 channel currents.
Docked Modeling of Interactions between Rg3 and the hKv1.4 Channel. To further examine the possible interaction mode between Rg3 and the hKv1.4 channel, we carried out homology modeling of wild-type and K531Y mutant hKv1.4 channels. Our model was generated using the MODELLER program and the crystal structure of hKv1.2. Virtual docking of Rg3 to the homology models was performed using the docking program GOLD. It is noteworthy that the best-fit docking results showed that Rg3 forms six hydrogen bonds with wild-type hKv1.4 channels but only two hydrogen bonds with K531Y mutant channels (Fig. 5 and Table 2). In the wild-type channel, the first carbohydrate coupled to the Rg3 backbone forms two hydrogen bonds with Lys531 of domain I (designated as Roman numeral I) and one hydrogen bond with His507 (IV). The second carbohydrate of Rg3 forms one hydrogen bond with Lys531 (I), one hydrogen bond with Thr505 (I), and one hydrogen bond with His507 (I). In the K531Y mutant channel, the second carbohydrate of Rg3 forms one hydrogen bond with Tyr531 (I), and the first carbohydrate of Rg3 forms a hydrogen bond with His507 (IV) (Fig. 5 and Table 2). It is noteworthy that the wild-type Kv1.4 channel pore is blocked by the hydrophobic triterpenoid backbone moiety of Rg3. The mutant channel is also blocked by Rg3, but the low affinity of Rg3 to the mutant channel (inferred from the small number of hydrogen bonds) might result in ineffective blocking of the mutant channel by Rg3, thus accounting for the inability of Rg3 to inhibit K531Y mutant channel currents.
Fig. 5. Theoretical docking of Rg3 to wild-type and K531Y mutant Kv1.4 channels. Hydrogen bonds are denoted as red broken lines. The Roman numerals in parenthesis indicate the number of the homotetramer subunit involved in the interaction. A, structure of Rg3. Glc, glucopyranoside. B, a docked model of wild-type Kv1.4 channel and Rg3. Six hydrogen bonds are predicted to form between key residues and Rg3, including three hydrogen bonds to Lys531 (I) one hydrogen bond each to Thr505 (I), His507 (I), and His507 (IV). The lower panel contains a 90° rotated view of the upper panel, showing that the Kv1.4 channel pore is predicted to be blocked by the hydrophobic triterpenoid moiety of the ginsenoside. C, a docked model showing binding of Rg3 to the K531Y mutant channel. Only two hydrogen bonds form, one each to Thr531 (I) and His507 (IV) (see Table 2).
TABLE 2 Modeled hydrogen bonds between hKv1.4 and Rg3
The Roman numerals in parentheses indicate the subunit number of the homotetramer involved in the interaction. The ' and '' markings indicate the first and second carbohydrates of Rg3, respectively (see Fig. 5A), and the adjacent numbers indicate the position of the relevant carbon in each carbohydrate ring.
The Kv1.4 channel is a transient A-type or rapidly inactivating Kv channel. Kv1.4 channels are mainly located at axon and presynaptic terminals (Alonso and Widmer, 1997; Cooper et al., 1998; Hoffman and Johnston, 1998; Adams et al., 2000) and function to modulate action potential waveforms and neurotransmitter release (Jackson et al., 1991; Debanne et al., 1997). They also affect the amplitude of the plateau phase and duration of action potentials in ventricular myocytes (Campbell et al., 1993; Patel and Campbell, 2005). Thus, these channels are one of the targets of drugs for treatment of pathologic conditions including cardiac arrhythmia. We previously demonstrated that Rg3 regulated hKv1.4 channel currents in a stereospecific manner (Jeong et al., 2004). However, very little was known of its molecular mechanism of action.
