点击显示 收起
The Johns Hopkins University School of Medicine, Baltimore, Maryland (S.Y.T., G.F.T., R.A.L.)
Departments of Physiology and Medicine, Heart & Stroke Richard Lewar Centre
Division of Cadiology University Health Network (R.G.T., P.H.B.), University of Toronto, Toronto, Ontario, Canada
Abstract
Voltage-gated Na+ (Nav) channels are responsible for initiating action potentials in excitable cells and are the targets of local anesthetics (LA). The LA receptor is localized to the cytoplasmic pore mouth formed by the S6 segments from all four domains (DIeCDIV) but several outer pore-lining residues have also been shown to influence LA block (albeit somewhat modestly). Many of the reported amino acid substitutions, however, also disrupt the inactivated conformations that favor LA binding, complicating the interpretation of their specific effects on drug block. In this article, we report that an externally accessible aromatic residue in the Nav channel pore, DIV-Trp1531, when substituted with cysteine, completely abolished LA block (e.g., 300 e mexiletine induced a use-dependent block with 65.0 ± 2.9% remaining current and eC11.0 ± 0.6 mV of steady-state inactivation shift of wild-type (WT) channels versus 97.4 ± 0.7% and eC2.4 ± 2.1 mV of W1531C, respectively; p < 0.05) without destabilizing fast inactivation (complete inactivation at 20 ms at eC20 mV; V1/2 = eC70.0 ± 1.6 mV versus eC48.6 ± 0.5 mV of WT). W1531C also abolished internal QX-222 block (200 e; 98.4 ± 3.4% versus 54.0 ± 3.2% of WT) without altering drug access. It is interesting that W1531Y restored WT blocking behavior, whereas W1531A channels exhibited an intermediate phenotype. Together, our results provide novel insights into the mechanism of drug action, and the structural relationship between the LA receptor and the outer pore vestibule.
Voltage-gated Na+ (Nav) channels mediate the rapid transmission of depolarizing electrical impulses in excitable tissues, such as nerve, heart, and skeletal muscle, to enable coordination of processes from muscle contraction to cognition. Genetic defects in Nav channels lead to various heritable excitability diseases such as muscle paralysis (e.g., hyperkalemic periodic paralysis, myotonia congenita), cardiac arrhythmias (e.g., LQT3 syndrome), and epilepsy, often as a result of altered channel inactivation, which subsequently causes misfiring of action potentials (Hudson et al., 1995; Keating and Sanguinetti, 1996; Hayward et al., 1997; Marban et al., 1998; Balser, 2001). It is not surprising that Nav channels are important clinically as the primary target of numerous drugs, such as local anesthetic (LA), anticonvulsant, antiarrhythmic, and antiepileptic agents. These Nav channel drugs exert their effects by preferentially binding to open and inactivated Nav channels (LA block is enhanced at depolarized potentials or with repetitive pulsing or both, i.e., use dependence), thereby inhibiting Na+ currents and ultimately reducing cellular excitability (Courtney, 1975; Hille, 1977; Hondeghem and Katzung, 1977; Bean et al., 1983). According to the modulated receptor hypothesis, state-dependent LA binding results from conformational changes of the drug receptor during gating (Hille, 1977): LA affinity is low when channels are in resting closed states but increases when they are open or inactivated (Hille, 1977; Hondeghem and Katzung, 1977). Therefore, amino acid substitutions that alter gating can indirectly modify LA block without affecting the intrinsic drug affinity. In addition, access to the binding site, which determines the rates of drug binding and unbinding from the receptor, is also state-dependent; both hydrophobic and hydrophilic pathways are available depending on the hydrophobicity of drugs (Hille, 1977; Hondeghem and Katzung, 1977; Schwartz et al., 1977; Ragsdale et al., 1994; Qu et al., 1995; Ragsdale et al., 1996). Hydrophobic LAs bind and unbind freely to Nav channels (Hille, 1977), but permanently charged drug derivatives (e.g., QX-222 and GEA-968) can gain access to the receptor site via the hydrophilic pathway, which is available only when channels are open (Strichartz, 1973; Courtney, 1975; Hille, 1977). Amphiphilic polar LAs such as mexiletine and lidocaine can block Nav channels via both the hydrophobic and hydrophilic pathways. Because both pathways can be modified by mutations, LA block depends on both the intrinsic binding affinity of the LA receptor as well as drug access.
Previous studies have localized the LA receptor of Nav channels to the cytoplasmic pore mouth (Strichartz, 1973) formed by the S6 segments from all four domains (DIeCDIV) (Ragsdale et al., 1994; Qu et al., 1995; Ragsdale et al., 1996; Wang et al., 2000; Yarov-Yarovoy et al., 2002; Kondratiev and Tomaselli, 2003). Single-channel recordings revealed that internally perfused QX-314, a permanently charged lidocaine congener, traverses 70% of the transmembrane electric field from the inside to reach its binding site (Gingrich et al., 1993; Kimbrough and Gingrich, 2000), approximately equivalent to the 21% electrical distance for external Cd2+ binding to cysteine-substituted P-loop residues (Backx et al., 1992; Chiamvimonvat et al., 1996a; Li et al., 2000). These observations strongly suggest that the external pore vestibule and the cytoplasmic LA receptor are in close proximity. Indeed, several P-loop residues (such as DI-Tyr401, DI-Trp402, DII-Glu755, DIII-Lys1237; rat skeletal muscle rNav1.4 numbering unless otherwise specified) have been shown to influence drug block (Kambouris et al., 1998; Sunami et al., 1997, 2000). However, the observed changes in drug block with these pore substitutions were relatively modest compared with those reported for the cytoplasmic S6 segments (Ragsdale et al., 1994). Furthermore, mutations could indirectly alter drug block by modifying gating and/or drug access (Tomaselli et al., 1995; Sunami et al., 1997, 2000; Chen et al., 2000; Lee et al., 2001). Therefore, it is often difficult to interpret the specific effect of a mutation on drug block. In this article, we report that an external aromatic pore residue, DIV-Trp1531, when substituted, uniquely abolished or attenuated LA block without destabilizing inactivation or altering drug access. The results are discussed in the context of novel insights into the mechanism of drug action and its relationship with the outer pore vestibule structure.
Materials and Methods
Site-Directed Mutagenesis and Heterologous Expression in Xenopus laevis oocytes. Nav1.4 was subcloned into the expression vector pGW1H (British Biotechnologies, Oxford, UK) or 64T (Krieg and Melton, 1984). Mutations were created using polymerase chain reaction with overlapping mutagenic primers, followed by full sequencing to ensure the presence and absence of desired and strayed mutations, respectively. cRNA was transcribed from SalI-linearized cDNA in 64T using T7 RNA polymerase (Promega, Madison, WI). Nav1.4 channel constructs were heterologously expressed and studied in X. laevis oocytes as described previously (Xue et al., 2002). In brief, stage IVeCVI oocytes were surgically removed from female frogs anesthetized by immersion in 0.2% tricaine, followed by digestion with 1.5 mg/ml collagenase (type IA) in OR-2 containing 88 mM NaCl, 2 mM KCl, 1 mM MgCl2, and 5 mM HEPES, pH 7.6, for 30 to 40 min. Wild-type or engineered cDNA (5 ng of cDNA; nuclear injection) or cRNA (50 ng of cRNA; cytoplasmic injection) together with the rat brain 1 subunit (/1 subunit ratio was 1:7) were injected into healthy, isolated, defolliculated oocytes. Injected oocytes were stored in ND96 (frog Ringer's solution) which contained 96 mM NaCl, 2 mM KCl, 1 mM BaCl2, 1 mM MgCl2, and 5 mM HEPES, pH 7.6, supplemented with 50 e蘥/ml gentamicin, 5 mM pyruvate, and 0.5 mM theophylline for 1 to 2 days for expression before experiments.
Electrophysiology and Experimental Protocols. All electrical recordings were performed at room temperature. The bath solution was ND96. Stock solutions of 100 mM mexiletine, lidocaine, QX-222, and QX-314 were made in deionized H2O. Drug stock was added to the bath solution at the concentrations indicated. A 5-min interval was allowed for equilibration whenever the drug concentration was changed. Intracellular application of QX-222 or QX-314 to oocytes was performed by injecting 100 nl ofa2mM drug solution 10 min before experiments. Assuming an oocyte volume of 1 e蘬 (Ragsdale et al., 1994; Qu et al., 1995), the final QX concentration applied intracellularly was 200 e. Data were collected using custom software. Capacity transients were removed using the p/4 leak subtraction protocol.
Use-dependent block was induced by applying a continuous train of 20 ms step depolarizations to eC10 mV from rest at a stimulation frequency of 1, 5, and 10 Hz in the presence of mexiletine or lidocaine. Test potentials for W1531C and W1531A channels were eC20 mV because the reversal potentials of these channels were close to 0 mV (Tsushima et al., 1997a). Holding potential was eC100 mV. Steady-state, use-dependent block was assessed as the fraction of block of the 15th pulse compared with the first pulse (i.e., 1 eC IPulse 15/IPulse 1, where IPulse 1 and IPulse 15 represent the peak currents measured during the 1st and 15th pulses, respectively).
Steady-state activation curves were constructed from current-voltage relationships recorded from a holding potential of eC120 mV using the equation m = g/gmax, where g was obtained from the current-voltage relationship by scaling the peak current (I) by the net driving force using the equation g = I/(Vt eC Erev). Vt is the test potential and Erev is the reversal potential. Current-voltage data were fit to the equation I = m x gmaxx(V eC Erev). Steady-state activation and inactivation curves were obtained by fitting the corresponding data with the Boltzmann function m or h = 1/{1 + exp[(Vt eC V1/2)/k]} where Vt is the test potential, V1/2 is the half-point of the relationship, and k (= RT/zF, where R is the gas constant, T is the absolute temperature, z is the charge number of the electrode reaction, and F is the Faraday constant) is the slope factor. Recovery from inactivation of Nav1.4 channel constructs was examined using a standard two-pulse protocol (500-ms depolarizing pulse to eC20 mV followed by repolarization to eC100 mV for a varying period before a second depolarizing "test" pulse to eC20 mV). A 500-ms pulse was chosen to ensure drug equilibrium without significantly inducing intermediate inactivation (Balser et al., 1996a; Featherstone et al., 1996). The holding potential was eC100 mV. Recovery data were fit with a biexponential function.
Tonic block was determined by measuring Nav currents in response to 50 ms depolarization from a holding potential of eC120 mV at 0.25 Hz stimulation in the presence and absence of drugs, and the data were fit with an isothermal binding equation with the Hill coefficient equals 1. All data are presented as mean ± S.E.M. Statistical significance was determined using the student-t test with p < 0.05 representing significance.
Results
Screening of Cysteine Pore Mutants for Lidocaine Sensitivity. DI-Trp402, DII-Trp756, DIII-Trp1239 and DIV-Trp1531, which are at homologous positions in the four domains, form a tryptophan ring in the extracellular pore vestibule of Nav1.4 channels (Fig. 1). Although W402C has been previously demonstrated to attenuate lidocaine block, presumably by accelerating exit from the intermediate inactivated state that has high affinity for lidocaine (Kambouris et al., 1998), the role of other pore-lining tryptophans (Peerez-Garce猘 et al., 1996; Tsushima et al., 1997b) has not been tested. Given the potential of the tryptophan ring to interact with the aromatic moiety of LA via -electrons (as a result of their proximity to the cytoplasmic drug receptor), we first screened for changes in lidocaine block of W756C, W1239C, and W1531C channels to test whether they participate in LA block. Indeed, significant reductions in use dependence were observed with W402C [percentage of current remaining at 15th pulse (i.e., IPulse 15/IPulse 1 = 65 ± 1.8% versus 49 ± 2.5% of WT with 300 e lidocaine at 10 Hz)] and W1239C (65 ± 1.9%) channels; W1531C was unique in that the substitution completely abolished use-dependent lidocaine block (95.5 ± 0.4%). Given the dramatic change in drug block observed with W1531C channels and the role of DIV in formation of the LA receptor (Ragsdale et al., 1994), we therefore focused on investigating the role of this DIV pore tryptophan in drug binding.