In the present study we observed that the K531A or K531Y mutation, in addition to raising extracellular [K+]o from 3 to 99 mM, attenuated or almost abolished Rg3 inhibition of hKv1.4 channel currents. These results show a possibility that Rg3 might interacts with the K+ activation sites His507, Lys531, and Ile533 (Pardo et al., 1992; Claydon et al., 2004). To test these possibilities, we examined the effects of Rg3 on H507Q, K531Y, I533M, and K531Y-I533M channels. As shown in Fig. 2, Rg3 did not inhibit K531Y and K531Y-I533M channels even at high concentrations, whereas its effect on H507Q and I533M mutant channels was similar to that seen with wild-type channels (Table 1). These results indicate that although both His507 and Ile533 are involved in K+ activation (Pardo et al., 1992; Claydon et al., 2004), they are not involved in Rg3 regulation of channel activity. This view was supported by the results from our double mutation experiments. Furthermore, mutations in the channel pore region, pH-sensitive sites, voltage sensor, and other regulatory sites did not affect Rg3 inhibition (Table 1). It is noteworthy that Rg3 is not structurally similar to TEA, a well known K+ channel blocker that can function on either side of the cell membrane. In contrast to TEA, Rg3 has no charged groups apart from the hydroxyls of its carbohydrate and backbone structures (Fig. 1A). Despite the structural difference between Rg3 and TEA, we found that the K531Y mutation of amino acid 531, which forms part of the external TEA interaction site, almost abolished Rg3 inhibition of the channel currents. We also showed that Rg3 competes with TEA for inhibition of K531Y channel currents (Fig. 4B) and vice versa (Fig. 4C). It is noteworthy that the rightward shift of the Rg3 concentration-response curve caused by TEA in wild-type channels was stronger than that of the TEA concentration-response curve caused by Rg3 in K531Y channels. Thus, by making use of the fact that K531Y channels are sensitive to TEA, whereas wild-type channels are not, we were able to demonstrate that Cys531 may be an allosteric interaction site for TEA and Rg3. However, it is unlikely that Rg3 exhibits an allosteric interaction with TEA on the intracellular surface of the channel, because we have shown, with the use of in outside-out patch-clamp experiments, that Rg3 regulates channels from the outside, not the inside (Lee et al., 2004).
MacDonald et al. (1998) showed that n-alkyl sulfate anions but not TEA inhibit wild-type rat Kv1.4 channel currents and that the mutation K533Y rendered channels sensitive to TEA but insensitive to n-alkyl sulfate anions. This indicates that Lys533 may play a role in n-alkyl sulfate anion-mediated Kv1.4 channel regulation via the external TEA interaction site. In addition, Zaks-Makhina et al. (2004) and Salvador-Recatala et al. (2006) have studied a neuroprotective compound, called 48F10, from yeast. They showed that 48F10 inhibited R476Y mutant rat Kv1.5 channel currents (Arg476 is analogous to Lys531 in hKv1.4 channel) and wild-type rat Kv2.1 channel currents via the external TEA interaction site, because the presence of external TEA greatly reduced 48F10 current inhibition. Taken together, the previous and present observations raise the possibility that a lysine or analogous amino acid residue in the outer pores of subsets of Kv channels not only forms part of the external TEA interaction site but also plays a role as an allosteric interaction or overlapping site for TEA and certain other compounds.
Ginsenosides have effects on multiple targets (Attele et al., 1999). We and others have reported that ginsenosides, including Rg3, also act on various ion channels at pre- and postsynaptic sites in the nervous systems and inhibit neurotransmitter release (Nah et al., 1995; Liu et al., 2001; Tachikawa et al., 2001; Kim et al., 2002; Sala et al., 2002; Choi et al., 2003; Lee et al., 2005). Thus, it seems that ginsenosides show a low degree of selectivity for ion channels compared with drugs or toxins that act on particular ion channels. However, we were not able to clearly define and comprehend the molecular mechanisms underlying ginsenoside-mediated regulation of multiple ion channels. We demonstrated that Rg3 regulates 5-HT3A receptor channel activity in the open state through interactions with amino acids such as Val291, Phe292, and Ile295 in the gating pore region of transmembrane domain 2 (Lee et al., 2007). In the present study, we found that Rg3 regulates Kv1.4 channel activity through interaction with the outer pore Lys531 residue and may interact allosterically with the external TEA binding site. Thus, Rg3 affects 5-HT3A receptors and Kv1.4 channel activity via different interaction sites and different modes of regulation.