Use-Dependent Block of W1531C by Mexiletine. As a first step to explore in detail the effect of W1531C on LA block, we next examined the effects of stimulation frequency and drug concentration on use-dependent block of Nav1.4 channels by mexiletine, a clinically used oral lidocaine analog, with local anesthetic and antiarrhythmic actions similar to lidocaine (Xu et al., 1992). Figure 2 shows the effect of 100 and 300 e mexiletine on Na+ currents through WT and W1531C channels during a train of depolarizing pulses. Under our control drug-free conditions, no significant use-dependent reduction was observed in either WT or W1531C channels (p > 0.05). Mexiletine (100 e) did not induce noticeable use-dependent block of WT at 1 Hz (p > 0.05) but did cause progressive reduction of Nav currents after depolarizations subsequent to pulse 1 at the higher stimulation frequencies of 5 and 10 Hz. Application of 300 e mexiletine further increased use-dependent block of WT Nav1.4 channels at 1, 5, and 10 Hz. In contrast to WT, the substitution W1531C significantly attenuated mexiletine-induced use-dependence at all the concentrations (including 500 e and 1 mM at 10 Hz) and frequencies tested (p > 0.05). Prolonging the depolarizing test pulse from 20 to 100 ms (to increase the population of inactivated channels) enhanced steady-state use-dependent block (assessed as IPulse 15, 300 e Mex/IPulse 1, 300 e Mex eC IPulse 15, drug-free/IPulse 1, drug-free) of WT Nav1.4 channels by 300 e mexiletine at 5 Hz to 32.3 ± 4.9% (n = 4) but caused only a modest 13.3 ± 1.1% (n = 4) block for W1531C channels under the same experimental conditions (Fig. 2B, center).
Effects of Mexiletine on Steady-State Fast-Inactivation of Nav1.4 Channels. Several mechanisms (alterations of drug affinity, drug access, gating, or a combination) may contribute to changes in LA block. Although the inactivation time course (h = 2.5 ± 0.1 ms at eC20 mV, n = 12) of W1531C Nav1.4 channels, an index of their rate of entry into fast-inactivation was notably slowed compared with that of WT (h = 1.3 ± 0.1 ms at the same voltage; n = 10), the DIV pore construct completely inactivated within 20 ms indicating no disruption of inactivation (compare Figure 2A). Indeed, Fig. 3 shows that the mid-point (V1/2) of steady-state inactivation for W1531C channels (eC70.0 ± 1.6 mV, n = 4) was negatively shifted compared with that of WT channels (eC48.6 ± 0.5 mV, n = 5; p < 0.05) suggesting that fast-inactivated W1531C channels were stabilized rather than destabilized by the substitution. Furthermore, steady-state activation was not altered by the substitution (V1/2: eC28.8 ± 0.7 mV versus eC33.7 ± 2.6 mV of WT; p > 0.05). The steady-state gating parameters are summarized in Table 1.
Recovery data were fit with a biexponential function (see Materials and Methods). Data represent the mean ± S.E.M. of three to five oocytes.
To explore the mechanisms by which W1531C affects LA block, we studied the changes in drug binding energetics by measuring shifts in V1/2 of inactivation or channel availability by mexiletine (Fig. 3, Table 1). For WT channels, mexiletine significantly shifted V1/2 in the hyperpolarizing direction as anticipated for preferential drug binding to the inactivated state with the addition of LA (Courtney, 1975; Hille, 1978) (V1/2 eC4.0 ± 0.4 mV, n = 3 and eC11.0 ± 0.6 mV, n = 3 for 100 and 300 e, respectively; p < 0.05). In contrast, mexiletine did not produce appreciable shift in V1/2 of W1531C channels (V1/2 eC2.1 ± 0.9 mV, n = 3 and eC2.4 ± 2.1 mV, n = 3 for 100 and 300 e, respectively; p > 0.05) suggesting disruption of drug binding to inactivated channels.
Effects of Mexiletine on Recovery from Fast-Inactivation of WT and W1531C channels. Next, we examined the effects of Trp1531 on the kinetics of LA block by examining recovery from inactivation in the presence and absence of mexiletine. Figure 4A shows that WT Nav1.4 channels recovered from inactivation with two time constants under control conditions. The fast component (fast, con) accounted for the majority (Afast>95%) of the total recovery; the comparatively minor slow component (slow, con) probably reflects channels recovering from the intermediate slow-inactivated states in response to 500-ms depolarizations (Featherstone et al., 1996). In the presence of 100 or 300 e mexiletine, the proportion of channels recovering quickly from fast-inactivation (i.e., Afast) was substantially reduced. However, fast, mex was very similar in magnitude to fast, con, consistent with the idea that this component represents the fast recovery of drug-free channels; in contrast, slow, mex reflects the recovery rate of mexiletine-bound channels (Bean et al., 1983). Therefore, Afast, mex and Aslow, mex can provide a measure of the proportion of drug-free and -bound channels, respectively. Figure 4B shows the effect of mexiletine on the recovery from inactivation for W1531C channels. Under control conditions, like WT, the recovery of W1531C channels from inactivation displayed two distinct phases, except that the proportion of W1531C channels recovering slowly made up a larger component of the total recovery (Aslow = 31.2 ± 8.6%; p < 0.05). In accordance with disrupted drug binding, application of 100 and 300 e mexiletine altered neither the time constants nor their relative amplitudes (p > 0.05). All fitted parameters for recovery from inaction of WT and W1531C channels with and without mexiletine are summarized in Table 1.
W1531C Abolished Mexiletine Binding to Intermediate Slow-Inactivated Channels. Because DI-W402C attenuates lidocaine block, presumably by accelerating exit from the intermediate inactivated state, which has high LA affinity (Kambouris et al., 1998), we tested whether the same mechanism also applies to W1531C channels by examining the effects of mexiletine and the substitution W1531C on entry into (intermediate) slow-inactivation of Nav1.4 channels (Fig. 5). The protocol employed is given in the inset. To examine the time-dependent development of intermediate slow-inactivation, Nav1.4 channels were prepulsed from a holding potential of eC120 to eC20 mV for a variable period of time from 2 ms to 10 s to induce inactivation. This depolarizing prepulse (P1) was followed by a short, 20-ms hyperpolarization to eC120 mV to allow recovery from fast- but not slow-inactivation. A test depolarizing pulse (P2) was then applied to assess any accumulation of intermediate slow-inactivation as current reduction relative to that recorded during the prepulse. Figure 5A shows that 300 e mexiletine substantially accelerated the entry of WT channels into the intermediate slow-inactivated state(s), consistent with the notion that LAs are allosteric effectors of inactivation (Balser et al., 1996b). It is interesting that entry of W1531C channels into intermediate slow-inactivation was enhanced (Fig. 5B) despite their rate of entry into the fast-inactivated state(s) was slowed by the substitution (compare Figure 2A). Unlike WT channels, however, mexiletine failed to facilitate the entry of W1531C channels into intermediate slow-inactivation. Taken together, these results suggest that mexiletine also did not bind intermediate slow-inactivated W1531C channels. Therefore, although Trp402 and Trp1531 are at the "equivalent" positions in the DI and DIV P-loops, respectively, they exert functionally distinctive effects on LA block.
Trp1531 Substitutions Altered Neither External Nor Cytoplasmic LA Access. To investigate whether drug access to the LA receptor was modified by the substitution W1531C, we employed the permanently and positively charged membrane-impermeant quaternary lidocaine derivative QX-222. QX-222 binds to the LA receptor only via the hydrophilic pathway, making it a useful probe for assessing access to the drug binding site (Frazier et al., 1970; Strichartz, 1973; Alpert et al., 1989). In addition, we also constructed W1531A and W1531Y channels to probe the role of the chemical structure of the side chain at position 1531 in LA block of Nav1.4 channels. In the absence of QX-222, neither WT, W1531C, W1531A, nor W1531Y channels exhibited use-dependence when a train of 100-ms depolarizing pulses was applied at 1 Hz (p > 0.05; data not shown). Figure 6, A and B, shows that application of 200 e intracellular QX-222 induced significant use-dependent reduction of WT Nav current (46.0 ± 3.2%, n = 4 at the 60th pulse). In contrast, no detectable block was observed when the same concentration of QX-222 was internally applied to W1531C channels (p > 0.05). It is interesting that, similar to WT, W1531Y channels were potently blocked by internal QX-222. W1531A channels had an intermediate phenotype. Therefore, use-dependent QX-222 block follows the sequence WTW1531Y >> W1531A > W1531C suggesting that both aromaticity and hydrophobicity of the native residue at position 1531 are important for drug block.
Although the results presented to date are entirely consistent with a disruption or destabilization of the drug-channel complex, it is still possible that the observed reduction in drug block arise from alterations in drug access to the LA receptor site (e.g., creation of an external drug access pathway for bound drug to escape) (Ragsdale et al., 1994; Qu et al., 1995). To differentiate between these possibilities, we performed the following experiments. Consistent with previous studies, extracellular application of 500 e QX-222 to WT Nav1.4 channels resulted in no significant Na+ current blockade 5 min after drug application (IQX/IControl = 100.0 ± 3.6%, n = 4, Fig. 6C), despite potent block by internal QX222. Likewise, 500 e external QX-222 also did not significantly block W1531C, W1531A, and W1531Y channels, which were insensitive, intermediate, and highly sensitive to internal QX-222, respectively, over the same period of time (Fig. 6C), suggesting that these substitutions did not create a low energy pathway between the LA receptor and the extracellular solution.
Unlike W1531A channels, the complete absence of external and internal QX-222 block of W1531C channels did not enable us to draw definitive conclusions about possible alterations of drug access (or the lack thereof). Therefore, we employed the more potent derivative QX-314. As anticipated, 500 e intracellular QX-314 blocked 83.1 ± 2.5% (n = 6) of WT currents but induced only a modest yet significant 16.2 ± 2.5% (n = 6) use-dependent reduction of peak INa at steady-state (IPulse 60) in W1531C channels (Fig. 7A). To assess the recovery of WT and W1531C channels from internal QX-314 block, a variable interval of hyperpolarization to eC140 mV was given to open the intracellular gate for the trapped drug molecule to unbind (Strichartz, 1973; Yeh and Tanguy, 1985; Starmer et al., 1986). Consistent with previous reports, recovery from internal QX-314 block of WT Nav1.4 channels was extremely slow (Strichartz, 1973; Yeh and Tanguy, 1985; Sunami et al., 1997) with 31.2 ± 9.4% (n = 3) of the blocked current recovered over a 240-s hyperpolarization period (Fig. 7B). In contrast, 58.4 ± 3.8% (n = 3) of W1531C recovered within 5 s. Complete recovery was achieved within 15 s with a recovery time constant of 5.9 ± 1.1 s (versus 81.0 ± 25.5 s for WT). It is noteworthy that the same concentration of QX-314, when applied externally, blocked neither WT nor W1531C channels (data not shown), as was the case for QX-222.