We next sought to examine the possible mechanisms underlying Rg3-induced hKv1.4 channel activity regulation. As shown in Fig. 1A, Rg3 consists of a carbohydrate portion, a steroid backbone, and an alkene side chain. To determine how Rg3 interacts with Kv1.4 channels, we performed docked modeling experiments using wild-type and K531Y mutant channels. Our docked modeling study revealed that the two carbohydrates of Rg3 could form six hydrogen bonds with residues Thr505 (I), His507 (I), His507 (IV), and Lys531 (I) in the wild-type channel. We demonstrated previously that Rg3 regulates ligand-gated ion channels at the extracellular but not the intracellular side using the outside-out patch-clamp method (Lee et al., 2004) and that modification or removal of the carbohydrate portion of Rg3 abolished Rg3-mediated ion channel regulations, but at the time we were unable to explain exactly how Rg3 regulates ion channel activity from the extracellular side and that the carbohydrate portion of Rg3 was involved in ion channel regulations (Kim et al., 2005). The modeling results in our present study suggest that the previously examined carbohydrate modifications might induce a conformational change in Rg3 and/or prevent the formation of hydrogen bonds between Rg3 and the critical residues. In the present study, the K531Y mutation was found to induce a conformational change in the channel protein (Fig. 5C), resulting in the formation of only two hydrogen bonds between Rg3 and amino acid residues at the pore entryway. Thus, loss of hydrogen bonding between Rg3 and the channel outer pore, whether through mutation or carbohydrate modification, seems to decrease the binding affinity of Rg3, resulting in loss of Rg3-induced channel regulation. Furthermore, as shown Fig. 5B, our modeling revealed that the triterpenoid backbone of Rg3 blocks the channel pore when the proper hydrogen bonds are formed; this may provide a secondary level of Rg3-induced inhibition of outward K+ currents after depolarization. Future studies will be necessary to determine the exact roles of the carbohydrates and/or triterpenoid backbone structures of Rg3 in terms of Kv1.4 channel regulation.
We might ask whether the in vitro Rg3 inhibition of the hKv1.4 channel applies also to its in vivo pharmacological effects. Ginseng has many beneficial effects on the cardiovascular systems (Gillis, 1997). Gao et al. (1992) have shown that ginsenoside administration via the intravenous route attenuates ischemic and reperfused arrhythmia in rats, and Yang et al. (1999) showed that ginsenoside administration via the intraperitoneal route attenuates myocardial reperfusion arrhythmia in rats fed a high cholesterol diet. Antiarrhythmic agents such as quinidine block antiarrhythmic effects on Kv1.4 and other cloned K+ channels (Wang et al., 2003), but we have no direct evidence that Rg3-mediated Kv1.4 channel regulation can be used prophylactically or therapeutically against arrhythmia, as quinidine can be. More investigation is needed about the potential application of Rg3 to heart dysfunction. In addition, Kim et al. (1999a,b) showed that Rg3 induced relaxation of the rat aorta via endothelium-dependent and -independent routes. They further showed that K+ channels in the rat aorta might be involved in the effect of Rg3, because Rg3-mediated aorta relaxation was achieved in a TEA-sensitive manner.
In summary, we have used site-directed mutagenesis, K+ activation experiments, and analysis of the external TEA interaction site to further characterize Rg3 regulation of hKv1.4 channel activity. We found that the Lys531 residue of the hKv1.4 channel is involved in Rg3-mediated Kv1.4 channel regulation and that Rg3 may interact allosterically with the external TEA binding site via residue Lys531. Furthermore, by molecular modeling, we showed for the first time that two carbohydrates of Rg3 interact with amino acid residues, including Lys531, through the formation of hydrogen bonds, which are decreased in K531Y mutant channels. These novel findings provide insight into the pharmacological basis of the beneficial effects of ginseng on cardiovascular systems.
ABBREVIATIONS: Kv, voltage-gated K+ channel; Rg3, ginsenoside 20-S-protopanaxadiol-3-[O-β-D-glucopyranosyl (12)-β-glucopyranoside]); PCR, polymerase chain reaction; I-V, current-voltage; TEA, tetraethylammonium.
【参考文献】
Adams JP, Anderson AE, Varga AW, Dineley KT, Cook RG, Pfaffinger PJ, and Sweatt JD (2000) The A-type potassium channel Kv4.2 is a substrate for the mitogen-activated protein kinase ERK. J Neurochem 75: 2277-2287.
Alonso G and Widmer H (1997) Clustering of Kv4.2 potassium channels in postsynaptic membrane of rat supraoptic neurons: an ultrastructural study. Neuroscience 77: 617-621.
Attele AS, Wu JA, Yuan CS (1999) Ginseng pharmacology: multiple constituents and multiple actions. Biochem Pharmacol 58: 1685-1693.
Bett GC and Rasmusson RL (2004) Inactivation and recovery in Kv1.4 K+ channels: lipophilic interactions at the intracellular mouth of the pore. J Physiol 556: 109-120.[Abstract/Free Full Text]
Campbell DL, Rasmusson RL, Qu Y, Strauss HC (1993) The calcium-independent transient outward potassium current in isolated ferret right ventricular myocytes. I. Basic characterization and kinetic analysis. J Gen Physiol 101: 571-601.[Abstract/Free Full Text]
Choi S, Lee JH, Oh S, Rhim H, Lee SM, and Nah SY (2003) Effects of ginsenoside Rg2 on the 5-HT3A receptor-mediated ion current in Xenopus oocytes. Mol Cells 15: 108-113.