To determine whether drug access to the LA receptor from the cytoplasmic side was altered in W1531C channels, we also examined the ability of QX-222 to bind to inactivated channels (Fig. 8). The protocol used is shown in the inset. Nav1.4 channels were prepulsed from a holding potential of eC120 to eC55 mV for a variable period of time from 1 ms to 10 s to induce inactivation without channel opening (Lawrence et al., 1991). At eC55 mV, more than 80% of the channels had entered the fast-inactivated state at steady state (compare Figure 3B). This prepulse was followed by a hyperpolarizing pulse of 300 ms at eC120 mV to allow recovery from fast-inactivation. The small reduction of current after depolarizing prepulses longer than 100 ms under control QX-222-free conditions probably reflects accumulation of intermediate inactivation (Bennett et al., 1995; Featherstone et al., 1996). After collecting control data, the same oocytes were then injected with 100 nl of 2 mM QX-222 for 10 min (final concentration, 200 e), and the protocol was repeated with a repetition interval of 25 s (to ensure complete recovery from use-dependent block). If internal QX-222 has access to the LA receptor when channels were inactivated, accumulated block would be expected. However, the onset of current reduction after QX-222 injection was not different from the control. Identical results were also obtained with Qx-314 (data not shown). Similar studies in WT Nav1.4 channels were not possible because of the extremely slow recovery of this channel from internal QX-222 or QX-314 block (compare Figure 7B) (Strichartz, 1973; Yeh and Tanguy, 1985; Sunami et al., 1997).
Tonic Block Was Enhanced by W1531C. Our results support the hypothesis that the LA receptor of W1531C channels is inaccessible to intracellular QX-222 when channels are inactivated or when QX-222 was applied externally. Although hydrophobic and amphiphilic LAs preferentially bind to inactivated Nav channels, they also bind to the resting state, thereby producing tonic block (Hille, 1977; Hondeghem and Katzung, 1977). Assessment of tonic block was accomplished by comparing the peak Na+ current amplitude (recorded after step depolarizations from a holding potential of eC120 mV at a low stimulation rate of 0.25 Hz) before and after application of various concentrations of mexiletine. Figure 9A shows the effects of 1 mM mexiletine on whole-cell currents of Nav1.4 and W1531C channels. The fraction of current remaining after mexiletine application was plotted against drug concentration to obtain the dose-response curves of WT and W1531C channels. Interestingly, W1531C channels were more sensitive to tonic block by mexiletine than WT. Assuming a 1:1 stoichiometry (Bean et al., 1983; Nuss et al., 1995; Balser et al., 1996b), the half-blocking concentrations (IC50) were 2974.3 ± 1.3 and 1286.0 ± 1.4 e for WT and W1531C channels, respectively (Fig. 9B).
Discussion
Trp1531 and LA Block. LAs are commonly used to treat a variety of diseases (including LQT3) that involve abnormal Nav channel function. The major finding of this study was that the replacement of the aromatic residue Trp1531, located in the ascending portion of the P-loop in DIV, with cysteine or alanine abolished or markedly attenuated LA-induced use dependence, shift in steady-state fast-inactivation, and delay in the recovery rate from inactivation. Because gating interacts profoundly with LA binding, mutagenesis experiments of Nav channels to study drug block are often difficult to interpret. For instance, gating was shown to be altered by numerous amino acid substitutions in the putative receptor site S6 and the P region, independent of their superimposed drug effects (Kondratiev and Tomaselli, 2003). In fact, LA block phenotype can be altered by mutations in channel regions that are distant from the putative receptor site (Fan et al., 1996; Li et al., 2002). Although several P-loop substitutions from other domains have also been shown to influence LA block (Sunami et al., 1997; Kambouris et al., 1998), Trp1531, however, is unique in that its substitution by cysteine completely abolished LA block without destabilizing inactivation and altering drug access. Indeed, the slower recovery from inactivation and the hyperpolarizing shifts of steady-state fast-inactivation of W1531C and W1531A channels suggest that these substitutions stabilize inactivation. Therefore, these changes in the gating properties of Trp1531 channels should enhance LA block, as predicted from a simple gating-drug block coupling effect [as has been described for the enhanced lidocaine block of the paramyotonia congenital mutation, R1448C, in skeletal muscle of Nav channels (Fan et al., 1996)]. However, this was obviously not the case for Trp1531 mutations. Modification of drug access can affect use-dependent block and recovery from inactivation by changing the rates for drug binding and unbinding without shifting steady-state inactivation (Bean et al., 1983). However, QX-222 access from the extracellular or cytoplasmic face was not altered by Trp1531 substitutions. Therefore, it is unlikely that the dramatic changes in LA block observed were secondary to changes in gating and drug access.
Molecular Picture of LA Block. Our data raise the intriguing possibility that Trp1531 constitutes part of the dynamic drug receptor. This hypothesis is appealing for several reasons. First, reduced drug block could not be secondary to alterations in gating because none of the Trp1531 substitutions studied disrupts inactivation. Second, Trp1531 is an aromatic residue capable of forming -electron interactions, which are known to be crucial for high-affinity LA block of Na channels (Zamponi and French, 1993; Ragsdale et al., 1994; Sheldon et al., 1995). Furthermore, Trp1531 is a determinant of ionic selectivity; it has long been thought that LAs bind deep in the pore from the cytoplasmic side up to the selectivity region (Hille, 2001; Sunami et al., 1997; Tsushima et al., 1997a). Although Trp1531 channels cannot discriminate among monovalent cations (Chiamvimonvat et al., 1996b; Tsushima et al., 1997a), they are impermeable to organic cations such as tetramethylammonium (R. A. Li and P. H. Backx, unpublished data) whose chemical structure mimics that of lidocaine, consistent with the finding that QX compounds do not block from the extracellular side. Finally, the external pore vestibule is adjacent to the cytoplasmic LA receptor pocket (e.g., benzocaine action remains potent even when the drug is covalently anchored to DI-Y401C) (Li et al., 1999): the electrical distance () for LA block is 0.7 to 0.8 from the inside (Gingrich et al., 1993), whereas the externally accessible Trp1531 has 0.25 from the outside for Cd2+ block (Chiamvimonvat et al., 1996a). At first glance, the external accessibility of Trp1531 does not seem to reconcile with the cytoplasmic location of the LA receptor (Ragsdale et al., 1994, 1996). Perhaps Trp1531 forms the "roof" of the cytoplasmic pore vestibule that contains the drug receptor (Fig. 10A). On the other hand, Trp1531 may translocate from the extracellular to the cytoplasmic face (or "flip" at the interface) during inactivation and/or upon drug binding (Fig. 10, B-C), which subsequently leads to intimate hydrophobic or -electron interactions of the aromatic drug portion with Trp1531 and other residues comprising the LA receptor (such as those in DIV S6). Previously, two distinct modes of fast and discrete-slow single-channel block of skeletal muscle and cardiac Nav channels by lidocaine and QX-314 have been observed (Gingrich et al., 1993; Zamponi and French, 1993). Based on these results, it has been suggested that low-affinity rapid QX-314 block serves as an intermediate to catalyze the formation of a high-affinity discrete block complex (Gingrich et al., 1993). According to our model, Trp1531 is critical for the transition from low- to high-affinity conformations, consistent with the notion that DIV P-loop is most flexible relative to P-loops from other domains (Tsushima et al., 1997b). Furthermore, our proposed drug binding mechanism is also in accordance with the model that LAs act as allosteric effectors of inactivation (Balser et al., 1996b). Indeed, dynamic rearrangements of the Na channel pore, similar to those reported for voltage-gated K+ (Kv) channels, have been demonstrated to link to slow inactivation and use dependence by molecular motions of the outer charge ring (Glu403, Glu758, Asp1241, and Asp1532) (Xiong et al., 2003) as well as state-dependent accessibility of F1236C in the DIII P-loop to externally applied methanethiosulfonate ethylammonium (Ong et al., 2000). When DIV Trp1531 is substituted, high-affinity LA binding is attenuated or abolished (Fig. 10C). In fact, ionic selectivity for monovalent cations is also altered (Tsushima et al., 1997a), presumably because of disruptions of the selectivity filter caused by these mutations. However, drug binding and selectivity can be retained with the conservative substitution W1531Y.
If drug molecules bind to the inactivated state, the slowed inactivation time course of W1531C channels (compare Fig. 2A) may contribute to the lack of use-dependence in that channels will be considerably less inactivated during each depolarizing pulse of the train. Furthermore, one might also argue that nonaromatic substitutions of Trp1531 intrinsically mimic LA action (or normal drug-bound channels) by negatively shifting the steady-state fast-inactivation (see Table 1) and decelerating the rate of recovery from inactivation, thereby saturating any possible drug effects. However, relative to WT channels, mexiletine-induced use-dependence of W1531C channels only very modestly increased even with 5-fold prolongation of depolarizing pulse (from 20 ms to 100 ms), suggesting that drug binding to fast-inactivated W1531C channels was indeed significantly attenuated. In addition, our preliminary experiments further indicate that substituting the equivalent tryptophan in the human cardiac Nav1.5 channels, Trp1712, with cysteine altered neither the inactivation time course nor steady-state inactivation yet also rendered the cardiac channels insensitive to drug block (F. Tsang and R. A. Li, unpublished data). Although the differences between Nav1.4-W1531C and Nav1.5-W1712C channels can be attributed to the fundamental differences in gating properties between the two isoforms, our observations collectively support the notion that the changes in drug effects observed were not secondary to changes in gating. Thus, it is possible that the DIV pore tryptophan constitutes part of the drug receptor.
W1531C enhanced tonic block despite its attenuating effects on LA-induced use dependence, V1/2, and recovery from inactivation. Similar changes have been previously observed in channels with residues replaced in DIV S6 (Ragsdale et al., 1994). Bean et al. (1983) suggest that tonic block antagonizes use-dependent block at depolarized holding potentials because channels with drug bound to the closed and/or open (i.e., rapid block upon channel opening) states are not available for further drug binding to the inactivated state, thus leading to an apparent reduced use dependence. Further experiments are needed to differentiate among the two possibilities.
Aromaticity Is Critical for Na+ Channel Function and Pharmacology. Remarkably, DIV-Trp1531 seems to govern several major Nav channel properties, whereas the analogous Trp402, Trp756, and Trp1239 in DI-III are not as significant. When the DIV aromatic residue is replaced with the nonaromatic amino acid cysteine or alanine, channels do not discriminate among monovalent cations (such as Na+, K+, NH4+, and Cs+). Trp1531 is also an important determinant for e?conotoxin (Li et al., 1997) as well as LA block. However, replacing Trp1531 with tyrosine or phenylalanine (i.e., W1531F or W1531Y) maintains WT phenotypes. These results suggest that Trp1531 plays a multifunctional role and that its aromaticity is crucial for normal channel function and pharmacology (Li et al., 1997; Tsushima et al., 1997a) and are consistent with the notion that DIV is a particularly important domain.
Conclusion
In summary, our study demonstrates that the DIV aromatic residue Trp1531 in the external pore vestibule of Nav1.4 channels, which has previously been shown to play a critical role in ionic selectivity (Chiamvimonvat et al., 1996a; Tsushima et al., 1997a) and e?conotoxin binding (Li et al., 1997), also prominently influences LA block. The observed reductions in drug sensitivity were not caused by changes in gating and drug access. We propose that Trp1531 dynamically constitutes part of the LA receptor. Together, these data provide novel structural and functional insights into the external pore of Nav channels in relation to LA block, which may be helpful for future molecular-based drug design.