Claydon TW, Makary SY, Dibb KM, and Boyett MR (2004) K+ activation of kir3.1/kir3.4 and kv1.4 K+ channels is regulated by extracellular charges. Biophys J 87: 2407-2418.
Cooper EC, Milroy A, Jan YN, Jan LY, and Lowenstein DH (1998) Presynaptic localization of Kv1.4-containing A-type potassium channels near excitatory synapses in the hippocampus. J Neurosci 18: 965-974.[Abstract/Free Full Text]
Debanne D, Guerineau NC, Gahwiler BH, and Thompson SM (1997) Action-potential propagation gated by an axonal I(A)-like K+ conductance in hippocampus. Nature 389: 286-289.
Fedida D and Hesketh JC (2001) Gating of voltage-dependent potassium channels. Prog Biophys Mol Biol 75: 165-199.
Gao BY, Li XJ, Liu L, and Zhang BH (1992) Effect of panaxatriol saponins isolated from Panax notoginseng (PTS) on myocardial ischemic arrhythmia in mice and rats. Yao Xue Xue Bao 27: 641-644.
Gillis CN (1997) Panax ginseng pharmacology: a nitric oxide link? Biochem Pharmacol 54: 1-8.
Gòmez-Hernandez JM, Lorra C, Pardo LA, Stuhmer W, Pongs O, Heinemann SH, and Elliott AA (1997) Molecular basis for different pore properties of potassium channels from the rat brain Kv1 gene family. Pflugers Arch 434: 661-668.
Hille B (2001) Ion channels of excitable membranes. Sinauer Associates, Inc, Sunderland, MA.
Heginbotham L and Mackinnon R (1992) The aromatic binding site for tetraethylammonium ion on potassium channels. Neuron 8: 483-491.
Hoffman DA and Johnston D (1998) Downregulation of transient K+ channels in dendrites of hippocampal CA1 pyramidal neurons by activation of PKA and PKC. J Neurosci 18: 3521-3528.[Abstract/Free Full Text]
Jackson MB, Konnerth A, and Augustine GJ (1991) Action potential broadening and frequency-dependent facilitation of calcium signals in pituitary nerve terminals. Proc Natl Acad Sci U S A 88: 380-384.[Abstract/Free Full Text]
Jeong SM, Lee JH, Kim JH, Lee BH, Yoon IS, Lee JH, Kim DH, Rhim H, Kim Y, and Nah SY (2004) Stereospecificity of ginsenoside Rg3 action on ion channels. Mol Cells 18: 383-389.
Judge SI, Monteiro MJ, Yeh JZ, and Bever CT (1999) Inactivation gating and 4-AP sensitivity in human brain Kv1.4 potassium channel. Brain Res 831: 43-54.
Kim JH, Hong YH, Lee JH, Kim DH, Jeong SM, Lee BH, Lee SM, and Nah SY (2005) A role for the carbohydrate portion of ginsenoside Rg3 in Na+ channel inhibition. Mol Cells 19: 137-142.
Kim ND, Kang SY, Kim MJ, Park JH, and Schini-Kerth VB (1999a) The ginsenoside Rg3 evokes endothelium-independent relaxation in rat aortic rings: role of K+ channels. Eur J Pharmacol 367: 51-57.
Kim ND, Kang SY, Park JH, and Schini-Kerth VB (1999b) Ginsenoside Rg3 mediates endothelium-dependent relaxation in response to ginsenosides in rat aorta: role of K+ channels. Eur J Pharmacol 367: 41-49.
Kim S, Ahn K, Oh TH, Nah SY, and Rhim H (2002) Inhibitory effect of ginsenosides on NMDA receptor-mediated signals in rat hippocampal neurons. Biochem Biophys Res Commun 296: 247-254.
Lee BH, Jeong SM, Ha TS, Park CS, Lee JH, Kim JH, Kim DH, Han JS, Kim HC, Ko SR, et al. (2004) Ginsenosides regulate ligand-gated ion channels from the outside. Mol Cells 18: 115-121.
Lee JH, Jeong SM, Kim JH, Lee BH, Yoon IS, Lee JH, Choi SH, Kim DH, Rhim H, Kim S, et al. (2005) Characteristics of ginsenoside Rg3-mediated brain Na+ current inhibition. Mol Pharmacol 68: 1114-1126.[Abstract/Free Full Text]
Lee BH, Lee JH, Lee SM, Yoon IS, Lee JH, Choi SH, Pyo MK, Rhim H, Kim HC, Jang CG, et al. (2007) Identification of ginsenoside interaction sites in 5-HT3A receptors. Neuropharmacology 52: 1139-1150.