References
Alpert LA, Fozzard HA, Hanck DA, and Makielski JC (1989) Is there a second external lidocaine binding site on mammalian cardiac cells Am J Physiol 257: H79eCH84.
Backx P, Yue D, Lawrence J, Marban E, and Tomaselli G (1992) Molecular localization of an ion-binding site within the pore of mammalian sodium channels. Science (Wash DC) 257: 248eC251.
Balser JR, Nuss HB, Chiamvirnonvat N, Peerez-Garce猘 MT, Marban E, and Tomaselli GF (1996a) External pore residues mediates slow inactivation in e蘬 rat skeletal muscle sodium channels. J Physiol 494: 431eC442.
Balser JR, Nuss HB, Orias DW, Johns DC, Marban E, and Lawrence JH (1996b) Local anesthetics as effectors of allosteric gating: Lidocaine effects on inactivation-deficient rat skeletal muscle Na channels. J Clin Investig 98: 2876eC2886.
Balser JR (2001) Inherited sodium channelopathies: novel therapeutic and proarrhythmic molecular mechanisms. Trends Cardiovasc Med 11: 229eC237.
Bean BP, Cohen CJ, and Tsien RW (1983) Lidocaine block of cardiac sodium channels. J Gen Physiol 81: 613eC642.
Bennett PB, Valenzuela C, Chen LQ, and Kallen RG (1995) On the molecular nature of the lidocaine receptor of cardiac Na+ channels. Modification of block by alterations in the alpha-subunit III-IV interdomain. Circ Res 77: 584eC592.
Chen Z, Ong BH, Kambouris NG, Marbe E, Tomaselli GF, and Balser JR (2000) Lidocaine induces a slow inactivated state in rat skeletal muscle sodium channels. J Physiol 524: 37eC49.
Chiamvimonvat N, Peerez-Garce猘 M, Ranjan R, Marban E, and Tomaselli GF (1996a) Depth asymmetries of the pore-lining segments of the sodium channel revealed by cysteine mutagenesis. Neuron 16: 1037eC1047.
Chiamvimonvat N, Peerez-Garce猘 MT, Tomaselli GF, and Marban E (1996b) Control of ion flux and selectivity by negatively charged residues in the outer mouth of rat sodium channels. J Physiol 491: 51eC59.
Courtney KR (1975) Mechanisms of frequency-dependent inhibition of sodium currents in frog myelinated berve by the lidocaine derivative GEA 968. J Pharmacol Exp Ther 195: 225eC236.
Fan Z, George Jr, AL, Kyle JW, and Makielski JC (1996) Two human paramyotonia congenital mutations have opposite effects on lidocaine block of Na+ channels expressed in a mammalian cell line. J Physiol 496: 275eC286.
Featherstone DE, Richmond JE, and Ruben PC (1996) Interaction between fast and slow inactivation in rSkM1 sodium channels. Biophys J 71: 3098eC3109.
Frazier DT, Narahashi T, and Yamada M (1970) The site of action and active form of local anesthetics. Experiments with quaternary compounds. J Pharmacol Exp Ther 171: 45eC51.
Gingrich KJ, Beardsley D, and Yue DT (1993) Ultra-deep blockade of Na+ channels by a quaternary ammonium ion: catalysis by a transition-intermediate state J Physiol (Lond) 471: 319eC341.
Hayward, LJ, Brown, RHJr. and Cannon SC (1997) Slow inactivation differs among mutant Na channels associated with myotonia and periodic paralysis. Biophys J 72: 1204eC1219.
Hille B (1977) Local anesthetics: hydrophilic and hydrophobic pathways for the drug receptor reaction. J Gen Physiol 69: 479eC515.
Hille B (1978) Local anesthetic action on inactivation of the Na+ channel in nerve and skeletal muscle, in Biophysical Aspects of Cardiac Muscle (Morad M, Smith S, eds) pp 55eC74, Academic Press Inc, New York.
Hille B (2001) Ionic Channels of Excitable Membranes, 3rd ed. Sinauer Associates, Inc., Sunderland, Massachusetts.
Hondeghem LM and Katzung BG (1977) Time- and voltage-dependent interactions of anti-arrhythmic drugs with cardiac sodium channels. Biochim Biophys Acta 472: 373eC398.
Hudson AJ, Ebers GC and DE Bulman (1995) Review: the skeletal muscle sodium and chloride channel diseases. Brain 118: 547eC563.
Kambouris KG, Hastings LA, Stepanovic S, Marban E, Tomaselli GF, and Balser JR (1998) Mechanistic link between lidocaine block and inactivation probed by outer pore mutations in the rat e? skeletal muscle sodium channel. J Physiol 512: 693eC705.
Keating MT and Sanguinetti MC (1996) Pathophysiology of ion channel mutations. Curr Opin Genet Dev 6: 326eC333.
Krieg PA and Melton DA (1984) Functional messenger RNAs are produced by SP6 in vitro transcription of cloned cDNAs. Nucleic Acids Res 12: 7057eC7070.
Kimbrough JT and Gingrich KJ (2000) Quaternary ammonium block of mutant Na+ channels lacking inactivation: features of a transition-intermediate mechanism. J Physiol 529: 93eC106.
Kondratiev A and Tomaselli GF (2003) Altered gating and local anesthetic block mediated by residues in the I-S6 and II-S6 transmembrane segments of voltage-dependent Na+ channels. Mol Pharmacol 64: 741eC752.
Lawrence JH, Yue DT, Rose WC, and Marban E (1991) Sodium channel inactivation from resting states in guinea-pig ventricular myocytes. J Physiol 443: 629eC650.
Lee PJ, Sunami A, and Fozzard HA (2001) Cardiac-specific external paths for lidocaine, defined by isoform-specific residues, accelerate recovery from use-dependent block. Circ Res 89: 1014eC1021.
Li RA, Tsushima RG, Himmeldirk K, Dime DS, and Backx PH (1999) Local anesthetic anchoring to cardiac sodium channels. Implications into tissue-selective drug targeting. Circ Res 85: 88eC98.
Li RA, Tsushima RG, Kallen RG, and Backx PH (1997) Pore residues critical for e?conotoxin binding to rat skeletal muscle Na+ channel. Biophys J 73: 1874eC1884.
Li RA, Ennis IL, Tomaselli GF, and Marban E (2002) Structural basis of differences in isoform-specific gating and lidocaine block between cardiac and skeletal muscle sodium channels. Mol Pharmacol 61: 136eC141.
Li RA, Velez P, Chiamvimonvat N, Tomaselli GF, and Marban E (2000) Charged residues between the selectivity filter and S6 segments contribute to the permeation phenotype of the sodium channel. J Gen Physiol 115: 81eC92.
Marban E, Yamagishi T, and Tomaselli GF (1998) Structure and function of voltage-gated sodium channels. J Physiol 508: 647eC657.
Nuss HB, Chiamvimonvat N, Peerez-Garce猘 MT, Tomaselli GF, and Marban E (1995) Functional association of the beta 1 subunit with human cardiac (hH1) and rat skeletal muscle (mu 1) sodium channel alpha subunits expressed in Xenopus oocytes. J Gen Physiol 106: 1171eC1191.
Ong BH, Tomaselli GF, and Balser JR (2000) A structural rearrangement in the sodium channel pore linked to slow inactivation and use dependence. J Gen Physiol 116: 653eC662.
Peerez-Garce猘 MT, Chiamvimonvat N, Marban E, and Tomaselli GF (1996) Structure of the sodium channel pore revealed by serial cysteine mutagenesis. Proc Natl Acad Sci USA 93: 300eC304.
Qu Y, Rogers J, Tanada T, Scheuer T, and Catterall WA (1995) Molecular determinant of drug access to the receptor site for antiarrhythmic drugs in the cardiac Na+ channel. Proc Natl Acad Sci USA 92: 11839eC11843.
Ragsdale DS, McPhee JC, Scheuer T, and Catterall WA (1994) Molecular determinants of state-dependent block of Na+ channels by local anesthetics. Science (Wash DC) 265: 1724eC1728.
Ragsdale DS, McPhee JC, Scheuer T, and Caterall WA (1996) Common molecular determinants of local anesthetic, antiarrhythmic and anticonvulsant block of voltage-gated Na+ channels. Proc Natl Acad Sci USA 93: 9270eC9275.
Sheldon R, Duff H, and Koshman ML (1995) Antiarrhythmic activity of quinine in humans. Circ 92: 2944eC2950.
Starmer CF, Yeh JZ, and Tanguy J (1986) A quantitative description of QX222 blockade of sodium channels in squid axons. Biophys J 49: 913eC920.
Strichartz GR (1973) The inhibition of sodium currents in myelinated nerve by quaternary derivatives of lidocaine. J Gen Physiol 62: 37eC57.
Sunami A, Dudley SC Jr, and Fozzard HA (1997) Sodium channel selectivity filter regulates antiarrhythmic drug binding. Proc Natl Acad Sci USA 94: 14126eC14131.
Sunami A, Glaaser IW, and Fozzard HA (2000) A critical residue for isoform difference in tetrodotoxin affinity is a molecular determinant of the external access path for local anesthetics in the cardiac sodium channel. Proc Natl Acad Sci USA 97: 2326eC2331.
Tomaselli GF, Chiamvimonvat N, Nuss HB, Balser JR, Peerez-Garce猘 MT, Xu RH, Orias DW, Backx PH, and Marban E (1995) A mutation in the pore of the sodium channel alters gating. Biophys J 68: 1814eC1827.
Tsushima RG, Li RA, and Backx PH (1997a) Altered ionic selectivity of Na+ channel revealed by cysteine mutations within the pore. J Gen Physiol 109: 463eC475.
Tsushima RG, Li RA, and Backx PH (1997b) P-loop flexibility in Na+ channel pores revealed by single- and double-cysteine replacements. J Gen Physiol 110: 59eC72.
Wang SY, Nau C, and Wang GK (2000) Residues in Na+ channel D3eCS6 segment modulate both batrachotoxin and local anesthetic affinities. Biophys J 79: 1379eC1387.
Xiong W, Li RA, Tian Y, and Tomaselli GF (2003) Molecular motions of the outer ring of charge of the sodium channel: do they couple to slow inactivation J Gen Physiol 122: 323eC332.
Xu XJ, Hao JX, Seiger A, Arner S, Lindblom U, and Wiesenfeld-Hallin Z (1992) Systemic mexiletine relieves chronic allodynialike symptoms in rats with ischemic spinal cord injury. Anesth Analg 74: 649eC652.
Xue T, Marban E and RA Li (2002) Dominant-negative suppression of HCN1- and HCN2-encoded pacemaker currents by an engineered HCN1 construct: insights into structure-function relationships and multimerization. Circ Res 90: 1267eC1273.
Yarov-Yarovoy V, McPhee JC, Idsvoog D, Pate C, Scheuer T, and Catterall WA (2002) Role of amino acid residues in transmembrane segments IS6 and IIS6 of the Na+ channel subunit in voltage-dependent gating and drug block. J Biol Chem 277: 35393eC35401.
Yeh JZ and Tanguy J (1985) Na channel activation gate modulates slow recovery from use-dependent block by local anesthetics in squid giant axons. Biophys J 47: 685eC694.
Zamponi GW and French RJ (1993) Dissecting lidocaine action: diethylamide and phenol mimic separate modes of lidocaine block of sodium channels from heart and skeletal muscle. Biophys J 65: 2335eC2347.