Liu D, Li B, Liu Y, Attele AS, Kyle JW, and Yuan CS (2001) Voltage-dependent inhibition of brain Na+ channels by American ginseng. Eur J Pharmacol 413: 47-54.
Lu G and Moriyama EN (2004) Vector NTI, a balanced all-in-one sequence analysis suite. Brief Bioinform 5: 378-388.[Abstract/Free Full Text]
MacDonald S, Elliott AA, Harrold JA, and Elliott JR (1998) Influence of outer pore residue K533 on the inhibition of Kv1.4 potassium channel s by n-alkyl sulphate anions. Pflugers Arch 436: 623-626.
Magidovich E and Yifrach O (2004) Conserved gating hinge in ligand- and voltage-dependent K+ channels. Biochemistry 43: 13242-13247.
Nah SY (1997) Ginseng: recent advances and trends. Kor J Ginseng Sci 21: 1-12.
Nah SY, Park HJ, and McCleskey EW (1995) A trace component of ginseng that inhibits Ca2+ channels through a pertussis toxin-sensitive G protein. Proc Natl Acad Sci U S A 92: 8739-8743.[Abstract/Free Full Text]
Patel SP and Campbell DL (2005) Transient outward potassium current, `Ito', phenotypes in the mammalian left ventricle: underlying molecular, cellular and biophysical mechanisms. J Physiol 569: 7-39.[Abstract/Free Full Text]
Pardo LA, Heinemann SH, Terlau H, Ludewig U, Lorra C, Pongs O, and Stuhmer W (1992) Extracellular K+ specifically modulates a rat brain K+ channel. Proc Natl Acad Sci U S A 89: 2466-2470.[Abstract/Free Full Text]
Sala F, Mulet J, Choi S, Jung SY, Nah SY, Rhim H, Valor LM, Criado M, and Sala S (2002) Effects of ginsenoside Rg2 on human neuronal nicotinic acetylcholine receptors. J Pharmacol Exp Ther 301: 1052-1059.[Abstract/Free Full Text]
Sali A and Blundell TL (1993) Comparative protein modelling by satisfaction of spatial restraints. J Mol Biol 234: 779-815.
Salvador-Recatala V, Kim Y, Zaks-Makhina E, and Levitan ES (2006) Voltage-gated K+ channel block by catechol derivatives: defining nonselective and selective pharmacophores. J Pharmacol Exp Ther 319: 758-764.[Abstract/Free Full Text]
Tachikawa E, Kudo K, Nunokawa M, Kashimoto T, Takahashi E, Kitakawa S (2001) Characterization of ginseng saponin ginsenoside-Rg3 inhibition of catecholamine secretion in bovine adrenal chromaffin cells. Biochem Pharmacol 62: 943-951.
Verdonk ML, Cole JC, Hartshorn MJ, Murray CW, and Taylor RD (2003) Improved protein-ligand docking using GOLD. Proteins 52: 609-623.
Wang S, Morales MJ, Qu YJ, Bett GCL, Strauss HC, and Rasmussion RL (2003) Kv1.4 channel block by quinidine: evidence for a drug-induced allosteric effect. J Physiol 546: 387-401.[Abstract/Free Full Text]
Watanabe I, Zhu J, Recio-Pinto E, and Thornhill WB (2004) Glycosylation affects the protein stability and cell surface expression of Kv1.4 but Not Kv1.1 potassium channels. A pore region determinant dictates the effect of glycosylation on trafficking. J Biol Chem 279: 8879-8885.[Abstract/Free Full Text]
Yang Y, He K, Wu T, Li Q, Zhang JS, and Fu ZG (1999) Effects of ginsenosides on myocardial reperfusion arrhythmia and lipid superoxidation in high cholesterol diet rats. Shi Yan Sheng Wu Xue Bao 32: 349-352.
Zaks-Makhina E, Kim Y, Aizenman E, and Levitan ES (2004) Novel neuroprotective K+ channel inhibitor identified by high-throughput screening in yeast. Mol. Pharmacol 65: 214-219.[Abstract/Free Full Text]
作者单位：Ginsentology Research Laboratory and Department of Physiology, College of Veterinary Medicine (J.-H.L., B.-H.L., S.-H.C., I.-S.Y., M.K.P., T.-J.S., W.-S.C., S.-Y.N.) and Bio/Molecular Informatics Center (Y.-H.L), Konkuk University, Seoul, Korea; Life Science Division, Korea Institute of Science and